AT32UC3A0/A1 Series - Complete

Features
• High Performance, Low Power 32-Bit Atmel® AVR® Microcontroller
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
– Compact Single-cycle RISC Instruction Set Including DSP Instruction Set
– Read-Modify-Write Instructions and Atomic Bit Manipulation
– Performing 1.49 DMIPS / MHz
Up to 91 DMIPS Running at 66 MHz from Flash (1 Wait-State)
Up to 49 DMIPS Running at 33MHz from Flash (0 Wait-State)
– Memory Protection Unit
Multi-hierarchy Bus System
– High-Performance Data Transfers on Separate Buses for Increased Performance
– 15 Peripheral DMA Channels Improves Speed for Peripheral Communication
Internal High-Speed Flash
– 512K Bytes, 256K Bytes, 128K Bytes Versions
– Single Cycle Access up to 33 MHz
– Prefetch Buffer Optimizing Instruction Execution at Maximum Speed
– 4ms Page Programming Time and 8ms Full-Chip Erase Time
– 100,000 Write Cycles, 15-year Data Retention Capability
– Flash Security Locks and User Defined Configuration Area
Internal High-Speed SRAM, Single-Cycle Access at Full Speed
– 64K Bytes (512KB and 256KB Flash), 32K Bytes (128KB Flash)
External Memory Interface on AT32UC3A0 Derivatives
– SDRAM / SRAM Compatible Memory Bus (16-bit Data and 24-bit Address Buses)
Interrupt Controller
– Autovectored Low Latency Interrupt Service with Programmable Priority
System Functions
– Power and Clock Manager Including Internal RC Clock and One 32KHz Oscillator
– Two Multipurpose Oscillators and Two Phase-Lock-Loop (PLL) allowing
Independant CPU Frequency from USB Frequency
– Watchdog Timer, Real-Time Clock Timer
Universal Serial Bus (USB)
– Device 2.0 Full Speed and On-The-Go (OTG) Low Speed and Full Speed
– Flexible End-Point Configuration and Management with Dedicated DMA Channels
– On-chip Transceivers Including Pull-Ups
Ethernet MAC 10/100 Mbps interface
– 802.3 Ethernet Media Access Controller
– Supports Media Independent Interface (MII) and Reduced MII (RMII)
One Three-Channel 16-bit Timer/Counter (TC)
– Three External Clock Inputs, PWM, Capture and Various Counting Capabilities
One 7-Channel 16-bit Pulse Width Modulation Controller (PWM)
Four Universal Synchronous/Asynchronous Receiver/Transmitters (USART)
– Independant Baudrate Generator, Support for SPI, IrDA and ISO7816 interfaces
– Support for Hardware Handshaking, RS485 Interfaces and Modem Line
Two Master/Slave Serial Peripheral Interfaces (SPI) with Chip Select Signals
One Synchronous Serial Protocol Controller
– Supports I2S and Generic Frame-Based Protocols
One Master/Slave Two-Wire Interface (TWI), 400kbit/s I2C-compatible
One 8-channel 10-bit Analog-To-Digital Converter
16-bit Stereo Audio Bitstream
– Sample Rate Up to 50 KHz
32-Bit Atmel AVR
Microcontroller
AT32UC3A0512
AT32UC3A0256
AT32UC3A0128
AT32UC3A1512
AT32UC3A1256
AT32UC3A1128
32058K-AVR32-01/12
AT32UC3A
• On-Chip Debug System (JTAG interface)
– Nexus Class 2+, Runtime Control, Non-Intrusive Data and Program Trace
• 100-pin TQFP (69 GPIO pins), 144-pin LQFP (109 GPIO pins) , 144 BGA (109 GPIO pins)
• 5V Input Tolerant I/Os
• Single 3.3V Power Supply or Dual 1.8V-3.3V Power Supply
2
32058K
AVR32-01/12
AT32UC3A
1. Description
The AT32UC3A is a complete System-On-Chip microcontroller based on the AVR32 UC RISC
processor running at frequencies up to 66 MHz. AVR32 UC is a high-performance 32-bit RISC
microprocessor core, designed for cost-sensitive embedded applications, with particular emphasis on low power consumption, high code density and high performance.
The processor implements a Memory Protection Unit (MPU) and a fast and flexible interrupt controller for supporting modern operating systems and real-time operating systems. Higher
computation capabilities are achievable using a rich set of DSP instructions.
The AT32UC3A incorporates on-chip Flash and SRAM memories for secure and fast access.
For applications requiring additional memory, an external memory interface is provided on
AT32UC3A0 derivatives.
The Peripheral Direct Memory Access controller (PDCA) enables data transfers between peripherals and memories without processor involvement. PDCA drastically reduces processing
overhead when transferring continuous and large data streams between modules within the
MCU.
The PowerManager improves design flexibility and security: the on-chip Brown-Out Detector
monitors the power supply, the CPU runs from the on-chip RC oscillator or from one of external
oscillator sources, a Real-Time Clock and its associated timer keeps track of the time.
The Timer/Counter includes three identical 16-bit timer/counter channels. Each channel can be
independently programmed to perform frequency measurement, event counting, interval measurement, pulse generation, delay timing and pulse width modulation.
The PWM modules provides seven independent channels with many configuration options
including polarity, edge alignment and waveform non overlap control. One PWM channel can
trigger ADC conversions for more accurate close loop control implementations.
The AT32UC3A also features many communication interfaces for communication intensive
applications. In addition to standard serial interfaces like UART, SPI or TWI, other interfaces like
flexible Synchronous Serial Controller, USB and Ethernet MAC are available.
The Synchronous Serial Controller provides easy access to serial communication protocols and
audio standards like I2S.
The Full-Speed USB 2.0 Device interface supports several USB Classes at the same time
thanks to the rich End-Point configuration. The On-The-GO (OTG) Host interface allows device
like a USB Flash disk or a USB printer to be directly connected to the processor.
The media-independent interface (MII) and reduced MII (RMII) 10/100 Ethernet MAC module
provides on-chip solutions for network-connected devices.
AT32UC3A integrates a class 2+ Nexus 2.0 On-Chip Debug (OCD) System, with non-intrusive
real-time trace, full-speed read/write memory access in addition to basic runtime control.
3
32058K
AVR32-01/12
AT32UC3A
2. Configuration Summary
The table below lists all AT32UC3A memory and package configurations:
Device
Flash
SRAM
Ext. Bus Interface
Ethernet
MAC
AT32UC3A0512
512 Kbytes
64 Kbytes
yes
yes
144 pin LQFP
144 pin BGA
AT32UC3A0256
256 Kbytes
64 Kbytes
yes
yes
144 pin LQFP
144 pin BGA
AT32UC3A0128
128 Kbytes
32 Kbytes
yes
yes
144 pin LQFP
144 pin BGA
AT32UC3A1512
512 Kbytes
64 Kbytes
no
yes
100 pin TQFP
AT32UC3A1256
256 Kbytes
64 Kbytes
no
yes
100 pin TQFP
AT32UC3A1128
128 Kbytes
32 Kbytes
no
yes
100 pin TQFP
Package
3. Abbreviations
• GCLK: Power Manager Generic Clock
• GPIO: General Purpose Input/Output
• HSB: High Speed Bus
• MPU: Memory Protection Unit
• OCD: On Chip Debug
• PB: Peripheral Bus
• PDCA: Peripheral Direct Memory Access Controller (PDC) version A
• USBB: USB On-The-GO Controller version B
4
32058K
AVR32-01/12
AT32UC3A
4. Blockdiagram
C ON FIGU RATIO N
HSB-PB
BRIDG E B
EXTERNAL
INTERRUPT
CO NTRO LLER
REAL TIM E
CO UNTER
W ATCHDO G
TIM ER
115 kHz
RCO SC
XIN 32
XO UT32
XIN0
XO U T0
XIN1
XO U T1
32 KHz
O SC
O SC0
O SC1
PLL0
PLL1
RESET_N
G CLK[3..0]
A[2..0]
B[2..0]
CLK[2..0]
PO W ER
M ANAG ER
CLO CK
G ENERATO R
REG ISTER S BUS
H SB
PERIPHERAL
DM A
CO NTRO LLER
HSB-PB
BRIDG E A
PB
RESET
CO NTRO LLER
SD A10
SD CK
SDC KE
SD CS0
SD W E
USART1
R XD
TXD
CLK
R TS, C TS
D SR, DTR, D CD , RI
USART0
USART2
USART3
R XD
TXD
CLK
R TS, C TS
SERIAL
PERIPHERAL
INTERFACE 0/1
M ISO , M O SI
NPC S0
NPC S[3..1]
SYNCHRO NO US
SERIAL
CO NTRO LLER
TW O -W IRE
INTERFACE
PULSE W IDTH
M O DULATIO N
CO NTRO LLER
ANALO G TO
DIG ITAL
CO NVERTER
AUDIO
BITSTREAM
DAC
CLO CK
CO NTRO LLER
SLEEP
CO NTRO LLER
NC S[3..0]
N RD
NW AIT
N W E0
N W E1
N W E3
RAS
CAS
PDC
HS
B
D ATA[15..0]
ADD R[23..0]
PDC
PB
EXTIN T[7..0]
KPS[7..0]
N M I_N
M
S
S
512 KB
FLASH
TIM ER /CO UNTER
SC K
GENERAL PURPOSE IOs
S
ETHERNET
M AC
S
S
HIG H SPEED
BUS M ATRIX
INTERRUPT
CO NTRO LLER
PA
PB
PC
PX
M
PDC
M
M D IO
M
64 KB
SRAM
PDC
DM A
M
PDC
M DC ,
TXD [3..0],
TX_C LK,
TX_EN ,
TX_ER ,
SPEED
S
M
PDC
GENERAL PURPOSE IOs
CO L,
CR S,
R XD[3..0],
R X_CLK,
R X_DV,
RX_ER
DM A
PDC
USB
INTERFACE
ID
VBO F
DATA
INTERFACE
PBB
VBU S
D+
D-
M EM O RY PRO TEC TIO N U NIT
INSTR
INTERFACE
FAST G PIO
FLASH
CONTROLLER
NEXUS
CLASS 2+
O CD
M C KO
M D O[5..0]
M SEO[1..0]
EVTI_N
EVTO_N
UC CPU
LOC AL BU S
INTERFACE
EXTERNAL BUS INTERFACE
(SDRAM & STATIC MEMORY
CONTROLLER)
JTAG
INTERFACE
PDC
TC K
TDO
TD I
TM S
MEMORY INTERFACE
Blockdiagram
PBA
Figure 4-1.
PA
PB
PC
PX
TX _C LO C K, TX_FRA ME _SYN C
TX_D ATA
RX _C LO C K, R X_FRA ME _SYN C
R X_D ATA
SC L
SD A
PW M [6..0]
AD[7..0]
AD VREF
DATA[1..0]
D ATAN [1..0]
5
32058K
AVR32-01/12
AT32UC3A
4.1
4.1.1
Processor and architecture
AVR32 UC CPU
• 32-bit load/store AVR32A RISC architecture.
–
–
–
–
–
15 general-purpose 32-bit registers.
32-bit Stack Pointer, Program Counter and Link Register reside in register file.
Fully orthogonal instruction set.
Privileged and unprivileged modes enabling efficient and secure Operating Systems.
Innovative instruction set together with variable instruction length ensuring industry leading
code density.
– DSP extention with saturating arithmetic, and a wide variety of multiply instructions.
• 3 stage pipeline allows one instruction per clock cycle for most instructions.
– Byte, half-word, word and double word memory access.
– Multiple interrupt priority levels.
• MPU allows for operating systems with memory protection.
4.1.2
Debug and Test system
• IEEE1149.1 compliant JTAG and boundary scan
• Direct memory access and programming capabilities through JTAG interface
• Extensive On-Chip Debug features in compliance with IEEE-ISTO 5001-2003 (Nexus 2.0) Class 2+
•
•
•
•
4.1.3
– Low-cost NanoTrace supported.
Auxiliary port for high-speed trace information
Hardware support for 6 Program and 2 data breakpoints
Unlimited number of software breakpoints supported
Advanced Program, Data, Ownership, and Watchpoint trace supported
Peripheral DMA Controller
• Transfers from/to peripheral to/from any memory space without intervention of the processor.
• Next Pointer Support, forbids strong real-time constraints on buffer management.
• Fifteen channels
–
–
–
–
–
4.1.4
Two for each USART
Two for each Serial Synchronous Controller
Two for each Serial Peripheral Interface
One for each ADC
Two for each TWI Interface
Bus system
• High Speed Bus (HSB) matrix with 6 Masters and 6 Slaves handled
– Handles Requests from the CPU Data Fetch, CPU Instruction Fetch, PDCA, USBB, Ethernet
Controller, CPU SAB, and to internal Flash, internal SRAM, Peripheral Bus A, Peripheral Bus
B, EBI.
– Round-Robin Arbitration (three modes supported: no default master, last accessed default
master, fixed default master)
– Burst Breaking with Slot Cycle Limit
– One Address Decoder Provided per Master
6
32058K
AVR32-01/12
AT32UC3A
• Peripheral Bus A able to run on at divided bus speeds compared to the High Speed Bus
Figure 4-1 gives an overview of the bus system. All modules connected to the same bus use the
same clock, but the clock to each module can be individually shut off by the Power Manager.
The figure identifies the number of master and slave interfaces of each module connected to the
High Speed Bus, and which DMA controller is connected to which peripheral.
7
32058K
AVR32-01/12
AT32UC3A
5. Signals Description
The following table gives details on the signal name classified by peripheral
The signals are multiplexed with GPIO pins as described in ”Peripheral Multiplexing on I/O lines”
on page 45.
Table 5-1.
Signal Description List
Signal Name
Function
Type
Active
Level
Comments
Power
VDDPLL
Power supply for PLL
Power
Input
1.65V to 1.95 V
VDDCORE
Core Power Supply
Power
Input
1.65V to 1.95 V
VDDIO
I/O Power Supply
Power
Input
3.0V to 3.6V
VDDANA
Analog Power Supply
Power
Input
3.0V to 3.6V
VDDIN
Voltage Regulator Input Supply
Power
Input
3.0V to 3.6V
VDDOUT
Voltage Regulator Output
Power
Output
1.65V to 1.95 V
GNDANA
Analog Ground
Ground
GND
Ground
Ground
Clocks, Oscillators, and PLL’s
XIN0, XIN1, XIN32
Crystal 0, 1, 32 Input
Analog
XOUT0, XOUT1,
XOUT32
Crystal 0, 1, 32 Output
Analog
JTAG
TCK
Test Clock
Input
TDI
Test Data In
Input
TDO
Test Data Out
TMS
Test Mode Select
Output
Input
Auxiliary Port - AUX
MCKO
Trace Data Output Clock
Output
MDO0 - MDO5
Trace Data Output
Output
8
32058K
AVR32-01/12
AT32UC3A
Table 5-1.
Signal Description List
Type
Active
Level
Signal Name
Function
MSEO0 - MSEO1
Trace Frame Control
Output
EVTI_N
Event In
Output
Low
EVTO_N
Event Out
Output
Low
Comments
Power Manager - PM
GCLK0 - GCLK3
Generic Clock Pins
RESET_N
Reset Pin
Output
Input
Low
Real Time Counter - RTC
RTC_CLOCK
RTC clock
Output
Watchdog Timer - WDT
WDTEXT
External Watchdog Pin
Output
External Interrupt Controller - EIC
EXTINT0 - EXTINT7
External Interrupt Pins
KPS0 - KPS7
Keypad Scan Pins
NMI_N
Non-Maskable Interrupt Pin
Input
Output
Input
Low
Ethernet MAC - MACB
COL
Collision Detect
Input
CRS
Carrier Sense and Data Valid
Input
MDC
Management Data Clock
MDIO
Management Data Input/Output
RXD0 - RXD3
Receive Data
Input
RX_CLK
Receive Clock
Input
RX_DV
Receive Data Valid
Input
RX_ER
Receive Coding Error
Input
SPEED
Speed
TXD0 - TXD3
Transmit Data
Output
TX_CLK
Transmit Clock or Reference Clock
Output
TX_EN
Transmit Enable
Output
TX_ER
Transmit Coding Error
Output
Output
I/O
9
32058K
AVR32-01/12
AT32UC3A
Table 5-1.
Signal Description List
Signal Name
Function
Type
Active
Level
Comments
External Bus Interface - HEBI
ADDR0 - ADDR23
Address Bus
Output
CAS
Column Signal
Output
DATA0 - DATA15
Data Bus
NCS0 - NCS3
Chip Select
Output
Low
NRD
Read Signal
Output
Low
NWAIT
External Wait Signal
Input
Low
NWE0
Write Enable 0
Output
Low
NWE1
Write Enable 1
Output
Low
NWE3
Write Enable 3
Output
Low
RAS
Row Signal
Output
Low
SDA10
SDRAM Address 10 Line
Output
SDCK
SDRAM Clock
Output
SDCKE
SDRAM Clock Enable
Output
SDCS0
SDRAM Chip Select
Output
Low
SDWE
SDRAM Write Enable
Output
Low
Low
I/O
General Purpose Input/Output 2 - GPIOA, GPIOB, GPIOC
P0 - P31
Parallel I/O Controller GPIOA
I/O
P0 - P31
Parallel I/O Controller GPIOB
I/O
P0 - P5
Parallel I/O Controller GPIOC
I/O
P0 - P31
Parallel I/O Controller GPIOX
I/O
Serial Peripheral Interface - SPI0, SPI1
MISO
Master In Slave Out
I/O
MOSI
Master Out Slave In
I/O
NPCS0 - NPCS3
SPI Peripheral Chip Select
I/O
SCK
Clock
Low
Output
Synchronous Serial Controller - SSC
RX_CLOCK
SSC Receive Clock
I/O
10
32058K
AVR32-01/12
AT32UC3A
Signal Description List
Table 5-1.
Signal Name
Function
Type
RX_DATA
SSC Receive Data
Input
RX_FRAME_SYNC
SSC Receive Frame Sync
I/O
TX_CLOCK
SSC Transmit Clock
I/O
TX_DATA
SSC Transmit Data
Output
TX_FRAME_SYNC
SSC Transmit Frame Sync
Active
Level
Comments
I/O
Timer/Counter - TIMER
A0
Channel 0 Line A
I/O
A1
Channel 1 Line A
I/O
A2
Channel 2 Line A
I/O
B0
Channel 0 Line B
I/O
B1
Channel 1 Line B
I/O
B2
Channel 2 Line B
I/O
CLK0
Channel 0 External Clock Input
Input
CLK1
Channel 1 External Clock Input
Input
CLK2
Channel 2 External Clock Input
Input
Two-wire Interface - TWI
SCL
Serial Clock
I/O
SDA
Serial Data
I/O
Universal Synchronous Asynchronous Receiver Transmitter - USART0, USART1, USART2, USART3
CLK
Clock
I/O
CTS
Clear To Send
DCD
Data Carrier Detect
Only USART1
DSR
Data Set Ready
Only USART1
DTR
Data Terminal Ready
Only USART1
RI
Ring Indicator
Only USART1
RTS
Request To Send
RXD
Receive Data
Input
TXD
Transmit Data
Output
Input
Output
11
32058K
AVR32-01/12
AT32UC3A
Table 5-1.
Signal Description List
Signal Name
Function
Type
Active
Level
Comments
Analog to Digital Converter - ADC
AD0 - AD7
Analog input pins
Analog
input
ADVREF
Analog positive reference voltage input
Analog
input
2.6 to 3.6V
Pulse Width Modulator - PWM
PWM0 - PWM6
PWM Output Pins
Output
Universal Serial Bus Device - USB
DDM
USB Device Port Data -
Analog
DDP
USB Device Port Data +
Analog
VBUS
USB VBUS Monitor and OTG Negociation
Analog
Input
USBID
ID Pin of the USB Bus
Input
USB_VBOF
USB VBUS On/off: bus power control port
output
Audio Bitstream DAC (ABDAC)
DATA0-DATA1
D/A Data out
Outpu
DATAN0-DATAN1
D/A Data inverted out
Outpu
12
32058K
AVR32-01/12
AT32UC3A
6. Power Considerations
6.1
Power Supplies
The AT32UC3A has several types of power supply pins:
•
•
•
•
•
VDDIO: Powers I/O lines. Voltage is 3.3V nominal.
VDDANA: Powers the ADC Voltage is 3.3V nominal.
VDDIN: Input voltage for the voltage regulator. Voltage is 3.3V nominal.
VDDCORE: Powers the core, memories, and peripherals. Voltage is 1.8V nominal.
VDDPLL: Powers the PLL. Voltage is 1.8V nominal.
The ground pins GND are common to VDDCORE, VDDIO, VDDPLL. The ground pin for
VDDANA is GNDANA.
Refer to ”Power Consumption” on page 767 for power consumption on the various supply pins.
Dual Power Supply
Single Power Supply
3.3V
3.3V
VDDANA
VDDIO
VDDIO
ADVREF
ADVREF
VDDIN
VDDIN
1.8V
Regulator
VDDPLL
1.8V
Regulator
VDDOUT
VDDOUT
VDDCORE
VDDANA
1.8V
VDDCORE
VDDPLL
13
32058K
AVR32-01/12
AT32UC3A
6.2
6.2.1
Voltage Regulator
Single Power Supply
The AT32UC3A embeds a voltage regulator that converts from 3.3V to 1.8V. The regulator takes
its input voltage from VDDIN, and supplies the output voltage on VDDOUT. VDDOUT should be
externally connected to the 1.8V domains.
Adequate input supply decoupling is mandatory for VDDIN in order to improve startup stability
and reduce source voltage drop. Two input decoupling capacitors must be placed close to the
chip.
Adequate output supply decoupling is mandatory for VDDOUT to reduce ripple and avoid oscillations. The best way to achieve this is to use two capacitors in parallel between VDDOUT and
GND as close to the chip as possible
3.3V
VDDIN
CIN2
1.8V
Regulator
CIN1
1.8V
VDDOUT
COUT2
COUT1
Refer to Section 38.3 on page 765 for decoupling capacitors values and regulator characteristics
6.2.2
Dual Power Supply
In case of dual power supply, VDDIN and VDDOUT should be connected to ground to prevent
from leakage current.
VDDIN
VDDOUT
14
32058K
AVR32-01/12
AT32UC3A
6.3
Analog-to-Digital Converter (A.D.C) reference.
The ADC reference (ADVREF) must be provided from an external source. Two decoupling
capacitors must be used to insure proper decoupling.
3.3V
ADVREF
C
VREF2
C
VREF1
Refer to Section 38.4 on page 765 for decoupling capacitors values and electrical
characteristics.
In case ADC is not used, the ADVREF pin should be connected to GND to avoid extra
consumption.
15
32058K
AVR32-01/12
AT32UC3A
7. Package and Pinout
The device pins are multiplexed with peripheral functions as described in ”Peripheral Multiplexing on I/O lines” on page 45.
TQFP100 Pinout
Figure 7-1.
75
51
76
50
100
26
1
25
TQFP100 Package Pinout
Table 7-1.
1
PB20
26
PA05
51
PA21
76
PB08
2
PB21
27
PA06
52
PA22
77
PB09
3
PB22
28
PA07
53
PA23
78
PB10
4
VDDIO
29
PA08
54
PA24
79
VDDIO
5
GND
30
PA09
55
PA25
80
GND
6
PB23
31
PA10
56
PA26
81
PB11
7
PB24
32
N/C
57
PA27
82
PB12
8
PB25
33
PA11
58
PA28
83
PA29
9
PB26
34
VDDCORE
59
VDDANA
84
PA30
10
PB27
35
GND
60
ADVREF
85
PC02
11
VDDOUT
36
PA12
61
GNDANA
86
PC03
12
VDDIN
37
PA13
62
VDDPLL
87
PB13
13
GND
38
VDDCORE
63
PC00
88
PB14
14
PB28
39
PA14
64
PC01
89
TMS
15
PB29
40
PA15
65
PB00
90
TCK
16
PB30
41
PA16
66
PB01
91
TDO
17
PB31
42
PA17
67
VDDIO
92
TDI
18
RESET_N
43
PA18
68
VDDIO
93
PC04
19
PA00
44
PA19
69
GND
94
PC05
20
PA01
45
PA20
70
PB02
95
PB15
21
GND
46
VBUS
71
PB03
96
PB16
22
VDDCORE
47
VDDIO
72
PB04
97
VDDCORE
16
32058K
AVR32-01/12
AT32UC3A
TQFP100 Package Pinout
Table 7-1.
23
PA02
48
DM
73
PB05
98
PB17
24
PA03
49
DP
74
PB06
99
PB18
25
PA04
50
GND
75
PB07
100
PB19
LQFP144 Pinout
Figure 7-2.
108
73
109
72
144
37
1
36
VQFP144 Package Pinout
Table 7-2.
1
PX00
37
GND
73
PA21
109
GND
2
PX01
38
PX10
74
PA22
110
PX30
3
PB20
39
PA05
75
PA23
111
PB08
4
PX02
40
PX11
76
PA24
112
PX31
5
PB21
41
PA06
77
PA25
113
PB09
6
PB22
42
PX12
78
PA26
114
PX32
7
VDDIO
43
PA07
79
PA27
115
PB10
8
GND
44
PX13
80
PA28
116
VDDIO
9
PB23
45
PA08
81
VDDANA
117
GND
10
PX03
46
PX14
82
ADVREF
118
PX33
11
PB24
47
PA09
83
GNDANA
119
PB11
12
PX04
48
PA10
84
VDDPLL
120
PX34
13
PB25
49
N/C
85
PC00
121
PB12
14
PB26
50
PA11
86
PC01
122
PA29
15
PB27
51
VDDCORE
87
PX20
123
PA30
16
VDDOUT
52
GND
88
PB00
124
PC02
17
VDDIN
53
PA12
89
PX21
125
PC03
18
GND
54
PA13
90
PB01
126
PB13
19
PB28
55
VDDCORE
91
PX22
127
PB14
20
PB29
56
PA14
92
VDDIO
128
TMS
21
PB30
57
PA15
93
VDDIO
129
TCK
17
32058K
AVR32-01/12
AT32UC3A
VQFP144 Package Pinout
Table 7-2.
22
PB31
58
PA16
94
GND
130
TDO
23
RESET_N
59
PX15
95
PX23
131
TDI
24
PX05
60
PA17
96
PB02
132
PC04
25
PA00
61
PX16
97
PX24
133
PC05
26
PX06
62
PA18
98
PB03
134
PB15
27
PA01
63
PX17
99
PX25
135
PX35
28
GND
64
PA19
100
PB04
136
PB16
29
VDDCORE
65
PX18
101
PX26
137
PX36
30
PA02
66
PA20
102
PB05
138
VDDCORE
31
PX07
67
PX19
103
PX27
139
PB17
32
PA03
68
VBUS
104
PB06
140
PX37
33
PX08
69
VDDIO
105
PX28
141
PB18
34
PA04
70
DM
106
PB07
142
PX38
35
PX09
71
DP
107
PX29
143
PB19
36
VDDIO
72
GND
108
VDDIO
144
PX39
Figure 7-3.
BGA144 Pinout
18
32058K
AVR32-01/12
AT32UC3A
Table 7-3.
BGA144 Package Pinout A1..M8
1
2
3
4
5
6
7
8
A
VDDIO
PB07
PB05
PB02
PB03
PB01
PC00
PA28
B
PB08
GND
PB06
PB04
VDDIO
PB00
PC01
VDDPLL
C
PB09
PX33
PA29
PC02
PX28
PX26
PX22
PX21
D
PB11
PB13
PB12
PX30
PX29
PX25
PX24
PX20
E
PB10
VDDIO
PX32
PX31
VDDIO
PX27
PX23
VDDANA
F
PA30
PB14
PX34
PB16
TCK
GND
GND
PX16
G
TMS
PC03
PX36
PX35
PX37
GND
GND
PA16
H
TDO
VDDCORE
PX38
PX39
VDDIO
PA01
PA10
VDDCORE
J
TDI
PB17
PB15
PX00
PX01
PA00
PA03
PA04
K
PC05
PC04
PB19
PB20
PX02
PB29
PB30
PA02
L
PB21
GND
PB18
PB24
VDDOUT
PX04
PB31
VDDIN
M
PB22
PB23
PB25
PB26
PX03
PB27
PB28
RESET_N
Table 7-4.
BGA144 Package Pinout A9..M12
9
10
11
12
A
PA26
PA25
PA24
PA23
B
PA27
PA21
GND
PA22
C
ADVREF
GNDANA
PX19
PA19
D
PA18
PA20
DP
DM
E
PX18
PX17
VDDIO
VBUS
F
PA17
PX15
PA15
PA14
G
PA13
PA12
PA11
NC
H
PX11
PA08
VDDCORE
VDDCORE
J
PX14
PA07
PX13
PA09
K
PX08
GND
PA05
PX12
L
PX06
PX10
GND
PA06
M
PX05
PX07
PX09
VDDIO
Note:
NC is not connected.
19
32058K
AVR32-01/12
AT32UC3A
8. I/O Line Considerations
8.1
JTAG pins
TMS, TDI and TCK have pull-up resistors. TDO is an output, driven at up to VDDIO, and has no
pull-up resistor.
8.2
RESET_N pin
The RESET_N pin is a schmitt input and integrates a permanent pull-up resistor to VDDIO. As
the product integrates a power-on reset cell, the RESET_N pin can be left unconnected in case
no reset from the system needs to be applied to the product.
8.3
TWI pins
When these pins are used for TWI, the pins are open-drain outputs with slew-rate limitation and
inputs with inputs with spike-filtering. When used as GPIO-pins or used for other peripherals, the
pins have the same characteristics as PIO pins.
8.4
GPIO pins
All the I/O lines integrate a programmable pull-up resistor. Programming of this pull-up resistor is
performed independently for each I/O line through the GPIO Controllers. After reset, I/O lines
default as inputs with pull-up resistors disabled, except when indicated otherwise in the column
“Reset State” of the GPIO Controller multiplexing tables.
20
32058K
AVR32-01/12
AT32UC3A
9. Processor and Architecture
This chapter gives an overview of the AVR32UC CPU. AVR32UC is an implementation of the
AVR32 architecture. A summary of the programming model, instruction set and MPU is presented. For further details, see the AVR32 Architecture Manual and the AVR32UC Technical
Reference Manual.
9.1
AVR32 Architecture
AVR32 is a new, high-performance 32-bit RISC microprocessor architecture, designed for costsensitive embedded applications, with particular emphasis on low power consumption and high
code density. In addition, the instruction set architecture has been tuned to allow a variety of
microarchitectures, enabling the AVR32 to be implemented as low-, mid- or high-performance
processors. AVR32 extends the AVR family into the world of 32- and 64-bit applications.
Through a quantitative approach, a large set of industry recognized benchmarks has been compiled and analyzed to achieve the best code density in its class. In addition to lowering the
memory requirements, a compact code size also contributes to the core’s low power characteristics. The processor supports byte and half-word data types without penalty in code size and
performance.
Memory load and store operations are provided for byte, half-word, word and double word data
with automatic sign- or zero extension of half-word and byte data. The C-compiler is closely
linked to the architecture and is able to exploit code optimization features, both for size and
speed.
In order to reduce code size to a minimum, some instructions have multiple addressing modes.
As an example, instructions with immediates often have a compact format with a smaller immediate, and an extended format with a larger immediate. In this way, the compiler is able to use
the format giving the smallest code size.
Another feature of the instruction set is that frequently used instructions, like add, have a compact format with two operands as well as an extended format with three operands. The larger
format increases performance, allowing an addition and a data move in the same instruction in a
single cycle. Load and store instructions have several different formats in order to reduce code
size and speed up execution.
The register file is organized as sixteen 32-bit registers and includes the Program Counter, the
Link Register, and the Stack Pointer. In addition, register R12 is designed to hold return values
from function calls and is used implicitly by some instructions.
9.2
The AVR32UC CPU
The AVR32 UC CPU targets low- and medium-performance applications, and provides an
advanced OCD system, no caches, and a Memory Protection Unit (MPU). Java acceleration
hardware is not implemented.
AVR32 UC provides three memory interfaces, one High Speed Bus master for instruction fetch,
one High Speed Bus master for data access, and one High Speed Bus slave interface allowing
other bus masters to access data RAMs internal to the CPU. Keeping data RAMs internal to the
CPU allows fast access to the RAMs, reduces latency and guarantees deterministic timing. Also,
power consumption is reduced by not needing a full High Speed Bus access for memory
accesses. A dedicated data RAM interface is provided for communicating with the internal data
RAMs.
21
32058K
AVR32-01/12
AT32UC3A
A local bus interface is provided for connecting the CPU to device-specific high-speed systems,
such as floating-point units and fast GPIO ports. This local bus has to be enabled by writing the
LOCEN bit in the CPUCR system register. The local bus is able to transfer data between the
CPU and the local bus slave in a single clock cycle. The local bus has a dedicated memory
range allocated to it, and data transfers are performed using regular load and store instructions.
Details on which devices that are mapped into the local bus space is given in the device-specific
“Peripherals” chapter of this data sheet.
Figure 9-1 on page 22 displays the contents of AVR32UC.
OCD interface
Reset interface
Overview of the AVR32UC CPU
Interrupt controller interface
Figure 9-1.
OCD
system
Power/
Reset
control
AVR32UC CPU pipeline
MPU
9.2.1
CPU Local
Bus
master
Data RAM interface
High
Speed
Bus slave
CPU Local Bus
High Speed Bus
High Speed Bus
High Speed Bus master
High
Speed
Bus
master
High Speed Bus
Data memory controller
Instruction memory controller
Pipeline Overview
AVR32 UC is a pipelined processor with three pipeline stages. There are three pipeline stages,
Instruction Fetch (IF), Instruction Decode (ID) and Instruction Execute (EX). The EX stage is
split into three parallel subsections, one arithmetic/logic (ALU) section, one multiply (MUL) section and one load/store (LS) section.
Instructions are issued and complete in order. Certain operations require several clock cycles to
complete, and in this case, the instruction resides in the ID and EX stages for the required number of clock cycles. Since there is only three pipeline stages, no internal data forwarding is
required, and no data dependencies can arise in the pipeline.
Figure 9-2 on page 23 shows an overview of the AVR32 UC pipeline stages.
22
32058K
AVR32-01/12
AT32UC3A
The AVR32UC Pipeline
Figure 9-2.
MUL
IF
ID
Pref etch unit
Decode unit
Regf ile
Read
A LU
LS
9.2.2
Multiply unit
Regf ile
w rite
A LU unit
Load-store
unit
AVR32A Microarchitecture Compliance
AVR32UC implements an AVR32A microarchitecture. The AVR32A microarchitecture is targeted at cost-sensitive, lower-end applications like smaller microcontrollers. This
microarchitecture does not provide dedicated hardware registers for shadowing of register file
registers in interrupt contexts. Additionally, it does not provide hardware registers for the return
address registers and return status registers. Instead, all this information is stored on the system
stack. This saves chip area at the expense of slower interrupt handling.
Upon interrupt initiation, registers R8-R12 are automatically pushed to the system stack. These
registers are pushed regardless of the priority level of the pending interrupt. The return address
and status register are also automatically pushed to stack. The interrupt handler can therefore
use R8-R12 freely. Upon interrupt completion, the old R8-R12 registers and status register are
restored, and execution continues at the return address stored popped from stack.
The stack is also used to store the status register and return address for exceptions and scall.
Executing the rete or rets instruction at the completion of an exception or system call will pop
this status register and continue execution at the popped return address.
9.2.3
Java Support
AVR32UC does not provide Java hardware acceleration.
9.2.4
Memory protection
The MPU allows the user to check all memory accesses for privilege violations. If an access is
attempted to an illegal memory address, the access is aborted and an exception is taken. The
MPU in AVR32UC is specified in the AVR32UC Technical Reference manual.
9.2.5
Unaligned reference handling
AVR32UC does not support unaligned accesses, except for doubleword accesses. AVR32UC is
able to perform word-aligned st.d and ld.d. Any other unaligned memory access will cause an
address exception. Doubleword-sized accesses with word-aligned pointers will automatically be
performed as two word-sized accesses.
23
32058K
AVR32-01/12
AT32UC3A
The following table shows the instructions with support for unaligned addresses. All other
instructions require aligned addresses.
Table 9-1.
9.2.6
Instructions with unaligned reference support
Instruction
Supported alignment
ld.d
Word
st.d
Word
Unimplemented instructions
The following instructions are unimplemented in AVR32UC, and will cause an Unimplemented
Instruction Exception if executed:
• All SIMD instructions
• All coprocessor instructions
• retj, incjosp, popjc, pushjc
• tlbr, tlbs, tlbw
• cache
9.2.7
CPU and Architecture revision
Two major revisions of the AVR32UC CPU currently exist. The device described in this
datasheet uses CPU revision 2.
The Architecture Revision field in the CONFIG0 system register identifies which architecture
revision is implemented in a specific device.
AVR32UC CPU revision 2 is fully backward-compatible with revision 1, ie. code compiled for
revision 1 is binary-compatible with revision 2 CPUs.
24
32058K
AVR32-01/12
AT32UC3A
9.3
9.3.1
Programming Model
Register file configuration
The AVR32UC register file is shown below.
The AVR32UC Register File
Figure 9-3.
9.3.2
Application
Supe rv isor
INT0
Bit 31
Bit 31
Bit 31
Bit 0
Bit 0
INT1
Bit 0
INT2
Bit 31
Bit 0
INT3
Bit 31
Bit 0
Bit 31
Bit 0
Exce ption
NMI
Bit 31
Bit 31
Bit 0
Bit 0
PC
LR
SP_APP
R12
PC
LR
SP_SYS
R12
PC
LR
SP_SYS
R12
PC
LR
SP_SYS
R12
PC
LR
SP_SYS
R12
PC
LR
SP_SYS
R12
PC
LR
SP_SYS
R12
PC
LR
SP_SYS
R12
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SM
R4PC
R3
R2
R1
R0
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SM
R4PC
R3
R2
R1
R0
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SM
R4PC
R3
R2
R1
R0
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SM
R4PC
R3
R2
R1
R0
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SM
R4PC
R3
R2
R1
R0
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SM
R4PC
R3
R2
R1
R0
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SM
R4PC
R3
R2
R1
R0
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SM
R4PC
R3
R2
R1
R0
SR
SR
SR
SR
SR
SR
SR
SR
Status register configuration
The Status Register (SR) is split into two halfwords, one upper and one lower, see Figure 9-4 on
page 25 and Figure 9-5 on page 26. The lower word contains the C, Z, N, V and Q condition
code flags and the R, T and L bits, while the upper halfword contains information about the
mode and state the processor executes in. Refer to the AVR32 Architecture Manual for details.
The Status Register High Halfword
Figure 9-4.
Bit 31
Bit 16
-
LC
1
-
-
DM
D
-
M2
M1
M0
EM
I3M
I2M
FE
I1M
I0M
GM
0
0
0
0
0
0
0
0
0
1
1
0
0
0
0
1
Bit name
Initial value
Global Interrupt Mask
Interrupt Level 0 Mask
Interrupt Level 1 Mask
Interrupt Level 2 Mask
Interrupt Level 3 Mask
Exception Mask
Mode Bit 0
Mode Bit 1
Mode Bit 2
Reserved
Debug State
Debug State Mask
Reserved
25
32058K
AVR32-01/12
AT32UC3A
Figure 9-5.
The Status Register Low Halfword
Bit 15
Bit 0
R
T
-
-
-
-
-
-
-
-
L
Q
V
N
Z
C
Bit name
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
0
Initial value
Carry
Zero
Sign
Overflow
Saturation
Lock
Reserved
Scratch
Register Remap Enable
9.3.3
Processor States
9.3.3.1
Normal RISC State
The AVR32 processor supports several different execution contexts as shown in Table 9-2 on
page 26.
Table 9-2.
Overview of execution modes, their priorities and privilege levels.
Priority
Mode
Security
Description
1
Non Maskable Interrupt
Privileged
Non Maskable high priority interrupt mode
2
Exception
Privileged
Execute exceptions
3
Interrupt 3
Privileged
General purpose interrupt mode
4
Interrupt 2
Privileged
General purpose interrupt mode
5
Interrupt 1
Privileged
General purpose interrupt mode
6
Interrupt 0
Privileged
General purpose interrupt mode
N/A
Supervisor
Privileged
Runs supervisor calls
N/A
Application
Unprivileged
Normal program execution mode
Mode changes can be made under software control, or can be caused by external interrupts or
exception processing. A mode can be interrupted by a higher priority mode, but never by one
with lower priority. Nested exceptions can be supported with a minimal software overhead.
When running an operating system on the AVR32, user processes will typically execute in the
application mode. The programs executed in this mode are restricted from executing certain
instructions. Furthermore, most system registers together with the upper halfword of the status
register cannot be accessed. Protected memory areas are also not available. All other operating
modes are privileged and are collectively called System Modes. They have full access to all privileged and unprivileged resources. After a reset, the processor will be in supervisor mode.
9.3.3.2
Debug State
The AVR32 can be set in a debug state, which allows implementation of software monitor routines that can read out and alter system information for use during application development. This
implies that all system and application registers, including the status registers and program
counters, are accessible in debug state. The privileged instructions are also available.
26
32058K
AVR32-01/12
AT32UC3A
All interrupt levels are by default disabled when debug state is entered, but they can individually
be switched on by the monitor routine by clearing the respective mask bit in the status register.
Debug state can be entered as described in the AVR32UC Technical Reference Manual.
Debug state is exited by the retd instruction.
9.3.4
System registers
The system registers are placed outside of the virtual memory space, and are only accessible
using the privileged mfsr and mtsr instructions. The table below lists the system registers specified in the AVR32 architecture, some of which are unused in AVR32UC. The programmer is
responsible for maintaining correct sequencing of any instructions following a mtsr instruction.
For detail on the system registers, refer to the AVR32UC Technical Reference Manual.
Table 9-3.
System Registers
Reg #
Address
Name
Function
0
0
SR
Status Register
1
4
EVBA
Exception Vector Base Address
2
8
ACBA
Application Call Base Address
3
12
CPUCR
CPU Control Register
4
16
ECR
Exception Cause Register
5
20
RSR_SUP
Unused in AVR32UC
6
24
RSR_INT0
Unused in AVR32UC
7
28
RSR_INT1
Unused in AVR32UC
8
32
RSR_INT2
Unused in AVR32UC
9
36
RSR_INT3
Unused in AVR32UC
10
40
RSR_EX
Unused in AVR32UC
11
44
RSR_NMI
Unused in AVR32UC
12
48
RSR_DBG
Return Status Register for Debug Mode
13
52
RAR_SUP
Unused in AVR32UC
14
56
RAR_INT0
Unused in AVR32UC
15
60
RAR_INT1
Unused in AVR32UC
16
64
RAR_INT2
Unused in AVR32UC
17
68
RAR_INT3
Unused in AVR32UC
18
72
RAR_EX
Unused in AVR32UC
19
76
RAR_NMI
Unused in AVR32UC
20
80
RAR_DBG
Return Address Register for Debug Mode
21
84
JECR
Unused in AVR32UC
22
88
JOSP
Unused in AVR32UC
23
92
JAVA_LV0
Unused in AVR32UC
24
96
JAVA_LV1
Unused in AVR32UC
25
100
JAVA_LV2
Unused in AVR32UC
27
32058K
AVR32-01/12
AT32UC3A
Table 9-3.
System Registers (Continued)
Reg #
Address
Name
Function
26
104
JAVA_LV3
Unused in AVR32UC
27
108
JAVA_LV4
Unused in AVR32UC
28
112
JAVA_LV5
Unused in AVR32UC
29
116
JAVA_LV6
Unused in AVR32UC
30
120
JAVA_LV7
Unused in AVR32UC
31
124
JTBA
Unused in AVR32UC
32
128
JBCR
Unused in AVR32UC
33-63
132-252
Reserved
Reserved for future use
64
256
CONFIG0
Configuration register 0
65
260
CONFIG1
Configuration register 1
66
264
COUNT
Cycle Counter register
67
268
COMPARE
Compare register
68
272
TLBEHI
Unused in AVR32UC
69
276
TLBELO
Unused in AVR32UC
70
280
PTBR
Unused in AVR32UC
71
284
TLBEAR
Unused in AVR32UC
72
288
MMUCR
Unused in AVR32UC
73
292
TLBARLO
Unused in AVR32UC
74
296
TLBARHI
Unused in AVR32UC
75
300
PCCNT
Unused in AVR32UC
76
304
PCNT0
Unused in AVR32UC
77
308
PCNT1
Unused in AVR32UC
78
312
PCCR
Unused in AVR32UC
79
316
BEAR
Bus Error Address Register
80
320
MPUAR0
MPU Address Register region 0
81
324
MPUAR1
MPU Address Register region 1
82
328
MPUAR2
MPU Address Register region 2
83
332
MPUAR3
MPU Address Register region 3
84
336
MPUAR4
MPU Address Register region 4
85
340
MPUAR5
MPU Address Register region 5
86
344
MPUAR6
MPU Address Register region 6
87
348
MPUAR7
MPU Address Register region 7
88
352
MPUPSR0
MPU Privilege Select Register region 0
89
356
MPUPSR1
MPU Privilege Select Register region 1
90
360
MPUPSR2
MPU Privilege Select Register region 2
91
364
MPUPSR3
MPU Privilege Select Register region 3
28
32058K
AVR32-01/12
AT32UC3A
Table 9-3.
9.4
System Registers (Continued)
Reg #
Address
Name
Function
92
368
MPUPSR4
MPU Privilege Select Register region 4
93
372
MPUPSR5
MPU Privilege Select Register region 5
94
376
MPUPSR6
MPU Privilege Select Register region 6
95
380
MPUPSR7
MPU Privilege Select Register region 7
96
384
MPUCRA
Unused in this version of AVR32UC
97
388
MPUCRB
Unused in this version of AVR32UC
98
392
MPUBRA
Unused in this version of AVR32UC
99
396
MPUBRB
Unused in this version of AVR32UC
100
400
MPUAPRA
MPU Access Permission Register A
101
404
MPUAPRB
MPU Access Permission Register B
102
408
MPUCR
MPU Control Register
103-191
412-764
Reserved
Reserved for future use
192-255
768-1020
IMPL
IMPLEMENTATION DEFINED
Exceptions and Interrupts
AVR32UC incorporates a powerful exception handling scheme. The different exception sources,
like Illegal Op-code and external interrupt requests, have different priority levels, ensuring a welldefined behavior when multiple exceptions are received simultaneously. Additionally, pending
exceptions of a higher priority class may preempt handling of ongoing exceptions of a lower priority class.
When an event occurs, the execution of the instruction stream is halted, and execution control is
passed to an event handler at an address specified in Table 9-4 on page 32. Most of the handlers are placed sequentially in the code space starting at the address specified by EVBA, with
four bytes between each handler. This gives ample space for a jump instruction to be placed
there, jumping to the event routine itself. A few critical handlers have larger spacing between
them, allowing the entire event routine to be placed directly at the address specified by the
EVBA-relative offset generated by hardware. All external interrupt sources have autovectored
interrupt service routine (ISR) addresses. This allows the interrupt controller to directly specify
the ISR address as an address relative to EVBA. The autovector offset has 14 address bits, giving an offset of maximum 16384 bytes. The target address of the event handler is calculated as
(EVBA | event_handler_offset), not (EVBA + event_handler_offset), so EVBA and exception
code segments must be set up appropriately. The same mechanisms are used to service all different types of events, including external interrupt requests, yielding a uniform event handling
scheme.
An interrupt controller does the priority handling of the external interrupts and provides the
autovector offset to the CPU.
9.4.1
System stack issues
Event handling in AVR32 UC uses the system stack pointed to by the system stack pointer,
SP_SYS, for pushing and popping R8-R12, LR, status register and return address. Since event
code may be timing-critical, SP_SYS should point to memory addresses in the IRAM section,
since the timing of accesses to this memory section is both fast and deterministic.
29
32058K
AVR32-01/12
AT32UC3A
The user must also make sure that the system stack is large enough so that any event is able to
push the required registers to stack. If the system stack is full, and an event occurs, the system
will enter an UNDEFINED state.
9.4.2
Exceptions and interrupt requests
When an event other than scall or debug request is received by the core, the following actions
are performed atomically:
1. The pending event will not be accepted if it is masked. The I3M, I2M, I1M, I0M, EM and
GM bits in the Status Register are used to mask different events. Not all events can be
masked. A few critical events (NMI, Unrecoverable Exception, TLB Multiple Hit and Bus
Error) can not be masked. When an event is accepted, hardware automatically sets the
mask bits corresponding to all sources with equal or lower priority. This inhibits acceptance of other events of the same or lower priority, except for the critical events listed
above. Software may choose to clear some or all of these bits after saving the necessary state if other priority schemes are desired. It is the event source’s responsability to
ensure that their events are left pending until accepted by the CPU.
2. When a request is accepted, the Status Register and Program Counter of the current
context is stored to the system stack. If the event is an INT0, INT1, INT2 or INT3, registers R8-R12 and LR are also automatically stored to stack. Storing the Status Register
ensures that the core is returned to the previous execution mode when the current
event handling is completed. When exceptions occur, both the EM and GM bits are set,
and the application may manually enable nested exceptions if desired by clearing the
appropriate bit. Each exception handler has a dedicated handler address, and this
address uniquely identifies the exception source.
3. The Mode bits are set to reflect the priority of the accepted event, and the correct register file bank is selected. The address of the event handler, as shown in Table 9-4, is
loaded into the Program Counter.
The execution of the event handler routine then continues from the effective address calculated.
The rete instruction signals the end of the event. When encountered, the Return Status Register
and Return Address Register are popped from the system stack and restored to the Status Register and Program Counter. If the rete instruction returns from INT0, INT1, INT2 or INT3,
registers R8-R12 and LR are also popped from the system stack. The restored Status Register
contains information allowing the core to resume operation in the previous execution mode. This
concludes the event handling.
9.4.3
Supervisor calls
The AVR32 instruction set provides a supervisor mode call instruction. The scall instruction is
designed so that privileged routines can be called from any context. This facilitates sharing of
code between different execution modes. The scall mechanism is designed so that a minimal
execution cycle overhead is experienced when performing supervisor routine calls from timecritical event handlers.
The scall instruction behaves differently depending on which mode it is called from. The behaviour is detailed in the instruction set reference. In order to allow the scall routine to return to the
correct context, a return from supervisor call instruction, rets, is implemented. In the AVR32UC
CPU, scall and rets uses the system stack to store the return address and the status register.
9.4.4
Debug requests
The AVR32 architecture defines a dedicated debug mode. When a debug request is received by
the core, Debug mode is entered. Entry into Debug mode can be masked by the DM bit in the
30
32058K
AVR32-01/12
AT32UC3A
status register. Upon entry into Debug mode, hardware sets the SR[D] bit and jumps to the
Debug Exception handler. By default, debug mode executes in the exception context, but with
dedicated Return Address Register and Return Status Register. These dedicated registers
remove the need for storing this data to the system stack, thereby improving debuggability. The
mode bits in the status register can freely be manipulated in Debug mode, to observe registers
in all contexts, while retaining full privileges.
Debug mode is exited by executing the retd instruction. This returns to the previous context.
9.4.5
Entry points for events
Several different event handler entry points exists. In AVR32 UC, the reset address is
0x8000_0000. This places the reset address in the boot flash memory area.
TLB miss exceptions and scall have a dedicated space relative to EVBA where their event handler can be placed. This speeds up execution by removing the need for a jump instruction placed
at the program address jumped to by the event hardware. All other exceptions have a dedicated
event routine entry point located relative to EVBA. The handler routine address identifies the
exception source directly.
AVR32UC uses the ITLB and DTLB protection exceptions to signal a MPU protection violation.
ITLB and DTLB miss exceptions are used to signal that an access address did not map to any of
the entries in the MPU. TLB multiple hit exception indicates that an access address did map to
multiple TLB entries, signalling an error.
All external interrupt requests have entry points located at an offset relative to EVBA. This
autovector offset is specified by an external Interrupt Controller. The programmer must make
sure that none of the autovector offsets interfere with the placement of other code. The autovector offset has 14 address bits, giving an offset of maximum 16384 bytes.
Special considerations should be made when loading EVBA with a pointer. Due to security considerations, the event handlers should be located in non-writeable flash memory, or optionally in
a privileged memory protection region if an MPU is present.
If several events occur on the same instruction, they are handled in a prioritized way. The priority
ordering is presented in Table 9-4. If events occur on several instructions at different locations in
the pipeline, the events on the oldest instruction are always handled before any events on any
younger instruction, even if the younger instruction has events of higher priority than the oldest
instruction. An instruction B is younger than an instruction A if it was sent down the pipeline later
than A.
The addresses and priority of simultaneous events are shown in Table 9-4. Some of the exceptions are unused in AVR32 UC since it has no MMU, coprocessor interface or floating-point unit.
31
32058K
AVR32-01/12
AT32UC3A
Priority and handler addresses for events
Table 9-4.
Priority
Handler Address
Name
Event source
Stored Return Address
1
0x8000_0000
Reset
External input
Undefined
2
Provided by OCD system
OCD Stop CPU
OCD system
First non-completed instruction
3
EVBA+0x00
Unrecoverable exception
Internal
PC of offending instruction
4
EVBA+0x04
TLB multiple hit
MPU
5
EVBA+0x08
Bus error data fetch
Data bus
First non-completed instruction
6
EVBA+0x0C
Bus error instruction fetch
Data bus
First non-completed instruction
7
EVBA+0x10
NMI
External input
First non-completed instruction
8
Autovectored
Interrupt 3 request
External input
First non-completed instruction
9
Autovectored
Interrupt 2 request
External input
First non-completed instruction
10
Autovectored
Interrupt 1 request
External input
First non-completed instruction
11
Autovectored
Interrupt 0 request
External input
First non-completed instruction
12
EVBA+0x14
Instruction Address
CPU
PC of offending instruction
13
EVBA+0x50
ITLB Miss
MPU
14
EVBA+0x18
ITLB Protection
MPU
PC of offending instruction
15
EVBA+0x1C
Breakpoint
OCD system
First non-completed instruction
16
EVBA+0x20
Illegal Opcode
Instruction
PC of offending instruction
17
EVBA+0x24
Unimplemented instruction
Instruction
PC of offending instruction
18
EVBA+0x28
Privilege violation
Instruction
PC of offending instruction
19
EVBA+0x2C
Floating-point
UNUSED
20
EVBA+0x30
Coprocessor absent
UNUSED
21
EVBA+0x100
Supervisor call
Instruction
PC(Supervisor Call) +2
22
EVBA+0x34
Data Address (Read)
CPU
PC of offending instruction
23
EVBA+0x38
Data Address (Write)
CPU
PC of offending instruction
24
EVBA+0x60
DTLB Miss (Read)
MPU
25
EVBA+0x70
DTLB Miss (Write)
MPU
26
EVBA+0x3C
DTLB Protection (Read)
MPU
PC of offending instruction
27
EVBA+0x40
DTLB Protection (Write)
MPU
PC of offending instruction
28
EVBA+0x44
DTLB Modified
UNUSED
32
32058K
AVR32-01/12
AT32UC3A
10. Memories
10.1
Embedded Memories
• Internal High-Speed Flash
– 512 KBytes (AT32UC3A0512, AT32UC3A1512)
– 256 KBytes (AT32UC3A0256, AT32UC3A1256)
– 128 KBytes (AT32UC3A1128, AT32UC3A2128)
- 0 Wait State Access at up to 33 MHz in Worst Case Conditions
- 1 Wait State Access at up to 66 MHz in Worst Case Conditions
- Pipelined Flash Architecture, allowing burst reads from sequential Flash locations, hiding
penalty of 1 wait state access
- Pipelined Flash Architecture typically reduces the cycle penalty of 1 wait state operation
to only 15% compared to 0 wait state operation
- 100 000 Write Cycles, 15-year Data Retention Capability
- 4 ms Page Programming Time, 8 ms Chip Erase Time
- Sector Lock Capabilities, Bootloader Protection, Security Bit
- 32 Fuses, Erased During Chip Erase
- User Page For Data To Be Preserved During Chip Erase
• Internal High-Speed SRAM, Single-cycle access at full speed
– 64 KBytes (AT32UC3A0512, AT32UC3A0256, AT32UC3A1512, AT32UC3A1256)
– 32KBytes (AT32UC3A1128)
10.2
Physical Memory Map
The system bus is implemented as a bus matrix. All system bus addresses are fixed, and they
are never remapped in any way, not even in boot. Note that AVR32 UC CPU uses unsegmented
translation, as described in the AVR32 Architecture Manual. The 32-bit physical address space
is mapped as follows:
Table 10-1.
AT32UC3A Physical Memory Map
Device
Start Address
Embedded SRAM
Size
AT32UC3A0512
AT32UC3A1512
AT32UC3A0256
AT32UC3A1256
AT32UC3A0128
AT32UC3A1128
0x0000_0000
64 Kbyte
64 Kbyte
64 Kbyte
64 Kbyte
32 Kbyte
32 Kbyte
Embedded Flash
0x8000_0000
512 Kbyte
512 Kbyte
256 Kbyte
256 Kbyte
128 Kbyte
128 Kbyte
EBI SRAM CS0
0xC000_0000
16 Mbyte
-
16 Mbyte
-
16 Mbyte
-
EBI SRAM CS2
0xC800_0000
16 Mbyte
-
16 Mbyte
-
16 Mbyte
-
EBI SRAM CS3
0xCC00_0000
16 Mbyte
-
16 Mbyte
-
16 Mbyte
-
EBI SRAM CS1
/SDRAM CS0
0xD000_0000
128 Mbyte
-
128 Mbyte
-
128 Mbyte
-
USB
Configuration
0xE000_0000
64 Kbyte
64 Kbyte
64 Kbyte
64 Kbyte
64 Kbyte
64 Kbyte
HSB-PB Bridge A
0xFFFE_0000
64 Kbyte
64 Kbyte
64 Kbyte
64 Kbyte
64 Kbyte
64 Kbyte
HSB-PB Bridge B
0xFFFF_0000
64 Kbyte
64 Kbyte
64 kByte
64 kByte
64 Kbyte
64 Kbyte
33
32058K
AVR32-01/12
AT32UC3A
Table 10-2.
Flash Memory Parameters
Part Number
Flash Size
(FLASH_PW)
Number of pages
(FLASH_P)
Page size
(FLASH_W)
General Purpose
Fuse bits
(FLASH_F)
AT32UC3A0512
512 Kbytes
1024
128 words
32 fuses
AT32UC3A1512
512 Kbytes
1024
128 words
32 fuses
AT32UC3A0256
256 Kbytes
512
128 words
32 fuses
AT32UC3A1256
256 Kbytes
512
128 words
32 fuses
AT32UC3A1128
128 Kbytes
256
128 words
32 fuses
AT32UC3A0128
128 Kbytes
256
128 words
32 fuses
10.3
Bus Matrix Connections
Accesses to unused areas returns an error result to the master requesting such an access.
The bus matrix has the several masters and slaves. Each master has its own bus and its own
decoder, thus allowing a different memory mapping per master. The master number in the table
below can be used to index the HMATRIX control registers. For example, MCFG0 is associated
with the CPU Data master interface.
Table 10-3.
High Speed Bus masters
Master 0
CPU Data
Master 1
CPU Instruction
Master 2
CPU SAB
Master 3
PDCA
Master 4
MACB DMA
Master 5
USBB DMA
Each slave has its own arbiter, thus allowing a different arbitration per slave. The slave number
in the table below can be used to index the HMATRIX control registers. For example, SCFG3 is
associated with the Internal SRAM Slave Interface.
Table 10-4.
High Speed Bus slaves
Slave 0
Internal Flash
Slave 1
HSB-PB Bridge 0
Slave 2
HSB-PB Bridge 1
Slave 3
Internal SRAM
Slave 4
USBB DPRAM
Slave 5
EBI
34
32058K
AVR32-01/12
AT32UC3A
Figure 10-1. HMatrix Master / Slave Connections
HMATRIX MASTERS
CPU Data
0
CPU
Instruction
1
CPU SAB
2
PDCA
3
MACB
4
USBB DMA
5
Internal Flash
HSB-PB
Bridge 0
HSB-PB
Bridge 1
Internal SRAM
Slave
USBB Slave
EBI
HMATRIX SLAVES
0
1
2
3
4
5
35
32058K
AVR32-01/12
AT32UC3A
11. Fuses Settings
The flash block contains a number of general purpose fuses. Some of these fuses have defined
meanings outside the flash controller and are described in this section.
The general purpose fuses are erase by a JTAG chip erase.
11.1
Flash General Purpose Fuse Register (FGPFRLO)
Table 11-1.
FGPFR Register Description
31
30
29
GPF31
GPF30
GPF29
23
22
21
28
27
BODEN
20
BODHYST
19
BODLEVEL[3:0]
15
14
13
26
18
25
BODLEVEL[5:4]
17
BOOTPROT
12
24
16
EPFL
11
10
9
8
3
2
1
0
LOCK[15:8]
7
6
5
4
LOCK[7:0]
BODEN: Brown Out Detector Enable
Table 11-2.
BODEN Field Description
BODEN
Description
0x0
BOD disabled
0x1
BOD enabled, BOD reset enabled
0x2
BOD enabled, BOD reset disabled
0x3
BOD disabled
BODHYST: Brown Out Detector Hysteresis
Table 11-3.
BODEN Field Description
BODHYST
Description
0b
The Brown out detector hysteresis is disabled
1b
he Brown out detector hysteresis is enabled.
BODLEVEL: Brown Out Detector Trigger Level
This controls the voltage trigger level for the Brown out detector. Refer to sectionTable 38-6 on
page 765 for values description. If the BODLEVEL is set higher than VDDCORE and enabled by
fuses, the part will be in constant reset. To recover from this situation, apply an external voltage
on VDDCORE that is higher than the BOD level and disable the BOD.
36
32058K
AVR32-01/12
AT32UC3A
LOCK, EPFL, BOOTPROT
These are Flash controller fuses and are described in the FLASHC section.
11.2
Default Fuse Value
The devices are shipped with the FGPFRLO register value: 0xFC07FFFF:
• GPF31 fuse set to 1b. This fuse is used by the pre-programmed USB bootloader.
• GPF30 fuse set to 1b. This fuse is used by the pre-programmed USB bootloader.
• GPF29 fuse set to 1b. This fuse is used by the pre-programmed USB bootloader.
• BODEN fuses set to 11b. BOD is disabled.
• BODHYST fuse set to 1b. The BOD hysteresis is enabled.
• BODLEVEL fuses set to 000000b. This is the minimum voltage trigger level for BOD.
• BOOTPROT fuses set to 011b. The bootloader protected size is 8 Ko.
• EPFL fuse set to 1b. External privileged fetch is not locked.
• LOCK fuses set to 1111111111111111b. No region locked.
See also the AT32UC3A Bootloader user guide document.
After the JTAG chip erase command, the FGPFRLO register value is 0xFFFFFFFF.
37
32058K
AVR32-01/12
AT32UC3A
12. Peripherals
12.1
Peripheral address map
Table 12-1.
Peripheral Address Mapping
Address
0xE0000000
0xFFFE0000
0xFFFE1000
0xFFFE1400
0xFFFE1800
0xFFFE1C00
0xFFFE2000
0xFFFF0000
0xFFFF0800
0xFFFF0C00
0xFFFF0D00
0xFFFF0D30
0xFFFF0D80
0xFFFF1000
0xFFFF1400
0xFFFF1800
Peripheral Name
Bus
USBB
USBB Slave Interface - USBB
HSB
USBB
USBB Configuration Interface - USBB
PBB
HMATRIX
HMATRIX Configuration Interface - HMATRIX
PBB
FLASHC
Flash Controller - FLASHC
PBB
MACB Configuration Interface - MACB
PBB
Static Memory Controller Configuration Interface SMC
PBB
SDRAM Controller Configuration Interface SDRAMC
PBB
PDCA
Peripheral DMA Interface - PDCA
PBA
INTC
Interrupt Controller Interface - INTC
PBA
PM
Power Manager - PM
PBA
RTC
Real Time Clock - RTC
PBA
WDT
WatchDog Timer - WDT
PBA
External Interrupt Controller - EIC
PBA
General Purpose IO Controller - GPIO
PBA
USART0
Universal Synchronous Asynchronous Receiver
Transmitter - USART0
PBA
USART1
Universal Synchronous Asynchronous Receiver
Transmitter - USART1
PBA
MACB
SMC
SDRAMC
EIC
GPIO
38
32058K
AVR32-01/12
AT32UC3A
Table 12-1.
Peripheral Address Mapping (Continued)
Address
0xFFFF1C00
0xFFFF2000
0xFFFF2400
0xFFFF2800
0xFFFF2C00
0xFFFF3000
0xFFFF3400
0xFFFF3800
0xFFFF3C00
12.2
Peripheral Name
Bus
USART2
Universal Synchronous Asynchronous Receiver
Transmitter - USART2
PBA
USART3
Universal Synchronous Asynchronous Receiver
Transmitter - USART3
PBA
SPI0
Serial Peripheral Interface - SPI0
PBA
SPI1
Serial Peripheral Interface - SPI1
PBA
TWI
Two Wire Interface - TWI
PBA
PWM
Pulse Width Modulation Controller - PWM
PBA
SSC
Synchronous Serial Controller - SSC
PBA
Timer/Counter - TC
PBA
Analog To Digital Converter - ADC
PBA
TC
ADC
CPU Local Bus Mapping
Some of the registers in the GPIO module are mapped onto the CPU local bus, in addition to
being mapped on the Peripheral Bus. These registers can therefore be reached both by
accesses on the Peripheral Bus, and by accesses on the local bus.
Mapping these registers on the local bus allows cycle-deterministic toggling of GPIO pins since
the CPU and GPIO are the only modules connected to this bus. Also, since the local bus runs at
CPU speed, one write or read operation can be performed per clock cycle to the local busmapped GPIO registers.
39
32058K
AVR32-01/12
AT32UC3A
The following GPIO registers are mapped on the local bus:
Table 12-2.
Local bus mapped GPIO registers
Port
Register
Mode
Local Bus
Address
Access
0
Output Driver Enable Register (ODER)
WRITE
0x4000_0040
Write-only
SET
0x4000_0044
Write-only
CLEAR
0x4000_0048
Write-only
TOGGLE
0x4000_004C
Write-only
WRITE
0x4000_0050
Write-only
SET
0x4000_0054
Write-only
CLEAR
0x4000_0058
Write-only
TOGGLE
0x4000_005C
Write-only
Pin Value Register (PVR)
-
0x4000_0060
Read-only
Output Driver Enable Register (ODER)
WRITE
0x4000_0140
Write-only
SET
0x4000_0144
Write-only
CLEAR
0x4000_0148
Write-only
TOGGLE
0x4000_014C
Write-only
WRITE
0x4000_0150
Write-only
SET
0x4000_0154
Write-only
CLEAR
0x4000_0158
Write-only
TOGGLE
0x4000_015C
Write-only
Pin Value Register (PVR)
-
0x4000_0160
Read-only
Output Driver Enable Register (ODER)
WRITE
0x4000_0240
Write-only
SET
0x4000_0244
Write-only
CLEAR
0x4000_0248
Write-only
TOGGLE
0x4000_024C
Write-only
WRITE
0x4000_0250
Write-only
SET
0x4000_0254
Write-only
CLEAR
0x4000_0258
Write-only
TOGGLE
0x4000_025C
Write-only
-
0x4000_0260
Read-only
Output Value Register (OVR)
1
Output Value Register (OVR)
2
Output Value Register (OVR)
Pin Value Register (PVR)
40
32058K
AVR32-01/12
AT32UC3A
Table 12-2.
Local bus mapped GPIO registers
Port
Register
Mode
Local Bus
Address
Access
3
Output Driver Enable Register (ODER)
WRITE
0x4000_0340
Write-only
SET
0x4000_0344
Write-only
CLEAR
0x4000_0348
Write-only
TOGGLE
0x4000_034C
Write-only
WRITE
0x4000_0350
Write-only
SET
0x4000_0354
Write-only
CLEAR
0x4000_0358
Write-only
TOGGLE
0x4000_035C
Write-only
-
0x4000_0360
Read-only
Output Value Register (OVR)
Pin Value Register (PVR)
12.3
Interrupt Request Signal Map
The various modules may output Interrupt request signals. These signals are routed to the Interrupt Controller (INTC), described in a later chapter. The Interrupt Controller supports up to 64
groups of interrupt requests. Each group can have up to 32 interrupt request signals. All interrupt
signals in the same group share the same autovector address and priority level. Refer to the
documentation for the individual submodules for a description of the semantics of the different
interrupt requests.
The interrupt request signals are connected to the INTC as follows.
Table 12-3.
Interrupt Request Signal Map
Group
Line
0
0
AVR32 UC CPU with optional MPU and
optional OCD
0
External Interrupt Controller
EIC 0
1
External Interrupt Controller
EIC 1
2
External Interrupt Controller
EIC 2
3
External Interrupt Controller
EIC 3
4
External Interrupt Controller
EIC 4
5
External Interrupt Controller
EIC 5
6
External Interrupt Controller
EIC 6
7
External Interrupt Controller
EIC 7
8
Real Time Counter
RTC
9
Power Manager
PM
10
Frequency Meter
FREQM
1
Module
Signal
SYSBLOCK
COMPARE
41
32058K
AVR32-01/12
AT32UC3A
Table 12-3.
Interrupt Request Signal Map
0
General Purpose Input/Output
GPIO 0
1
General Purpose Input/Output
GPIO 1
2
General Purpose Input/Output
GPIO 2
3
General Purpose Input/Output
GPIO 3
4
General Purpose Input/Output
GPIO 4
5
General Purpose Input/Output
GPIO 5
6
General Purpose Input/Output
GPIO 6
7
General Purpose Input/Output
GPIO 7
8
General Purpose Input/Output
GPIO 8
9
General Purpose Input/Output
GPIO 9
10
General Purpose Input/Output
GPIO 10
11
General Purpose Input/Output
GPIO 11
12
General Purpose Input/Output
GPIO 12
13
General Purpose Input/Output
GPIO 13
0
Peripheral DMA Controller
PDCA 0
1
Peripheral DMA Controller
PDCA 1
2
Peripheral DMA Controller
PDCA 2
3
Peripheral DMA Controller
PDCA 3
4
Peripheral DMA Controller
PDCA 4
5
Peripheral DMA Controller
PDCA 5
6
Peripheral DMA Controller
PDCA 6
7
Peripheral DMA Controller
PDCA 7
8
Peripheral DMA Controller
PDCA 8
9
Peripheral DMA Controller
PDCA 9
10
Peripheral DMA Controller
PDCA 10
11
Peripheral DMA Controller
PDCA 11
12
Peripheral DMA Controller
PDCA 12
13
Peripheral DMA Controller
PDCA 13
14
Peripheral DMA Controller
PDCA 14
4
0
Flash Controller
FLASHC
5
0
Universal Synchronous/Asynchronous
Receiver/Transmitter
USART0
6
0
Universal Synchronous/Asynchronous
Receiver/Transmitter
USART1
7
0
Universal Synchronous/Asynchronous
Receiver/Transmitter
USART2
8
0
Universal Synchronous/Asynchronous
Receiver/Transmitter
USART3
2
3
42
32058K
AVR32-01/12
AT32UC3A
Table 12-3.
Interrupt Request Signal Map
9
0
Serial Peripheral Interface
SPI0
10
0
Serial Peripheral Interface
SPI1
11
0
Two-wire Interface
TWI
12
0
Pulse Width Modulation Controller
PWM
13
0
Synchronous Serial Controller
SSC
0
Timer/Counter
TC0
1
Timer/Counter
TC1
2
Timer/Counter
TC2
15
0
Analog to Digital Converter
ADC
16
0
Ethernet MAC
MACB
17
0
USB 2.0 OTG Interface
USBB
18
0
SDRAM Controller
19
0
Audio Bitstream DAC
14
12.4
SDRAMC
DAC
Clock Connections
12.4.1
Timer/Counters
Each Timer/Counter channel can independently select an internal or external clock source for its
counter:
Table 12-4.
Timer/Counter clock connections
Source
Name
Connection
Internal
TIMER_CLOCK1
32 KHz Oscillator
TIMER_CLOCK2
PBA clock / 2
TIMER_CLOCK3
PBA clock / 8
TIMER_CLOCK4
PBA clock / 32
TIMER_CLOCK5
PBA clock / 128
XC0
See Section 12.7
External
XC1
XC2
12.4.2
USARTs
Each USART can be connected to an internally divided clock:
Table 12-5.
USART clock connections
USART
Source
Name
Connection
0
Internal
CLK_DIV
PBA clock / 8
1
2
3
43
32058K
AVR32-01/12
AT32UC3A
12.4.3
SPIs
Each SPI can be connected to an internally divided clock:
Table 12-6.
SPI clock connections
SPI
Source
Name
Connection
0
Internal
CLK_DIV
PBA clock or
PBA clock / 32
1
12.5
Nexus OCD AUX port connections
If the OCD trace system is enabled, the trace system will take control over a number of pins, irrespectively of the PIO configuration. Two different OCD trace pin mappings are possible,
depending on the configuration of the OCD AXS register. For details, see the AVR32 UC Technical Reference Manual.
Table 12-7.
12.6
Nexus OCD AUX port connections
Pin
AXS=0
AXS=1
EVTI_N
PB19
PA08
MDO[5]
PB16
PA27
MDO[4]
PB14
PA26
MDO[3]
PB13
PA25
MDO[2]
PB12
PA24
MDO[1]
PB11
PA23
MDO[0]
PB10
PA22
EVTO_N
PB20
PB20
MCKO
PB21
PA21
MSEO[1]
PB04
PA07
MSEO[0]
PB17
PA28
PDC handshake signals
The PDC and the peripheral modules communicate through a set of handshake signals. The following table defines the valid settings for the Peripheral Identifier (PID) in the PDC Peripheral
Select Register (PSR).
Table 12-8.
PDC Handshake Signals
PID Value
Peripheral module & direction
0
ADC
1
SSC - RX
2
USART0 - RX
3
USART1 - RX
44
32058K
AVR32-01/12
AT32UC3A
Table 12-8.
12.7
PDC Handshake Signals
PID Value
Peripheral module & direction
4
USART2 - RX
5
USART3 - RX
6
TWI - RX
7
SPI0 - RX
8
SPI1 - RX
9
SSC - TX
10
USART0 - TX
11
USART1 - TX
12
USART2 - TX
13
USART3 - TX
14
TWI - TX
15
SPI0 - TX
16
SPI1 - TX
17
ABDAC
Peripheral Multiplexing on I/O lines
Each GPIO line can be assigned to one of 3 peripheral functions; A, B or C. The following table
define how the I/O lines on the peripherals A, B and C are multiplexed by the GPIO.
Table 12-9.
GPIO Controller Function Multiplexing
TQFP100
VQFP144
PIN
GPIO Pin
Function A
Function B
Function C
19
25
PA00
GPIO 0
USART0 - RXD
TC - CLK0
20
27
PA01
GPIO 1
USART0 - TXD
TC - CLK1
23
30
PA02
GPIO 2
USART0 - CLK
TC - CLK2
24
32
PA03
GPIO 3
USART0 - RTS
EIM - EXTINT[4]
DAC - DATA[0]
25
34
PA04
GPIO 4
USART0 - CTS
EIM - EXTINT[5]
DAC - DATAN[0]
26
39
PA05
GPIO 5
USART1 - RXD
PWM - PWM[4]
27
41
PA06
GPIO 6
USART1 - TXD
PWM - PWM[5]
28
43
PA07
GPIO 7
USART1 - CLK
PM - GCLK[0]
SPI0 - NPCS[3]
29
45
PA08
GPIO 8
USART1 - RTS
SPI0 - NPCS[1]
EIM - EXTINT[7]
30
47
PA09
GPIO 9
USART1 - CTS
SPI0 - NPCS[2]
MACB - WOL
31
48
PA10
GPIO 10
SPI0 - NPCS[0]
EIM - EXTINT[6]
33
50
PA11
GPIO 11
SPI0 - MISO
USB - USB_ID
36
53
PA12
GPIO 12
SPI0 - MOSI
USB - USB_VBOF
37
54
PA13
GPIO 13
SPI0 - SCK
39
56
PA14
GPIO 14
SSC TX_FRAME_SYNC
SPI1 - NPCS[0]
EBI - NCS[0]
40
57
PA15
GPIO 15
SSC - TX_CLOCK
SPI1 - SCK
EBI - ADDR[20]
45
32058K
AVR32-01/12
AT32UC3A
Table 12-9.
GPIO Controller Function Multiplexing
41
58
PA16
GPIO 16
SSC - TX_DATA
SPI1 - MOSI
EBI - ADDR[21]
42
60
PA17
GPIO 17
SSC - RX_DATA
SPI1 - MISO
EBI - ADDR[22]
43
62
PA18
GPIO 18
SSC - RX_CLOCK
SPI1 - NPCS[1]
MACB - WOL
44
64
PA19
GPIO 19
SSC RX_FRAME_SYNC
SPI1 - NPCS[2]
45
66
PA20
GPIO 20
EIM - EXTINT[8]
SPI1 - NPCS[3]
51
73
PA21
GPIO 21
ADC - AD[0]
EIM - EXTINT[0]
USB - USB_ID
52
74
PA22
GPIO 22
ADC - AD[1]
EIM - EXTINT[1]
USB - USB_VBOF
53
75
PA23
GPIO 23
ADC - AD[2]
EIM - EXTINT[2]
DAC - DATA[1]
54
76
PA24
GPIO 24
ADC - AD[3]
EIM - EXTINT[3]
DAC - DATAN[1]
55
77
PA25
GPIO 25
ADC - AD[4]
EIM - SCAN[0]
EBI - NCS[0]
56
78
PA26
GPIO 26
ADC - AD[5]
EIM - SCAN[1]
EBI - ADDR[20]
57
79
PA27
GPIO 27
ADC - AD[6]
EIM - SCAN[2]
EBI - ADDR[21]
58
80
PA28
GPIO 28
ADC - AD[7]
EIM - SCAN[3]
EBI - ADDR[22]
83
122
PA29
GPIO 29
TWI - SDA
USART2 - RTS
84
123
PA30
GPIO 30
TWI - SCL
USART2 - CTS
65
88
PB00
GPIO 32
MACB - TX_CLK
USART2 - RTS
USART3 - RTS
66
90
PB01
GPIO 33
MACB - TX_EN
USART2 - CTS
USART3 - CTS
70
96
PB02
GPIO 34
MACB - TXD[0]
DAC - DATA[0]
71
98
PB03
GPIO 35
MACB - TXD[1]
DAC - DATAN[0]
72
100
PB04
GPIO 36
MACB - CRS
USART3 - CLK
73
102
PB05
GPIO 37
MACB - RXD[0]
DAC - DATA[1]
74
104
PB06
GPIO 38
MACB - RXD[1]
DAC - DATAN[1]
75
106
PB07
GPIO 39
MACB - RX_ER
76
111
PB08
GPIO 40
MACB - MDC
77
113
PB09
GPIO 41
MACB - MDIO
78
115
PB10
GPIO 42
MACB - TXD[2]
USART3 - RXD
EBI - SDCK
81
119
PB11
GPIO 43
MACB - TXD[3]
USART3 - TXD
EBI - SDCKE
82
121
PB12
GPIO 44
MACB - TX_ER
TC - CLK0
EBI - RAS
87
126
PB13
GPIO 45
MACB - RXD[2]
TC - CLK1
EBI - CAS
88
127
PB14
GPIO 46
MACB - RXD[3]
TC - CLK2
EBI - SDWE
95
134
PB15
GPIO 47
MACB - RX_DV
96
136
PB16
GPIO 48
MACB - COL
USB - USB_ID
EBI - SDA10
98
139
PB17
GPIO 49
MACB - RX_CLK
USB - USB_VBOF
EBI - ADDR[23]
99
141
PB18
GPIO 50
MACB - SPEED
ADC - TRIGGER
PWM - PWM[6]
100
143
PB19
GPIO 51
PWM - PWM[0]
PM - GCLK[0]
EIM - SCAN[4]
1
3
PB20
GPIO 52
PWM - PWM[1]
PM - GCLK[1]
EIM - SCAN[5]
2
5
PB21
GPIO 53
PWM - PWM[2]
PM - GCLK[2]
EIM - SCAN[6]
3
6
PB22
GPIO 54
PWM - PWM[3]
PM - GCLK[3]
EIM - SCAN[7]
6
9
PB23
GPIO 55
TC - A0
USART1 - DCD
EBI - NCS[3]
46
32058K
AVR32-01/12
AT32UC3A
Table 12-9.
GPIO Controller Function Multiplexing
7
11
PB24
GPIO 56
TC - B0
USART1 - DSR
8
13
PB25
GPIO 57
TC - A1
USART1 - DTR
9
14
PB26
GPIO 58
TC - B1
USART1 - RI
10
15
PB27
GPIO 59
TC - A2
PWM - PWM[4]
14
19
PB28
GPIO 60
TC - B2
PWM - PWM[5]
15
20
PB29
GPIO 61
USART2 - RXD
PM - GCLK[1]
EBI - NCS[2]
16
21
PB30
GPIO 62
USART2 - TXD
PM - GCLK[2]
EBI - SDCS
17
22
PB31
GPIO 63
USART2 - CLK
PM - GCLK[3]
EBI - NWAIT
63
85
PC00
GPIO 64
64
86
PC01
GPIO 65
85
124
PC02
GPIO 66
86
125
PC03
GPIO 67
93
132
PC04
GPIO 68
94
133
PC05
GPIO 69
1
PX00
GPIO 100
EBI - DATA[10]
USART0 - RXD
2
PX01
GPIO 99
EBI - DATA[9]
USART0 - TXD
4
PX02
GPIO 98
EBI - DATA[8]
USART0 - CTS
10
PX03
GPIO 97
EBI - DATA[7]
USART0 - RTS
12
PX04
GPIO 96
EBI - DATA[6]
USART1 - RXD
24
PX05
GPIO 95
EBI - DATA[5]
USART1 - TXD
26
PX06
GPIO 94
EBI - DATA[4]
USART1 - CTS
31
PX07
GPIO 93
EBI - DATA[3]
USART1 - RTS
33
PX08
GPIO 92
EBI - DATA[2]
USART3 - RXD
35
PX09
GPIO 91
EBI - DATA[1]
USART3 - TXD
38
PX10
GPIO 90
EBI - DATA[0]
USART2 - RXD
40
PX11
GPIO 109
EBI - NWE1
USART2 - TXD
42
PX12
GPIO 108
EBI - NWE0
USART2 - CTS
44
PX13
GPIO 107
EBI - NRD
USART2 - RTS
46
PX14
GPIO 106
EBI - NCS[1]
59
PX15
GPIO 89
EBI - ADDR[19]
USART3 - RTS
TC - B0
61
PX16
GPIO 88
EBI - ADDR[18]
USART3 - CTS
TC - A1
63
PX17
GPIO 87
EBI - ADDR[17]
TC - B1
65
PX18
GPIO 86
EBI - ADDR[16]
TC - A2
67
PX19
GPIO 85
EBI - ADDR[15]
EIM - SCAN[0]
TC - B2
87
PX20
GPIO 84
EBI - ADDR[14]
EIM - SCAN[1]
TC - CLK0
89
PX21
GPIO 83
EBI - ADDR[13]
EIM - SCAN[2]
TC - CLK1
91
PX22
GPIO 82
EBI - ADDR[12]
EIM - SCAN[3]
TC - CLK2
95
PX23
GPIO 81
EBI - ADDR[11]
EIM - SCAN[4]
97
PX24
GPIO 80
EBI - ADDR[10]
EIM - SCAN[5]
TC - A0
47
32058K
AVR32-01/12
AT32UC3A
Table 12-9.
12.8
GPIO Controller Function Multiplexing
99
PX25
GPIO 79
EBI - ADDR[9]
EIM - SCAN[6]
101
PX26
GPIO 78
EBI - ADDR[8]
EIM - SCAN[7]
103
PX27
GPIO 77
EBI - ADDR[7]
SPI0 - MISO
105
PX28
GPIO 76
EBI - ADDR[6]
SPI0 - MOSI
107
PX29
GPIO 75
EBI - ADDR[5]
SPI0 - SCK
110
PX30
GPIO 74
EBI - ADDR[4]
SPI0 - NPCS[0]
112
PX31
GPIO 73
EBI - ADDR[3]
SPI0 - NPCS[1]
114
PX32
GPIO 72
EBI - ADDR[2]
SPI0 - NPCS[2]
118
PX33
GPIO 71
EBI - ADDR[1]
SPI0 - NPCS[3]
120
PX34
GPIO 70
EBI - ADDR[0]
SPI1 - MISO
135
PX35
GPIO 105
EBI - DATA[15]
SPI1 - MOSI
137
PX36
GPIO 104
EBI - DATA[14]
SPI1 - SCK
140
PX37
GPIO 103
EBI - DATA[13]
SPI1 - NPCS[0]
142
PX38
GPIO 102
EBI - DATA[12]
SPI1 - NPCS[1]
144
PX39
GPIO 101
EBI - DATA[11]
SPI1 - NPCS[2]
Oscillator Pinout
The oscillators are not mapped to the normal A,B or C functions and their muxings are controlled
by registers in the Power Manager (PM). Please refer to the power manager chapter for more
information about this.
Table 12-10. Oscillator pinout
12.9
TQFP100 pin
VQFP144 pin
Pad
Oscillator pin
85
124
PC02
xin0
93
132
PC04
xin1
63
85
PC00
xin32
86
125
PC03
xout0
94
133
PC05
xout1
64
86
PC01
xout32
USART Configuration
Table 12-11. USART Supported Mode
SPI
RS485
ISO7816
IrDA
Modem
Manchester
Encoding
USART0
Yes
No
No
No
No
No
USART1
Yes
Yes
Yes
Yes
Yes
Yes
USART2
Yes
No
No
No
No
No
USART3
Yes
No
No
No
No
No
48
32058K
AVR32-01/12
AT32UC3A
12.10 GPIO
The GPIO open drain feature (GPIO ODMER register (Open Drain Mode Enable Register)) is
not available for this device.
12.11 Peripheral overview
12.11.1
External Bus Interface
• Optimized for Application Memory Space support
• Integrates Two External Memory Controllers:
– Static Memory Controller
– SDRAM Controller
• Optimized External Bus:
– 16-bit Data Bus
– 24-bit Address Bus, Up to 16-Mbytes Addressable
– Optimized pin multiplexing to reduce latencies on External Memories
• 4 SRAM Chip Selects, 1SDRAM Chip Select:
– Static Memory Controller on NCS0
– SDRAM Controller or Static Memory Controller on NCS1
– Static Memory Controller on NCS2
– Static Memory Controller on NCS3
12.11.2
Static Memory Controller
12.11.3
• 4 Chip Selects Available
• 64-Mbyte Address Space per Chip Select
• 8-, 16-bit Data Bus
• Word, Halfword, Byte Transfers
• Byte Write or Byte Select Lines
• Programmable Setup, Pulse And Hold Time for Read Signals per Chip Select
• Programmable Setup, Pulse And Hold Time for Write Signals per Chip Select
• Programmable Data Float Time per Chip Select
• Compliant with LCD Module
• External Wait Request
• Automatic Switch to Slow Clock Mode
• Asynchronous Read in Page Mode Supported: Page Size Ranges from 4 to 32 Bytes
SDRAM Controller
• Numerous Configurations Supported
– 2K, 4K, 8K Row Address Memory Parts
– SDRAM with Two or Four Internal Banks
– SDRAM with 16-bit Data Path
• Programming Facilities
– Word, Half-word, Byte Access
– Automatic Page Break When Memory Boundary Has Been Reached
– Multibank Ping-pong Access
– Timing Parameters Specified by Software
– Automatic Refresh Operation, Refresh Rate is Programmable
• Energy-saving Capabilities
– Self-refresh, Power-down and Deep Power Modes Supported
49
32058K
AVR32-01/12
AT32UC3A
– Supports Mobile SDRAM Devices
• Error Detection
– Refresh Error Interrupt
• SDRAM Power-up Initialization by Software
• CAS Latency of 1, 2, 3 Supported
• Auto Precharge Command Not Used
12.11.4
USB Controller
12.11.5
• USB 2.0 Compliant, Full-/Low-Speed (FS/LS) and On-The-Go (OTG), 12 Mbit/s
• 7 Pipes/Endpoints
• 960 bytes of Embedded Dual-Port RAM (DPRAM) for Pipes/Endpoints
• Up to 2 Memory Banks per Pipe/Endpoint (Not for Control Pipe/Endpoint)
• Flexible Pipe/Endpoint Configuration and Management with Dedicated DMA Channels
• On-Chip Transceivers Including Pull-Ups
Serial Peripheral Interface
• Supports communication with serial external devices
– Four chip selects with external decoder support allow communication with up to 15
peripherals
– Serial memories, such as DataFlash and 3-wire EEPROMs
– Serial peripherals, such as ADCs, DACs, LCD Controllers, CAN Controllers and Sensors
– External co-processors
• Master or slave serial peripheral bus interface
– 8- to 16-bit programmable data length per chip select
– Programmable phase and polarity per chip select
– Programmable transfer delays between consecutive transfers and between clock and data
per chip select
– Programmable delay between consecutive transfers
– Selectable mode fault detection
• Very fast transfers supported
– Transfers with baud rates up to Peripheral Bus A (PBA) max frequency
– The chip select line may be left active to speed up transfers on the same device
12.11.6
Two-wire Interface
•
•
•
•
12.11.7
High speed up to 400kbit/s
Compatibility with standard two-wire serial memory
One, two or three bytes for slave address
Sequential read/write operations
USART
• Programmable Baud Rate Generator
• 5- to 9-bit full-duplex synchronous or asynchronous serial communications
–
–
–
–
–
–
–
–
–
1, 1.5 or 2 stop bits in Asynchronous Mode or 1 or 2 stop bits in Synchronous Mode
Parity generation and error detection
Framing error detection, overrun error detection
MSB- or LSB-first
Optional break generation and detection
By 8 or by-16 over-sampling receiver frequency
Hardware handshaking RTS-CTS
Receiver time-out and transmitter timeguard
Optional Multi-drop Mode with address generation and detection
50
32058K
AVR32-01/12
AT32UC3A
– Optional Manchester Encoding
• RS485 with driver control signal
• ISO7816, T = 0 or T = 1 Protocols for interfacing with smart cards
– NACK handling, error counter with repetition and iteration limit
• IrDA modulation and demodulation
– Communication at up to 115.2 Kbps
• Test Modes
– Remote Loopback, Local Loopback, Automatic Echo
• SPI Mode
– Master or Slave
– Serial Clock Programmable Phase and Polarity
– SPI Serial Clock (SCK) Frequency up to Internal Clock Frequency PBA/4
• Supports Connection of Two Peripheral DMA Controller Channels (PDC)
– Offers Buffer Transfer without Processor Intervention
12.11.8
Serial Synchronous Controller
• Provides serial synchronous communication links used in audio and telecom applications (with
CODECs in Master or Slave Modes, I2S, TDM Buses, Magnetic Card Reader, etc.)
• Contains an independent receiver and transmitter and a common clock divider
• Offers a configurable frame sync and data length
• Receiver and transmitter can be programmed to start automatically or on detection of different
event on the frame sync signal
• Receiver and transmitter include a data signal, a clock signal and a frame synchronization signal
12.11.9
Timer Counter
• Three 16-bit Timer Counter Channels
• Wide range of functions including:
– Frequency Measurement
– Event Counting
– Interval Measurement
– Pulse Generation
– Delay Timing
– Pulse Width Modulation
– Up/down Capabilities
• Each channel is user-configurable and contains:
– Three external clock inputs
– Five internal clock inputs
– Two multi-purpose input/output signals
• Two global registers that act on all three TC Channels
12.11.10 Pulse Width Modulation Controller
• 7 channels, one 20-bit counter per channel
• Common clock generator, providing Thirteen Different Clocks
– A Modulo n counter providing eleven clocks
– Two independent Linear Dividers working on modulo n counter outputs
• Independent channel programming
– Independent Enable Disable Commands
– Independent Clock
– Independent Period and Duty Cycle, with Double Bufferization
– Programmable selection of the output waveform polarity
– Programmable center or left aligned output waveform
51
32058K
AVR32-01/12
AT32UC3A
12.11.11 Ethernet 10/100 MAC
•
•
•
•
•
•
•
•
•
•
•
•
Compatibility with IEEE Standard 802.3
10 and 100 Mbits per second data throughput capability
Full- and half-duplex operations
MII or RMII interface to the physical layer
Register Interface to address, data, status and control registers
DMA Interface, operating as a master on the Memory Controller
Interrupt generation to signal receive and transmit completion
28-byte transmit and 28-byte receive FIFOs
Automatic pad and CRC generation on transmitted frames
Address checking logic to recognize four 48-bit addresses
Support promiscuous mode where all valid frames are copied to memory
Support physical layer management through MDIO interface control of alarm and update
time/calendar data
12.11.12 Audio Bitstream DAC
• Digital Stereo DAC
• Oversampled D/A conversion architecture
– Oversampling ratio fixed 128x
– FIR equalization filter
– Digital interpolation filter: Comb4
– 3rd Order Sigma-Delta D/A converters
• Digital bitstream outputs
• Parallel interface
• Connected to Peripheral DMA Controller for background transfer without CPU intervention
52
32058K
AVR32-01/12
AT32UC3A
13. Power Manager (PM)
Rev: 2.0.0.1
13.1
Features
•
•
•
•
•
•
•
•
•
•
•
•
•
13.2
Controls integrated oscillators and PLLs
Generates clocks and resets for digital logic
Supports 2 crystal oscillators 450 kHz-16 MHz
Supports 2 PLLs 80-240 MHz
Supports 32 KHz ultra-low power oscillator
Integrated low-power RC oscillator
On-the fly frequency change of CPU, HSB, PBA, and PBB clocks
Sleep modes allow simple disabling of logic clocks, PLLs, and oscillators
Module-level clock gating through maskable peripheral clocks
Wake-up from internal or external interrupts
Generic clocks with wide frequency range provided
Automatic identification of reset sources
Controls brownout detector (BOD), RC oscillator, and bandgap voltage reference through control
and calibration registers
Description
The Power Manager (PM) controls the oscillators and PLLs, and generates the clocks and
resets in the device. The PM controls two fast crystal oscillators, as well as two PLLs, which can
multiply the clock from either oscillator to provide higher frequencies. Additionally, a low-power
32 KHz oscillator is used to generate the real-time counter clock for high accuracy real-time
measurements. The PM also contains a low-power RC oscillator with fast start-up time, which
can be used to clock the digital logic.
The provided clocks are divided into synchronous and generic clocks. The synchronous clocks
are used to clock the main digital logic in the device, namely the CPU, and the modules and
peripherals connected to the HSB, PBA, and PBB buses. The generic clocks are asynchronous
clocks, which can be tuned precisely within a wide frequency range, which makes them suitable
for peripherals that require specific frequencies, such as timers and communication modules.
The PM also contains advanced power-saving features, allowing the user to optimize the power
consumption for an application. The synchronous clocks are divided into three clock domains,
one for the CPU and HSB, one for modules on the PBA bus, and one for modules on the PBB
bus.The three clocks can run at different speeds, so the user can save power by running peripherals at a relatively low clock, while maintaining a high CPU performance. Additionally, the
clocks can be independently changed on-the-fly, without halting any peripherals. This enables
the user to adjust the speed of the CPU and memories to the dynamic load of the application,
without disturbing or re-configuring active peripherals.
Each module also has a separate clock, enabling the user to switch off the clock for inactive
modules, to save further power. Additionally, clocks and oscillators can be automatically
switched off during idle periods by using the sleep instruction on the CPU. The system will return
to normal on occurrence of interrupts.
The Power Manager also contains a Reset Controller, which collects all possible reset sources,
generates hard and soft resets, and allows the reset source to be identified by software.
53
32058K
AVR32-01/12
AT32UC3A
13.3
Block Diagram
RCOSC
Oscillator 0
Synchronous
Clock Generator
Synchronous
clocks
CPU, HSB,
PBA, PBB
G eneric Clock
Generator
Generic clocks
PLL0
PLL1
Oscillator 1
32 KHz
Oscillator
OSC/PLL
Control signals
32 KHz clock
for RTC
RC
Oscillator
Slow clock
Oscillator and
PLL Control
Startup
Counter
Voltage Regulator
Interrupts
fuses
Sleep Controller
Sleep
instruction
Calibration
Registers
Brown-Out
Detector
Reset Controller
resets
Power-O n
Detector
Other reset
sources
External Reset Pad
Figure 13-1. Power Manager block diagram
54
32058K
AVR32-01/12
AT32UC3A
13.4
Product Dependencies
13.4.1
I/O Lines
The PM provides a number of generic clock outputs, which can be connected to output pins,
multiplexed with GPIO lines. The programmer must first program the GPIO controller to assign
these pins to their peripheral function. If the I/O pins of the PM are not used by the application,
they can be used for other purposes by the GPIO controller.
13.4.2
Interrupt
The PM interrupt line is connected to one of the internal sources of the interrupt controller. Using
the PM interrupt requires the interrupt controller to be programmed first.
13.4.3
Clock implementation
In AT32UC3A, the HSB shares the source clock with the CPU. This means that writing to the
HSBDIV and HSBSEL bits in CKSEL has no effect. These bits will always read the same as
CPUDIV and CPUSEL.
13.5
Functional Description
13.5.1
Slow clock
The slow clock is generated from an internal RC oscillator which is always running, except in
Static mode. The slow clock can be used for the main clock in the device, as described in ”Synchronous clocks” on page 58. The slow clock is also used for the Watchdog Timer and
measuring various delays in the Power Manager.
The RC oscillator has a 3 cycles startup time, and is always available when the CPU is running.
The RC oscillator operates at approximately 115 kHz, and can be calibrated to a narrow range
by the RCOSCCAL fuses. Software can also change RC oscillator calibration through the use of
the RCCR register. Please see the Electrical Characteristics section for details.
RC oscillator can also be used as the RTC clock when crystal accuracy is not required.
13.5.2
Oscillator 0 and 1 operation
The two main oscillators are designed to be used with an external 450 kHz to 16 MHz crystal
and two biasing capacitors, as shown in Figure 13-2. Oscillator 0 can be used for the main clock
in the device, as described in ”Synchronous clocks” on page 58. Both oscillators can be used as
source for the generic clocks, as described in ”Generic clocks” on page 61.
The oscillators are disabled by default after reset. When the oscillators are disabled, the XIN and
XOUT pins can be used as general purpose I/Os. When the oscillators are configured to use an
external clock, the clock must be applied to the XIN pin while the XOUT pin can be used as a
general purpose I/O.
The oscillators can be enabled by writing to the OSCnEN bits in MCCTRL. Operation mode
(external clock or crystal) is chosen by writing to the MODE field in OSCCTRLn. Oscillators are
automatically switched off in certain sleep modes to reduce power consumption, as described in
Section 13.5.7 on page 60.
After a hard reset, or when waking up from a sleep mode that disabled the oscillators, the oscillators may need a certain amount of time to stabilize on the correct frequency. This start-up time
can be set in the OSCCTRLn register.
55
32058K
AVR32-01/12
AT32UC3A
The PM masks the oscillator outputs during the start-up time, to ensure that no unstable clocks
propagate to the digital logic. The OSCnRDY bits in POSCSR are automatically set and cleared
according to the status of the oscillators. A zero to one transition on these bits can also be configured to generate an interrupt, as described in ”Interrupt Enable/Disable/Mask/Status/Clear” on
page 76.
C2
XO U T
XIN
C1
Figure 13-2. Oscillator connections
13.5.3
32 KHz oscillator operation
The 32 KHz oscillator operates as described for Oscillator 0 and 1 above. The 32 KHz oscillator
is used as source clock for the Real-Time Counter.
The oscillator is disabled by default, but can be enabled by writing OSC32EN in OSCCTRL32.
The oscillator is an ultra-low power design and remains enabled in all sleep modes except Static
mode.
While the 32 KHz oscillator is disabled, the XIN32 and XOUT32 pins are available as general
purpose I/Os. When the oscillator is configured to work with an external clock (MODE field in
OSCCTRL32 register), the external clock must be connected to XIN32 while the XOUT32 pin
can be used as a general purpose I/O.
The startup time of the 32 KHz oscillator can be set in the OSCCTRL32, after which OSC32RDY
in POSCSR is set. An interrupt can be generated on a zero to one transition of OSC32RDY.
As a crystal oscillator usually requires a very long startup time (up to 1 second), the 32 KHz
oscillator will keep running across resets, except Power-On-Reset.
13.5.4
PLL operation
The device contains two PLLs, PLL0 and PLL1. These are disabled by default, but can be
enabled to provide high frequency source clocks for synchronous or generic clocks. The PLLs
can take either Oscillator 0 or 1 as reference clock. The PLL output is divided by a multiplication
factor, and the PLL compares the resulting clock to the reference clock. The PLL will adjust its
output frequency until the two compared clocks are equal, thus locking the output frequency to a
multiple of the reference clock frequency.
The Voltage Controlled Oscillator inside the PLL can generate frequencies from 80 to 240 MHz.
To make the PLL output frequencies under 80 MHz the OTP[1] bitfield could be set. This will
56
32058K
AVR32-01/12
AT32UC3A
divide the output of the PLL by two and bring the clock in range of the max frequency of the
CPU.
When the PLL is switched on, or when changing the clock source or multiplication factor for the
PLL, the PLL is unlocked and the output frequency is undefined. The PLL clock for the digital
logic is automatically masked when the PLL is unlocked, to prevent connected digital logic from
receiving a too high frequency and thus become unstable.
PLLM UL
P L L O P T [1 ]
fvco
O u tp u t
D iv id e r
0
1 /2
O sc0
c lo c k
0
O sc1
c lo c k
In p u t
D iv id e r
1
PLLO SC
P hase
D e te c to r
VCO
Lock
D e te c to r
fP LL
M ask
P L L c lo c k
1
L o c k b it
PLLO PT
P L L D IV
Figure 13-3. PLL with control logic and filters
13.5.4.1
Enabling the PLL
PLLn is enabled by writing the PLLEN bit in the PLLn register. PLLOSC selects Oscillator 0 or 1
as clock source. The PLLMUL and PLLDIV bitfields must be written with the multiplication and
division factors, respectively, creating the voltage controlled ocillator frequency fVCO and the PLL
frequency fPLL :
fVCO = (PLLMUL+1)/(PLLDIV) • fOSC if PLLDIV > 0.
fVCO = 2*(PLLMUL+1) • fOSC if PLLDIV = 0.
If PLLOPT[1] field is set to 0:
fPLL = fVCO.
If PLLOPT[1] field is set to 1:
fPLL = fVCO / 2.
The PLLn:PLLOPT field should be set to proper values according to the PLL operating frequency. The PLLOPT field can also be set to divide the output frequency of the PLLs by 2.
The lock signal for each PLL is available as a LOCKn flag in POSCSR. An interrupt can be generated on a 0 to 1 transition of these bits.
57
32058K
AVR32-01/12
AT32UC3A
13.5.5
Synchronous clocks
The slow clock (default), Oscillator 0, or PLL0 provide the source for the main clock, which is the
common root for the synchronous clocks for the CPU/HSB, PBA, and PBB modules. The main
clock is divided by an 8-bit prescaler, and each of these four synchronous clocks can run from
any tapping of this prescaler, or the undivided main clock, as long as fCPU fPBA,B,. The synchronous clock source can be changed on-the fly, responding to varying load in the application. The
clock domains can be shut down in sleep mode, as described in ”Sleep modes” on page 60.
Additionally, the clocks for each module in the four domains can be individually masked, to avoid
power consumption in inactive modules.
Sleep
Controller
Sleep
instruction
0
Main clock
Slow clock
Osc0 clock
PLL0 clock
Prescaler
CPUDIV
MCSEL
Mask
1
CPU clocks
HSB clocks
CPUMASK
PBAclocks
PBB clocks
CPUSEL
Figure 13-4. Synchronous clock generation
13.5.5.1
Selecting PLL or oscillator for the main clock
The common main clock can be connected to the slow clock, Oscillator 0, or PLL0. By default,
the main clock will be connected to the slow clock. The user can connect the main clock to Oscillator 0 or PLL0 by writing the MCSEL bitfield in the Main Clock Control Register (MCCTRL). This
must only be done after that unit has been enabled, otherwise a deadlock will occur. Care
should also be taken that the new frequency of the synchronous clocks does not exceed the
maximum frequency for each clock domain.
13.5.5.2
Selecting synchronous clock division ratio
The main clock feeds an 8-bit prescaler, which can be used to generate the synchronous clocks.
By default, the synchronous clocks run on the undivided main clock. The user can select a pres-
58
32058K
AVR32-01/12
AT32UC3A
caler division for the CPU clock by writing CKSEL:CPUDIV to 1 and CPUSEL to the prescaling
value, resulting in a CPU clock frequency:
fCPU = fmain / 2(CPUSEL+1)
Similarly, the clock for the PBA, and PBB can be divided by writing their respective bitfields. To
ensure correct operation, frequencies must be selected so that fCPU fPBA,B. Also, frequencies
must never exceed the specified maximum frequency for each clock domain.
CKSEL can be written without halting or disabling peripheral modules. Writing CKSEL allows a
new clock setting to be written to all synchronous clocks at the same time. It is possible to keep
one or more clocks unchanged by writing the same value a before to the xxxDIV and xxxSEL bitfields. This way, it is possible to e.g. scale CPU and HSB speed according to the required
performance, while keeping the PBA and PBB frequency constant.
13.5.5.3
Clock Ready flag
There is a slight delay from CKSEL is written and the new clock setting becomes effective. During this interval, the Clock Ready (CKRDY) flag in ISR will read as 0. If IER:CKRDY is written to
1, the Power Manager interrupt can be triggered when the new clock setting is effective. CKSEL
must not be re-written while CKRDY is 0, or the system may become unstable or hang.
13.5.6
Peripheral clock masking
By default, the clock for all modules are enabled, regardless of which modules are actually being
used. It is possible to disable the clock for a module in the CPU, HSB, PBA, or PBB clock
domain by writing the corresponding bit in the Clock Mask register (CPU/HSB/PBA/PBB) to 0.
When a module is not clocked, it will cease operation, and its registers cannot be read or written.
The module can be re-enabled later by writing the corresponding mask bit to 1.
A module may be connected to several clock domains, in which case it will have several mask
bits.
Table 13-5 contains a list of implemented maskable clocks.
13.5.6.1
Cautionary note
Note that clocks should only be switched off if it is certain that the module will not be used.
Switching off the clock for the internal RAM will cause a problem if the stack is mapped there.
Switching off the clock to the Power Manager (PM), which contains the mask registers, or the
corresponding PBx bridge, will make it impossible to write the mask registers again. In this case,
they can only be re-enabled by a system reset.
13.5.6.2
Mask Ready flag
Due to synchronization in the clock generator, there is a slight delay from a mask register is written until the new mask setting goes into effect. When clearing mask bits, this delay can usually
be ignored. However, when setting mask bits, the registers in the corresponding module must
not be written until the clock has actually be re-enabled. The status flag MSKRDY in ISR provides the required mask status information. When writing either mask register with any value,
this bit is cleared. The bit is set when the clocks have been enabled and disabled according to
the new mask setting. Optionally, the Power Manager interrupt can be enabled by writing the
MSKRDY bit in IER.
59
32058K
AVR32-01/12
AT32UC3A
13.5.7
Sleep modes
In normal operation, all clock domains are active, allowing software execution and peripheral
operation. When the CPU is idle, it is possible to switch off the CPU clock and optionally other
clock domains to save power. This is activated by the sleep instruction, which takes the sleep
mode index number as argument.
13.5.7.1
Entering and exiting sleep modes
The sleep instruction will halt the CPU and all modules belonging to the stopped clock domains.
The modules will be halted regardless of the bit settings of the mask registers.
Oscillators and PLLs can also be switched off to save power. Some of these modules have a relatively long start-up time, and are only switched off when very low power consumption is
required.
The CPU and affected modules are restarted when the sleep mode is exited. This occurs when
an interrupt triggers. Note that even if an interrupt is enabled in sleep mode, it may not trigger if
the source module is not clocked.
13.5.7.2
Supported sleep modes
The following sleep modes are supported. These are detailed in Table 13-1.
•Idle: The CPU is stopped, the rest of the chip is operating. Wake-up sources are any interrupt.
•Frozen: The CPU and HSB modules are stopped, peripherals are operating. Wake-up sources
are any interrupt from PB modules.
•Standby: All synchronous clocks are stopped, but oscillators and PLLs are running, allowing
quick wake-up to normal mode. Wake-up sources are RTC or external interrupt (EIC).
•Stop: As Standby, but Oscillator 0 and 1, and the PLLs are stopped. 32 KHz (if enabled) and
RC oscillators and RTC/WDT still operate. Wake-up sources are RTC, external interrupt (EIC),
or external reset pin.
•DeepStop: All synchronous clocks, Oscillator 0 and 1 and PLL 0 and 1 are stopped. 32 KHz
oscillator can run if enabled. RC oscillator still operates. Bandgap voltage reference and BOD is
turned off. Wake-up sources are RTC, external interrupt (EIC) or external reset pin.
•Static: All oscillators, including 32 KHz and RC oscillator are stopped. Bandgap voltage reference BOD detector is turned off. Wake-up sources are external interrupt (EIC) in asynchronous
mode only or external reset pin.
Sleep modes
Table 13-1.
Sleep Mode
CPU
HSB
PBA,B
GCLK
Osc0,1
PLL0,1
Osc32
RCOsc
BOD &
Bandgap
Voltage
Regulator
0
Idle
Stop
Run
Run
Run
Run
Run
On
Full power
1
Frozen
Stop
Stop
Run
Run
Run
Run
On
Full power
2
Standby
Stop
Stop
Stop
Run
Run
Run
On
Full power
3
Stop
Stop
Stop
Stop
Stop
Run
Run
On
Low power
4
DeepStop
Stop
Stop
Stop
Stop
Run
Run
Off
Low power
5
Static
Stop
Stop
Stop
Stop
Stop
Stop
Off
Low power
Index
60
32058K
AVR32-01/12
AT32UC3A
The power level of the internal voltage regulator is also adjusted according to the sleep mode to
reduce the internal regulator power consumption.
13.5.7.3
Precautions when entering sleep mode
Modules communicating with external circuits should normally be disabled before entering a
sleep mode that will stop the module operation. This prevents erratic behavior when entering or
exiting sleep mode. Please refer to the relevant module documentation for recommended
actions.
Communication between the synchronous clock domains is disturbed when entering and exiting
sleep modes. This means that bus transactions are not allowed between clock domains affected
by the sleep mode. The system may hang if the bus clocks are stopped in the middle of a bus
transaction.
The CPU is automatically stopped in a safe state to ensure that all CPU bus operations are complete when the sleep mode goes into effect. Thus, when entering Idle mode, no further action is
necessary.
When entering a sleep mode (except Idle mode), all HSB masters must be stopped before
entering the sleep mode. Also, if there is a chance that any PB write operations are incomplete,
the CPU should perform a read operation from any register on the PB bus before executing the
sleep instruction. This will stall the CPU while waiting for any pending PB operations to
complete.
13.5.7.4
Wake up
The USB can be used to wake up the part from sleep modes through register PM_AWEN of the
Power Manager.
13.5.8
Generic clocks
Timers, communication modules, and other modules connected to external circuitry may require
specific clock frequencies to operate correctly. The Power Manager contains an implementation
defined number of generic clocks that can provide a wide range of accurate clock frequencies.
Each generic clock module runs from either Oscillator 0 or 1, or PLL0 or 1. The selected source
can optionally be divided by any even integer up to 512. Each clock can be independently
enabled and disabled, and is also automatically disabled along with peripheral clocks by the
Sleep Controller.
61
32058K
AVR32-01/12
AT32UC3A
Sleep
Controller
0
Osc0 clock
Osc1 clock
PLL0 clock
PLL1 clock
0
Divider
Mask
Generic Clock
1
1
PLLSEL
OSCSEL
DIV
DIVEN
CEN
Figure 13-5. Generic clock generation
13.5.8.1
Enabling a generic clock
A generic clock is enabled by writing the CEN bit in GCCTRL to 1. Each generic clock can use
either Oscillator 0 or 1 or PLL0 or 1 as source, as selected by the PLLSEL and OSCSEL bits.
The source clock can optionally be divided by writing DIVEN to 1 and the division factor to DIV,
resulting in the output frequency:
fGCLK = fSRC / (2*(DIV+1))
13.5.8.2
Disabling a generic clock
The generic clock can be disabled by writing CEN to 0 or entering a sleep mode that disables
the PB clocks. In either case, the generic clock will be switched off on the first falling edge after
the disabling event, to ensure that no glitches occur. If CEN is written to 0, the bit will still read as
1 until the next falling edge occurs, and the clock is actually switched off. When writing CEN to 0,
the other bits in GCCTRL should not be changed until CEN reads as 0, to avoid glitches on the
generic clock.
When the clock is disabled, both the prescaler and output are reset.
13.5.8.3
Changing clock frequency
When changing generic clock frequency by writing GCCTRL, the clock should be switched off by
the procedure above, before being re-enabled with the new clock source or division setting. This
prevents glitches during the transition.
62
32058K
AVR32-01/12
AT32UC3A
13.5.8.4
Generic clock implementation
In AT32UC3A, there are 6 generic clocks. These are allocated to different functions as shown in
Table 13-2.
Table 13-2.
Generic clock allocation
Clock number
13.5.9
Function
0
GCLK0 pin
1
GCLK1 pin
2
GCLK2 pin
3
GCLK3 pin
4
USBB
5
ABDAC
Divided PB clocks
The clock generator in the Power Manager provides divided PBA and PBB clocks for use by
peripherals that require a prescaled PBx clock. This is described in the documentation for the
relevant modules.
The divided clocks are not directly maskable, but are stopped in sleep modes where the PBx
clocks are stopped.
13.5.10
Debug operation
During a debug session, the user may need to halt the system to inspect memory and CPU registers. The clocks normally keep running during this debug operation, but some peripherals may
require the clocks to be stopped, e.g. to prevent timer overflow, which would cause the program
to fail. For this reason, peripherals on the PBA and PBB buses may use “debug qualified” PBx
clocks. This is described in the documentation for the relevant modules. The divided PBx clocks
are always debug qualified clocks.
Debug qualified PB clocks are stopped during debug operation. The debug system can optionally keep these clocks running during the debug operation. This is described in the
documentation for the On-Chip Debug system.
13.5.11
Reset Controller
The Reset Controller collects the various reset sources in the system and generates hard and
soft resets for the digital logic.
The device contains a Power-On Detector, which keeps the system reset until power is stable.
This eliminates the need for external reset circuitry to guarantee stable operation when powering
up the device.
63
32058K
AVR32-01/12
AT32UC3A
It is also possible to reset the device by asserting the RESET_N pin. This pin has an internal pullup, and does not need to be driven externally when negated. Table 13-4 lists these and other
reset sources supported by the Reset Controller.
RC _RCAUSE
RESET_N
P o w e r-O n
D e te c to r
CPU, HSB,
PBA, PBB
R eset
C o n tro lle r
B ro w n o u t
D e te c to r
O C D , R T C /W D T
C lo c k G e n e ra to
JT A G
OCD
W a tc h d o g R e s e t
Figure 13-6. Reset Controller block diagram
In addition to the listed reset types, the JTAG can keep parts of the device statically reset
through the JTAG Reset Register. See JTAG documentation for details.
Table 13-3.
Reset description
Reset source
Description
Power-on Reset
Supply voltage below the power-on reset detector threshold
voltage
External Reset
RESET_N pin asserted
Brownout Reset
Supply voltage below the brownout reset detector threshold
voltage
CPU Error
Caused by an illegal CPU access to external memory while
in Supervisor mode
Watchdog Timer
See watchdog timer documentation.
OCD
See On-Chip Debug documentation
When a Reset occurs, some parts of the chip are not necessarily reset, depending on the reset
source. Only the Power On Reset (POR) will force a reset of the whole chip.
64
32058K
AVR32-01/12
AT32UC3A
Table 13-4 lists parts of the device that are reset, depending on the reset source.
Effect of the different reset events
Table 13-4.
Power-On
Reset
External
Reset
Watchdog
Reset
BOD
Reset
CPU Error
Reset
OCD
Reset
CPU/HSB/PBA/PBB
(excluding Power Manager)
Y
Y
Y
Y
Y
Y
32 KHz oscillator
Y
N
N
N
N
N
RTC control register
Y
N
N
N
N
N
GPLP registers
Y
N
N
N
N
N
Watchdog control register
Y
Y
N
Y
Y
Y
Voltage Calibration register
Y
N
N
N
N
N
RC Oscillator Calibration register
Y
N
N
N
N
N
BOD control register
Y
Y
N
N
N
N
Bandgap control register
Y
Y
N
N
N
N
Clock control registers
Y
Y
Y
Y
Y
Y
Osc0/Osc1 and control registers
Y
Y
Y
Y
Y
Y
PLL0/PLL1 and control registers
Y
Y
Y
Y
Y
Y
OCD system and OCD registers
Y
Y
N
Y
Y
N
The cause of the last reset can be read from the RCAUSE register. This register contains one bit
for each reset source, and can be read during the boot sequence of an application to determine
the proper action to be taken.
13.5.11.1
Power-On Detector
The Power-On Detector monitors the VDDCORE supply pin and generates a reset when the
device is powered on. The reset is active until the supply voltage from the linear regulator is
above the power-on threshold level. The reset will be re-activated if the voltage drops below the
power-on threshold level. See Electrical Characteristics for parametric details.
13.5.11.2
Brown-Out Detector
The Brown-Out Detector (BOD) monitors the VDDCORE supply pin and compares the supply
voltage to the brown-out detection level, as set in BOD:LEVEL. The BOD is disabled by default,
but can be enabled either by software or by flash fuses. The Brown-Out Detector can either generate an interrupt or a reset when the supply voltage is below the brown-out detection level. In
any case, the BOD output is available in bit POSCR:BODET bit.
Note that any change to the BOD:LEVEL field of the BOD register should be done with the BOD
deactivated to avoid spurious reset or interrupt.
See Electrical Characteristics for parametric details.
65
32058K
AVR32-01/12
AT32UC3A
13.5.11.3
External Reset
The external reset detector monitors the state of the RESET_N pin. By default, a low level on
this pin will generate a reset.
13.5.12
Calibration registers
The Power Manager controls the calibration of the RC oscillator, voltage regulator, bandgap
voltage reference through several calibrations registers.
Those calibration registers are loaded after a Power On Reset with default values stored in factory-programmed flash fuses.
Although it is not recommended to override default factory settings, it is still possible to override
these default values by writing to those registers. To prevent unexpected writes due to software
bugs, write access to these registers is protected by a “key”. First, a write to the register must be
made with the field “KEY” equal to 0x55 then a second write must be issued with the “KEY” field
equal to 0xAA
13.6
User Interface
Offset
Register
Name
Access
Reset State
0x0000
Main Clock Control
MCCTRL
Read/Write
0x00000000
0x0004
Clock Select
CKSEL
Read/Write
0x00000000
0x0008
CPU Mask
CPUMASK
Read/Write
0x00000003
0x000C
HSB Mask
HSBMASK
Read/Write
0x0000007F
0x0010
PBA Mask
PBAMASK
Read/Write
0x0000FFFF
0x0014
PBB Mask
PBBMASK
Read/Write
0x0000003F
0x0018 - 0x001C
Reserved
0x0020
PLL0 Control
PLL0
Read/Write
0x00000000
0x0024
PLL1 Control
PLL1
Read/Write
0x00000000
0x0028
Oscillator 0 Control Register
OSCCTRL0
Read/Write
0x00000000
0x002C
Oscillator 1 Control Register
OSCCTRL1
Read/Write
0x00000000
0x0030
Oscillator 32 Control Register
OSCCTRL32
Read/Write
0x00000000
0x0034
Reserved
0x0038
Reserved
0x003C
Reserved
0x0040
PM Interrupt Enable Register
IER
Write Only
0x00000000
0x0044
PM Interrupt Disable Register
IDR
Write Only
0x00000000
0x0048
PM Interrupt Mask Register
IMR
Read Only
0x00000000
0x004C
PM Interrupt Status Register
ISR
Read Only
0x00000000
0x0050
PM Interrupt Clear Register
ICR
Write Only
0x00000000
0x0054
Power and Oscillators Status Register
POSCSR
Read/Write
0x00000000
0x0058 - 0x005C
Reserved
66
32058K
AVR32-01/12
AT32UC3A
0x0060
Generic Clock Control
GCCTRL
Read/Write
0x00000000
0x0064 - 0x00BC
Reserved
0x00C0
RC Oscillator Calibration Register
RCCR
Read/Write
Factory settings
0x00C4
Bandgap Calibration Register
BGCR
Read/Write
Factory settings
0x00C8
Linear Regulator Calibration Register
VREGCR
Read/Write
Factory settings
0x00CC
Reserved
0x00D0
BOD Level Register
BOD
Read/Write
BOD fuses in Flash
0x00D4 - 0x013C
Reserved
0x0140
Reset Cause Register
RCAUSE
Read Only
Latest Reset Source
0x0144 - 0x01FC
Reserved
0x0200
General Purpose Low-Power register 0
GPLP0
Read/Write
0x00000000
0x0204
General Purpose Low-Power register 1
GPLP1
Read/Write
0x00000000
67
32058K
AVR32-01/12
AT32UC3A
13.6.1
Main Clock Control
Name:
MCCTRL
Access Type:
Read/Write
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
OSC1EN
OSC0EN
MCSEL
• MCSEL: Main Clock Select
0: The slow clock is the source for the main clock
1: Oscillator 0 is source for the main clock
2: PLL0 is source for the main clock
3: Reserved
• OSC0EN: Oscillator 0 Enable
0: Oscillator 0 is disabled
1: Oscillator 0 is enabled
• OSC1EN: Oscillator 1 Enable
0: Oscillator 1is disabled
1: Oscillator 1is enabled
68
32058K
AVR32-01/12
AT32UC3A
13.6.2
Clock Select
Name:
CKSEL
Access Type:
Read/Write
31
30
29
28
27
PBBDIV
-
-
-
-
23
22
21
20
19
PBADIV
-
-
-
-
15
14
13
12
11
HSBDIV
-
-
-
-
7
6
5
4
3
CPUDIV
-
-
-
-
26
25
24
PBBSEL
18
17
16
PBASEL
10
9
8
HSBSEL
2
1
0
CPUSEL
• PBBDIV, PBBSEL: PBB Division and Clock Select
PBBDIV = 0: PBB clock equals main clock.
PBBDIV = 1: PBB clock equals main clock divided by 2(PBBSEL+1).
• PBADIV, PBASEL: PBA Division and Clock Select
PBADIV = 0: PBA clock equals main clock.
PBADIV = 1: PBA clock equals main clock divided by 2(PBASEL+1).
• HSBDIV, HSBSEL: HSB Division and Clock Select
For the AT32UC3A, HSBDIV always equals CPUDIV, and HSBSEL always equals CPUSEL, as the HSB clock is always equal to
the CPU clock.
• CPUDIV, CPUSEL: CPU Division and Clock Select
CPUDIV = 0: CPU clock equals main clock.
CPUDIV = 1: CPU clock equals main clock divided by 2(CPUSEL+1).
Note that if xxxDIV is written to 0, xxxSEL should also be written to 0 to ensure correct operation.
Also note that writing this register clears POSCSR:CKRDY. The register must not be re-written until CKRDY goes high.
69
32058K
AVR32-01/12
AT32UC3A
13.6.3
Clock Mask
Name:
CPU/HSB/PBA/PBBMASK
Access Type:
Read/Write
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
MASK[31:24]
23
22
21
20
MASK[23:16]
15
14
13
12
MASK[15:8]
7
6
5
4
MASK[7:0]
• MASK: Clock Mask
If bit n is cleared, the clock for module n is stopped. If bit n is set, the clock for module n is enabled according to the current
power mode. The number of implemented bits in each mask register, as well as which module clock is controlled by each bit, is
shown in Table 13-5.
Maskable module clocks in AT32UC3A.
Table 13-5.
Bit
CPUMASK
HSBMASK
PBAMASK
PBBMASK
0
-
FLASHC
INTC
HMATRIX
1
OCD
PBA bridge
GPIO
USBB
2
-
PBB bridge
PDCA
FLASHC
3
-
USBB
PM/RTC/EIC
MACB
4
-
MACB
ADC
SMC
5
-
PDCA
SPI0
SDRAMC
6
-
EBI
SPI1
-
7
-
-
TWI
-
8
-
-
USART0
-
9
-
-
USART1
-
10
-
-
USART2
-
11
-
-
USART3
-
12
-
-
PWM
-
13
-
-
SSC
-
70
32058K
AVR32-01/12
AT32UC3A
Maskable module clocks in AT32UC3A.
Table 13-5.
Bit
CPUMASK
HSBMASK
PBAMASK
PBBMASK
14
-
-
TC
-
15
-
-
ABDAC
-
16
SYSTIMER
(COMPARE/COUNT
REGISTERS CLK)
-
-
-
31:
17
-
-
-
-
71
32058K
AVR32-01/12
AT32UC3A
13.6.4
PLL Control
Name:
PLL0,1
Access Type:
Read/Write
31
30
29
28
27
RESERVED
23
26
25
24
18
17
16
9
8
1
0
PLLOSC
PLLEN
PLLCOUNT
22
21
20
19
RESERVED
15
14
PLLMUL
13
12
11
10
RESERVED
PLLDIV
7
6
5
-
-
-
4
3
PLLOPT
2
• RESERVED: Reserved bitfields
Reserved for internal use. Always write to 0.
• PLLCOUNT: PLL Count
Specifies the number of slow clock cycles before ISR:LOCKn will be set after PLLn has been written, or after PLLn has been
automatically re-enabled after exiting a sleep mode.
• PLLMUL: PLL Multiply Factor
• PLLDIV: PLL Division Factor
These bitfields determine the ratio of the PLL output frequency (voltage controlled oscillator frequency fVCO) to the source
oscillator frequency:
fVCO = (PLLMUL+1)/(PLLDIV) • fOSC if PLLDIV > 0.
fVCO = 2*(PLLMUL+1) • fOSC if PLLDIV = 0.
If PLLOPT[1] field is set to 0:
fPLL = fVCO.
If PLLOPT[1] field is set to 1:
fPLL = fVCO / 2.
Note that the MUL field cannot be equal to 0 or 1, or the behavior of the PLL will be undefined.
• PLLOPT: PLL Option
Select the operating range for the PLL.
PLLOPT[0]: Select the VCO frequency range.
PLLOPT[1]: Enable the extra output divider.
PLLOPT[2]: Disable the Wide-Bandwidth mode (Wide-Bandwidth mode allows a faster startup time and out-of-lock time).
72
32058K
AVR32-01/12
AT32UC3A
Table 13-6.
PLLOPT Fields Description in AT32UC3A
Description
PLLOPT[0]: VCO frequency
0
160MHz<fvco<240MHz
1
80MHz<fvco<180MHz
PLLOPT[1]: Output divider
0
fPLL = fvco
1
fPLL = fvco/2
0
Wide Bandwidth Mode enabled
1
Wide Bandwidth Mode disabled
PLLOPT[2]
• PLLOSC: PLL Oscillator Select
0: Oscillator 0 is the source for the PLL.
1: Oscillator 1 is the source for the PLL.
• PLLEN: PLL Enable
0: PLL is disabled.
1: PLL is enabled.
73
32058K
AVR32-01/12
AT32UC3A
13.6.5
PM Oscillator 0/1 Control
Register name
OSCCTRL0,1
Register access
Read/Write
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
7
6
5
4
3
-
-
-
-
-
STARTUP
2
1
0
MODE
• MODE: Oscillator Mode
Choose between crystal, or external clock
0: External clock connected on XIN, XOUT can be used as an I/O (no crystal)
1 to 3: reserved
4: Crystal is connected to XIN/XOUT - Oscillator is used with gain G0 ( XIN from 0.4
5: Crystal is connected to XIN/XOUT - Oscillator is used with gain G1 ( XIN from 0.9
6: Crystal is connected to XIN/XOUT - Oscillator is used with gain G2 ( XIN from 3.0
7: Crystal is connected to XIN/XOUT - Oscillator is used with gain G3 ( XIN from 8.0
MHz to 0.9 MHz ).
MHz to 3.0 MHz ).
MHz to 8.0 MHz ).
Mhz).
• STARTUP: Oscillator Startup Time
Select startup time for the oscillator.
Table 13-7.
Startup time for oscillators 0 and 1
Number of RC oscillator
clock cycle
Approximative Equivalent time
(RCOsc = 115 kHz)
0
0
0
1
64
560 us
2
128
1.1 ms
3
2048
18 ms
4
4096
36 ms
5
8192
71 ms
6
16384
142 ms
7
Reserved
Reserved
STARTUP
74
32058K
AVR32-01/12
AT32UC3A
13.6.6
PM 32 KHz Oscillator Control Register
Register name
OSCCTRL32
Register access
Read/Write
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
15
14
13
12
11
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
OSC32EN
STARTUP
10
9
8
MODE
Note: This register is only reset by Power-On Reset
• OSC32EN: Enable the 32 KHz oscillator
0: 32 KHz Oscillator is disabled
1: 32 KHz Oscillator is enabled
• MODE: Oscillator Mode
Choose between crystal, or external clock
0: External clock connected on XIN32, XOUT32 can be used as a I/O (no crystal)
1: Crystal is connected to XIN32/XOUT32
2 to 7: reserved
• STARTUP: Oscillator Startup Time
Select startup time for 32 KHz oscillator
Table 13-8.
Startup time for 32 KHz oscillator
Number of RC oscillator
clock cycle
Approximative Equivalent time
(RCOsc = 115 kHz)
0
0
0
1
128
1.1 ms
2
8192
72.3 ms
3
16384
143 ms
4
65536
570 ms
5
131072
1.1 s
6
262144
2.3 s
7
524288
4.6 s
STARTUP
75
32058K
AVR32-01/12
AT32UC3A
13.6.7
Interrupt Enable/Disable/Mask/Status/Clear
Name:
IER/IDR/IMR/ISR/ICR
Access Type:
IER/IDR/ICR: Write-only
IMR/ISR: Read-only
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
BODDET
15
14
13
12
11
10
9
8
-
-
-
-
-
-
OSC32RDY
OSC1RDY
7
6
5
4
3
2
1
0
OSC0RDY
MSKRDY
CKRDY
-
-
-
LOCK1
LOCK0
• BODDET: Brown out detection
Set to 1 when 0 to 1 transition on POSCSR:BODDET bit is detected: BOD has detected that power supply is going below
BOD reference value.
• OSC32RDY: 32 KHz oscillator Ready
Set to 1 when 0 to 1 transition on the POSCSR:OSC32RDY bit is detected: The 32 KHz oscillator is stable and ready to be
used as clock source.
• OSC1RDY: Oscillator 1 Ready
Set to 1 when 0 to 1 transition on the POSCSR:OSC1RDY bit is detected: Oscillator 1 is stable and ready to be used as
clock source.
• OSC0RDY: Oscillator 0 Ready
Set to 1 when 0 to 1 transition on the POSCSR:OSC1RDY bit is detected: Oscillator 1 is stable and ready to be used as
clock source.
• MSKRDY: Mask Ready
Set to 1 when 0 to 1 transition on the POSCSR:MSKRDY bit is detected: Clocks are now masked according to the
(CPU/HSB/PBA/PBB)_MASK registers.
• CKRDY: Clock Ready
0: The CKSEL register has been written, and the new clock setting is not yet effective.
1: The synchronous clocks have frequencies as indicated in the CKSEL register.
Note: Writing ICR:CKRDY to 1 has no effect.
• LOCK1: PLL1 locked
Set to 1 when 0 to 1 transition on the POSCSR:LOCK1 bit is detected: PLL 1 is locked and ready to be selected as clock
source.
• LOCK0: PLL0 locked
Set to 1 when 0 to 1 transition on the POSCSR:LOCK0 bit is detected: PLL 0 is locked and ready to be selected as clock
source.
76
32058K
AVR32-01/12
AT32UC3A
The effect of writing or reading the bits listed above depends on which register is being accessed:
• IER (Write-only)
•
•
•
•
0: No effect
1: Enable Interrupt
IDR (Write-only)
0: No effect
1: Disable Interrupt
IMR (Read-only)
0: Interrupt is disabled
1: Interrupt is enabled
ISR (Read-only)
0: An interrupt event has not occurred or has been previously cleared
1: An interrupt event has not occurred
ICR (Write-only)
0: No effect
1: Clear corresponding event
77
32058K
AVR32-01/12
AT32UC3A
13.6.8
Power and Oscillators Status
Name:
POSCSR
Access Type:
Read-only
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
BODDET
15
14
13
12
11
10
9
8
-
-
-
-
-
-
OSC32RDY
OSC1RDY
7
6
5
4
3
2
1
0
OSC0RDY
MSKRDY
CKRDY
-
-
-
LOCK1
LOCK0
• BODDET: Brown out detection
0: No BOD event
1: BOD has detected that power supply is going below BOD reference value.
• OSC32RDY: 32 KHz oscillator Ready
0: The 32 KHz oscillator is not enabled or not ready.
1: The 32 KHz oscillator is stable and ready to be used as clock source.
• OSC1RDY: OSC1 ready
•
•
•
•
•
0: Oscillator 1 not enabled or not ready.
1: Oscillator 1 is stable and ready to be used as clock source.
OSC0RDY: OSC0 ready
0: Oscillator 0 not enabled or not ready.
1: Oscillator 0 is stable and ready to be used as clock source.
MSKRDY: Mask ready
0: Mask register has been changed, masking in progress.
1: Clock are masked according to xxx_MASK
CKRDY:
0: The CKSEL register has been written, and the new clock setting is not yet effective.
1: The synchronous clocks have frequencies as indicated in the CKSEL register.
LOCK1: PLL 1 locked
0:PLL 1 is unlocked
1:PLL 1 is locked, and ready to be selected as clock source.
LOCK0: PLL 0 locked
0: PLL 0 is unlocked
1: PLL 0 is locked, and ready to be selected as clock source.
78
32058K
AVR32-01/12
AT32UC3A
13.6.9
Generic Clock Control
Name:
GCCTRL
Access Type:
Read/Write
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
DIV[7:0]
7
6
5
4
3
2
1
0
-
-
-
DIVEN
-
CEN
PLLSEL
OSCSEL
There is one GCCTRL register per generic clock in the design.
• DIV: Division Factor
• DIVEN: Divide Enable
0: The generic clock equals the undivided source clock.
1: The generic clock equals the source clock divided by 2*(DIV+1).
• CEN: Clock Enable
0: Clock is stopped.
1: Clock is running.
• PLLSEL: PLL Select
0: Oscillator is source for the generic clock.
1: PLL is source for the generic clock.
• OSCSEL: Oscillator Select
0: Oscillator (or PLL) 0 is source for the generic clock.
1: Oscillator (or PLL) 1 is source for the generic clock.
79
32058K
AVR32-01/12
AT32UC3A
13.6.10
Reset Cause
Name:
RCAUSE
Access Type:
Read-only
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
JTAGHARD
OCDRST
7
6
5
4
3
2
1
0
CPUERR
-
-
JTAG
WDT
EXT
BOD
POR
• POR Power-on Reset
The CPU was reset due to the supply voltage being lower than the power-on threshold level.
• BOD: Brown-out Reset
•
•
•
•
•
•
The CPU was reset due to the supply voltage being lower than the brown-out threshold level.
EXT: External Reset Pin
The CPU was reset due to the RESET pin being asserted.
WDT: Watchdog Reset
The CPU was reset because of a watchdog timeout.
JTAG: JTAG reset
The CPU was reset by setting the bit RC_CPU in the JTAG reset register.
CPUERR: CPU Error
The CPU was reset because it had detected an illegal access.
OCDRST: OCD Reset
The CPU was reset because the RES strobe in the OCD Development Control register has been written to one.
JTAGHARD: JTAG Hard Reset
The chip was reset by setting the bit RC_OCD in the JTAG reset register or by using the JTAG HALT instruction.
80
32058K
AVR32-01/12
AT32UC3A
13.6.11
BOD Control
BOD Level register
Register name
BOD
Register access
Read/Write
31
30
29
28
27
26
25
24
KEY
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
FCD
15
14
13
12
11
10
9
8
-
-
-
-
-
-
7
6
5
4
3
2
-
HYST
CTRL
1
0
LEVEL
• KEY: Register Write protection
This field must be written twice, first with key value 0x55, then 0xAA, for a write operation to have an effect.
• FCD: BOD Fuse calibration done
Set to 1 when CTRL, HYST and LEVEL fields has been updated by the Flash fuses after power-on reset or Flash fuses update
If one, the CTRL, HYST and LEVEL values will not be updated again by Flash fuses
Can be cleared to allow subsequent overwriting of the value by Flash fuses
• CTRL: BOD Control
0: BOD is off
1: BOD is enabled and can reset the chip
2: BOD is enabled and but cannot reset the chip. Only interrupt will be sent to interrupt controller, if enabled in the IMR register.
3: BOD is off
• HYST: BOD Hysteresis
0: No hysteresis
1: Hysteresis On
• LEVEL: BOD Level
This field sets the triggering threshold of the BOD. See Electrical Characteristics for actual voltage levels.
Note that any change to the LEVEL field of the BOD register should be done with the BOD deactivated to avoid spurious reset
or interrupt.
81
32058K
AVR32-01/12
AT32UC3A
13.6.12
RC Oscillator Calibration
Register name
RCCR
Register access
Read/Write
31
30
29
28
27
26
25
24
KEY
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
FCD
15
14
13
12
11
10
9
8
-
-
-
-
-
-
7
6
5
4
3
2
CALIB
1
0
CALIB
• CALIB: Calibration Value
Calibration Value for the RC oscillator.
• FCD: Flash Calibration Done
Set to 1 when CTRL, HYST, and LEVEL fields have been updated by the Flash fuses after power-on reset, or after Flash fuses
are reprogrammed. The CTRL, HYST and LEVEL values will not be updated again by the Flash fuses until a new power-on
reset or the FCD field is written to zero.
• KEY: Register Write protection
This field must be written twice, first with key value 0x55, then 0xAA, for a write operation to have an effect.
82
32058K
AVR32-01/12
AT32UC3A
13.6.13
Bandgap Calibration
Register name
BGCR
Register access
Read/Write
31
30
29
28
27
26
25
24
KEY
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
FCD
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
CALIB
• KEY: Register Write protection
This field must be written twice, first with key value 0x55, then 0xAA, for a write operation to have an effect.
• CALIB: Calibration value
Calibration value for Bandgap. See Electrical Characteristics for voltage values.
• FCD: Flash Calibration Done
Set to 1 when the CALIB field has been updated by the Flash fuses after power-on reset or when the Flash fuses are
reprogrammed. The CALIB field will not be updated again by the Flash fuses until a new power-on reset or the FCD field is
written to zero.
83
32058K
AVR32-01/12
AT32UC3A
13.6.14
PM Voltage Regulator Calibration Register
Register name
VREGCR
Register access
Read/Write
31
30
29
28
27
26
25
24
KEY
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
FCD
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
CALIB
• KEY: Register Write protection
This field must be written twice, first with key value 0x55, then 0xAA, for a write operation to have an effect.
• CALIB: Calibration value
Calibration value for Voltage Regulator. See Electrical Characteristics for voltage values.
• FCD: Flash Calibration Done
Set to 1 when the CALIB field has been updated by the Flash fuses after power-on reset or when the Flash fuses are
reprogrammed. The CALIB field will not be updated again by the Flash fuses until a new power-on reset or the FCD field is
written to zero.
84
32058K
AVR32-01/12
AT32UC3A
13.6.15
General Purpose Low-power register 0/1
Register name
GPLP0,1
Register access
Read/Write
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
GPLP
23
22
21
20
GPLP
15
14
13
12
GPLP
7
6
5
4
GPLP
These registers are general purpose 32-bit registers that are reset only by power-on-reset. Any other reset will keep the
content of these registers untouched.
85
32058K
AVR32-01/12
AT32UC3A
14. Real Time Counter (RTC)
Rev: 2.3.0.1
14.1
Features
•
•
•
•
32-bit real-time counter with 16-bit prescaler
Clocked from RC oscillator or 32 KHz oscillator
High resolution: Max count frequency 16 KHz
Long delays
– Max timeout 272 years
• Extremely low power consumption
• Available in all sleep modes except Static
• Interrupt on wrap
14.2
Description
The Real Time Counter (RTC) enables periodic interrupts at long intervals, or accurate measurement of real-time sequences. The RTC is fed from a 16-bit prescaler, which is clocked from
the RC oscillator or the 32 KHz oscillator. Any tapping of the prescaler can be selected as clock
source for the RTC, enabling both high resolution and long timeouts. The prescaler cannot be
written directly, but can be cleared by the user.
The RTC can generate an interrupt when the counter wraps around the value stored in the top
register, producing accurate periodic interrupts.
86
32058K
AVR32-01/12
AT32UC3A
14.3
Block Diagram
Figure 14-1. Real Time Counter module block diagram
RTC_CTRL
EN PCLR
CLK32
32 kHz
1
RC OSC
0
RTC_TOP
16-bit Prescaler
32-bit counter
TOPI
IRQ
RTC_VAL
14.4
Product Dependencies
14.4.1
Power Management
The RTC is continuously clocked, and remains operating in all sleep modes except Static. Interrupts are not available in DeepStop mode.
14.4.2
Interrupt
The RTC interrupt line is connected to one of the internal sources of the interrupt controller.
Using the RTC interrupt requires the interrupt controller to be programmed first.
14.4.3
Debug Operation
The RTC prescaler is frozen during debug operation, unless the OCD system keeps peripherals
running in debug operation.
14.4.4
Clocks
The RTC can use the internal RC oscillator as clock source. This oscillator is always enabled
whenever these modules are active. Please refer to the Electrical Characteristics chapter for the
characteristic frequency of this oscillator (fRC).
The RTC can also use the 32 KHz crystal oscillator as clock source. This oscillator must be
enabled before use. Please refer to the Power Manager chapter for details.
14.5
Functional Description
14.5.1
14.5.1.1
RTC operation
Source clock
The RTC is enabled by writing the EN bit in the CTRL register to 1. The 16-bit prescaler will then
increment on the selected clock. The prescaler cannot be read or written, but it can be reset by
writing the PCLR strobe.
87
32058K
AVR32-01/12
AT32UC3A
The CLK32 bit selects either the RC oscillator or the 32 KHz oscillator as clock source for the
prescaler.
The PSEL bitfield selects the prescaler tapping, selecting the source clock for the RTC:
fRTC = 2-(PSEL+1) * (fRC or 32 KHz)
14.5.1.2
Counter operation
When enabled, the RTC will increment until it reaches TOP, and then wrap to 0x0. The status bit
TOPI in ISR is set when this occurs. From 0x0 the counter will count TOP+1 cycles of the source
clock before it wraps back to 0x0.
The RTC count value can be read from or written to the register VAL. Due to synchronization,
continuous reading of the VAL with the lowest prescaler setting will skip every other value.
14.5.1.3
RTC Interrupt
Writing the TOPI bit in IER enables the RTC interrupt, while writing the corresponding bit in IDR
disables the RTC interrupt. IMR can be read to see whether or not the interrupt is enabled. If
enabled, an interrupt will be generated if the TOPI flag in ISR is set. The flag can be cleared by
writing TOPI in ICR to one.
The RTC interrupt can wake the CPU from all sleep modes except DeepStop and Static mode.
14.5.1.4
RTC wakeup
The RTC can also wake up the CPU directly without triggering an interrupt when the TOPI flag in
ISR is set. In this case, the CPU will continue executing from the instruction following the sleep
instruction.
This direct RTC wakeup is enabled by writing the WAKE_EN bit in the CTRL register to one.
When the CPU wakes from sleep, the WAKE_EN bit must be written to zero to clear the internal
wake signal to the sleep controller, otherwise a new sleep instruction will have no effect.
The RTC wakeup is available in all sleep modes except Static mode. The RTC wakeup can be
configured independently of the RTC interrupt.
14.5.1.5
Busy bit
Due to the crossing of clock domains, the RTC uses a few clock cycles to propagate the values
stored in CTRL, TOP, and VAL to the RTC. The BUSY bit in CTRL indicates that a register write
is still going on and all writes to TOP, CTRL, and VAL will be discarded until BUSY goes low
again.
88
32058K
AVR32-01/12
AT32UC3A
14.6
User Interface
Offset
Register
Register Name
Access
Reset
0x00
RTC Control
CTRL
Read/Write
0x0
0x04
RTC Value
VAL
Read/Write
0x0
0x08
RTC Top
TOP
Read/Write
0x0
0x10
RTC Interrupt Enable
IER
Write-only
0x0
0x14
RTC Interrupt Disable
IDR
Write-only
0x0
0x18
RTC Interrupt Mask
IMR
Read-only
0x0
0x1C
RTC Interrupt Status
ISR
Read-only
0x0
0x20
RTC Interrupt Clear
ICR
Write-only
0x0
14.6.1
RTC Control
Name:
CTRL
Access Type:
Read/Write
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
CLKEN
15
14
13
12
11
10
9
8
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
BUSY
CLK32
WAKE_EN
PCLR
EN
PSEL
• CLKEN: Clock enable
0: The clock is disabled
1: The clockis enabled
• PSEL: Prescale Select
Selects prescaler bit PSEL as source clock for the RTC.
• BUSY: RTC busy
0: The RTC accepts writes to TOP, VAL, and CTRL.
1: The RTC is busy and will discard writes to TOP, VAL, and CTRL.
• CLK32: 32 KHz oscillator select
0: The RTC uses the RC oscillator as clock source
1: The RTC uses the 32 KHz oscillator as clock source
89
32058K
AVR32-01/12
AT32UC3A
• WAKE_EN: Wakeup enable
0: The RTC does not wake up the CPU from sleep modes
1: The RTC wakes up the CPU from sleep modes.
• PCLR: Prescaler Clear
Writing 1 to this strobe clears the prescaler.
• EN: Enable
0: The RTC is disabled
1: The RTC is enabled
90
32058K
AVR32-01/12
AT32UC3A
14.6.2
RTC Value
Name:
VAL
Access Type:
Read/Write
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
VAL[31:24]
23
22
21
20
VAL[23:16]
15
14
13
12
VAL[15:8]
7
6
5
4
VAL[7:0]
• VAL: RTC Value
This value is incremented on every rising edge of the source clock.
91
32058K
AVR32-01/12
AT32UC3A
14.6.3
RTC Top
Name:
TOP
Access Type:
Read/Write
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
TOP[31:24]
23
22
21
20
TOP[23:16]
15
14
13
12
TOP[15:8]
7
6
5
4
TOP[7:0]
• TOP: RTC Top Value
VAL wraps at this value.
92
32058K
AVR32-01/12
AT32UC3A
14.6.4
RTC Interrupt Enable/Disable/Mask/Status/Clear
Name:
IER/IDR/IMR/ISR/ICR
Access Type:
IER/IDR/ICR: Write-only
IMR/ISR: Read-only
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
TOPI
• TOPI: Top Interrupt
VAL has wrapped at its top value.
The effect of writing or reading this bit depends on which register is being accessed:
• IER (Write-only)
•
•
•
•
0: No effect
1: Enable Interrupt
IDR (Write-only)
0: No effect
1: Disable Interrupt
IMR (Read-only)
0: Interrupt is disabled
1: Interrupt is enabled
ISR (Read-only)
0: An interrupt event has occurred
1: An interrupt even has not occurred
ICR (Write-only)
0: No effect
1: Clear interrupt even
93
32058K
AVR32-01/12
AT32UC3A
15. Watchdog Timer (WDT)
Rev: 2.3.0.1
15.1
Features
• Watchdog Timer counter with 16-bit prescaler
• Clocked from RC oscillator
15.2
Description
The Watchdog Timer (WDT) has a prescaler generating a timeout period. This prescaler is
clocked from the RC oscillator. The watchdog timer must be periodically reset by software within
the timeout period, otherwise, the device is reset and starts executing from the boot vector. This
allows the device to recover from a condition that has caused the system to be unstable.
15.3
Block Diagram
Figure 15-1. Watchdog Timer module block diagram
WDT_CLR
RCOSC
32-bit
Prescaler
EN
15.4
Watchdog
Detector
Watchdog Reset
WDT_CTRL
Product Dependencies
15.4.1
Power Management
When the WDT is enabled, the WDT remains clocked in all sleep modes, and it is not possible to
enter Static mode.
15.4.2
Debug Operation
The WDT prescaler is frozen during debug operation, unless the OCD system keeps peripherals
running in debug operation.
15.4.3
Clocks
The WDT can use the internal RC oscillator as clock source. This oscillator is always enabled
whenever these modules are active. Please refer to the Electrical Characteristics chapter for the
characteristic frequency of this oscillator (fRC).
94
32058K
AVR32-01/12
AT32UC3A
15.5
Functional Description
The WDT is enabled by writing the EN bit in the CTRL register to one. This also enables the RC
clock for the prescaler. The PSEL bitfield in the same register selects the watchdog timeout
period:
TWDT = 2(PSEL+1) / fRC
The next timeout period will begin as soon as the watchdog reset has occured and count down
during the reset sequence. Care must be taken when selecting the PSEL value so that the timeout period is greater than the startup time of the chip, otherwise a watchdog reset can reset the
chip before any code has been run.
To avoid accidental disabling of the watchdog, the CTRL register must be written twice, first with
the KEY field set to 0x55, then 0xAA without changing the other bitfields. Failure to do so will
cause the write operation to be ignored, and CTRL does not change value.
The CLR register must be written with any value with regular intervals shorter than the watchdog
timeout period. Otherwise, the device will receive a soft reset, and the code will start executing
from the boot vector.
When the WDT is enabled, it is not possible to enter Static mode. Attempting to do so will result
in entering Shutdown mode, leaving the WDT operational.
95
32058K
AVR32-01/12
AT32UC3A
15.6
User Interface
Offset
Register
Register Name
Access
Reset
0x00
WDT Control
CTRL
Read/Write
0x0
0x04
WDT Clear
CLR
Write-only
0x0
96
32058K
AVR32-01/12
AT32UC3A
15.6.1
WDT Control
Name:
CTRL
Access Type:
Read/Write
31
30
29
28
27
26
25
24
KEY[7:0]
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
EN
PSEL
• KEY
This bitfield must be written twice, first with key value 0x55, then 0xAA, for a write operation to be effective. This bitfield always
reads as zero.
• PSEL: Prescale Select
Prescaler bit PSEL is used as watchdog timeout period.
• EN: WDT Enable
0: WDT is disabled.
1: WDT is enabled.
97
32058K
AVR32-01/12
AT32UC3A
15.6.2
WDT Clear
Name:
CLR
Access Type:
Write-only
When the watchdog timer is enabled, this register must be periodically written, with any value, within the watchdog timeout
period, to prevent a watchdog reset.
98
32058K
AVR32-01/12
AT32UC3A
16. Interrupt Controller (INTC)
Rev: 1.0.1.1
16.1
Description
The INTC collects interrupt requests from the peripherals, prioritizes them, and delivers an interrupt request and an autovector to the CPU. The AVR32 architecture supports 4 priority levels for
regular, maskable interrupts, and a Non-Maskable Interrupt (NMI).
The INTC supports up to 64 groups of interrupts. Each group can have up to 32 interrupt request
lines, these lines are connected to the peripherals. Each group has an Interrupt Priority Register
(IPR) and an Interrupt Request Register (IRR). The IPRs are used to assign a priority level and
an autovector to each group, and the IRRs are used to identify the active interrupt request within
each group. If a group has only one interrupt request line, an active interrupt group uniquely
identifies the active interrupt request line, and the corresponding IRR is not needed. The INTC
also provides one Interrupt Cause Register (ICR) per priority level. These registers identify the
group that has a pending interrupt of the corresponding priority level. If several groups have an
pending interrupt of the same level, the group with the lowest number takes priority.
16.2
Block Diagram
Figure 16-1 on page 99 gives an overview of the INTC. The grey boxes represent registers that
can be accessed via the Peripheral Bus (PB). The interrupt requests from the peripherals
(IREQn) and the NMI are input on the left side of the figure. Signals to and from the CPU are on
the right side of the figure.
Figure 16-1. Overview of the Interrupt Controller
Interrupt Controller
CPU
NMIREQ
Masks
OR
ValReqN
GrpReqN
IPRn
IRRn
IREQ34
IREQ33
IREQ32
OR
GrpReq1
Request
masking
INT_level, offset
INTLEVEL
Prioritizer
IREQ63
SREG
masks
I[3-0]M
GM
ValReq1
IPR1
AUTOVECTOR
INT_level, offset
IRR1
IREQ31
IREQ2
IREQ1
IREQ0
OR
GrpReq0
ValReq0
IPR0
INT_level, offset
IRR0
IRR registers
16.3
IPR registers
ICR registers
Operation
All of the incoming interrupt requests (IREQs) are sampled into the corresponding Interrupt
Request Register (IRR). The IRRs must be accessed to identify which IREQ within a group that
is active. If several IREQs within the same group is active, the interrupt service routine must pri-
99
32058K
AVR32-01/12
AT32UC3A
oritize between them. All of the input lines in each group are logically-ORed together to form the
GrpReqN lines, indicating if there is a pending interrupt in the corresponding group.
The Request Masking hardware maps each of the GrpReq lines to a priority level from INT0 to
INT3 by associating each group with the INTLEVEL field in the corresponding IPR register. The
GrpReq inputs are then masked by the I0M, I1M, I2M, I3M and GM mask bits from the CPU status register. Any interrupt group that has a pending interrupt of a priority level that is not masked
by the CPU status register, gets its corresponding ValReq line asserted.
The Prioritizer hardware uses the ValReq lines and the INTLEVEL field in the IPRs to select the
pending interrupt of the highest priority. If a NMI interrupt is pending, it automatically gets highest priority of any pending interrupt. If several interrupt groups of the highest pending interrupt
level have pending interrupts, the interrupt group with the highest number is selected.
Interrupt level (INTLEVEL) and handler autovector offset (AUTOVECTOR) of the selected interrupt are transmitted to the CPU for interrupt handling and context switching. The CPU doesn't
need to know which interrupt is requesting handling, but only the level and the offset of the handler address. The IRR registers contain the interrupt request lines of the groups and can be read
via PB for checking which interrupts of the group are actually active.
Masking of the interrupt requests is done based on five interrupt mask bits of the CPU status
register, namely interrupt level 3 mask (I3M) to interrupt level 0 mask (I0M), and Global interrupt
mask (GM). An interrupt request is masked if either the Global interrupt mask or the corresponding interrupt level mask bit is set.
16.3.1
Non maskable interrupts
A NMI request has priority over all other interrupt requests. NMI has a dedicated exception vector address defined by the AVR32 architecture, so AUTOVECTOR is undefined when
INTLEVEL indicates that an NMI is pending.
16.3.2
CPU response
When the CPU receives an interrupt request it checks if any other exceptions are pending. If no
exceptions of higher priority are pending, interrupt handling is initiated. When initiating interrupt
handling, the corresponding interrupt mask bit is set automatically for this and lower levels in status register. E.g, if interrupt on level 3 is approved for handling the interrupt mask bits I3M, I2M,
I1M, and I0M are set in status register. If interrupt on level 1 is approved the masking bits I1M,
and I0M are set in status register. The handler offset is calculated from AUTOVECTOR and
EVBA and a change-of-flow to this address is performed.
Setting of the interrupt mask bits prevents the interrupts from the same and lower levels to be
passed trough the interrupt controller. Setting of the same level mask bit prevents also multiple
request of the same interrupt to happen.
It is the responsibility of the handler software to clear the interrupt request that caused the interrupt before returning from the interrupt handler. If the conditions that caused the interrupt are not
cleared, the interrupt request remains active.
16.3.3
Clearing an interrupt request
Clearing of the interrupt request is done by writing to registers in the corresponding peripheral
module, which then clears the corresponding NMIREQ/IREQ signal.
The recommended way of clearing an interrupt request is a store operation to the controlling
peripheral register, followed by a dummy load operation from the same register. This causes a
100
32058K
AVR32-01/12
AT32UC3A
pipeline stall, which prevents the interrupt from accidentally re-triggering in case the handler is
exited and the interrupt mask is cleared before the interrupt request is cleared.
16.4
User Interface
This chapter lists the INTC registers are accessible through the PB bus. The registers are used
to control the behaviour and read the status of the INTC.
16.4.1
Memory Map
The following table shows the address map of the INTC registers, relative to the base address of
the INTC.
INTC address map
Table 16-1.
16.4.2
Offset
Register
Name
Access
Reset Value
0
Interrupt Priority Register 0
IPR0
Read/Write
0x0000_0000
4
Interrupt Priority Register 1
IPR1
Read/Write
0x0000_0000
...
...
...
...
...
252
Interrupt Priority Register 63
IPR63
Read/Write
0x0000_0000
256
Interrupt Request Register 0
IRR0
Read-only
N/A
260
Interrupt Request Register 1
IRR1
Read-only
N/A
...
...
...
...
...
508
Interrupt Request Register 63
IRR63
Read-only
N/A
512
Interrupt Cause Register 3
ICR3
Read-only
N/A
516
Interrupt Cause Register 2
ICR2
Read-only
N/A
520
Interrupt Cause Register 1
ICR1
Read-only
N/A
524
Interrupt Cause Register 0
ICR0
Read-only
N/A
Interrupt Request Map
The mapping of interrupt requests from peripherals to INTREQs is presented in the Peripherals
Section.
101
32058K
AVR32-01/12
AT32UC3A
16.4.3
Interrupt Request Registers
Register Name:
IRR0...IRR63
Access Type:
Read-only
31
IRR(32*x+31)
30
IRR(32*x+30)
29
IRR(32*x+29)
28
IRR(32*x+28)
27
IRR(32*x+27)
26
IRR(32*x+26)
25
IRR(32*x+25)
24
IRR(32*x+24)
23
IRR(32*x+23)
22
IRR(32*x+22)
21
IRR(32*x+21)
20
IRR(32*x+20)
19
IRR(32*x+19)
18
IRR(32*x+18)
17
IRR(32*x+17)
16
IRR(32*x+16)
15
IRR(32*x+15)
14
IRR(32*x+14)
13
IRR(32*x+13)
12
IRR(32*x+12)
11
IRR(32*x+11)
10
IRR(32*x+10)
9
IRR(32*x+9)
8
IRR(32*x+8)
7
IRR(32*x+7)
6
IRR(32*x+6)
5
IRR(32*x+5)
4
IRR(32*x+4)
3
IRR(32*x+3)
2
IRR(32*x+2)
1
IRR(32*x+1)
0
IRR(32*x+0)
• IRR: Interrupt Request line
0 = No interrupt request is pending on this input request input.
1 = An interrupt request is pending on this input request input.
The are 64 IRRs, one for each group. Each IRR has 32 bits, one for each possible interrupt request, for a total of 2048 possible input lines. The IRRs are read by the software interrupt handler in order to determine which interrupt request is
pending. The IRRs are sampled continuously, and are read-only.
102
32058K
AVR32-01/12
AT32UC3A
16.4.4
Interrupt Priority Registers
Register Name:
IPR0...IPR63
Access Type:
Read/Write
31
30
INTLEVEL[1:0]
29
-
28
-
27
-
26
-
25
-
24
-
23
-
22
-
21
-
20
-
19
-
18
-
17
-
16
-
15
-
14
-
13
12
9
8
7
6
5
4
3
AUTOVECTOR[7:0]
1
0
11
10
AUTOVECTOR[13:8]
2
• INTLEVEL: Interrupt level associated with this group
Indicates the EVBA-relative offset of the interrupt handler of the corresponding group:
INTLEVEL[1:0]
Priority
0
0
INT0
0
1
INT1
1
0
INT2
1
1
INT3
• AUTOVECTOR: Autovector address for this group
Handler offset is used to give the address of the interrupt handler. The least significant bit should be written to zero to give
halfword alignment
103
32058K
AVR32-01/12
AT32UC3A
16.4.5
Interrupt Cause Registers
Register Name:
ICR0...ICR3
Access Type:
Read-only
31
-
30
-
29
-
28
-
27
-
26
-
25
-
24
-
23
-
22
-
21
-
20
-
19
-
18
-
17
-
16
-
15
-
14
-
13
-
12
-
11
-
10
-
9
-
8
-
7
-
6
-
5
4
3
2
1
0
CAUSE
• CAUSE: Interrupt group causing interrupt of priority n
ICRn identifies the group with the highest priority that has a pending interrupt of level n. If no interrupts of level n are pending, or the priority level is masked, the value of ICRn is UNDEFINED.
104
32058K
AVR32-01/12
AT32UC3A
17. External Interrupts Controller (EIC)
Rev: 2.3.0.2
17.1
Features
•
•
•
•
•
•
•
•
17.2
Dedicated interrupt requests for each interrupt
Individually maskable interrupts
Interrupt on rising or falling edge
Interrupt on high or low level
Asynchronous interrupts for sleep modes without clock
Filtering of interrupt lines
Keypad scan support
Maskable NMI interrupt
Description
The External Interrupt Module allows pins to be configured as external interrupts. Each pin has
its own interrupt request and can be individually masked. Each pin can generate an interrupt on
rising or falling edge, or high or low level. Every line has a configurable filter too remove spikes
on the interrupt lines. Every interrupt pin can also be configured to be asynchronous to wake up
the part from sleep modes where the clock has been disabled.
A Non-Maskable Interrupt (NMI) is also supported. This has the same properties as the other
external interrupts, but is connected to the NMI request of the CPU, enabling it to interrupt any
other interrupt mode.
The External Interrupt Module has support for keypad scanning for keypads laid out in rows and
columns. Columns are driven by a separate set of scanning outputs, while rows are sensed by
the external interrupt lines. The pressed key will trigger an interrupt, which can be identified
through the user registers of the module.
The External Interrupt Module can wake up the part from sleep modes without triggering an
interrupt. In this mode, code execution starts from the instruction following the sleep instruction.
105
32058K
AVR32-01/12
AT32UC3A
17.3
Block Diagram
Figure 17-1. External Interrupt Module block diagram
EIM_LEVEL
EIM_MODE
EIM_EDGE
Polarity
control
EIM_EN
EIM_DIS
EXTINTn
NMI
Enable
Asynchronus
detector
EIM_FILTER
EIM_LEVEL
EIM_MODE
EIM_EDGE
INTn
Filter
Edge/Level
Detector
EIM_ISR
EIM_CTRL
EIM_IER
EIM_IDR
EIM_ICR
Wake
detect
Mask
IRQn
EIM_IMR
EIM_WAKE
RC clk
Prescaler
PRESC
Shifter
EN
SCAN
PIN
EIM_SCAN
17.4
Product Dependencies
17.4.1
I/O Lines
The External Interrupt and keypad scan pins are multiplexed with PIO lines. To act as external
interrupts, these pins must be configured as inputs pins by the PIO controller. It is also possible
to trigger the interrupt by driving these pins from registers in the PIO controller, or another
peripheral output connected to the same pin.
17.4.2
Power Management
All interrupts are available in every sleep mode. However, in sleep modes where the clock is
stopped, asynchronous interrupts must be selected.
17.4.3
Interrupt
The external interrupt lines are connected to internal sources of the interrupt controller. Using
the external interrutps requires the interrupt controller to be programmed first.
Using the Non-Maskable Interrupt does not require the interrupt controller to be programmed.
106
32058K
AVR32-01/12
AT32UC3A
17.5
Functional Description
17.5.1
External Interrupts
To enable an external interrupt EXTINTn must be written to 1 in register EN. Similarly, writing
EXTINTn to 1 in register DIS disables the interrupt. The status of each Interrupt line can be
observed in the CTRL register.
Each external interrupt pin EXTINTn can be configured to produce an interrupt on rising or falling edge, or high or low level. External interrupts are configured by the MODE, EDGE, and
LEVEL registers. Each interrupt n has a bit INTn in each of these registers.
Similarly, each interrupt has a corresponding bit in each of the interrupt control and status registers. Writing 1 to the INTn strobe in IER enables the external interrupt on pin EXTINTn, while
writing 1 to INTn in IDR disables the external interrupt. IMR can be read to check which interrupts are enabled. When the interrupt triggers, the corresponding bit in ISR will be set. The flag
remains set until the corresponding strobe bit in ICR is written to 1.
Writing INTn in MODE to 0 enables edge triggered interrupts, while writing the bit to 1 enables
level triggered interrupts.
If EXTINTn is configured as an edge triggered interrupt, writing INTn in EDGE to 0 will trigger the
interrupt on falling edge, while writing the bit to 1 will trigger the interrupt on rising edge.
If EXTINTn is configured as a level triggered interrupt, writing INTn in LEVEL to 0 will trigger the
interrupt on low level, while writing the bit to 1 will trigger the interrupt on high level.
To remove spikes that are longer than the clock period in the current mode each external interrupt contains a filter that can be enabled by writing 1 to INTn to FILTER.
Each interrupt line can be made asynchronous by writing 1 to INTn in the ASYNC register. This
will route the interrupt signal through the asynchronous path of the module. All edge interrupts
will be interpreted as level interrupts and the filter is disabled.
17.5.1.1
Synchronization of external interrupts
The pin value of the EXTINTn pins is normally synchronized to the CPU clock, so spikes shorter
than a CPU clock cycle are not guaranteed to produce an interrupt. In Stop mode, spikes shorter
than a 32 KHz clock cycle are not guaranteed to produce an interrupt.
In Static mode, only unsynchronized interrupts remain active, and any short spike on this interrupt will wake up the device.
17.5.1.2
Wakeup
The External interrupts can be used to wake up the part from sleep modes. The wakeup can be
interpreted in two ways. If the corresponding bit in IMR is set, then the execution starts at the
interrupt handler for this interrupt. If the bit in IMR is not set, then the execution starts from the
next instruction after the sleep instruction.
17.5.2
Non-Maskable Interrupt
The NMI supports the same features as the external interrupts, and is accessed through the
same registers. The description in Section 17.5.1 should be followed, accessing the NMI bit
instead of the INTn bits.
107
32058K
AVR32-01/12
AT32UC3A
The NMI is non-maskable within the CPU in the sense that it can interrupt any other execution
mode. Still, as for the other external interrupts, the actual NMI input line can be enabled and disabled by accessing the registers in the External Interrupt Module. These interrupts are not
enabled by default, allowing the proper interrupt vectors to be set up by the CPU before the
interrupts are enabled.
17.5.3
Keypad scan support
The External Interrupt Module also includes support for keypad scanning. The keypad scan feature is compatible with keypads organized as rows and columns, where a row is shorted against
a column when a key is pressed.
The rows should be connected to the external interrupt pins with pullups enabled in the GPIO
module. These external interrupts should be enabled as low level or falling edge interrupts. The
columns should be connected to the available scan pins. The GPIO must be configured to let the
required scan pins be controlled by the EIC module. Unused external interrupt or scan pins can
be left controlled by the GPIO or other peripherals.
The Keypad Scan function is enabled by writing :EN to 1, which starts the keypad scan counter.
The SCAN outputs are tristated, except SCAN[0], which is driven to zero. After 2(SCAN:PRESC+1)
RC clock cycles this pattern is left shifted, so that SCAN[1] is driven to zero while the other outputs are tristated. This sequence repeats infinitely, wrapping from the most significant SCAN pin
to SCAN[0].
When a key is pressed, the pulled-up row is driven to zero by the column, and an external interrupt triggers. The scanning stops, and the software can then identify the key pressed by the
interrupt status register and the SCAN:PINS value.
The scanning stops whenever there is an active interrupt request from the EIC to the CPU.
When the CPU clears the interrupt flags, scanning resumes.
108
32058K
AVR32-01/12
AT32UC3A
17.6
User Interface
Offset
Register
Register Name
Access
Reset
0x00
EIC Interrupt Enable
IER
Write-only
0x0
0x04
EIC Interrupt Disable
IDR
Write-only
0x0
0x08
EIC Interrupt Mask
IMR
Read-only
0x0
0x0C
EIC Interrupt Status
ISR
Read-only
0x0
0x10
EIC Interrupt Clear
ICR
Write-only
0x0
0x14
External Interrupt Mode
MODE
Read/Write
0x0
0x18
External Interrupt Edge
EDGE
Read/Write
0x0
0x1C
External Interrupt Level
LEVEL
Read/Write
0x0
0x20
External Interrupt Filter
FILTER
Read/Write
0x0
0x24
External Interrupt Test
TEST
Read/Write
0x0
0x28
External Interrupt Asynchronous
ASYNC
Read/Write
0x0
0x2C
External Interrupt Scan
SCAN
Read/Write
0x0
0x30
External Interrupt Enable
EN
Write-only
0x0
0x34
External Interrupt Disable
DIS
Write-only
0x0
0x38
External Interrupt Control
CTRL
Read/Write
0x0
109
32058K
AVR32-01/12
AT32UC3A
17.6.1
EIC Interrupt Enable/Disable/Mask/Status/Clear
Name:
IER/IDR/IMR/ISR/ICR
Access Type:
IER/IDR/ICR: Write-only
IMR/ISR: Read-only
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
NMI
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
INT0
The effect of writing or reading the bits listed above depends on which register is being accessed:
• IER (Write-only)
•
•
•
•
0: No effect
1: Enable Interrupt
IDR (Write-only)
0: No effect
1: Disable Interrupt
IMR (Read-only)
0: Interrupt is disabled
1: Interrupt is enabled
ISR (Read-only)
0: An interrupt event has occurred
1: An interrupt even has not occurred
ICR (Write-only)
0: No effect
1: Clear interrupt event
110
32058K
AVR32-01/12
AT32UC3A
17.6.2
External Interrupt Mode/Edge/Level/Filter/Async
Name:
MODE/EDGE/LEVEL/FILTER/ASYNC
Access Type:
Read/Write
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
NMI
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
INT0
The bit interpretation is register specific:
• MODE
0: Interrupt is edge triggered
1: Interrupt is level triggered
• EDGE
0: Interrupt triggers on falling edge
1: Interrupt triggers on rising edge
• LEVEL
0: Interrupt triggers on low level
1: Interrupt triggers on high level
• FILTER
0: Interrupt is not filtered
1: Interrupt is filtered
• ASYNC
0: Interrupt is synchronized to the clock
1: Interrupt is asynchronous
111
32058K
AVR32-01/12
AT32UC3A
17.6.3
External Interrupt Test
Name:
TEST
Access Type:
Read/Write
31
30
29
28
27
26
25
24
TEST_EN
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
NMI
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
INT0
• NMI
If TEST_EN is 1, the value of this bit will be the value to the interrupt detector and the value on the pad will be ignored.
• INTn
If TEST_EN is 1, the value of this bit will be the value to the interrupt detector and the value on the pad will be ignored.
• TEST_EN
0: External interrupt test is disabled
1: External interrupt test is enabled
112
32058K
AVR32-01/12
AT32UC3A
17.6.4
External Interrupt Scan
Name:
SCAN
Access Type:
Read/Write
31
30
29
28
27
26
25
24
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
EN
PIN[2:0]
PRESC[4:0]
• EN
0: Keypad scanning is disabled
1: Keypad scanning is enabled
• PRESC
Prescale select for the keypad scan rate:
Scan rate = 2(SCAN:PRESC+1) TRC
The RC clock period can be found in the Electrical Characteristics section.
• PIN
The index of the currently active scan pin. Writing to this bitfield has no effect.
113
32058K
AVR32-01/12
AT32UC3A
17.6.5
External Interrupt Enable/Disable/Control
Name:
EN/DIS/CTRL
Access Type:
EN/DIS: Write-only
CTRL: Read-only
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
NMI
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
INT0
The bit interpretation is register specific:
• EN
0: No effect
1: Interrupt is enabled
• DIS
0: No effect
1: Interrupt is disabled
• CTRL
0: Interrupt is disabled
1: Interrupt is enabled
114
32058K
AVR32-01/12
AT32UC3A
18. Flash Controller (FLASHC)
Rev: 2.0.0.2
18.1
Features
• Controls flash block with dual read ports allowing staggered reads.
• Supports 0 and 1 wait state bus access.
• Allows interleaved burst reads for systems with one wait state, outputting one 32-bit word per
clock cycle.
• 32-bit HSB interface for reads from flash array and writes to page buffer.
• 32-bit PB interface for issuing commands to and configuration of the controller.
• 16 lock bits, each protecting a region consisting of (total number of pages in the flash block /
16) pages.
Regions can be individually protected or unprotected.
Additional protection of the Boot Loader pages.
Supports reads and writes of general-purpose NVM bits.
Supports reads and writes of additional NVM pages.
Supports device protection through a security bit.
Dedicated command for chip-erase, first erasing all on-chip volatile memories before erasing
flash and clearing security bit.
• Interface to Power Manager for power-down of flash-blocks in sleep mode.
•
•
•
•
•
•
18.2
Description
The flash controller (FLASHC) interfaces a flash block with the 32-bit internal HSB bus. Performance for uncached systems with high clock-frequency and one wait state is increased by
placing words with sequential addresses in alternating flash subblocks. Having one read interface per subblock allows them to be read in parallel. While data from one flash subblock is
being output on the bus, the sequential address is being read from the other flash subblock
and will be ready in the next clock cycle.
The controller also manages the programming, erasing, locking and unlocking sequences with
dedicated commands.
18.3
Product dependencies
18.3.1
Power management
The HFLASHC has two bus clocks connected: One High speed bus clock
(CLK_FLASHC_HSB) and one Peripheral bus clock (CLK_FLASHC_PB). These clocks are
generated by the Power manager. Both clocks are turned on by default, but the user has to
ensure that CLK_FLASHC_HSB is not turned off before reading the flash or writing the pagebuffer and that CLK_FLASHC_PB is not turned of before accessing the FLASHC configuration
and control registers.
18.3.2
Interrupt
The FLASHC interrupt lines are connected to internal sources of the interrupt controller. Using
FLASHC interrutps requires the interrupt controller to be programmed first.
115
32058K
AVR32-01/12
AT32UC3A
18.4
Functional description
18.4.1
Bus interfaces
The FLASHC has two bus interfaces, one High-Speed Bus (HSB) interface for reads from the
flash array and writes to the page buffer, and one Peripheral Bus (PB) interface for writing
commands and control to and reading status from the controller.
18.4.2
Memory organization
To maximize performance for high clock-frequency systems, FLASHC interfaces to a flash
block with two read ports. The flash block has several parameters, given by the design of the
flash block. Refer to the “Memories” chapter for the device-specific values of the parameters.
• p pages (FLASH_P)
• w words in each page and in the page buffer (FLASH_W)
• pw words in total (FLASH_PW)
• f general-purpose fuse bits (FLASH_F)
• 1 security fuse bit
• 1 User Page
18.4.3
User page
The User page is an additional page, outside the regular flash array, that can be used to store
various data, like calibration data and serial numbers. This page is not erased by regular chip
erase. The User page can only be written and erased by proprietary commands. Read
accesses to the User page is performed just as any other read access to the flash. The
address map of the User page is given in Figure 18-1.
18.4.4
Read operations
The FLASHC provides two different read modes:
• 0 wait state (0ws) for clock frequencies < (access time of the flash plus the bus delay)
• 1 wait state (1ws) for clock frequencies < (access time of the flash plus the bus delay)/2
Higher clock frequencies that would require more wait states are not supported by the flash
controller.
The programmer can select the wait states required by writing to the FWS field in the flash
control register (FCR). It is the responsibility of the programmer to select a number of wait
states compatible with the clock frequency and timing characteristics of the flash block.
In 0ws mode, only one of the two flash read ports is accessed. The other flash read port is idle.
In 1ws mode, both flash read ports are active. One read port reading the addressed word, and
the other reading the next sequential word.
If the clock frequency allows, the user should use 0ws mode, because this gives the lowest
power consumption for low-frequency systems as only one flash read port is read. Using 1ws
mode has a power/performance ratio approaching 0ws mode as the clock frequency
approaches twice the max frequency of 0ws mode. Using two flash read ports use twice the
power, but also give twice the performance.
116
32058K
AVR32-01/12
AT32UC3A
The flash controller supports flash blocks with up to 2^21 word addresses, as displayed in Figure 18-1. Reading the memory space between address pw and 2^21-1 returns an undefined
result. The User page is permanently mapped to word address 2^21.
Table 18-1.
User row addresses
Memory type
Start address, byte sized
Size
Main array
0
pw words = 4pw bytes
User
2^23 = 8388608
128 words = 512 bytes
Figure 18-1. Memory map for the Flash memories
A ll a d d r e s s e s a r e w o r d a d d r e s s e s
U nused
U ser page
Unused
2^21+128
2^21
Flash data array
pw
p w -1
0
F la s h w it h
e x tra p a g e
18.4.5
Quick Page Read
A dedicated command, Quick Page Read (QPR), is provided to read all words in an
addressed page. All bits in all words in this page are AND’ed together, returning a 1-bit result.
This result is placed in the Quick Page Read Result (QPRR) bit in Flash Status Register
(FSR). The QPR command is useful to check that a page is in an erased state. The QPR
instruction is much faster than performing the erased-page check using a regular software
subroutine.
18.4.6
Write page buffer operations
The internal memory area reserved for the embedded flash can also be written through a
write-only page buffer. The page buffer is addressed only by the address bits required to
address w words (since the page buffer is word addressable) and thus wrap around within the
internal memory area address space and appear to be repeated within it.
When writing to the page buffer, the PAGEN field in the FCMD register is updated with the
page number corresponding to page address of the latest word written into the page buffer.
117
32058K
AVR32-01/12
AT32UC3A
The page buffer is also used for writes to the User page.
Write operations can be prevented by programming the Memory Protection Unit of the CPU.
Writing 8-bit and 16-bit data to the page buffer is not allowed and may lead to unpredictable
data corruption.
Page buffer write operations are performed with 4 wait states.
Writing to the page buffer can only change page buffer bits from one to zero, ie writing
0xaaaaaaaa to a page buffer location that has the value 0x00000000, will not change the page
buffer value. The only way to change a bit from zero to one, is to reset the entire page buffer
with the Clear Page Buffer command.
The page buffer is not automatically reset after a page write. The programmer should do this
manually by issuing the Clear Page Buffer flash command. This can be done after a page
write, or before the page buffer is loaded with data to be stored to the flash page.
Example: Writing a word into word address 130 of a flash with 128 words in the page buffer.
PAGEN will be updated with the value 1, and the word will be written into word 2 in the page
buffer.
18.4.7
Writing words to a page that is not completely erased
This can be used for EEPROM emulation, i.e. writes with granularity of one word instead of an
entire page. Only words that are in an completely erased state (0xFFFFFFFF) can be
changed. The procedure is as follows:
1. Clear page buffer
2. Write to the page buffer the result of the logical bitwise AND operation between the
contents of the flash page and the new data to write. Only words that were in an
erased state can be changed from the original page.
3. Write Page.
18.5
Flash commands
The FLASHC offers a command set to manage programming of the flash memory, locking and
unlocking of regions, and full flash erasing. See chapter 18.8.3 for a complete list of
commands.
To run a command, the field CMD of the Flash Command Register (FCMD) has to be written
with the command number. As soon as the FCMD register is written, the FRDY flag is automatically cleared. Once the current command is complete, the FRDY flag is automatically set.
If an interrupt has been enabled by setting the bit FRDY in FCR, the interrupt line of the flash
controller is activated. All flash commands except for Quick Page Read (QPR) will generate an
interrupt request upon completion if FRDY is set.
After a command has been written to FCMD, the programming algorithm should wait until the
command has been executed before attempting to read instructions or data from the flash or
writing to the page buffer, as the flash will be busy. The waiting can be performed either by
polling the Flash Status Register (FSR) or by waiting for the flash ready interrupt. The command written to FCMD is initiated on the first clock cycle where the HSB bus interface in
FLASHC is IDLE. The user must make sure that the access pattern to the FLASHC HSB interface contains an IDLE cycle so that the command is allowed to start. Make sure that no bus
masters such as DMA controllers are performing endless burst transfers from the flash. Also,
make sure that the CPU does not perform endless burst transfers from flash. This is done by
118
32058K
AVR32-01/12
AT32UC3A
letting the CPU enter sleep mode after writing to FCMD, or by polling FSR for command completion. This polling will result in an access pattern with IDLE HSB cycles.
All the commands are protected by the same keyword, which has to be written in the eight
highest bits of the FCMD register. Writing FCMD with data that does not contain the correct
key and/or with an invalid command has no effect on the flash memory; however, the PROGE
flag is set in the Flash Status Register (FSR). This flag is automatically cleared by a read
access to the FSR register.
Writing a command to FCMD while another command is being executed has no effect on the
flash memory; however, the PROGE flag is set in the Flash Status Register (FSR). This flag is
automatically cleared by a read access to the FSR register.
If the current command writes or erases a page in a locked region, or a page protected by the
BOOTPROT fuses, the command has no effect on the flash memory; however, the LOCKE
flag is set in the FSR register. This flag is automatically cleared by a read access to the FSR
register.
18.5.1
Write/erase page operation
Flash technology requires that an erase must be done before programming. The entire flash
can be erased by an Erase All command. Alternatively, pages can be individually erased by
the Erase Page command.
The User page can be written and erased using the mechanisms described in this chapter.
After programming, the page can be locked to prevent miscellaneous write or erase
sequences. Locking is performed on a per-region basis, so locking a region locks all pages
inside the region. Additional protection is provided for the lowermost address space of the
flash. This address space is allocated for the Boot Loader, and is protected both by the lock
bit(s) corresponding to this address space, and the BOOTPROT[2:0] fuses.
Data to be written are stored in an internal buffer called page buffer. The page buffer contains
w words. The page buffer wraps around within the internal memory area address space and
appears to be repeated by the number of pages in it. Writing of 8-bit and 16-bit data to the
page buffer is not allowed and may lead to unpredictable data corruption.
Data must be written to the page buffer before the programming command is written to the
Flash Command Register FCMD. The sequence is as follows:
• Reset the page buffer with the Clear Page Buffer command.
• Fill the page buffer with the desired contents, using only 32-bit access.
• Programming starts as soon as the programming key and the programming command are
written to the Flash Command Register. The PAGEN field in the Flash Command Register
(FCMD) must contain the address of the page to write. PAGEN is automatically updated
when writing to the page buffer, but can also be written to directly. The FRDY bit in the
Flash Status Register (FSR) is automatically cleared when the page write operation starts.
• When programming is completed, the bit FRDY in the Flash Status Register (FSR) is set. If
an interrupt was enabled by setting the bit FRDY in FCR, the interrupt line of the flash
controller is set.
Two errors can be detected in the FSR register after a programming sequence:
• Programming Error: A bad keyword and/or an invalid command have been written in the
FCMD register.
119
32058K
AVR32-01/12
AT32UC3A
• Lock Error: The page to be programmed belongs to a locked region. A command must be
executed to unlock the corresponding region before programming can start.
18.5.2
Erase All operation
The entire memory is erased if the Erase All command (EA) is written to the Flash Command
Register (FCMD). Erase All erases all bits in the flash array. The User page is not erased. All
flash memory locations, the general-purpose fuse bits, and the security bit are erased (reset to
0xFF) after an Erase All.
The EA command also ensures that all volatile memories, such as register file and RAMs, are
erased before the security bit is erased.
Erase All operation is allowed only if no regions are locked, and the BOOTPROT fuses are
programmed with a region size of 0. Thus, if at least one region is locked, the bit LOCKE in
FSR is set and the command is cancelled. If the bit LOCKE has been written to 1 in FCR, the
interrupt line rises.
When the command is complete, the bit FRDY bit in the Flash Status Register (FSR) is set. If
an interrupt has been enabled by setting the bit FRDY in FCR, the interrupt line of the flash
controller is set. Two errors can be detected in the FSR register after issuing the command:
• Programming Error: A bad keyword and/or an invalid command have been written in the
FCMD register.
• Lock Error: At least one lock region to be erased is protected, or BOOTPROT is different
from 0. The erase command has been refused and no page has been erased. A Clear Lock
Bit command must be executed previously to unlock the corresponding lock regions.
18.5.3
Region lock bits
The flash block has p pages, and these pages are grouped into 16 lock regions, each region
containing p/16 pages. Each region has a dedicated lock bit preventing writing and erasing
pages in the region. After production, the device may have some regions locked. These locked
regions are reserved for a boot or default application. Locked regions can be unlocked to be
erased and then programmed with another application or other data.
To lock or unlock a region, the commands Lock Region Containing Page (LP) and Unlock
Region Containing Page (UP) are provided. Writing one of these commands, together with the
number of the page whose region should be locked/unlocked, performs the desired operation.
One error can be detected in the FSR register after issuing the command:
• Programming Error: A bad keyword and/or an invalid command have been written in the
FCMD register.
The lock bits are implemented using the lowest 16 general-purpose fuse bits. This means that
lock bits can also be set/cleared using the commands for writing/erasing general-purpose fuse
bits, see chapter 18.6. The general-purpose bit being in an erased (1) state means that the
region is unlocked.
The lowermost pages in the Flash can additionally be protected by the BOOTPROT fuses, see
Section 18.6.
18.6
General-purpose fuse bits
Each flash block has a number of general-purpose fuse bits that the application programmer
can use freely. The fuse bits can be written and erased using dedicated commands, and read
120
32058K
AVR32-01/12
AT32UC3A
through a dedicated Peripheral Bus address. Some of the general-purpose fuse bits are
reserved for special purposes, and should not be used for other functions.:
Table 18-2.
General-purpose fuses with special functions
GeneralPurpose fuse
number
Name
Usage
15:0
LOCK
Region lock bits.
EPFL
External Privileged Fetch Lock. Used to prevent the CPU from
fetching instructions from external memories when in privileged
mode. This bit can only be changed when the security bit is
cleared. The address range corresponding to external
memories is device-specific, and not known to the flash
controller. This fuse bit is simply routed out of the CPU or bus
system, the flash controller does not treat this fuse in any
special way, except that it can not be altered when the security
bit is set.
If the security bit is set, only an external JTAG Chip Erase can
clear EPFL. No internal commands can alter EPFL if the
security bit is set.
When the fuse is erased (i.e. "1"), the CPU can execute
instructions fetched from external memories. When the fuse is
programmed (i.e. "0"), instructions can not be executed from
external memories.
BOOTPROT
Used to select one of four different bootloader sizes. Pages
included in the bootloader area can not be erased or
programmed except by a JTAG chip erase. BOOTPROT can
only be changed when the security bit is cleared.
If the security bit is set, only an external JTAG Chip Erase can
clear BOOTPROT, and thereby allow the pages protected by
BOOTPROT to be programmed. No internal commands can
alter BOOTPROT or the pages protected by BOOTPROT if the
security bit is set.
16
19:17
The BOOTPROT fuses protects the following address space for the Boot Loader:
Table 18-3.
Boot Loader area specified by BOOTPROT
BOOTPROT
Pages protected by
BOOTPROT
Size of protected
memory
7
None
0
6
0-1
1kByte
5
0-3
2kByte
4
0-7
4kByte
3
0-15
8kByte
2
0-31
16kByte
1
0-63
32kByte
0
0-127
64kByte
To erase or write a general-purpose fuse bit, the commands Write General-Purpose Fuse Bit
(WGPB) and Erase General-Purpose Fuse Bit (EGPB) are provided. Writing one of these
121
32058K
AVR32-01/12
AT32UC3A
commands, together with the number of the fuse to write/erase, performs the desired
operation.
An entire General-Purpose Fuse byte can be written at a time by using the Program GP Fuse
Byte (PGPFB) instruction. A PGPFB to GP fuse byte 2 is not allowed if the flash is locked by
the security bit. The PFB command is issued with a parameter in the PAGEN field:
• PAGEN[2:0] - byte to write
• PAGEN[10:3] - Fuse value to write
All General-Purpose fuses can be erased by the Erase All General-Purpose fuses (EAGP)
command. An EAGP command is not allowed if the flash is locked by the security bit.
Two errors can be detected in the FSR register after issuing these commands:
• Programming Error: A bad keyword and/or an invalid command have been written in the
FCMD register.
• Lock Error: A write or erase of any of the special-function fuse bits in Table 18-3 was
attempted while the flash is locked by the security bit.
The lock bits are implemented using the lowest 16 general-purpose fuse bits. This means that
the 16 lowest general-purpose fuse bits can also be written/erased using the commands for
locking/unlocking regions, see Section 18.5.3.
18.7
Security bit
The security bit allows the entire chip to be locked from external JTAG or other debug access
for code security. The security bit can be written by a dedicated command, Set Security Bit
(SSB). Once set, the only way to clear the security bit is through the JTAG Chip Erase
command.
Once the Security bit is set, the following Flash controller commands will be unavailable and
return a lock error if attempted:
• Write General-Purpose Fuse Bit (WGPB) to BOOTPROT or EPFL fuses
• Erase General-Purpose Fuse Bit (EGPB) to BOOTPROT or EPFL fuses
• Program General-Purpose Fuse Byte (PGPFB) of fuse byte 2
• Erase All General-Purpose Fuses (EAGPF)
One error can be detected in the FSR register after issuing the command:
• Programming Error: A bad keyword and/or an invalid command have been written in the
FCMD register.
122
32058K
AVR32-01/12
AT32UC3A
18.8
User interface
18.8.1
Address map
The following addresses are used by the FLASHC. All offsets are relative to the base address
allocated to the flash controller.
Table 18-4.
Flash controller register mapping
Offset
Register
Name
Access
Reset
state
0x0
Flash Control Register
FCR
R/W
0
0x4
Flash Command Register
FCMD
R/W
0
0x8
Flash Status Register
FSR
R/W
0 (*)
0xc
Flash General Purpose Fuse Register Hi
FGPFRHI
R
NA (*)
0x10
Flash General Purpose Fuse Register Lo
FGPFRLO
R
NA (*)
(*) The value of the Lock bits is dependent of their programmed state. All other bits in FSR are
0. All bits in FGPFR and FCFR are dependent on the programmed state of the fuses they map
to. Any bits in these registers not mapped to a fuse read 0.
123
32058K
AVR32-01/12
AT32UC3A
18.8.2
Flash Control Register (FCR)
Offset: 0x0
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
SASD
7
6
5
4
3
2
1
0
-
FWS
-
-
PROGE
LOCKE
-
FRDY
FRDY: Flash Ready Interrupt Enable
0: Flash Ready does not generate an interrupt.
1: Flash Ready generates an interrupt.
LOCKE: Lock Error Interrupt Enable
0: Lock Error does not generate an interrupt.
1: Lock Error generates an interrupt.
PROGE: Programming Error Interrupt Enable
0: Programming Error does not generate an interrupt.
1: Programming Error generates an interrupt.
FWS: Flash Wait State
0: The flash is read with 0 wait states.
1: The flash is read with 1 wait state.
SASD: Sense Amplifier Sample Disable
0: The sense amplifiers in the flash are in sampling mode.
1: The sense amplifiers in the flash are permanently enabled. Consumes more power.
124
32058K
AVR32-01/12
AT32UC3A
18.8.3
Flash Command Register (FCMD)
Offset: 0x4
The FCMD can not be written if the flash is in the process of performing a flash command.
Doing so will cause the FCR write to be ignored, and the PROGE bit to be set.
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
KEY
23
22
21
20
PAGEN [15:8]
15
14
13
12
PAGEN [7:0]
7
6
-
-
5
4
CMD
CMD: Command
This field defines the flash command. Issuing any unused command will cause the Programming Error flag to be set, and the corresponding interrupt to be requested if the PROGE bit in
FCR is set.
Table 18-5.
Set of commands
Command
Value
Mnemonic
No operation
0
NOP
Write Page
1
WP
Erase Page
2
EP
Clear Page Buffer
3
CPB
Lock region containing given Page
4
LP
Unlock region containing given Page
5
UP
Erase All
6
EA
Write General-Purpose Fuse Bit
7
WGPB
Erase General-Purpose Fuse Bit
8
EGPB
Set Security Bit
9
SSB
Program GP Fuse Byte
10
PGPFB
Erase All GPFuses
11
EAGPF
Quick Page Read
12
QPR
Write User Page
13
WUP
Erase User Page
14
EUP
Quick Page Read User Page
15
QPRUP
125
32058K
AVR32-01/12
AT32UC3A
PAGEN: Page number
The PAGEN field is used to address a page or fuse bit for certain operations. In order to simplify programming, the PAGEN field is automatically updated every time the page buffer is
written to. For every page buffer write, the PAGEN field is updated with the page number of
the address being written to. Hardware automatically masks writes to the PAGEN field so that
only bits representing valid page numbers can be written, all other bits in PAGEN are always
0. As an example, in a flash with 1024 pages (page 0 - page 1023), bits 15:10 will always be 0.
Table 18-6.
Semantic of PAGEN field in different commands
Command
PAGEN description
No operation
Not used
Write Page
The number of the page to write
Clear Page Buffer
Not used
Lock region containing given Page
Page number whose region should be locked
Unlock region containing given Page
Page number whose region should be unlocked
Erase All
Not used
Write General-Purpose Fuse Bit
GPFUSE #
Erase General-Purpose Fuse Bit
GPFUSE #
Set Security Bit
Not used
Program GP Fuse Byte
WriteData[7:0], ByteAddress[2:0]
Erase All GP Fuses
Not used
Quick Page Read
Page number
Write User Page
Not used
Erase User Page
Not used
Quick Page Read User Page
Not used
KEY: Write protection key
This field should be written with the value 0xA5 to enable the command defined by the bits of
the register. If the field is written with a different value, the write is not performed and no action
is started.
This field always reads as 0.
126
32058K
AVR32-01/12
AT32UC3A
18.8.4
Flash Status Register (FSR)
Offset: 0x08
31
30
29
28
27
26
25
24
LOCK15
LOCK14
LOCK13
LOCK12
LOCK11
LOCK10
LOCK9
LOCK8
23
22
21
20
19
18
17
16
LOCK7
LOCK6
LOCK5
LOCK4
LOCK3
LOCK2
LOCK1
LOCK0
15
14
13
12
11
10
9
8
-
-
-
-
FSZ
7
6
5
4
3
2
1
0
-
-
QPRR
SECURITY
PROGE
LOCKE
-
FRDY
FRDY: Flash Ready Status
0: The flash controller is busy and the application must wait before running a new command.
1: The flash controller is ready to run a new command.
LOCKE: Lock Error Status
Automatically cleared when FSR is read.
0: No programming of at least one locked lock region has happened since the last read of
FSR.
1: Programming of at least one locked lock region has happened since the last read of FSR.
PROGE: Programming Error Status
Automatically cleared when FSR is read.
0: No invalid commands and no bad keywords were written in the Flash Command Register
FCMD.
1: An invalid command and/or a bad keyword was/were written in the Flash Command Register FCMD.
SECURITY: Security Bit Status
0: The security bit is inactive.
1: The security bit is active.
QPRR: Quick Page Read Result
0: The result is zero, i.e. the page is not erased.
1: The result is one, i.e. the page is erased.
Automatically cleared when FSR is read.
127
32058K
AVR32-01/12
AT32UC3A
FSZ: Flash Size
The size of the flash. Not all device families will provide all flash sizes indicated in the table.
Table 18-7.
Flash size
FSZ
Flash Size
0
32 KByte
1
64 kByte
2
128 kByte
3
256 kByte
4
384 kByte
5
512 kByte
6
768 kByte
7
1024 kByte
LOCKx: Lock Region x Lock Status
0: The corresponding lock region is not locked.
1: The corresponding lock region is locked.
128
32058K
AVR32-01/12
AT32UC3A
18.8.5
Flash General Purpose Fuse Register High (FGPFRHI)
Offset: 0x0C
31
30
29
28
27
26
25
24
GPF63
GPF62
GPF61
GPF60
GPF59
GPF58
GPF57
GPF56
23
22
21
20
19
18
17
16
GPF55
GPF54
GPF53
GPF52
GPF51
GPF50
GPF49
GPF48
15
14
13
12
11
10
9
8
GPF47
GPF46
GPF45
GPF44
GPF43
GPF42
GPF41
GPF40
7
6
5
4
3
2
1
0
GPF39
GPF38
GPF37
GPF36
GPF35
GPF34
GPF33
GPF32
This register is only used in systems with more than 32 GP fuses.
GPFxx: General Purpose Fuse xx
0: The fuse has a written/programmed state.
1: The fuse has an erased state.
129
32058K
AVR32-01/12
AT32UC3A
18.8.6
Flash General Purpose Fuse Register Low (FGPFRLO)
Offset: 0x10
31
30
29
28
27
26
25
24
GPF31
GPF30
GPF29
GPF28
GPF27
GPF26
GPF25
GPF24
23
22
21
20
19
18
17
16
GPF23
GPF22
GPF21
GPF20
GPF19
GPF18
GPF17
GPF16
15
14
13
12
11
10
9
8
GPF15
GPF14
GPF13
GPF12
GPF11
GPF10
GPF09
GPF08
7
6
5
4
3
2
1
0
GPF07
GPF06
GPF05
GPF04
GPF03
GPF02
GPF01
GPF00
GPFxx: General Purpose Fuse xx
0: The fuse has a written/programmed state.
1: The fuse has an erased state.
130
32058K
AVR32-01/12
AT32UC3A
131
32058K
AVR32-01/12
AT32UC3A
19. HSB Bus Matrix (HMATRIX)
Rev: 2.3.0.1
19.1
Features
•
•
•
•
•
•
•
•
•
•
•
19.2
User Interface on peripheral bus
Configurable Number of Masters (Up to sixteen)
Configurable Number of Slaves (Up to sixteen)
One Decoder for Each Master
Three Different Memory Mappings for Each Master (Internal and External boot, Remap)
One Remap Function for Each Master
Programmable Arbitration for Each Slave
– Round-Robin
– Fixed Priority
Programmable Default Master for Each Slave
– No Default Master
– Last Accessed Default Master
– Fixed Default Master
One Cycle Latency for the First Access of a Burst
Zero Cycle Latency for Default Master
One Special Function Register for Each Slave (Not dedicated)
Description
The Bus Matrix implements a multi-layer bus structure, that enables parallel access paths
between multiple High Speed Bus (HSB) masters and slaves in a system, thus increasing the
overall bandwidth. The Bus Matrix interconnects up to 16 HSB Masters to up to 16 HSB Slaves.
The normal latency to connect a master to a slave is one cycle except for the default master of
the accessed slave which is connected directly (zero cycle latency). The Bus Matrix provides 16
Special Function Registers (SFR) that allow the Bus Matrix to support application specific
features.
19.3
Memory Mapping
The Bus Matrix provides one decoder for every HSB Master Interface. The decoder offers each
HSB Master several memory mappings. In fact, depending on the product, each memory area
may be assigned to several slaves. Booting at the same address while using different HSB
slaves (i.e. external RAM, internal ROM or internal Flash, etc.) becomes possible.
The Bus Matrix user interface provides Master Remap Control Register (MRCR) that performs
remap action for every master independently.
19.4
Special Bus Granting Mechanism
The Bus Matrix provides some speculative bus granting techniques in order to anticipate access
requests from some masters. This mechanism reduces latency at first access of a burst or single
transfer. This bus granting mechanism sets a different default master for every slave.
At the end of the current access, if no other request is pending, the slave remains connected to
its associated default master. A slave can be associated with three kinds of default masters: no
default master, last access master and fixed default master.
132
32058K
AVR32-01/12
AT32UC3A
19.4.1
No Default Master
At the end of the current access, if no other request is pending, the slave is disconnected from
all masters. No Default Master suits low-power mode.
19.4.2
Last Access Master
At the end of the current access, if no other request is pending, the slave remains connected to
the last master that performed an access request.
19.4.3
Fixed Default Master
At the end of the current access, if no other request is pending, the slave connects to its fixed
default master. Unlike last access master, the fixed master does not change unless the user
modifies it by a software action (field FIXED_DEFMSTR of the related SCFG).
To change from one kind of default master to another, the Bus Matrix user interface provides the
Slave Configuration Registers, one for each slave, that set a default master for each slave. The
Slave Configuration Register contains two fields: DEFMSTR_TYPE and FIXED_DEFMSTR. The
2-bit DEFMSTR_TYPE field selects the default master type (no default, last access master, fixed
default master), whereas the 4-bit FIXED_DEFMSTR field selects a fixed default master provided that DEFMSTR_TYPE is set to fixed default master. Please refer to the Bus Matrix user
interface description.
19.5
Arbitration
The Bus Matrix provides an arbitration mechanism that reduces latency when conflict cases
occur, i.e. when two or more masters try to access the same slave at the same time. One arbiter
per HSB slave is provided, thus arbitrating each slave differently.
The Bus Matrix provides the user with the possibility of choosing between 2 arbitration types for
each slave:
1. Round-Robin Arbitration (default)
2. Fixed Priority Arbitration
This choice is made via the field ARBT of the Slave Configuration Registers (SCFG).
Each algorithm may be complemented by selecting a default master configuration for each
slave.
When a re-arbitration must be done, specific conditions apply. See Section 19.5.1 ”Arbitration
Rules” on page 133.
19.5.1
Arbitration Rules
Each arbiter has the ability to arbitrate between two or more different master requests. In order
to avoid burst breaking and also to provide the maximum throughput for slave interfaces, arbitration may only take place during the following cycles:
1. Idle Cycles: When a slave is not connected to any master or is connected to a master
which is not currently accessing it.
2. Single Cycles: When a slave is currently doing a single access.
3. End of Burst Cycles: When the current cycle is the last cycle of a burst transfer. For
defined length burst, predicted end of burst matches the size of the transfer but is managed differently for undefined length burst. See Section “19.5.1.1” on page 134.
133
32058K
AVR32-01/12
AT32UC3A
4. Slot Cycle Limit: When the slot cycle counter has reached the limit value indicating that
the current master access is too long and must be broken. See Section “19.5.1.2” on
page 134.
19.5.1.1
Undefined Length Burst Arbitration
In order to avoid long slave handling during undefined length bursts (INCR), the Bus Matrix provides specific logic in order to re-arbitrate before the end of the INCR transfer. A predicted end
of burst is used as a defined length burst transfer and can be selected from among the following
five possibilities:
1. Infinite: No predicted end of burst is generated and therefore INCR burst transfer will
never be broken.
2. One beat bursts: Predicted end of burst is generated at each single transfer inside the
INCP transfer.
3. Four beat bursts: Predicted end of burst is generated at the end of each four beat
boundary inside INCR transfer.
4. Eight beat bursts: Predicted end of burst is generated at the end of each eight beat
boundary inside INCR transfer.
5. Sixteen beat bursts: Predicted end of burst is generated at the end of each sixteen beat
boundary inside INCR transfer.
This selection can be done through the field ULBT of the Master Configuration Registers
(MCFG).
19.5.1.2
Slot Cycle Limit Arbitration
The Bus Matrix contains specific logic to break long accesses, such as very long bursts on a
very slow slave (e.g., an external low speed memory). At the beginning of the burst access, a
counter is loaded with the value previously written in the SLOT_CYCLE field of the related Slave
Configuration Register (SCFG) and decreased at each clock cycle. When the counter reaches
zero, the arbiter has the ability to re-arbitrate at the end of the current byte, half word or word
transfer.
19.5.2
Round-Robin Arbitration
This algorithm allows the Bus Matrix arbiters to dispatch the requests from different masters to
the same slave in a round-robin manner. If two or more master requests arise at the same time,
the master with the lowest number is first serviced, then the others are serviced in a round-robin
manner.
There are three round-robin algorithms implemented:
• Round-Robin arbitration without default master
• Round-Robin arbitration with last default master
• Round-Robin arbitration with fixed default master
19.5.2.1
Round-Robin Arbitration without Default Master
This is the main algorithm used by Bus Matrix arbiters. It allows the Bus Matrix to dispatch
requests from different masters to the same slave in a pure round-robin manner. At the end of
the current access, if no other request is pending, the slave is disconnected from all masters.
This configuration incurs one latency cycle for the first access of a burst. Arbitration without
default master can be used for masters that perform significant bursts.
134
32058K
AVR32-01/12
AT32UC3A
19.5.2.2
Round-Robin Arbitration with Last Default Master
This is a biased round-robin algorithm used by Bus Matrix arbiters. It allows the Bus Matrix to
remove the one latency cycle for the last master that accessed the slave. In fact, at the end of
the current transfer, if no other master request is pending, the slave remains connected to the
last master that performed the access. Other non privileged masters still get one latency cycle if
they want to access the same slave. This technique can be used for masters that mainly perform
single accesses.
19.5.2.3
Round-Robin Arbitration with Fixed Default Master
This is another biased round-robin algorithm. It allows the Bus Matrix arbiters to remove the one
latency cycle for the fixed default master per slave. At the end of the current access, the slave
remains connected to its fixed default master. Every request attempted by this fixed default master will not cause any latency whereas other non privileged masters will still get one latency
cycle. This technique can be used for masters that mainly perform single accesses.
19.5.3
Fixed Priority Arbitration
This algorithm allows the Bus Matrix arbiters to dispatch the requests from different masters to
the same slave by using the fixed priority defined by the user. If two or more master requests are
active at the same time, the master with the highest priority number is serviced first. If two or
more master requests with the same priority are active at the same time, the master with the
highest number is serviced first.
For each slave, the priority of each master may be defined through the Priority Registers for
Slaves (PRAS and PRBS).
19.6
Slave and Master assignation
The index number assigned to Bus Matrix slaves and masters are described in Memories
chapter.
135
32058K
AVR32-01/12
AT32UC3A
19.7
User Interface
Table 19-1.
Register Mapping
Offset
Register
Name
Access
Reset Value
0x0000
Master Configuration Register 0
MCFG0
Read/Write
0x00000002
0x0004
Master Configuration Register 1
MCFG1
Read/Write
0x00000002
0x0008
Master Configuration Register 2
MCFG2
Read/Write
0x00000002
0x000C
Master Configuration Register 3
MCFG3
Read/Write
0x00000002
0x0010
Master Configuration Register 4
MCFG4
Read/Write
0x00000002
0x0014
Master Configuration Register 5
MCFG5
Read/Write
0x00000002
0x0018
Master Configuration Register 6
MCFG6
Read/Write
0x00000002
0x001C
Master Configuration Register 7
MCFG7
Read/Write
0x00000002
0x0020
Master Configuration Register 8
MCFG8
Read/Write
0x00000002
0x0024
Master Configuration Register 9
MCFG9
Read/Write
0x00000002
0x0028
Master Configuration Register 10
MCFG10
Read/Write
0x00000002
0x002C
Master Configuration Register 11
MCFG11
Read/Write
0x00000002
0x0030
Master Configuration Register 12
MCFG12
Read/Write
0x00000002
0x0034
Master Configuration Register 13
MCFG13
Read/Write
0x00000002
0x0038
Master Configuration Register 14
MCFG14
Read/Write
0x00000002
0x003C
Master Configuration Register 15
MCFG15
Read/Write
0x00000002
0x0040
Slave Configuration Register 0
SCFG0
Read/Write
0x00000010
0x0044
Slave Configuration Register 1
SCFG1
Read/Write
0x00000010
0x0048
Slave Configuration Register 2
SCFG2
Read/Write
0x00000010
0x004C
Slave Configuration Register 3
SCFG3
Read/Write
0x00000010
0x0050
Slave Configuration Register 4
SCFG4
Read/Write
0x00000010
0x0054
Slave Configuration Register 5
SCFG5
Read/Write
0x00000010
0x0058
Slave Configuration Register 6
SCFG6
Read/Write
0x00000010
0x005C
Slave Configuration Register 7
SCFG7
Read/Write
0x00000010
0x0060
Slave Configuration Register 8
SCFG8
Read/Write
0x00000010
0x0064
Slave Configuration Register 9
SCFG9
Read/Write
0x00000010
0x0068
Slave Configuration Register 10
SCFG10
Read/Write
0x00000010
0x006C
Slave Configuration Register 11
SCFG11
Read/Write
0x00000010
0x0070
Slave Configuration Register 12
SCFG12
Read/Write
0x00000010
0x0074
Slave Configuration Register 13
SCFG13
Read/Write
0x00000010
0x0078
Slave Configuration Register 14
SCFG14
Read/Write
0x00000010
0x007C
Slave Configuration Register 15
SCFG15
Read/Write
0x00000010
0x0080
Priority Register A for Slave 0
PRAS0
Read/Write
0x00000000
0x0084
Priority Register B for Slave 0
PRBS0
Read/Write
0x00000000
0x0088
Priority Register A for Slave 1
PRAS1
Read/Write
0x00000000
136
32058K
AVR32-01/12
AT32UC3A
Table 19-1.
Register Mapping (Continued)
Offset
Register
Name
Access
Reset Value
0x008C
Priority Register B for Slave 1
PRBS1
Read/Write
0x00000000
0x0090
Priority Register A for Slave 2
PRAS2
Read/Write
0x00000000
0x0094
Priority Register B for Slave 2
PRBS2
Read/Write
0x00000000
0x0098
Priority Register A for Slave 3
PRAS3
Read/Write
0x00000000
0x009C
Priority Register B for Slave 3
PRBS3
Read/Write
0x00000000
0x00A0
Priority Register A for Slave 4
PRAS4
Read/Write
0x00000000
0x00A4
Priority Register B for Slave 4
PRBS4
Read/Write
0x00000000
0x00A8
Priority Register A for Slave 5
PRAS5
Read/Write
0x00000000
0x00AC
Priority Register B for Slave 5
PRBS5
Read/Write
0x00000000
0x00B0
Priority Register A for Slave 6
PRAS6
Read/Write
0x00000000
0x00B4
Priority Register B for Slave 6
PRBS6
Read/Write
0x00000000
0x00B8
Priority Register A for Slave 7
PRAS7
Read/Write
0x00000000
0x00BC
Priority Register B for Slave 7
PRBS7
Read/Write
0x00000000
0x00C0
Priority Register A for Slave 8
PRAS8
Read/Write
0x00000000
0x00C4
Priority Register B for Slave 8
PRBS8
Read/Write
0x00000000
0x00C8
Priority Register A for Slave 9
PRAS9
Read/Write
0x00000000
0x00CC
Priority Register B for Slave 9
PRBS9
Read/Write
0x00000000
0x00D0
Priority Register A for Slave 10
PRAS10
Read/Write
0x00000000
0x00D4
Priority Register B for Slave 10
PRBS10
Read/Write
0x00000000
0x00D8
Priority Register A for Slave 11
PRAS11
Read/Write
0x00000000
0x00DC
Priority Register B for Slave 11
PRBS11
Read/Write
0x00000000
0x00E0
Priority Register A for Slave 12
PRAS12
Read/Write
0x00000000
0x00E4
Priority Register B for Slave 12
PRBS12
Read/Write
0x00000000
0x00E8
Priority Register A for Slave 13
PRAS13
Read/Write
0x00000000
0x00EC
Priority Register B for Slave 13
PRBS13
Read/Write
0x00000000
0x00F0
Priority Register A for Slave 14
PRAS14
Read/Write
0x00000000
0x00F4
Priority Register B for Slave 14
PRBS14
Read/Write
0x00000000
0x00F8
Priority Register A for Slave 15
PRAS15
Read/Write
0x00000000
0x00FC
Priority Register B for Slave 15
PRBS15
Read/Write
0x00000000
0x0100
Master Remap Control Register
MRCR
Read/Write
0x00000000
–
–
0x0104 - 0x010C
Reserved
–
0x0110
Special Function Register 0
SFR0
Read/Write
–
0x0114
Special Function Register 1
SFR1
Read/Write
–
0x0118
Special Function Register 2
SFR2
Read/Write
–
0x011C
Special Function Register 3
SFR3
Read/Write
–
0x0120
Special Function Register 4
SFR4
Read/Write
–
137
32058K
AVR32-01/12
AT32UC3A
Table 19-1.
Register Mapping (Continued)
Offset
Register
Name
Access
Reset Value
0x0124
Special Function Register 5
SFR5
Read/Write
–
0x0128
Special Function Register 6
SFR6
Read/Write
–
0x012C
Special Function Register 7
SFR7
Read/Write
–
0x0130
Special Function Register 8
SFR8
Read/Write
–
0x0134
Special Function Register 9
SFR9
Read/Write
–
0x0138
Special Function Register 10
SFR10
Read/Write
–
0x013C
Special Function Register 11
SFR11
Read/Write
–
0x0140
Special Function Register 12
SFR12
Read/Write
–
0x0144
Special Function Register 13
SFR13
Read/Write
–
0x0148
Special Function Register 14
SFR14
Read/Write
–
0x014C
Special Function Register 15
SFR15
Read/Write
–
–
–
0x0150 - 0x01F8
Reserved
–
138
32058K
AVR32-01/12
AT32UC3A
19.7.1
Bus Matrix Master Configuration Registers
Register Name:
MCFG0...MCFG15
Access Type:
Read/Write
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
–
15
14
13
12
11
10
9
8
–
–
–
–
–
–
–
–
7
6
5
4
3
2
1
0
–
–
–
–
–
ULBT
• ULBT: Undefined Length Burst Type
0: Infinite Length Burst
No predicted end of burst is generated and therefore INCR bursts coming from this master cannot be broken.
1: Single Access
The undefined length burst is treated as a succession of single accesses, allowing re-arbitration at each beat of the INCR
burst.
2: Four Beat Burst
The undefined length burst is split into a four-beat burst, allowing re-arbitration at each four-beat burst end.
3: Eight Beat Burst
The undefined length burst is split into an eight-beat burst, allowing re-arbitration at each eight-beat burst end.
4: Sixteen Beat Burst
The undefined length burst is split into a sixteen-beat burst, allowing re-arbitration at each sixteen-beat burst end.
139
32058K
AVR32-01/12
AT32UC3A
19.7.2
Bus Matrix Slave Configuration Registers
Register Name:
SCFG0...SCFG15
Access Type:
Read/Write
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
ARBT
23
22
21
20
19
18
17
16
–
–
15
14
13
12
11
10
9
8
–
–
–
–
–
–
–
–
7
6
5
4
3
2
1
0
FIXED_DEFMSTR
DEFMSTR_TYPE
SLOT_CYCLE
• SLOT_CYCLE: Maximum Number of Allowed Cycles for a Burst
When the SLOT_CYCLE limit is reached for a burst, it may be broken by another master trying to access this slave.
This limit has been placed to avoid locking a very slow slave when very long bursts are used.
This limit must not be very small. Unreasonably small values break every burst and the Bus Matrix arbitrates without performing any data transfer. 16 cycles is a reasonable value for SLOT_CYCLE.
• DEFMSTR_TYPE: Default Master Type
0: No Default Master
At the end of the current slave access, if no other master request is pending, the slave is disconnected from all masters.
This results in a one cycle latency for the first access of a burst transfer or for a single access.
1: Last Default Master
At the end of the current slave access, if no other master request is pending, the slave stays connected to the last master
having accessed it.
This results in not having one cycle latency when the last master tries to access the slave again.
2: Fixed Default Master
At the end of the current slave access, if no other master request is pending, the slave connects to the fixed master the
number that has been written in the FIXED_DEFMSTR field.
This results in not having one cycle latency when the fixed master tries to access the slave again.
• FIXED_DEFMSTR: Fixed Default Master
This is the number of the Default Master for this slave. Only used if DEFMSTR_TYPE is 2. Specifying the number of a master which is not connected to the selected slave is equivalent to setting DEFMSTR_TYPE to 0.
• ARBT: Arbitration Type
0: Round-Robin Arbitration
1: Fixed Priority Arbitration
140
32058K
AVR32-01/12
AT32UC3A
19.7.3
Bus Matrix Priority Registers A For Slaves
Register Name:
PRAS0...PRAS15
Access Type:
Read/Write
31
30
29
28
27
26
M7PR
23
22
21
20
19
18
M5PR
15
14
13
6
17
16
12
11
10
9
8
1
0
M2PR
5
M1PR
24
M4PR
M3PR
7
25
M6PR
4
3
2
M0PR
• MxPR: Master x Priority
Fixed priority of Master x for accessing the selected slave. The higher the number, the higher the priority.
141
32058K
AVR32-01/12
AT32UC3A
19.7.4
Bus Matrix Priority Registers B For Slaves
Register Name:
PRBS0...PRBS15
Access Type:
Read/Write
31
30
29
28
27
26
M15PR
23
22
21
20
19
18
M13PR
15
14
13
6
17
16
12
11
10
9
8
1
0
M10PR
5
M9PR
24
M12PR
M11PR
7
25
M14PR
4
3
2
M8PR
• MxPR: Master x Priority
Fixed priority of Master x for accessing the selected slave. The higher the number, the higher the priority.
142
32058K
AVR32-01/12
AT32UC3A
19.7.5
Bus Matrix Master Remap Control Register
Register Name:
MRCR
Access Type:
Read/Write
Reset:
0x0000_0000
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
–
15
14
13
12
11
10
9
8
RCB15
RCB14
RCB13
RCB12
RCB11
RCB10
RCB9
RCB8
7
6
5
4
3
2
1
0
RCB7
RCB6
RCB5
RCB4
RCB3
RCB2
RCB1
RCB0
• RCB: Remap Command Bit for Master x
0: Disable remapped address decoding for the selected Master
1: Enable remapped address decoding for the selected Master
143
32058K
AVR32-01/12
AT32UC3A
19.7.6
Bus Matrix Special Function Registers
Register Name:
SFR0...SFR15
Access Type:
Read/Write
Reset:
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
SFR
23
22
21
20
SFR
15
14
13
12
SFR
7
6
5
4
SFR
• SFR: Special Function Register Fields
The bitfields of these registers are described in the Peripherals chapter.
144
32058K
AVR32-01/12
AT32UC3A
20. External Bus Interface (EBI)
Rev: 1.0.0.1
20.1
Features
• Present only on AT32UC3A0512 and AT32UC3A0256
• Optimized for Application Memory Space support
• Integrates Two External Memory Controllers:
– Static Memory Controller
– SDRAM Controller
• Optimized External Bus:
– 16-bit Data Bus
– 24-bit Address Bus, Up to 16-Mbytes Addressable
– Optimized pin multiplexing to reduce latencies on External Memories
• 4 SRAM Chip Selects, 1 SDRAM Chip Selects:
– Static Memory Controller on NCS0
– SDRAM Controller or Static Memory Controller on NCS1
– Static Memory Controller on NCS2
– Static Memory Controller on NCS3
20.2
Description
The External Bus Interface (EBI) is designed to ensure the successful data transfer between
several external devices and the AT32UC3A device. The Static Memory and SDRAM Controllers are all featured external Memory Controllers on the EBI. These external Memory Controllers
are capable of handling several types of external memory and peripheral devices, such as
SRAM, PROM, EPROM, EEPROM, Flash, and SDRAM.
The EBI handles data transfers with up to five external devices, each assigned to five address
spaces defined by the embedded Memory Controller. Data transfers are performed through a
16-bit data bus, an address bus of up to 24 bits, up to four chip select lines (NCS[3:0]) and several control pins that are generally multiplexed between the different external Memory
Controllers.
145
32058K
AVR32-01/12
AT32UC3A
20.3
Block Diagram
20.3.1
External Bus Interface
Figure 20-1 shows the organization of the External Bus Interface.
Figure 20-1. Organization of the External Bus Interface
146
32058K
AVR32-01/12
AT32UC3A
20.4
I/O Lines Description
Table 20-1.
EBI I/O Lines Description
Name
Function
Type
Active Level
EBI
D0 - D15
Data Bus
I/O
A0 - A23
Address Bus
NWAIT
External Wait Signal
Output
Input
Low
SMC
NCS0 - NCS3
Chip Select Lines
Output
Low
NWR0 - NWR3
Write Signals
Output
Low
NOE
Output Enable
Output
Low
NRD
Read Signal
Output
Low
NWE
Write Enable
Output
Low
NBS0 - NBS3
Byte Mask Signals
Output
Low
SDRAM Controller
SDCK
SDRAM Clock
Output
SDCKE
SDRAM Clock Enable
Output
BA0 - BA1
Bank Select
Output
SDWE
SDRAM Write Enable
Output
Low
RAS - CAS
Row and Column Signal
Output
Low
NWR0 - NWR3
Write Signals
Output
Low
NBS0 - NBS3
Byte Mask Signals
Output
Low
SDA10
SDRAM Address 10 Line
Output
High
Depending on the Memory Controller in use, all signals are not connected directly through the
Mux Logic.
Table 20-2 on page 147 details the connections between the two Memory Controllers and the
EBI pins.
Table 20-2.
EBI Pins and Memory Controllers I/O Lines Connections
EBI Pins
SDRAMC I/O Lines
SMC I/O Lines
NWR1/NBS1
NBS1
NWR1/NUB
A0/NBS0
Not Supported
SMC_A0/NLB
A1/NBS2/NWR2
Not Supported
SMC_A1
A[11:2]
SDRAMC_A[9:0]
SMC_A[11:2]
SDA10
SDRAMC_A10
Not Supported
A12
Not Supported
SMC_A12
A[14:13]
SDRAMC_A[12:11]
SMC_A[14:13]
147
32058K
AVR32-01/12
AT32UC3A
Table 20-2.
EBI Pins and Memory Controllers I/O Lines Connections
EBI Pins
20.5
SDRAMC I/O Lines
SMC I/O Lines
A[22:15]
Not Supported
SMC_A[22:15]
A[23]
Not Supported
SMC_A[23]
D[15:0]
D[15:0]
D[15:0]
Application Example
20.5.1
Hardware Interface
Table 20-3 on page 148 details the connections to be applied between the EBI pins and the
external devices for each Memory Controller.
Table 20-3.
EBI Pins and External Static Devices Connections
Signals
Pins of the Interfaced Device
8-bit Static
Device
Controller
2 x 8-bit
Static
Devices
16-bit Static
Device
SMC
D0 - D7
D0 - D7
D0 - D7
D0 - D7
D8 - D15
–
D8 - D15
D8 - D15
A0/NBS0
A0
–
NLB
A1/NWR2/NBS2
A1
A0
A0
A[2:22]
A[1:21]
A[1:21]
A[23]
A[22]
A[22]
NCS0
CS
CS
CS
NCS1/SDCS0
CS
CS
CS
NCS2
CS
CS
CS
NCS3
CS
CS
CS
NRD/NOE
OE
OE
OE
A2 - A22
A23
NWR0/NWE
WE
WE
(1)
WE
(1)
NUB
NWR1/NBS1
–
WE
NWR3/NBS3
–
–
Notes:
1.
2.
3.
4.
5.
–
NWR1 enables upper byte writes. NWR0 enables lower byte writes.
NWRx enables corresponding byte x writes. (x = 0,1,2 or 3)
NBS0 and NBS1 enable respectively lower and upper bytes of the lower 16-bit word.
NBS2 and NBS3 enable respectively lower and upper bytes of the upper 16-bit word.
BEx: Byte x Enable (x = 0,1,2 or 3)
148
32058K
AVR32-01/12
AT32UC3A
Table 20-4.
EBI Pins and External SDRAM Devices Connections
Pins of the Interfaced
Device
Signals
Controller
SDRAM
SDRAMC
D0 - D15
D0 - D15
A0/NBS0
DQM0
A1/NWR2/NBS2
DQM2
A2 - A10
A[0:8]
A11
A9
SDA10
A10
A12
A13 - A14
A15
–
A[11:12]
–
A16/BA0
BA0
A17/BA1
BA1
A18 - A23
–
NCS0
–
NCS1/SDCS0
CS[0]
NCS2
–
NCS2
–
NCS3
–
NRD/NOE
–
NWR0/NWE
–
NWR1/NBS1
DQM1
NWR3/NBS3
DQM3
SDCK
CLK
SDCKE
CKE
RAS
RAS
CAS
CAS
SDWE
WE
NWAIT
–
149
32058K
AVR32-01/12
AT32UC3A
20.5.2
Connection Examples
Figure 20-2 shows an example of connections between the EBI and external devices.
Figure 20-2. EBI Connections to Memory Devices
EBI
D0-D15
RAS
CAS
SDCK
SDCKE
SDWE
A0/NBS0
NWR1/NBS1
A1/NWR2/NBS2
NWR3/NBS3
NRD/NOE
NWR0/NWE
D0-D7
D0-D7
CS
CLK
CKE
SDWE WE
RAS
CAS
DQM
NBS0
2M x 8
SDRAM
A0-A9, A11
A10
BA0
BA1
D8-D15
D0-D7
CS
CLK
CKE
SDWE
WE
RAS
CAS
DQM
NBS1
A2-A11, A13
SDA10
A16/BA0
A17/BA1
2M x 8
SDRAM
A0-A9, A11
A10
BA0
BA1
A2-A11, A13
SDA10
A16/BA0
A17/BA1
SDA10
A2-A15
A16/BA0
A17/BA1
A18-A23
NCS0
NCS1/SDCS
NCS2
NCS3
128K x 8
SRAM
D0-D7
D0-D7
CS
OE
NRD/NOE
WE
A0/NWR0/NBS0
A0-A16
128K x 8
SRAM
A1-A17
D8-D15
D0-D7
A0-A16
A1-A17
CS
OE
NRD/NOE
WE
NWR1/NBS1
150
32058K
AVR32-01/12
AT32UC3A
20.6
Product Dependencies
20.6.1
I/O Lines
The pins used for interfacing the External Bus Interface may be multiplexed with the GPIO lines.
The programmer must first program the GPIO controller to assign the External Bus Interface
pins to their peripheral function. If I/O lines of the External Bus Interface are not used by the
application, they can be used for other purposes by the GPIO Controller.
20.6.2
Power Management
The EBI HSB clock and SDRAMC, SMC and ECC PB clocks are generated by the Power Manager. Before using the EBI, the programmer must ensure that these clocks are enabled in the
Power Manager.
To prevent bus errors EBI operation must be terminated before entering sleep mode
20.6.3
Interrupt
The EBI interface has an interrupt line connected to the Interrupt Controller. Handling the EBI
interrupt requires programming the interrupt controller before configuring the EBI.
20.7
Functional Description
The EBI transfers data between the internal HSB Bus (handled by the HMatrix) and the external
memories or peripheral devices. It controls the waveforms and the parameters of the external
address, data and control busses and is composed of the following elements:
• The Static Memory Controller (SMC)
• The SDRAM Controller (SDRAMC)
• A chip select assignment feature that assigns an HSB address space to the external devices
• A multiplex controller circuit that shares the pins between the different Memory Controllers
20.7.1
Bus Multiplexing
The EBI offers a complete set of control signals that share the 16-bit data lines, the address
lines of up to 24 bits and the control signals through a multiplex logic operating in function of the
memory area requests.
Multiplexing is specifically organized in order to guarantee the maintenance of the address and
output control lines at a stable state while no external access is being performed. Multiplexing is
also designed to respect the data float times defined in the Memory Controllers. Furthermore,
refresh cycles of the SDRAM are executed independently by the SDRAM Controller without
delaying the other external Memory Controller accesses.
20.7.2
Pull-up Control
A specific HMATRIX_SFR register in the Matrix User Interface permit enabling of on-chip pull-up
resistors on the data bus lines not multiplexed with the GPIO Controller lines. For details on this
register, refer to the Peripherals Section. The pull-up resistors are enabled after reset. Setting
the EBI_DBPUC bit disables the pull-up resistors on lines not muxed with GPIO. Enabling the
pull-up resistor on lines multiplexed with GPIO lines can be performed by programming the
appropriate GPIO controller.
151
32058K
AVR32-01/12
AT32UC3A
20.7.3
Static Memory Controller
For information on the Static Memory Controller, refer to the Static Memory Controller Section.
20.7.4
SDRAM Controller
For information on the SDRAM Controller, refer to the SDRAM Section.
152
32058K
AVR32-01/12
AT32UC3A
21. Peripheral DMA Controller (PDCA)
rev: 1.0.0.0
21.1
Features
21.2
Overview
• Generates Transfers to/from Peripherals such as USART, SSC and SPI
• Two address pointers/counters per channel allowing double buffering
The Peripheral DMA controller (PDCA) transfers data between on-chip peripheral modules such
as USART, SPI, SSC and on- and off-chip memories. Using the PDCA avoids CPU intervention
for data transfers, improving the performance of the microcontroller. The PDCA can transfer
data from memory to a peripheral or from a peripheral to memory.
The PDCA consists of a number of DMA channels. Each channel has:
• A 32-bit memory pointer
• A 16-bit transfer counter
• A 32-bit memory pointer reload value
• A 16-bit transfer counter reload value
The PDCA communicates with the peripheral modules over a number of handshake interfaces.
The peripheral signals to the PDCA when it is ready to receive or transmit data. The PDCA
acknowledges the request when the transmission has started.
The number of handshake-interfaces may be higher than the number of DMA channels. If this is
the case, the DMA channel must be programmed to use the desired interface.
When a transmit buffer is empty or a receive buffer is full, an interrupt request can be signalled.
153
32058K
AVR32-01/12
AT32UC3A
21.3
Block Diagram
Peripheral
0
HSB to PB
Bridge
Peripheral Bus
HSB
Bus Matrix
HSB
Peripheral
1
Peripheral
2
Peripheral DMA
Controller
(PDCA)
Interrupt
Controller
IRQ
Peripheral
(n-1)
Handshake interfaces
21.4
Functional Description
21.4.1
Configuration
Each channel in the PDCA has a set of configuration registers. Among these are the Memory
Address Register (MAR), the Peripheral Select Register (PSR) and the Transfer Counter Register (TCR). The 32-bit Memory Address Register must be programmed with the start address of
the memory buffer. The register will be automatically updated after each transfer to point to the
next location in memory. The Peripheral Select Register must be programmed to select the
desired peripheral/handshake interface. The Transfer Counter Register determines the number
of data items to be transferred. The counter will be decreased by one for each data item that has
been transferred.
Both the Memory Address Register and the Transfer Counter Register can be read at any time
to check the progress of the transfer.
Each channel has also reload registers for the Memory Address Register and the Transfer
Counter Register. When the TCR reaches zero, the values in the reload registers are loaded into
MAR and TCR. In this way, the PDCA can operate on two buffers for each channel.
21.4.2
Memory Pointer
Each channel has a 32-bit Memory Pointer Register (MAR). This register holds the memory
address for the next transfer to be performed. The register is automatically updated after each
154
32058K
AVR32-01/12
AT32UC3A
transfer. The address will be increased by either 1, 2 or 4 depending on the size of the DMA
transfer (Byte, Half-Word or Word). The Memory Address Register can be read at any time during transfer.
21.4.3
Transfer Counter
Each channel has a 16-bit Transfer Counter Register (TCR). This register must be programmed
with the number of transferred to be performed. TCR should contain the number of data items to
be transferred independently of the transfer size. The Transfer Counter Register can be read at
any time during transfer to see the number of remaining transfers.
21.4.4
Reload Registers
Both the Memory Address Register and the Transfer Counter Register have a reload register,
respectively Memory Address Reload Register (MARR) and Transfer Counter Reload Register
(TCRR). These registers provide the possibility for the PDCA to work on two memory buffers for
each channel. When one buffer has completed, MAR and TCR will be reloaded with the values
in MARR and TCRR. The reload logic is always enabled and will trigger if the TCR reaches zero
while TCRR holds a non-zero value.
21.4.5
Peripheral Selection
The Peripheral Select Register decides which peripheral should be connected to the PDCA
channel. Configuring PSR will both select the direction of the transfer (memory to peripheral or
peripheral to memory), which handshake interface to use, and the address of the peripheral
holding register.
21.4.6
Transfer Size
The transfer size can be set individually for each channel to be either Byte, Half-Word or Word
(8-bit, 16-bit or 32-bit respectively). Transfer size is set by programming the SIZE bit-field in the
Mode Register (MR).
21.4.7
Enabling and Disabling
Each DMA channel is enabled by writing ‘1’ to the Transfer Enable bit (TEN) in the Control Register (CR) and disabled by writing ‘1’ to the Transfer Disable bit (TDIS). The current status can
be read from the Status Register (SR).
21.4.8
Interrupts
Interrupts can be enabled by writing to the Interrupt Enable Register (IER) and disabled by writing to Interrupt Disable Register (IDR). The Interrupt Mask Register (IMR) can be read to see
whether an interrupt is enabled or not. The current status of an interrupt source can be read
through the Interrupt Status Register (ISR).
The PDCA has three interrupt sources:
• Reload Counter Zero - The Transfer Counter Reload Register is zero.
• Transfer Finished - Both the Transfer Counter Register and Transfer Counter Reload Register
are zero.
• Transfer Error - An error has occurred in accessing memory.
155
32058K
AVR32-01/12
AT32UC3A
21.4.9
Priority
If more then one PDCA channel is requesting transfer at a given time, the PDCA channels are
prioritized by their channel number. Channels with lower numbers have priority over channels
with higher numbers, giving channel 0 the highest priority.
21.4.10
Error Handling
If the memory address is set to point to an invalid location in memory, an error will occur when
the PDCA tries to perform a transfer. When an error occurs, the Transfer Error flag (TERR) in
the Interrupt Status Register will be set and the DMA channel that caused the error will be
stopped. In order to restart the channel, the user must program the Memory Address Register to
a valid address and then write the Error Clear bit (ECLR) in the Control Register (CR) to ‘1’. An
interrupt can optionally be triggered on errors by writing the TERR-bit in the Interrupt Enable
Register (IER) to ‘1’.
21.5
User Interface
21.5.1
Memory Map Overview
Table 21-1.
Register Map Overview
Address Range
Contents
0x0000 - 0x003F
DMA channel 0 configuration registers
0x0040 - 0x007F
DMA channel 1 configuration registers
0x0080 - 0x00BF
DMA channel 2 configuration registers
0x00C0 - 0x00FF
DMA channel 3 configuration registers
0x0100 - 0x013F
DMA channel 4 configuration registers
-
-
-
DMA channel n-1 configuration registers
Note:
21.5.2
The number of channels is implementation specific. See part documentation for details.
Channel Memory Map
Offset
Register
Register Name
Access
Reset
0x00
Memory Address Register
MAR
Read/Write
0x00000000
0x04
Peripheral Select Register
PSR
Read/Write
*
0x08
Transfer Counter Register
TCR
Read/Write
0x00000000
0x0C
Memory Address Reload Register
MARR
Read/Write
0x00000000
0x10
Transfer Counter Reload Register
TCRR
Read/Write
0x00000000
0x14
Control Register
CR
Write-only
-
0x18
Mode Register
MR
Read/Write
0x00000000
0x1C
Status Register
SR
Read-only
0x00000000
0x20
Interrupt Enable Register
IER
Write-only
-
156
32058K
AVR32-01/12
AT32UC3A
Offset
Register
Register Name
Access
Reset
0x24
Interrupt Disable Register
IDR
Write-only
-
0x28
Interrupt Mask Register
IMR
Read-only
0x00000000
0x2C
Interrupt Status Register
ISR
Read-only
0x00000000
157
32058K
AVR32-01/12
AT32UC3A
21.5.3
Name:
PDCA Memory Address Register
MAR
Access Type:
31
Read/Write
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
MADDR
23
22
21
20
MADDR
15
14
13
12
MADDR
7
6
5
4
MADDR
• MADDR: Memory Address
Address of memory buffer. MADDR should be programmed to point to the start of the memory buffer when configuring the
PDCA. During transfer, MADDR will point to the next memory location to be read/written.
158
32058K
AVR32-01/12
AT32UC3A
21.5.4
Name:
PDCA Peripheral Select Register
PSR
Access Type:
Read/Write
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
PID
• PID: Peripheral Identifier
The Peripheral Identifier selects which peripheral should be connected to the DMA channel. Programming PID will select
both which handshake interface to use, the direction of the transfer and also the address of the Receive/Transfer Holding
Register for the peripheral. The PID values for the different peripheral modules are implementation specific. See the part
specific documentation for details.
The width of the PID bitfield is implementation specific and dependent on the number of peripheral modules in the
microcontroller.
159
32058K
AVR32-01/12
AT32UC3A
21.5.5PDCA Transfer Counter Register
Name:
TCR
Access Type:
Read/Write
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
3
2
1
0
TCV
7
6
5
4
TCV
• TCV: Transfer Counter Value
Number of data items to be transferred by PDCA. TCV must be programmed with the total number of transfers to be made.
During transfer, TCV contains the number of remaining transfers to be done.
160
32058K
AVR32-01/12
AT32UC3A
21.5.6
Name:
PDCA Memory Address Reload Register
MARR
Access Type:
31
Read/Write
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
MARV
23
22
21
20
MARV
15
14
13
12
MARV
7
6
5
4
MARV
• MARV: Memory Address Reload Value
Reload Value for the Memory Address Register (MAR). This value will be loaded into MAR when TCR reaches zero if the
TCRR has a non-zero value.
161
32058K
AVR32-01/12
AT32UC3A
21.5.7
Name:
PDCA Transfer Counter Reload Register
TCRR
Access Type:
Read/Write
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
3
2
1
0
TCRV
7
6
5
4
TCRV
• TCRV: Transfer Counter Reload Value
Reload value for the Transfer Counter Register (TCR). When TCR reaches zero, it will be reloaded with TCRV if TCRV has
a positive value. If TCRV is zero, no more transfers will be performed for the channel. When TCR is realoaded, the Transfer
Counter Reload Register is cleared.
162
32058K
AVR32-01/12
AT32UC3A
21.5.8
Name:
PDCA Control Register
CR
Access Type:
Write-only
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
ECLR
7
6
5
4
3
2
1
0
-
-
-
-
-
-
TDIS
TEN
• ECLR: Error Clear
0 = No Effect.
1 = Clear Transfer Error (TERR) flag in the Status Register (SR). Clearing the Transfer Error flag will allow the channel to
transmit data. The memory address must first be set to point to a valid location.
• TEN: Transfer Enable
0 = No Effect.
1 = Enable transfer for DMA channel.
• TDIS: Transfer Disable
0 = No Effect.
1 = Disable transfer for DMA channel.
163
32058K
AVR32-01/12
AT32UC3A
21.5.9
Name:
PDCA Mode Register
MR
Read/Write
Access Type:
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
-
SIZE
• SIZE: Size of transfer
SIZE
Size of Transfer
0
0
Byte
0
1
Half-Word
1
0
Word
1
1
Reserved
164
32058K
AVR32-01/12
AT32UC3A
21.5.10
Name:
PDCA Status Register
SR
Access Type:
Read
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
TEN
• TEN: Transfer Enabled
0 = Transfer is disabled for the DMA channel
1 = Transfer is enabled for the DMA channel.
165
32058K
AVR32-01/12
AT32UC3A
21.5.11
Name:
PDCA Interrupt Enable Register
IER
Write-only
Access Type:
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
TERR
TRC
RCZ
• TERR: Transfer Error
0 = No effect.
1 = Enable Transfer Error interrupt.
• TRC: Transfer Complete
0 = No effect.
1 = Enable Transfer Complete interrupt.
• RCZ: Reload Counter Zero
0 = No effect.
1 = Enable Reload Counter Zero interrupt.
166
32058K
AVR32-01/12
AT32UC3A
21.5.12
Name:
PDCA Interrupt Disable Register
IDR
Write-only
Access Type:
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
TERR
TRC
RCZ
• TERR: Transfer Error
0 = No effect.
1 = Disable Transfer Error interrupt.
• TRC: Transfer Complete
0 = No effect.
1 = Disable Transfer Complete interrupt.
• RCZ: Reload Counter Zero
0 = No effect.
1 = Disable Reload Counter Zero interrupt.
167
32058K
AVR32-01/12
AT32UC3A
21.5.13
Name:
PDCA Interrupt Mask Register
IMR
Read-only
Access Type:
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
TERR
TRC
RCZ
• TERR: Transfer Error
0 = Transfer Error interrupt is disabled.
1 = Transfer Error interrupt is enabled.
• TRC: Transfer Complete
0 = Transfer Complete interrupt is disabled.
1 = Transfer Complete interrupt is enabled.
• RCZ: Reload Counter Zero
0 = Reload Counter Zero interrupt is disabled.
1 = Reload Counter Zero interrupt is enabled.
168
32058K
AVR32-01/12
AT32UC3A
21.5.14
Name:
PDCA Interrupt Status Register
ISR
Read-only
Access Type:
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
TERR
TRC
RCZ
• TERR: Transfer Error
0 = No transfer errors have occurred.
1 = A transfer error has occurred.
• TRC: Transfer Complete
0 = The Transfer Counter Register (TCR) and/or the Transfer Counter Reload Register (TCRR) hold a non-zero value.
1 = Both the Transfer Counter Register (TCR) and the Transfer Counter Reload Register (TCRR) are zero.
• RCZ: Reload Counter Zero
0 = The Transfer Counter Reload Register (TCRR) holds a non-zero value.
1 = The Transfer Counter Reload Register (TCRR) is zero.
169
32058K
AVR32-01/12
AT32UC3A
22. General-Purpose Input/Output Controller (GPIO)
Rev. 1.1.0.2
22.1
Features
Each I/O line of the GPIO features:
•
•
•
•
•
•
22.2
Configurable pin-change, rising-edge or falling-edge interrupt on any I/O line.
A glitch filter providing rejection of pulses shorter than one clock cycle.
Open Drain mode enabling sharing of an I/O line between the MCU and external components.
Input visibility and output control.
Multiplexing of up to four peripheral functions per I/O line.
Programmable internal pull-up resistor.
Overview
The General Purpose Input/Output manages the I/O pins of the microcontroller. Each I/O line
may be dedicated as a general-purpose I/O or be assigned to a function of an embedded peripheral. This assures effective optimization of the pins of a product.
Table 22-1.
Overview of the GPIO system
PBA Configuration
Interface
Interrupt Controller
GPIO Interrupt Request
PIN
General Purpose
Input/Output - GPIO
Power Manager
GPIO Clock
PIN
PIN
MCU
I/O Pins
PIN
PIN
Embedded
Peripheral
22.3
Pin Control
Signals
Product dependencies
22.3.1
Module Configuration
Most of the features of the GPIO are configurable for each product. The programmer must refer
to the Peripherals Section for these settings.
Product specific settings includes:
170
32058K
AVR32-01/12
AT32UC3A
•
•
•
•
22.3.2
Number of I/O pins.
Functions implemented on each pin.
Peripheral function(s) multiplexed on each I/O pin.
Reset state of registers.
Interrupt Lines
The GPIO interrupt lines are connected to the interrupt controller. Using the GPIO interrupt
requires the interrupt controller to be programmed first.
22.3.3
22.4
Power and Clock Management
The clock for the GPIO is controlled by the power manager. The programmer must ensure that
the GPIO clock is enabled in the power manager before using the GPIO. The clock must be
enabled in order to access the configuration registers of the GPIO and when interrupts are
enabled. After configuring the GPIO, the clock can be disabled if interrupts are not enabled.
Functional Description
The GPIO controls the I/O lines of the microcontroller. The control logic associated with each pin
is represented in the figure below:
Figure 22-1. Overview of the GPIO pad connections
GPIO_ODER
GPIO_PUER
1
Periph. A output enable
0
Periph. B output enable
0
Periph. C output enable
Periph. D output enable
GPIO_PMR1
GPIO_PMR0
1
GPIO_GPER
GPIO_ODMER
Periph. A output data
Periph. B output data
0
Periph. C output data
Periph. D output data
0
1
GPIO_OVR
PAD
1
Periph. A input data
Periph. B input data
Periph. C input data
GPIO_PVR
Periph. D input data
GPIO_IER
0
Edge Detector
Glitch Filter
GPIO_IMR1
GPIO_GFER
22.4.1
1
1
0
Interrupt Request
GPIO_IMR0
Pull-up Resistor Control
Each I/O line is designed with an embedded pull-up resistor. The pull-up resistor can be enabled
or disabled by accessing the corresponding bit in PUER (Pull-up Enable Register). Control of the
pull-up resistor is possible whether an I/O line is controlled by a peripheral or the GPIO.
171
32058K
AVR32-01/12
AT32UC3A
22.4.2
I/O Line or Peripheral Function Selection
When a pin is multiplexed with one or more peripheral functions, the selection is controlled with
the register GPER. If a bit in the register is set, the corresponding pin is controlled by the GPIO.
If a bit is cleared, the corresponding pin is controlled by a peripheral function.
22.4.3
Peripheral Selection
The GPIO provides multiplexing of up to four peripheral functions on a single pin. The selection
is performed by accessing PMR0 (Peripheral Mux Register 0) and PMR1 (Peripheral Mux Register 1).
22.4.4
Output Control
When the I/O line is assigned to a peripheral function, i.e. the corresponding bit in GPER is at 0,
the drive of the I/O line is controlled by the peripheral. The peripheral, depending on the value in
PMR0 and PMR1, determines whether the pin is driven or not.
When the I/O line is controlled by the GPIO, the value of ODER (Output Driver Enable Register)
determines if the pin is driven or not. When a bit in this register is at 1, the corresponding I/O line
is driven by the GPIO. When the bit is at 0, the GPIO does not drive the line.
The level driven on an I/O line can be determined by writing OVR (Output Value Register).
22.4.5
Open Drain Mode
Each I/O line can be independently programmed to operate in open drain mode. This feature
permits several drivers to be connected on the I/O line. The drivers should only actively drive the
line low. An external pull-up resistor (or enabling the internal one) is generally required to guarantee a high level on the line when no driver is active.
The Open Drain feature is controlled by ODMER (Open Drain Mode Enable Register). The Open
Drain mode can be selected whether the I/O line is controlled by the GPIO or assigned to a
peripheral function.
22.4.6
Inputs
The level on each I/O line can be read through PVR (Pin Value Register). This register indicates
the level of the I/O lines regardless of whether the lines are driven by the GPIO or by an external
component. Note that due to power saving measures, PVR register can only be read when
GPER is set for the corresponding pin or if interrupt is enabled for the pin.
Output Line Timings
The figure below shows the timing of the I/O line when setting and clearing the Output Value
Register by accessing OVR. The same timing applies when performing a ‘set’ or ‘clear’ access
i.e. writing to OVRS or OVRC. The timing of PVR (Pin Value Register) is also shown.
172
32058K
AVR32-01/12
AT32UC3A
Figure 22-2. Output line timings
clock
Write GPIO_OVR to 1
Write GPIO_OVR to 0
PBA Access
PBA Access
GPIO_OVR / I/O Line
GPIO_PVR
22.4.7
Interrupts
The GPIO can be programmed to generate an interrupt when it detects an input change on an
I/O line. The module can be configured to signal an interrupt whenever a pin changes value or
only to trigger on rising edges or falling edges. Interrupt is enabled on a pin by setting the corresponding bit in IER (Interrupt Enable Register). The interrupt mode is set by accessing IMR0
(Interrupt Mode Register 0) and IMR1 (Interrupt Mode Register 1). Interrupt can be enabled on a
pin, regardless of the configuration the I/O line, i.e. controlled by the GPIO or assigned to a
peripheral function.
In every port there are four interrupt lines connected to the interrupt controller. Every eigth interrupts in the port are ored together to form an interrupt line.
When an interrupt event is detected on an I/O line, and the corresponding bit in IER is set, the
GPIO interrupt request line is asserted. A number of interrupt signals are ORed-wired together
to generate a single interrupt signal to the interrupt controller.
IFR (Interrupt Flag Register) can by read by software to determine which pin(s) caused the interrupt. The interrupt flag must be manually cleared by writing to IFR.
GPIO interrupts can only be triggered when the GPIO clock is enabled.
22.4.8
Input Glitch Filter
Optional input glitch filters can be enabled on each I/O line. When the glitch filter is enabled, a
glitch with duration of less than 1 clock cycle is automatically rejected, while a pulse with duration of 2 clock cycles or more is accepted. For pulse durations between 1 clock cycle and 2 clock
cycles, the pulse may or may not be taken into account, depending on the precise timing of its
occurrence. Thus for a pulse to be guaranteed visible it must exceed 2 clock cycles, whereas for
a glitch to be reliably filtered out, its duration must not exceed 1 clock cycle. The filter introduces
2 clock cycles latency.
The glitch filters are controlled by the register GFER (Glitch Filter Enable Register). When a bit is
set in GFER, the glitch filter on the corresponding pin is enabled. The glitch filter affects only
interrupt inputs. Inputs to peripherals or the value read through PVR are not affected by the
glitch filters.
22.4.9
Interrupt Timings
The figure below shows the timing for rising edge (or pin-change) interrupts when the glitch filter
is disabled. For the pulse to be registered, it must be sampled at the rising edge of the clock. In
this example, this is not the case for the first pulse. The second pulse is however sampled on a
rising edge and will trigger an interrupt request.
173
32058K
AVR32-01/12
AT32UC3A
Figure 22-3. Interrupt timing with glitch filter disabled
clock
Pin Level
GPIO_IFR
The figure below shows the timing for rising edge (or pin-change) interrupts when the glitch filter
is enabled. For the pulse to be registered, it must be sampled on two subsequent rising edges.
In the example, the first pulse is rejected while the second pulse is accepted and causes an
interrupt request.
Figure 22-4. Interrupt timing with glitch filter enabled
clock
Pin Level
GPIO_IFR
174
32058K
AVR32-01/12
AT32UC3A
22.5
General Purpose Input/Output (GPIO) User Interface
The GPIO controls all the I/O pins on the AVR32 microcontroller. The pins are managed as 32bit ports that are configurable through an PB interface. Each port has a set of configuration registers. The overall memory map of the GPIO is shown below. The number of pins and hence the
number of ports is product specific.
0x0000
Port 0 Configuration Registers
0x0100
Port 1 Configuration Registers
0x0200
Port 2 Configuration Registers
0x0300
Port 3 Configuration Registers
0x0400
Port 4 Configuration Registers
In the Peripheral muxing table in the Peripherals chapter each GPIO line has a unique number.
Note that the PA, PB, PC and PX ports do not directly correspond to the GPIO ports. To find the
corresponding port and pin the following formulas can be used:
GPIO port = floor((GPIO number) / 32), example: floor((36)/32) = 1
GPIO pin = GPIO number mod 32, example: 36 mod 32 = 4
The table below shows the configuration registers for one port. Addresses shown are relative to
the port address offset. The specific address of a configuration register is found by adding the
register offset and the port offset to the GPIO start address. One bit in each of the configuration
registers corresponds to an I/O pin.
GPIO Register Map
Table 22-2.
Offset
Register
Function
Name
Access
Reset value
0x00
GPIO Enable Register
Read/Write
GPER
Read/Write
1b for each
implemented
GPIO pin in port
0x04
GPIO Enable Register
Set
GPERS
Write-Only
0x08
GPIO Enable Register
Clear
GPERC
Write-Only
0x0C
GPIO Enable Register
Toggle
GPERT
Write-Only
0x10
Peripheral Mux Register 0
Read/Write
PMR0
Read/Write
0x00000000
175
32058K
AVR32-01/12
AT32UC3A
GPIO Register Map
Table 22-2.
Offset
Register
Function
Name
Access
0x14
Peripheral Mux Register 0
Set
PMR0S
Write-Only
0x18
Peripheral Mux Register 0
Clear
PMR0C
Write-Only
0x1C
Peripheral Mux Register 0
Toggle
PMR0T
Write-Only
0x20
Peripheral Mux Register 1
Read/Write
PMR1
Read/Write
0x24
Peripheral Mux Register 1
Set
PMR1S
Write-Only
0x28
Peripheral Mux Register 1
Clear
PMR1C
Write-Only
0x2C
Peripheral Mux Register 1
Toggle
PMR1T
Write-Only
0x30
RESERVED
-
-
-
0x34
RESERVED
-
-
-
0x38
RESERVED
-
-
-
0x3C
RESERVED
-
-
-
0x40
Output Driver Enable Register
Read/Write
ODER
Read/Write
0x44
Output Driver Enable Register
Set
ODERS
Write-Only
0x48
Output Driver Enable Register
Clear
ODERC
Write-Only
0x4C
Output Driver Enable Register
Toggle
ODERT
Write-Only
0x50
Output Value Register
Read/Write
OVR
Read/Write
0x54
Output Value Register
Set
OVRS
Write-Only
0x58
Output Value Register
Clear
OVRC
Write-Only
0x5c
Output Value Register
Toggle
OVRT
Write-Only
0x60
Pin Value Register
Read
PVR
Read-Only
0x64
Pin Value Register
-
-
-
0x68
Pin Value Register
-
-
-
0x6c
Pin Value Register
-
-
-
0x70
Pull-up Enable Register
Read/Write
PUER
Read/Write
0x74
Pull-up Enable Register
Set
PUERS
Write-Only
0x78
Pull-up Enable Register
Clear
PUERC
Write-Only
0x7C
Pull-up Enable Register
Toggle
PUERT
Write-Only
0x80
Open Drain Mode Enable Register
Read/Write
ODMER
Read/Write
0x84
Open Drain Mode Enable Register
Set
ODMERS
Write-Only
0x88
Open Drain Mode Enable Register
Clear
ODMERC
Write-Only
0x8C
Open Drain Mode Enable Register
Toggle
ODMERT
Write-Only
0x90
Interrupt Enable Register
Read/Write
IER
Read/Write
0x94
Interrupt Enable Register
Set
IERS
Write-Only
0x98
Interrupt Enable Register
Clear
IERC
Write-Only
0x9C
Interrupt Enable Register
Toggle
IERT
Write-Only
Reset value
0x00000000
0x00000000
0x00000000
depending on
pin states
0x00000000
0x00000000
0x00000000
176
32058K
AVR32-01/12
AT32UC3A
GPIO Register Map
Table 22-2.
Offset
Register
Function
Name
Access
Reset value
0xA0
Interrupt Mode Register 0
Read/Write
IMR0
Read/Write
0x00000000
0xA4
Interrupt Mode Register 0
Set
IMR0S
Write-Only
0xA8
Interrupt Mode Register 0
Clear
IMR0C
Write-Only
0xAC
Interrupt Mode Register 0
Toggle
IMR0T
Write-Only
0xB0
Interrupt Mode Register 1
Read/Write
IMR1
Read/Write
0xB4
Interrupt Mode Register 1
Set
IMR1S
Write-Only
0xB8
Interrupt Mode Register 1
Clear
IMR1C
Write-Only
0xBC
Interrupt Mode Register 1
Toggle
IMR1T
Write-Only
0xC0
Glitch Filter Enable Register
Read/Write
GFER
Read/Write
0xC4
Glitch Filter Enable Register
Set
GFERS
Write-Only
0xC8
Glitch Filter Enable Register
Clear
GFERC
Write-Only
0xCC
Glitch Filter Enable Register
Toggle
GFERT
Write-Only
0xD0
Interrupt Flag Register
Read
IFR
Read-Only
0xD4
Interrupt Flag Register
-
-
-
0xD8
Interrupt Flag Register
Clear
IFRC
Write-Only
0xDC
Interrupt Flag Register
-
-
-
0xE00xFF
RESERVED
-
-
-
22.5.1
0x00000000
1b for each
implemented
GPIO pin in port
0x00000000
Access Types
Each configuration register can be accessed in four different ways. The first address location
can be used to write the register directly. This address can also be used to read the register
value. The following addresses facilitate three different types of write access to the register. Performing a “set” access, all bits written to ‘1’ will be set. Bits written to ‘0’ will be unchanged by the
operation. Performing a “clear” access, all bits written to ‘1’ will be cleared. Bits written to ‘0’ will
be unchanged by the operation. Finally, a toggle access will toggle the value of all bits written to
‘1’. Again all bits written to ‘0’ remain unchanged. Note that for some registers (e.g. IFR), not all
access methods are permitted.
Note that for ports with less than 32 bits, the corresponding control registers will have unused
bits. This is also the case for features that are not implemented for a specific pin. Writing to an
unused bit will have no effect. Reading unused bits will always return 0.
177
32058K
AVR32-01/12
AT32UC3A
22.5.2
Name:
GPIO Enable Register
GPER
Access:
Read, Write, Set, Clear, Toggle
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-P31: GPIO Enable
0 = A peripheral function controls the corresponding pin.
1 = The GPIO controls the corresponding pin.
178
32058K
AVR32-01/12
AT32UC3A
22.5.3
Name:
Peripheral Mux Register 0
PMR0
Read, Write, Set, Clear, Toggle
Access:
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-31: Peripheral Multiplexer Select bit 0
22.5.4
Name:
Peripheral Mux Register 1
PMR1
Read, Write, Set, Clear, Toggle
Access:
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-31: Peripheral Multiplexer Select bit 1
{PMR1, PMR0}
00
01
10
11
Selected Peripheral Function
A
B
C
D
179
32058K
AVR32-01/12
AT32UC3A
22.5.5
Name:
Output Driver Enable Register
ODER
Access:
Read, Write, Set, Clear, Toggle
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-31: Output Driver Enable
0 = The output driver is disabled for the corresponding pin.
1 = The output driver is enabled for the corresponding pin.
180
32058K
AVR32-01/12
AT32UC3A
22.5.6
Name:
Output Value Register
OVR
Access:
Read, Write, Set, Clear, Toggle
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-31: Output Value
0 = The value to be driven on the I/O line is 0.
1 = The value to be driven on the I/O line is 1.
181
32058K
AVR32-01/12
AT32UC3A
22.5.7
Name:
Pin Value Register
PVR
Access:
Read
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-31: Pin Value
0 = The I/O line is at level ‘0’.
1 = The I/O line is at level ‘1’.
Note that the level of a pin can only be read when GPER is set or interrupt is enabled for the pin.
182
32058K
AVR32-01/12
AT32UC3A
22.5.8
Name:
Pull-up Enable Register
PUER
Access:
Read, Write, Set, Clear, Toggle
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-31: Pull-up Enable
0 = The internal pull-up resistor is disabled for the corresponding pin.
1 = The internal pull-up resistor is enabled for the corresponding pin.
183
32058K
AVR32-01/12
AT32UC3A
22.5.9
Name:
Open Drain Mode Enable Register
ODMER
Access:
Read, Write, Set, Clear, Toggle
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-31: Open Drain Mode Enable
0 = Open drain mode is disabled for the corresponding pin.
1 = Open drain mode is enabled for the corresponding pin.
184
32058K
AVR32-01/12
AT32UC3A
22.5.10
Name:
Interrupt Enable Register
IER
Access:
Read, Write, Set, Clear, Toggle
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-31: Interrupt Enable
0 = Interrupt is disabled for the corresponding pin.
1 = Interrupt is enabled for the corresponding pin.
185
32058K
AVR32-01/12
AT32UC3A
22.5.11
Name:
Interrupt Mode Register 0
IMR0
Read, Write, Set, Clear, Toggle
Access:
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-31: Interrupt Mode Bit 0
22.5.12
Name:
Interrupt Mode Register 1
IMR1
Read, Write, Set, Clear, Toggle
Access:
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-31: Interrupt Mode Bit 1
{IMR1, IMR0}
00
01
10
11
Interrupt Mode
Pin Change
Rising Edge
Falling Edge
Reserved
186
32058K
AVR32-01/12
AT32UC3A
22.5.13
Name:
Glitch Filter Enable Register
GFER
Access:
Read, Write, Set, Clear, Toggle
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-31: Glitch Filter Enable
0 = Glitch filter is disabled for the corresponding pin.
1 = Glitch filter is enabled for the corresponding pin.
NOTE! The value of this register should only be changed when IER is ‘0’. Updating this GFER while interrupt on the corresponding pin is enabled can cause an unintentional interrupt to be triggered.
187
32058K
AVR32-01/12
AT32UC3A
22.5.14
Name:
Interrupt Flag Register
IFR
Access:
Read, Clear
31
30
29
28
27
26
25
24
P31
P30
P29
P28
P27
P26
P25
P24
23
22
21
20
19
18
17
16
P23
P22
P21
P20
P19
P18
P17
P16
15
14
13
12
11
10
9
8
P15
P14
P13
P12
P11
P10
P9
P8
7
6
5
4
3
2
1
0
P7
P6
P5
P4
P3
P2
P1
P0
• P0-31: Interrupt Flag
0 = An interrupt condition has been detected on the corresponding pin.
1 = No interrupt condition has been detected on the corresponding pin.
The number of interrupt request lines is dependant on the number of I/O pins on the MCU. Refer to the product specific
data for details. Note also that a bit in the Interrupt Flag register is only valid if the corresponding bit in IER is set.
188
32058K
AVR32-01/12
AT32UC3A
22.6
Programming Examples
22.6.1
8-bit LED-Chaser
// Set R0 to GPIO base address
mov
R0, LO(AVR32_GPIO_BASE_ADDRESS)
orh
R0, HI(AVR32_GPIO_BASE_ADDRESS)
// Enable GPIO control of pin 0-8
mov
R1, 0xFF
st.w
R0[AVR32_GPIO_GPERS], R1
// Set initial value of port
mov
R2, 0x01
st.w
R0[AVR32_GPIO_OVRS], R2
// Set up toggle value. Two pins are toggled
// in each round. The bit that is currently set,
// and the next bit to be set.
mov
R2, 0x0303
orh
R2, 0x0303
loop:
// Only change 8 LSB
mov
R3, 0x00FF
and
R3, R2
st.w
R0[AVR32_GPIO_OVRT], R3
rol
R2
rcall
delay
rjmp
loop
It is assumed in this example that a subroutine "delay" exists that returns after a given time.
22.6.2
Configuration of USART pins
The example below shows how to configure a peripheral module to control I/O pins. It assumed
in this example that the USART receive pin (RXD) is connected to PC16 and that the USART
transmit pin (TXD) is connected to PC17. For both pins, the USART is peripheral B. In this
example, the state of the GPIO registers is assumed to be unknown. The two USART pins are
therefore first set to be controlled by the GPIO with output drivers disabled. The pins can then be
assured to be tri-stated while changing the Peripheral Mux Registers.
// Set up pointer to GPIO, PORTC
mov
R0, LO(AVR32_GPIO_BASE_ADDRESS + PORTC_OFFSET)
orh
R0, HI(AVR32_GPIO_BASE_ADDRESS + PORTC_OFFSET)
// Disable output drivers
189
32058K
AVR32-01/12
AT32UC3A
mov
R1, 0x0000
orh
R1, 0x0003
st.w
R0[AVR32_GPIO_ODERC], R1
// Make the GPIO control the pins
st.w
R0[AVR32_GPIO_GPERS], R1
// Select peripheral B on PC16-PC17
st.w
R0[AVR32_GPIO_PMR0S], R1
st.w
R0[GPIO_PMR1C], R1
// Enable peripheral control
st.w
R0[AVR32_GPIO_GPERC], R1
190
32058K
AVR32-01/12
AT32UC3A
23. Serial Peripheral Interface (SPI)
Rev: 1.9.9.3
23.1
Features
• Supports Communication with Serial External Devices
– Four Chip Selects with External Decoder Support Allow Communication with Up to 15
Peripherals
– Serial Memories, such as DataFlash and 3-wire EEPROMs
– Serial Peripherals, such as ADCs, DACs, LCD Controllers, CAN Controllers and Sensors
– External Co-processors
• Master or Slave Serial Peripheral Bus Interface
– 8- to 16-bit Programmable Data Length Per Chip Select
– Programmable Phase and Polarity Per Chip Select
– Programmable Transfer Delays Between Consecutive Transfers and Between Clock and Data
Per Chip Select
– Programmable Delay Between Consecutive Transfers
– Selectable Mode Fault Detection
• Connection to PDC Channel Capabilities Optimizes Data Transfers
– One Channel for the Receiver, One Channel for the Transmitter
– Next Buffer Support
23.2
Description
The Serial Peripheral Interface (SPI) circuit is a synchronous serial data link that provides communication with external devices in Master or Slave Mode. It also enables communication
between processors if an external processor is connected to the system.
The Serial Peripheral Interface is essentially a shift register that serially transmits data bits to
other SPIs. During a data transfer, one SPI system acts as the “master”' which controls the data
flow, while the other devices act as “slaves'' which have data shifted into and out by the master.
Different CPUs can take turn being masters (Multiple Master Protocol opposite to Single Master
Protocol where one CPU is always the master while all of the others are always slaves) and one
master may simultaneously shift data into multiple slaves. However, only one slave may drive its
output to write data back to the master at any given time.
A slave device is selected when the master asserts its NSS signal. If multiple slave devices
exist, the master generates a separate slave select signal for each slave (NPCS).
The SPI system consists of two data lines and two control lines:
• Master Out Slave In (MOSI): This data line supplies the output data from the master shifted
into the input(s) of the slave(s).
• Master In Slave Out (MISO): This data line supplies the output data from a slave to the input of
the master. There may be no more than one slave transmitting data during any particular
transfer.
• Serial Clock (SPCK): This control line is driven by the master and regulates the flow of the data
bits. The master may transmit data at a variety of baud rates; the SPCK line cycles once for
each bit that is transmitted.
• Slave Select (NSS): This control line allows slaves to be turned on and off by hardware.
191
32058K
AVR32-01/12
AT32UC3A
23.3
Block Diagram
Figure 23-1. Block Diagram
PDC
eral Bus
SPCK
MISO
Power
Manager
MOSI
MCK
SPI Interface
PIO
NPCS0/NSS
NPCS1
DIV
NPCS2
MCK (1)
32
Interrupt Control
NPCS3
SPI Interrupt
192
32058K
AVR32-01/12
AT32UC3A
23.4
Application Block Diagram
Figure 23-2. Application Block Diagram: Single Master/Multiple Slave Implementation
SPI Master
SPCK
SPCK
MISO
MISO
MOSI
MOSI
NPCS0
NSS
SPCK
NPCS1
NPCS2
NPCS3
Slave 0
NC
MISO
Slave 1
MOSI
NSS
SPCK
MISO
Slave 2
MOSI
NSS
193
32058K
AVR32-01/12
AT32UC3A
23.5
Signal Description
Table 23-1.
Signal Description
Type
Pin Name
Pin Description
Master
Slave
MISO
Master In Slave Out
Input
Output
MOSI
Master Out Slave In
Output
Input
SPCK
Serial Clock
Output
Input
NPCS1-NPCS3
Peripheral Chip Selects
Output
Unused
NPCS0/NSS
Peripheral Chip Select/Slave Select
Output
Input
194
32058K
AVR32-01/12
AT32UC3A
23.6
Product Dependencies
23.6.1
I/O Lines
The pins used for interfacing the compliant external devices may be multiplexed with PIO lines.
The programmer must first program the PIO controllers to assign the SPI pins to their peripheral
functions. To use the local loopback function the SPI pins must be controlled by the SPI.
23.6.2
Power Management
The SPI clock is generated by the Power Manager. Before using the SPI, the programmer must
ensure that the SPI clock is enabled in the Power Manager.
In the SPI description, Master Clock (MCK) is the clock of the peripheral bus to which the SPI is
connected.
23.6.3
Interrupt
The SPI interface has an interrupt line connected to the Interrupt Controller. Handling the SPI
interrupt requires programming the interrupt controller before configuring the SPI.
195
32058K
AVR32-01/12
AT32UC3A
23.7
Functional Description
23.7.1
Modes of Operation
The SPI operates in Master Mode or in Slave Mode.
Operation in Master Mode is programmed by writing at 1 the MSTR bit in the Mode Register.
The pins NPCS0 to NPCS3 are all configured as outputs, the SPCK pin is driven, the MISO line
is wired on the receiver input and the MOSI line driven as an output by the transmitter.
If the MSTR bit is written at 0, the SPI operates in Slave Mode. The MISO line is driven by the
transmitter output, the MOSI line is wired on the receiver input, the SPCK pin is driven by the
transmitter to synchronize the receiver. The NPCS0 pin becomes an input, and is used as a
Slave Select signal (NSS). The pins NPCS1 to NPCS3 are not driven and can be used for other
purposes.
The data transfers are identically programmable for both modes of operations. The baud rate
generator is activated only in Master Mode.
23.7.2
Data Transfer
Four combinations of polarity and phase are available for data transfers. The clock polarity is
programmed with the CPOL bit in the Chip Select Register. The clock phase is programmed with
the NCPHA bit. These two parameters determine the edges of the clock signal on which data is
driven and sampled. Each of the two parameters has two possible states, resulting in four possible combinations that are incompatible with one another. Thus, a master/slave pair must use the
same parameter pair values to communicate. If multiple slaves are used and fixed in different
configurations, the master must reconfigure itself each time it needs to communicate with a different slave.
Table 23-2 shows the four modes and corresponding parameter settings.
Table 23-2.
SPI Bus Protocol Mode
SPI Mode
CPOL
NCPHA
0
0
1
1
0
0
2
1
1
3
1
0
Figure 23-3 and Figure 23-4 show examples of data transfers.
196
32058K
AVR32-01/12
AT32UC3A
Figure 23-3. SPI Transfer Format (NCPHA = 1, 8 bits per transfer)
1
SPCK cycle (for reference)
2
3
4
6
5
7
8
SPCK
(CPOL = 0)
SPCK
(CPOL = 1)
MOSI
(from master)
MSB
MISO
(from slave)
MSB
6
5
4
3
2
1
LSB
6
5
4
3
2
1
LSB
*
NSS
(to slave)
* Not defined, but normally MSB of previous character received.
Figure 23-4. SPI Transfer Format (NCPHA = 0, 8 bits per transfer)
1
SPCK cycle (for reference)
2
3
4
5
8
7
6
SPCK
(CPOL = 0)
SPCK
(CPOL = 1)
MOSI
(from master)
MISO
(from slave)
*
MSB
6
5
4
3
2
1
MSB
6
5
4
3
2
1
LSB
LSB
NSS
(to slave)
* Not defined but normally LSB of previous character transmitted.
197
32058K
AVR32-01/12
AT32UC3A
23.7.3
Master Mode Operations
When configured in Master Mode, the SPI uses the internal programmable baud rate generator
as clock source. It fully controls the data transfers to and from the slave(s) connected to the SPI
bus. The SPI drives the chip select line to the slave and the serial clock signal (SPCK).
The SPI features two holding registers, the Transmit Data Register and the Receive Data Register, and a single Shift Register. The holding registers maintain the data flow at a constant rate.
After enabling the SPI, a data transfer begins when the processor writes to the TDR (Transmit
Data Register). The written data is immediately transferred in the Shift Register and transfer on
the SPI bus starts. While the data in the Shift Register is shifted on the MOSI line, the MISO line
is sampled and shifted in the Shift Register. Transmission cannot occur without reception.
Before writing the TDR, the PCS field must be set in order to select a slave.
If new data is written in TDR during the transfer, it stays in it until the current transfer is completed. Then, the received data is transferred from the Shift Register to RDR, the data in TDR is
loaded in the Shift Register and a new transfer starts.
The transfer of a data written in TDR in the Shift Register is indicated by the TDRE bit (Transmit
Data Register Empty) in the Status Register (SR). When new data is written in TDR, this bit is
cleared. The TDRE bit is used to trigger the Transmit PDC channel.
The end of transfer is indicated by the TXEMPTY flag in the SR register. If a transfer delay (DLYBCT) is greater than 0 for the last transfer, TXEMPTY is set after the completion of said delay.
The master clock (MCK) can be switched off at this time.
The transfer of received data from the Shift Register in RDR is indicated by the RDRF bit
(Receive Data Register Full) in the Status Register (SR). When the received data is read, the
RDRF bit is cleared.
If the RDR (Receive Data Register) has not been read before new data is received, the Overrun
Error bit (OVRES) in SR is set. When this bit is set the SPI will continue to update RDR when
data is received, overwriting the previously received data. The user has to read the status register to clear the OVRES bit.
Figure 23-5 on page 199 shows a block diagram of the SPI when operating in Master Mode. Figure 23-6 on page 200 shows a flow chart describing how transfers are handled.
198
32058K
AVR32-01/12
AT32UC3A
23.7.3.1
Master Mode Block Diagram
Figure 23-5. Master Mode Block Diagram
FDIV
SPI_CSR0..3
SCBR
MCK
0
Baud Rate Generator
MCK/N
SPCK
1
SPI
Clock
SPI_CSR0..3
BITS
NCPHA
CPOL
LSB
MISO
SPI_RDR
RDRF
OVRES
RD
MSB
Shift Register
MOSI
SPI_TDR
TDRE
TD
SPI_CSR0..3
CSAAT
SPI_RDR
PCS
PS
NPCS3
PCSDEC
SPI_MR
PCS
0
NPCS2
Current
Peripheral
NPCS1
SPI_TDR
NPCS0
PCS
1
MSTR
MODF
NPCS0
MODFDIS
199
32058K
AVR32-01/12
AT32UC3A
23.7.3.2
Master Mode Flow Diagram
Figure 23-6. Master Mode Flow Diagram S
SPI Enable
- NPCS defines the current Chip Select
- CSAAT, DLYBS, DLYBCT refer to the fields of the
Chip Select Register corresponding to the Current Chip Select
- When NPCS is 0xF, CSAAT is 0.
1
TDRE ?
0
1
CSAAT ?
1
0
0
PS ?
1
Fixed
peripheral
Variable
peripheral
Variable
peripheral
SPI_TDR(PCS)
= NPCS ?
no
NPCS = SPI_TDR(PCS)
NPCS = SPI_MR(PCS)
Fixed
peripheral
0
PS ?
yes
SPI_MR(PCS)
= NPCS ?
no
NPCS = 0xF
NPCS = 0xF
Delay DLYBCS
Delay DLYBCS
NPCS = SPI_TDR(PCS)
NPCS = SPI_MR(PCS),
SPI_TDR(PCS)
Delay DLYBS
Serializer = SPI_TDR(TD)
TDRE = 1
Data Transfer
SPI_RDR(RD) = Serializer
RDRF = 1
Delay DLYBCT
TDRE ?
0
1
1
CSAAT ?
0
NPCS = 0xF
Delay DLYBCS
200
32058K
AVR32-01/12
AT32UC3A
23.7.3.3
Clock Generation
The SPI Baud rate clock is generated by dividing the Master Clock (MCK) or the Master Clock
divided by 32, by a value between 1 and 255. The selection between Master Clock or Master
Clock divided by 32 is done by the FDIV value set in the Mode Register
This allows a maximum operating baud rate at up to Master Clock and a minimum operating
baud rate of MCK divided by 255*32.
Programming the SCBR field at 0 is forbidden. Triggering a transfer while SCBR is at 0 can lead
to unpredictable results.
At reset, SCBR is 0 and the user has to program it at a valid value before performing the first
transfer.
The divisor can be defined independently for each chip select, as it has to be programmed in the
SCBR field of the Chip Select Registers. This allows the SPI to automatically adapt the baud
rate for each interfaced peripheral without reprogramming.
23.7.3.4
Transfer Delays
Figure 23-7 shows a chip select transfer change and consecutive transfers on the same chip
select. Three delays can be programmed to modify the transfer waveforms:
• The delay between chip selects, programmable only once for all the chip selects by writing the
DLYBCS field in the Mode Register. Allows insertion of a delay between release of one chip
select and before assertion of a new one.
• The delay before SPCK, independently programmable for each chip select by writing the field
DLYBS. Allows the start of SPCK to be delayed after the chip select has been asserted.
• The delay between consecutive transfers, independently programmable for each chip select by
writing the DLYBCT field. Allows insertion of a delay between two transfers occurring on the
same chip select
These delays allow the SPI to be adapted to the interfaced peripherals and their speed and bus
release time.
Figure 23-7. Programmable Delays
Chip Select 1
Chip Select 2
SPCK
DLYBCS
DLYBS
DLYBCT
DLYBCT
201
32058K AVR32-01/12
AT32UC3A
23.7.3.5
Peripheral Selection
The serial peripherals are selected through the assertion of the NPCS0 to NPCS3 signals. By
default, all the NPCS signals are high before and after each transfer.
The peripheral selection can be performed in two different ways:
• Fixed Peripheral Select: SPI exchanges data with only one peripheral
• Variable Peripheral Select: Data can be exchanged with more than one peripheral
Fixed Peripheral Select is activated by writing the PS bit to zero in MR (Mode Register). In this
case, the current peripheral is defined by the PCS field in MR and the PCS field in TDR have no
effect.
Variable Peripheral Select is activated by setting PS bit to one. The PCS field in TDR is used to
select the current peripheral. This means that the peripheral selection can be defined for each
new data.
The Fixed Peripheral Selection allows buffer transfers with a single peripheral. Using the PDC is
an optimal means, as the size of the data transfer between the memory and the SPI is either 8
bits or 16 bits. However, changing the peripheral selection requires the Mode Register to be
reprogrammed.
The Variable Peripheral Selection allows buffer transfers with multiple peripherals without reprogramming the Mode Register. Data written in TDR is 32 bits wide and defines the real data to be
transmitted and the peripheral it is destined to. Using the PDC in this mode requires 32-bit wide
buffers, with the data in the LSBs and the PCS and LASTXFER fields in the MSBs, however the
SPI still controls the number of bits (8 to16) to be transferred through MISO and MOSI lines with
the chip select configuration registers. This is not the optimal means in term of memory size for
the buffers, but it provides a very effective means to exchange data with several peripherals
without any intervention of the processor.
23.7.3.6
Peripheral Chip Select Decoding
The user can program the SPI to operate with up to 15 peripherals by decoding the four Chip
Select lines, NPCS0 to NPCS3 with an external logic. This can be enabled by writing the PCSDEC bit at 1 in the Mode Register (MR).
When operating without decoding, the SPI makes sure that in any case only one chip select line
is activated, i.e. driven low at a time. If two bits are defined low in a PCS field, only the lowest
numbered chip select is driven low.
When operating with decoding, the SPI directly outputs the value defined by the PCS field of
either the Mode Register or the Transmit Data Register (depending on PS).
As the SPI sets a default value of 0xF on the chip select lines (i.e. all chip select lines at 1) when
not processing any transfer, only 15 peripherals can be decoded.
The SPI has only four Chip Select Registers, not 15. As a result, when decoding is activated,
each chip select defines the characteristics of up to four peripherals. As an example, CRS0
defines the characteristics of the externally decoded peripherals 0 to 3, corresponding to the
PCS values 0x0 to 0x3. Thus, the user has to make sure to connect compatible peripherals on
the decoded chip select lines 0 to 3, 4 to 7, 8 to 11 and 12 to 14.
202
32058K AVR32-01/12
AT32UC3A
23.7.3.7
Peripheral Deselection
When operating normally, as soon as the transfer of the last data written in TDR is completed,
the NPCS lines all rise. This might lead to runtime error if the processor is too long in responding
to an interrupt, and thus might lead to difficulties for interfacing with some serial peripherals
requiring the chip select line to remain active during a full set of transfers.
To facilitate interfacing with such devices, the Chip Select Register can be programmed with the
CSAAT bit (Chip Select Active After Transfer) at 1. This allows the chip select lines to remain in
their current state (low = active) until transfer to another peripheral is required.
Figure 23-8 shows different peripheral deselection cases and the effect of the CSAAT bit.
Figure 23-8. Peripheral Deselection
CSAAT = 0
TDRE
NPCS[0..3]
CSAAT = 1
DLYBCT
DLYBCT
A
A
A
A
DLYBCS
A
DLYBCS
PCS = A
PCS = A
Write SPI_TDR
TDRE
NPCS[0..3]
DLYBCT
DLYBCT
A
A
A
A
DLYBCS
A
DLYBCS
PCS=A
PCS = A
Write SPI_TDR
TDRE
NPCS[0..3]
DLYBCT
DLYBCT
A
B
A
B
DLYBCS
PCS = B
DLYBCS
PCS = B
Write SPI_TDR
203
32058K AVR32-01/12
AT32UC3A
23.7.3.8
Mode Fault Detection
A mode fault is detected when the SPI is programmed in Master Mode and a low level is driven
by an external master on the NPCS0/NSS signal. NPCS0, MOSI, MISO and SPCK must be configured in open-drain through the PIO controller, so that external pull up resistors are needed to
guarantee high level.
When a mode fault is detected, the MODF bit in the SR is set until the SR is read and the SPI is
automatically disabled until re-enabled by writing the SPIEN bit in the CR (Control Register) at 1.
By default, the Mode Fault detection circuitry is enabled. The user can disable Mode Fault
detection by setting the MODFDIS bit in the SPI Mode Register (MR).
23.7.4
SPI Slave Mode
When operating in Slave Mode, the SPI processes data bits on the clock provided on the SPI
clock pin (SPCK).
The SPI waits for NSS to go active before receiving the serial clock from an external master.
When NSS falls, the clock is validated on the serializer, which processes the number of bits
defined by the BITS field of the Chip Select Register 0 (CSR0). These bits are processed following a phase and a polarity defined respectively by the NCPHA and CPOL bits of the CSR0. Note
that BITS, CPOL and NCPHA of the other Chip Select Registers have no effect when the SPI is
programmed in Slave Mode.
The bits are shifted out on the MISO line and sampled on the MOSI line.
When all the bits are processed, the received data is transferred in the Receive Data Register
and the RDRF bit rises. If RDRF is already high when the data is transferred, the Overrun bit
rises and the data transfer to RDR is aborted.
When a transfer starts, the data shifted out is the data present in the Shift Register. If no data
has been written in the Transmit Data Register (TDR), the last data received is transferred. If no
data has been received since the last reset, all bits are transmitted low, as the Shift Register
resets at 0.
When a first data is written in TDR, it is transferred immediately in the Shift Register and the
TDRE bit rises. If new data is written, it remains in TDR until a transfer occurs, i.e. NSS falls and
there is a valid clock on the SPCK pin. When the transfer occurs, the last data written in TDR is
transferred in the Shift Register and the TDRE bit rises. This enables frequent updates of critical
variables with single transfers.
Then, a new data is loaded in the Shift Register from the Transmit Data Register. In case no
character is ready to be transmitted, i.e. no character has been written in TDR since the last load
from TDR to the Shift Register, the Shift Register is not modified and the last received character
is retransmitted.
Figure 23-9 shows a block diagram of the SPI when operating in Slave Mode.
204
32058K AVR32-01/12
AT32UC3A
Figure 23-9. Slave Mode Functional Block Diagram
SPCK
NSS
SPIEN
SPI
Clock
SPIENS
SPIDIS
SPI_CSR0
BITS
NCPHA
CPOL
MOSI
LSB
SPI_RDR
MSB
Shift Register
FLOAD
RDRF
OVRES
RD
SPI_TDR
TD
MISO
TDRE
205
32058K AVR32-01/12
AT32UC3A
23.8
Serial Peripheral Interface (SPI) User Interface
Table 23-3.
SPI Register Mapping
Offset
Register
Register Name
Access
Reset
0x00
Control Register
CR
Write-only
---
0x04
Mode Register
MR
Read/Write
0x0
0x08
Receive Data Register
RDR
Read-only
0x0
0x0C
Transmit Data Register
TDR
Write-only
---
0x10
Status Register
SR
Read-only
0x000000F0
0x14
Interrupt Enable Register
IER
Write-only
---
0x18
Interrupt Disable Register
IDR
Write-only
---
0x1C
Interrupt Mask Register
IMR
Read-only
0x0
0x20 - 0x2C
Reserved
0x30
Chip Select Register 0
CSR0
Read/Write
0x0
0x34
Chip Select Register 1
CSR1
Read/Write
0x0
0x38
Chip Select Register 2
CSR2
Read/Write
0x0
0x3C
Chip Select Register 3
CSR3
Read/Write
0x0
206
32058K AVR32-01/12
AT32UC3A
23.8.1
SPI Control Register
Name:
CR
Access Type:
Write-only
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
LASTXFER
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
–
15
14
13
12
11
10
9
8
–
–
–
–
–
–
–
–
7
6
5
4
3
2
1
0
SWRST
–
–
–
–
–
SPIDIS
SPIEN
• SPIEN: SPI Enable
0 = No effect.
1 = Enables the SPI to transfer and receive data.
• SPIDIS: SPI Disable
0 = No effect.
1 = Disables the SPI.
As soon as SPDIS is set, SPI finishes its transfer.
All pins are set in input mode and no data is received or transmitted.
If a transfer is in progress, the transfer is finished before the SPI is disabled.
If both SPIEN and SPIDIS are equal to one when the control register is written, the SPI is disabled.
• SWRST: SPI Software Reset
0 = No effect.
1 = Reset the SPI. A software-triggered hardware reset of the SPI interface is performed.
The SPI is in slave mode after a software reset.
PDC channels are not affected by software reset.
• LASTXFER: Last Transfer
0 = No effect.
1 = The current NPCS will be deasserted after the character written in TD has been transferred. When CSAAT is set, this
allows to close the communication with the current serial peripheral by raising the corresponding NPCS line as soon as TD
transfer has completed.
207
32058K AVR32-01/12
AT32UC3A
23.8.2
SPI Mode Register
Name:
MR
Access Type:
Read/Write
31
30
29
28
27
26
19
18
25
24
17
16
DLYBCS
23
22
21
20
–
–
–
–
15
14
13
12
11
10
9
8
–
–
–
–
–
–
–
–
PCS
7
6
5
4
3
2
1
0
LLB
–
–
MODFDIS
FDIV
PCSDEC
PS
MSTR
• MSTR: Master/Slave Mode
0 = SPI is in Slave mode.
1 = SPI is in Master mode.
• PS: Peripheral Select
0 = Fixed Peripheral Select.
1 = Variable Peripheral Select.
• PCSDEC: Chip Select Decode
0 = The chip selects are directly connected to a peripheral device.
1 = The four chip select lines are connected to a 4- to 16-bit decoder.
When PCSDEC equals one, up to 15 Chip Select signals can be generated with the four lines using an external 4- to 16-bit
decoder. The Chip Select Registers define the characteristics of the 15 chip selects according to the following rules:
CSR0 defines peripheral chip select signals 0 to 3.
CSR1 defines peripheral chip select signals 4 to 7.
CSR2 defines peripheral chip select signals 8 to 11.
CSR3 defines peripheral chip select signals 12 to 14.
• FDIV: Clock Selection
0 = The SPI operates at MCK.
1 = The SPI operates at MCK/32.
• MODFDIS: Mode Fault Detection
0 = Mode fault detection is enabled.
1 = Mode fault detection is disabled.
• LLB: Local Loopback Enable
0 = Local loopback path disabled.
1 = Local loopback path enabled.
LLB controls the local loopback on the data serializer for testing in Master Mode only. MISO is internally connected to
MOSI.
208
32058K AVR32-01/12
AT32UC3A
• PCS: Peripheral Chip Select
This field is only used if Fixed Peripheral Select is active (PS = 0).
If PCSDEC = 0:
PCS = xxx0
NPCS[3:0] = 1110
PCS = xx01
NPCS[3:0] = 1101
PCS = x011
NPCS[3:0] = 1011
PCS = 0111
NPCS[3:0] = 0111
PCS = 1111
forbidden (no peripheral is selected)
(x = don’t care)
If PCSDEC = 1:
NPCS[3:0] output signals = PCS.
• DLYBCS: Delay Between Chip Selects
This field defines the delay from NPCS inactive to the activation of another NPCS. The DLYBCS time guarantees non-overlapping chip selects and solves bus contentions in case of peripherals having long data float times.
If DLYBCS is less than or equal to six, six MCK periods (or 6*N MCK periods if FDIV is set) will be inserted by default.
Otherwise, the following equation determines the delay:
If FDIV is 0:
DLYBCS
Delay Between Chip Selects = ----------------------MCK
If FDIV is 1:
DLYBCS × NDelay Between Chip Selects = -------------------------------MCK
209
32058K AVR32-01/12
AT32UC3A
23.8.3
SPI Receive Data Register
Name:
RDR
Access Type:
Read-only
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
15
14
13
12
PCS
11
10
9
8
3
2
1
0
RD
7
6
5
4
RD
• RD: Receive Data
Data received by the SPI Interface is stored in this register right-justified. Unused bits read zero.
• PCS: Peripheral Chip Select
In Master Mode only, these bits indicate the value on the NPCS pins at the end of a transfer. Otherwise, these bits read
zero.
210
32058K AVR32-01/12
AT32UC3A
23.8.4
SPI Transmit Data Register
Name:
TDR
Access Type:
Write-only
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
LASTXFER
23
22
21
20
19
18
17
16
–
–
–
–
15
14
13
12
PCS
11
10
9
8
3
2
1
0
TD
7
6
5
4
TD
• TD: Transmit Data
Data to be transmitted by the SPI Interface is stored in this register. Information to be transmitted must be written to the
transmit data register in a right-justified format.
PCS: Peripheral Chip Select
This field is only used if Variable Peripheral Select is active (PS = 1).
If PCSDEC = 0:
PCS = xxx0
NPCS[3:0] = 1110
PCS = xx01
NPCS[3:0] = 1101
PCS = x011
NPCS[3:0] = 1011
PCS = 0111
NPCS[3:0] = 0111
PCS = 1111
forbidden (no peripheral is selected)
(x = don’t care)
If PCSDEC = 1:
NPCS[3:0] output signals = PCS
• LASTXFER: Last Transfer
0 = No effect.
1 = The current NPCS will be deasserted after the character written in TD has been transferred. When CSAAT is set, this
allows to close the communication with the current serial peripheral by raising the corresponding NPCS line as soon as TD
transfer has completed.
This field is only used if Variable Peripheral Select is active (PS = 1).
211
32058K AVR32-01/12
AT32UC3A
23.8.5
SPI Status Register
Name:
SR
Access Type:
Read-only
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
SPIENS
15
14
13
12
11
10
9
8
–
–
–
–
–
–
TXEMPTY
NSSR
7
6
5
4
3
2
1
0
TXBUFE
RXBUFF
ENDTX
ENDRX
OVRES
MODF
TDRE
RDRF
• RDRF: Receive Data Register Full
0 = No data has been received since the last read of RDR
1 = Data has been received and the received data has been transferred from the serializer to RDR since the last read of
RDR.
• TDRE: Transmit Data Register Empty
0 = Data has been written to TDR and not yet transferred to the serializer.
1 = The last data written in the Transmit Data Register has been transferred to the serializer.
TDRE equals zero when the SPI is disabled or at reset. The SPI enable command sets this bit to one.
• MODF: Mode Fault Error
0 = No Mode Fault has been detected since the last read of SR.
1 = A Mode Fault occurred since the last read of the SR.
• OVRES: Overrun Error Status
0 = No overrun has been detected since the last read of SR.
1 = An overrun has occurred since the last read of SR.
An overrun occurs when RDR is loaded at least twice from the serializer since the last read of the RDR.
• ENDRX: End of RX buffer
0 = The Receive Counter Register has not reached 0 since the last write in RCR or RNCR.
1 = The Receive Counter Register has reached 0 since the last write in RCR or RNCR.
• ENDTX: End of TX buffer
0 = The Transmit Counter Register has not reached 0 since the last write in TCR or TNCR.
1 = The Transmit Counter Register has reached 0 since the last write in TCR or TNCR.
• RXBUFF: RX Buffer Full
0 = RCR or RNCR has a value other than 0.
1 = Both RCR and RNCR has a value of 0.
• TXBUFE: TX Buffer Empty
0 = TCR or TNCR has a value other than 0.
212
32058K AVR32-01/12
AT32UC3A
1 = Both TCR and TNCR has a value of 0.
• NSSR: NSS Rising
0 = No rising edge detected on NSS pin since last read.
1 = A rising edge occurred on NSS pin since last read.
• TXEMPTY: Transmission Registers Empty
0 = As soon as data is written in TDR.
1 = TDR and internal shifter are empty. If a transfer delay has been defined, TXEMPTY is set after the completion of such
delay.
• SPIENS: SPI Enable Status
0 = SPI is disabled.
1 = SPI is enabled.
213
32058K AVR32-01/12
AT32UC3A
23.8.6
SPI Interrupt Enable Register
Name:
IER
Access Type:
Write-only
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
–
15
14
13
12
11
10
9
8
–
–
–
–
–
–
TXEMPTY
NSSR
7
6
5
4
3
2
1
0
TXBUFE
RXBUFF
ENDTX
ENDRX
OVRES
MODF
TDRE
RDRF
• RDRF: Receive Data Register Full Interrupt Enable
• TDRE: SPI Transmit Data Register Empty Interrupt Enable
• MODF: Mode Fault Error Interrupt Enable
•
•
•
•
•
•
•
OVRES: Overrun Error Interrupt Enable
ENDRX: End of Receive Buffer Interrupt Enable
ENDTX: End of Transmit Buffer Interrupt Enable
RXBUFF: Receive Buffer Full Interrupt Enable
TXBUFE: Transmit Buffer Empty Interrupt Enable
TXEMPTY: Transmission Registers Empty Enable
NSSR: NSS Rising Interrupt Enable
0 = No effect.
1 = Enables the corresponding interrupt.
214
32058K AVR32-01/12
AT32UC3A
23.8.7
SPI Interrupt Disable Register
Name:
IDR
Access Type:
Write-only
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
–
15
14
13
12
11
10
9
8
–
–
–
–
–
–
TXEMPTY
NSSR
7
6
5
4
3
2
1
0
TXBUFE
RXBUFF
ENDTX
ENDRX
OVRES
MODF
TDRE
RDRF
• RDRF: Receive Data Register Full Interrupt Disable
• TDRE: SPI Transmit Data Register Empty Interrupt Disable
• MODF: Mode Fault Error Interrupt Disable
•
•
•
•
•
•
•
OVRES: Overrun Error Interrupt Disable
ENDRX: End of Receive Buffer Interrupt Disable
ENDTX: End of Transmit Buffer Interrupt Disable
RXBUFF: Receive Buffer Full Interrupt Disable
TXBUFE: Transmit Buffer Empty Interrupt Disable
TXEMPTY: Transmission Registers Empty Disable
NSSR: NSS Rising Interrupt Disable
0 = No effect.
1 = Disables the corresponding interrupt.
215
32058K AVR32-01/12
AT32UC3A
23.8.8
SPI Interrupt Mask Register
Name:
IMR
Access Type:
Read-only
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
–
15
14
13
12
11
10
9
8
–
–
–
–
–
–
TXEMPTY
NSSR
7
6
5
4
3
2
1
0
TXBUFE
RXBUFF
ENDTX
ENDRX
OVRES
MODF
TDRE
RDRF
• RDRF: Receive Data Register Full Interrupt Mask
• TDRE: SPI Transmit Data Register Empty Interrupt Mask
• MODF: Mode Fault Error Interrupt Mask
•
•
•
•
•
•
•
OVRES: Overrun Error Interrupt Mask
ENDRX: End of Receive Buffer Interrupt Mask
ENDTX: End of Transmit Buffer Interrupt Mask
RXBUFF: Receive Buffer Full Interrupt Mask
TXBUFE: Transmit Buffer Empty Interrupt Mask
TXEMPTY: Transmission Registers Empty Mask
NSSR: NSS Rising Interrupt Mask
0 = The corresponding interrupt is not enabled.
1 = The corresponding interrupt is enabled.
216
32058K AVR32-01/12
AT32UC3A
23.8.9
SPI Chip Select Register
Name:
CSR0... CSR3
Access Type:
Read/Write
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
DLYBCT
23
22
21
20
DLYBS
15
14
13
12
SCBR
7
6
5
4
BITS
3
2
1
0
CSAAT
CSNAAT
NCPHA
CPOL
• CPOL: Clock Polarity
0 = The inactive state value of SPCK is logic level zero.
1 = The inactive state value of SPCK is logic level one.
CPOL is used to determine the inactive state value of the serial clock (SPCK). It is used with NCPHA to produce the
required clock/data relationship between master and slave devices.
• NCPHA: Clock Phase
0 = Data is changed on the leading edge of SPCK and captured on the following edge of SPCK.
1 = Data is captured on the leading edge of SPCK and changed on the following edge of SPCK.
NCPHA determines which edge of SPCK causes data to change and which edge causes data to be captured. NCPHA is
used with CPOL to produce the required clock/data relationship between master and slave devices.
• CSNAAT: Chip Select Not Active After Transfer
0 = The Peripheral Chip Select Line rises as soon as the last transfer is acheived
1 = The Peripheral Chip Select Line rises after every transfer
CSNAAT can be used to force the Peripheral Chip Select Line to go inactive after every transfer. This allows successful
interfacing to SPI slave devices that require this behavior.
• CSAAT: Chip Select Active After Transfer
0 = The Peripheral Chip Select Line rises as soon as the last transfer is achieved.
1 = The Peripheral Chip Select does not rise after the last transfer is achieved. It remains active until a new transfer is
requested on a different chip select.
• BITS: Bits Per Transfer
The BITS field determines the number of data bits transferred. Reserved values should not be used, see Table 23-4 on
page 218.
217
32058K AVR32-01/12
AT32UC3A
.
Table 23-4.
BITS, Bits Per Transfer
BITS
Bits Per Transfer
0000
8
0001
9
0010
10
0011
11
0100
12
0101
13
0110
14
0111
15
1000
16
1001
Reserved
1010
Reserved
1011
Reserved
1100
Reserved
1101
Reserved
1110
Reserved
1111
Reserved
• SCBR: Serial Clock Baud Rate
In Master Mode, the SPI Interface uses a modulus counter to derive the SPCK baud rate from the Master Clock MCK. The
Baud rate is selected by writing a value from 1 to 255 in the SCBR field. The following equations determine the SPCK baud
rate:
If FDIV is 0:
MCK
SPCK Baudrate = --------------SCBR
If FDIV is 1:
MCK SPCK Baudrate = ----------------------------( N × SCBR )
Note:
N = 32
Programming the SCBR field at 0 is forbidden. Triggering a transfer while SCBR is at 0 can lead to unpredictable results.
At reset, SCBR is 0 and the user has to program it at a valid value before performing the first transfer.
• DLYBS: Delay Before SPCK
This field defines the delay from NPCS valid to the first valid SPCK transition.
When DLYBS equals zero, the NPCS valid to SPCK transition is 1/2 the SPCK clock period.
218
32058K AVR32-01/12
AT32UC3A
Otherwise, the following equations determine the delay:
If FDIV is 0:
DLYBS
Delay Before SPCK = ------------------MCK
If FDIV is 1:
N × DLYBS
Delay Before SPCK = ----------------------------MCK
Note:
N = 32
• DLYBCT: Delay Between Consecutive Transfers
This field defines the delay between two consecutive transfers with the same peripheral without removing the chip select.
The delay is always inserted after each transfer and before removing the chip select if needed.
When DLYBCT equals zero, no delay between consecutive transfers is inserted and the clock keeps its duty cycle over the
character transfers.
Otherwise, the following equation determines the delay:
If FDIV is 0:
SCBR× DLYBCT- + ---------------Delay Between Consecutive Transfers = 32
----------------------------------MCK
2MCK
If FDIV is 1:
N = 32
32 × N × DLYBCT N × SCBR
Delay Between Consecutive Transfers = ----------------------------------------------- + ------------------------MCK
2MCK
219
32058K AVR32-01/12
AT32UC3A
24. Two-Wire Interface (TWI)
2.1.1.0
24.1
Features
•
•
•
•
•
•
•
Compatible with Atmel Two-wire Interface Serial Memory and I²C Compatible Devices(1)
One, Two or Three Bytes for Slave Address
Sequential Read-write Operations
Master, Multi-master and Slave Mode Operation
Bit Rate: Up to 400 Kbits
General Call Supported in Slave mode
Connection to Peripheral DMA Controller (PDC) Channel Capabilities Optimizes Data Transfers in
Master Mode Only
– One Channel for the Receiver, One Channel for the Transmitter
– Next Buffer Support
Note:
1. See Table 24-1 below for details on compatibility with I²C Standard.
24.2
Overview
The Atmel Two-wire Interface (TWI) interconnects components on a unique two-wire bus, made
up of one clock line and one data line with speeds of up to 400 Kbits per second, based on a
byte-oriented transfer format. It can be used with any Atmel Two-wire Interface bus Serial
EEPROM and I²C compatible device such as Real Time Clock (RTC), Dot Matrix/Graphic LCD
Controllers and Temperature Sensor, to name but a few. The TWI is programmable as a master
or a slave with sequential or single-byte access. Multiple master capability is supported. Arbitration of the bus is performed internally and puts the TWI in slave mode automatically if the bus
arbitration is lost.
A configurable baud rate generator permits the output data rate to be adapted to a wide range of
core clock frequencies.
Below, Table 24-1 lists the compatibility level of the Atmel Two-wire Interface in Master Mode and
a full I2C compatible device.
Atmel TWI compatibility with I2C Standard
Table 24-1.
I2C Standard
Atmel TWI
Standard Mode Speed (100 KHz)
Supported
Fast Mode Speed (400 KHz)
Supported
7 or 10 bits Slave Addressing
Supported
(1)
START BYTE
Not Supported
Repeated Start (Sr) Condition
Supported
ACK and NACK Management
Supported
Slope control and input filtering (Fast mode)
Not Supported
Clock stretching
Supported
Note:
1. START + b000000001 + Ack + Sr
220
32058K AVR32-01/12
AT32UC3A
24.3
List of Abbreviations
Table 24-2.
24.4
Abbreviations
Abbreviation
Description
TWI
Two-wire Interface
A
Acknowledge
NA
Non Acknowledge
P
Stop
S
Start
Sr
Repeated Start
SADR
Slave Address
ADR
Any address except SADR
R
Read
W
Write
Block Diagram
Figure 24-1. Block Diagram
Peripheral Bus
Bridge
TWCK
PIO
PM
MCK
TWD
Two-wire
Interface
TWI
Interrupt
INTC
221
32058K AVR32-01/12
AT32UC3A
24.5
Application Block Diagram
Figure 24-2. Application Block Diagram
VDD
Rp
Host with
TWI
Interface
Rp
TWD
TWCK
Atmel TWI
Serial EEPROM
Slave 1
I²C RTC
I²C LCD
Controller
I²C Temp.
Sensor
Slave 2
Slave 3
Slave 4
Rp: Pull up value as given by the I²C Standard
24.6
I/O Lines Description
Table 24-3.
I/O Lines Description
Pin Name
Pin Description
TWD
Two-wire Serial Data
Input/Output
TWCK
Two-wire Serial Clock
Input/Output
24.7
24.7.1
Type
Product Dependencies
I/O Lines
Both TWD and TWCK are bidirectional lines, connected to a positive supply voltage via a current
source or pull-up resistor (see Figure 24-2 on page 222). When the bus is free, both lines are
high. The output stages of devices connected to the bus must have an open-drain or open-collector to perform the wired-AND function.
TWD and TWCK pins may be multiplexed with GPIO lines. To enable the TWI, the programmer
must perform the following steps:
• Program the GPIO controller to:
– Dedicate TWD and TWCK as peripheral lines.
– Define TWD and TWCK as open-drain.
24.7.2
Power Management
The TWI clock is generated by the Power Manager (PM). Before using the TWI, the programmer
must ensure that the TWI clock is enabled in the PM.
In the TWI description, Master Clock (MCK) is the clock of the peripheral bus to which the TWI is
connected.
222
32058K AVR32-01/12
AT32UC3A
24.7.3
Interrupt
The TWI interface has an interrupt line connected to the Interrupt Controller (INTC). In order to
handle interrupts, the INTC must be programmed before configuring the TWI.
24.8
24.8.1
Functional Description
Transfer Format
The data put on the TWD line must be 8 bits long. Data is transferred MSB first; each byte must
be followed by an acknowledgement. The number of bytes per transfer is unlimited (see Figure
24-4).
Each transfer begins with a START condition and terminates with a STOP condition (see Figure
24-3).
• A high-to-low transition on the TWD line while TWCK is high defines the START condition.
• A low-to-high transition on the TWD line while TWCK is high defines a STOP condition.
Figure 24-3.
START and STOP Conditions
TWD
TWCK
Start
Stop
Figure 24-4. Transfer Format
TWD
TWCK
Start
24.9
Address
R/W
Ack
Data
Ack
Data
Ack
Stop
Modes of Operation
The TWI has six modes of operations:
• Master transmitter mode
• Master receiver mode
• Multi-master transmitter mode
• Multi-master receiver mode
• Slave transmitter mode
• Slave receiver mode
These modes are described in the following chapters.
223
32058K AVR32-01/12
AT32UC3A
24.10 Master Mode
24.10.1
Definition
The Master is the device which starts a transfer, generates a clock and stops it.
24.10.2
Application Block Diagram
Figure 24-5. Master Mode Typical Application Block Diagram
VDD
Rp
Host with
TWI
Interface
Rp
TWD
TWCK
Atmel TWI
Serial EEPROM
Slave 1
I²C RTC
I²C LCD
Controller
I²C Temp.
Sensor
Slave 2
Slave 3
Slave 4
Rp: Pull up value as given by the I²C Standard
24.10.3
Programming Master Mode
The following registers have to be programmed before entering Master mode:
1. DADR (+ IADRSZ + IADR if a 10 bit device is addressed): The device address is used to
access slave devices in read or write mode.
2. CKDIV + CHDIV + CLDIV: Clock Waveform.
3. SVDIS: Disable the slave mode.
4. MSEN: Enable the master mode.
24.10.4
Master Transmitter Mode
After the master initiates a Start condition when writing into the Transmit Holding Register, THR,
it sends a 7-bit slave address, configured in the Master Mode register (DADR in MMR), to notify
the slave device. The bit following the slave address indicates the transfer direction, 0 in this
case (MREAD = 0 in MMR).
The TWI transfers require the slave to acknowledge each received byte. During the acknowledge clock pulse (9th pulse), the master releases the data line (HIGH), enabling the slave to pull
it down in order to generate the acknowledge. The master polls the data line during this clock
pulse and sets the NACK in the status register if the slave does not acknowledge the byte. As
with the other status bits, an interrupt can be generated if enabled in the interrupt enable register
(IER). If the slave acknowledges the byte, the data written in the THR, is then shifted in the internal shifter and transferred. When an acknowledge is detected, the TXRDY bit is set until a new
write in the THR. When no more data is written into the THR, the master generates a stop condition to end the transfer. The end of the complete transfer is marked by the TXCOMP bit set to
one. See Figure 24-6, Figure 24-7, and Figure 24-8 on page 225.
TXRDY is used as Transmit Ready for the PDC transmit channel.
224
32058K AVR32-01/12
AT32UC3A
Figure 24-6. Master Write with One Data Byte
S
TWD
DADR
W
A
DATA
A
P
TXCOMP
TXRDY
STOP sent automaticaly
(ACK received and TXRDY = 1)
Write THR (DATA)
Figure 24-7. Master Write with Multiple Data Byte
S
TWD
DADR
W
A
DATA n
A
DATA n+5
A
DATA n+x
A
P
TXCOMP
TXRDY
Write THR (Data n)
Write THR (Data n+1)
Write THR (Data n+x)
Last data sent
STOP sent automaticaly
(ACK received and TXRDY = 1)
Figure 24-8. Master Write with One Byte Internal Address and Multiple Data Bytes
TWD S
DADR
W
A
IADR(7:0)
A
DATA n
A
DATA n+5
A
DATA n+x
A
P
TXCOMP
TXRDY
Write THR (Data n)
24.10.5
Write THR (Data n+1)
Write THR (Data n+x) STOP sent automaticaly
Last data sent (ACK received and TXRDY = 1)
Master Receiver Mode
The read sequence begins by setting the START bit. After the start condition has been sent, the
master sends a 7-bit slave address to notify the slave device. The bit following the slave address
indicates the transfer direction, 1 in this case (MREAD = 1 in MMR). During the acknowledge
clock pulse (9th pulse), the master releases the data line (HIGH), enabling the slave to pull it
down in order to generate the acknowledge. The master polls the data line during this clock
pulse and sets the NACK bit in the status register if the slave does not acknowledge the byte.
If an acknowledge is received, the master is then ready to receive data from the slave. After data
has been received, the master sends an acknowledge condition to notify the slave that the data
has been received except for the last data, after the stop condition. See Figure 24-9. When the
225
32058K AVR32-01/12
AT32UC3A
RXRDY bit is set in the status register, a character has been received in the receive-holding register (RHR). The RXRDY bit is reset when reading the RHR.
When a single data byte read is performed, with or without internal address (IADR), the START
and STOP bits must be set at the same time. See Figure 24-9. When a multiple data byte read is
performed, with or without IADR, the STOP bit must be set after the next-to-last data received.
See Figure 24-10. For Internal Address usage see ”Internal Address” on page 226.
Figure 24-9. Master Read with One Data Byte
TWD
S
DADR
R
A
DATA
N
P
TXCOMP
Write START &
STOP Bit
RXRDY
Read RHR
Figure 24-10. Master Read with Multiple Data Bytes
TWD
S
DADR
R
A
DATA n
A
DATA (n+1)
A
DATA (n+m)-1
A
DATA (n+m)
N
P
TXCOMP
Write START Bit
RXRDY
Read RHR
DATA n
Read RHR
DATA (n+1)
Read RHR
DATA (n+m)-1
Read RHR
DATA (n+m)
Write STOP Bit
after next-to-last data read
RXRDY is used as Receive Ready for the PDC receive channel.
24.10.6
Internal Address
The TWI interface can perform various transfer formats: Transfers with 7-bit slave address
devices and 10-bit slave address devices.
24.10.6.1
7-bit Slave Addressing
When Addressing 7-bit slave devices, the internal address bytes are used to perform random
address (read or write) accesses to reach one or more data bytes, within a memory page location in a serial memory, for example. When performing read operations with an internal address,
the TWI performs a write operation to set the internal address into the slave device, and then
switch to Master Receiver mode. Note that the second start condition (after sending the IADR) is
sometimes called “repeated start” (Sr) in I2C fully-compatible devices. See Figure 24-12. See
Figure 24-11 and Figure 24.11 for Master Write operation with internal address.
The three internal address bytes are configurable through the Master Mode register (MMR).
226
32058K AVR32-01/12
AT32UC3A
If the slave device supports only a 7-bit address, i.e. no internal address, IADRSZ must be set to
0.
n the figures below the following abbreviations are used:I
•S
Start
• Sr
Repeated Start
•P
Stop
•W
Write
•R
Read
•A
Acknowledge
•N
Not Acknowledge
• DADR
Device Address
• IADR
Internal Address
Figure 24-11. Master Write with One, Two or Three Bytes Internal Address and One Data Byte
Three bytes internal address
S
TWD
DADR
W
A
IADR(23:16)
A
IADR(15:8)
A
IADR(7:0)
A
W
A
IADR(15:8)
A
IADR(7:0)
A
DATA
A
W
A
IADR(7:0)
A
DATA
A
DATA
A
P
Two bytes internal address
S
TWD
DADR
P
One byte internal address
S
TWD
DADR
P
Figure 24-12. Master Read with One, Two or Three Bytes Internal Address and One Data Byte
Three bytes internal address
TWD
S
DADR
W
A
IADR(23:16)
A
IADR(15:8)
A
IADR(7:0)
A
Sr
DADR
R
A
DATA
N
P
Two bytes internal address
TWD
S
DADR
W
A
IADR(15:8)
A
IADR(7:0)
A
Sr
W
A
IADR(7:0)
A
Sr
R
A
DADR
R
A
DATA
N
P
One byte internal address
TWD
S
DADR
DADR
DATA
N
P
227
32058K AVR32-01/12
AT32UC3A
24.10.6.2
10-bit Slave Addressing
For a slave address higher than 7 bits, the user must configure the address size (IADRSZ) and
set the other slave address bits in the internal address register (IADR). The two remaining Internal address bytes, IADR[15:8] and IADR[23:16] can be used the same as in 7-bit Slave
Addressing.
Example: Address a 10-bit device:
(10-bit device address is b1 b2 b3 b4 b5 b6 b7 b8 b9 b10)
1. Program IADRSZ = 1,
2. Program DADR with 1 1 1 1 0 b1 b2 (b1 is the MSB of the 10-bit address, b2, etc.)
3. Program IADR with b3 b4 b5 b6 b7 b8 b9 b10 (b10 is the LSB of the 10-bit address)
Figure 24.11 below shows a byte write to an Atmel AT24LC512 EEPROM. This demonstrates
the use of internal addresses to access the device.
24.11 Internal Address Usage Using the Peripheral DMA Controller (PDC)
S
T
A
R
T
Device
Address
W
R
I
T
E
FIRST
WORD ADDRESS
SECOND
WORD ADDRESS
S
T
O
P
DATA
0
M
S
B
LR A
S / C
BW K
M
S
B
A
C
K
LA
SC
BK
A
C
K
The use of the PDC significantly reduces the CPU load.
To assure correct implementation, respect the following programming sequences:
24.11.1
Data Transmit with the PDC
1. Initialize the transmit PDC (memory pointers, size, etc.).
2. Configure the master mode (DADR, CKDIV, etc.).
3. Start the transfer by setting the PDC TXTEN bit.
4. Wait for the PDC end TX flag.
5. Disable the PDC by setting the PDC TXDIS bit.
24.11.2
Data Receive with the PDC
1. Initialize the receive PDC (memory pointers, size - 1, etc.).
2. Configure the master mode (DADR, CKDIV, etc.).
3. Start the transfer by setting the PDC RXTEN bit.
4. Wait for the PDC end RX flag.
5. Disable the PDC by setting the PDC RXDIS bit.
228
32058K AVR32-01/12
AT32UC3A
24.11.3
Read-write Flowcharts
The following flowcharts shown in Figure 24-13 to Figure 24-18 on page 234 give examples for
read and write operations. A polling or interrupt method can be used to check the status bits.
The interrupt method requires that the interrupt enable register (IER) be configured first.
Figure 24-13. TWI Write Operation with Single Data Byte without Internal Address.
BEGIN
Set TWI clock
(CLDIV, CHDIV, CKDIV) in TWI_CWGR
(Needed only once)
Set the Control register:
- Master enable
TWI_CR = MSEN + SVDIS
Set the Master Mode register:
- Device slave address (DADR)
- Transfer direction bit
Write ==> bit MREAD = 0
Load Transmit register
TWI_THR = Data to send
Read Status register
TXRDY = 1?
No
Yes
Read Status register
TXCOMP = 1?
No
Yes
Transfer finished
229
32058K AVR32-01/12
AT32UC3A
Figure 24-14. TWI Write Operation with Single Data Byte and Internal Address
BEGIN
Set TWI clock
(CLDIV, CHDIV, CKDIV) in TWI_CWGR
(Needed only once)
Set the Control register:
- Master enable
TWI_CR = MSEN + SVDIS
Set the Master Mode register:
- Device slave address (DADR)
- Internal address size (IADRSZ)
- Transfer direction bit
Write ==> bit MREAD = 0
Set the internal address
TWI_IADR = address
Load transmit register
TWI_THR = Data to send
Read Status register
TXRDY = 1?
No
Yes
Read Status register
TXCOMP = 1?
No
Yes
Transfer finished
230
32058K AVR32-01/12
AT32UC3A
Figure 24-15. TWI Write Operation with Multiple Data Bytes with or without Internal Address
BEGIN
Set TWI clock
(CLDIV, CHDIV, CKDIV) in TWI_CWGR
(Needed only once)
Set the Control register:
- Master enable
TWI_CR = MSEN + SVDIS
Set the Master Mode register:
- Device slave address
- Internal address size (if IADR used)
- Transfer direction bit
Write ==> bit MREAD = 0
No
Internal address size = 0?
Set the internal address
TWI_IADR = address
Yes
Load Transmit register
TWI_THR = Data to send
Read Status register
TWI_THR = data to send
TXRDY = 1?
No
Yes
Yes
Data to send?
Read Status register
Yes
TXCOMP = 1?
No
END
231
32058K AVR32-01/12
AT32UC3A
Figure 24-16. TWI Read Operation with Single Data Byte without Internal Address
BEGIN
Set TWI clock
(CLDIV, CHDIV, CKDIV) in TWI_CWGR
(Needed only once)
Set the Control register:
- Master enable
TWI_CR = MSEN + SVDIS
Set the Master Mode register:
- Device slave address
- Transfer direction bit
Read ==> bit MREAD = 1
Start the transfer
TWI_CR = START | STOP
Read status register
RXRDY = 1?
No
Yes
Read Receive Holding Register
Read Status register
TXCOMP = 1?
No
Yes
END
232
32058K AVR32-01/12
AT32UC3A
Figure 24-17. TWI Read Operation with Single Data Byte and Internal Address
BEGIN
Set TWI clock
(CLDIV, CHDIV, CKDIV) in TWI_CWGR
(Needed only once)
Set the Control register:
- Master enable
TWI_CR = MSEN + SVDIS
Set the Master Mode register:
- Device slave address
- Internal address size (IADRSZ)
- Transfer direction bit
Read ==> bit MREAD = 1
Set the internal address
TWI_IADR = address
Start the transfer
TWI_CR = START | STOP
Read Status register
RXRDY = 1?
No
Yes
Read Receive Holding register
Read Status register
TXCOMP = 1?
No
Yes
END
233
32058K AVR32-01/12
AT32UC3A
Figure 24-18. TWI Read Operation with Multiple Data Bytes with or without Internal Address
BEGIN
Set TWI clock
(CLDIV, CHDIV, CKDIV) in TWI_CWGR
(Needed only once)
Set the Control register:
- Master enable
TWI_CR = MSEN + SVDIS
Set the Master Mode register:
- Device slave address
- Internal address size (if IADR used)
- Transfer direction bit
Read ==> bit MREAD = 1
Internal address size = 0?
Set the internal address
TWI_IADR = address
Yes
Start the transfer
TWI_CR = START
Read Status register
RXRDY = 1?
No
Yes
Read Receive Holding register (TWI_RHR)
No
Last data to read
but one?
Yes
Stop the transfer
TWI_CR = STOP
Read Status register
RXRDY = 1?
No
Yes
Read Receive Holding register (TWI_RHR)
Read status register
TXCOMP = 1?
No
Yes
END
234
32058K AVR32-01/12
AT32UC3A
24.12 Multi-master Mode
24.12.1
Definition
More than one master may handle the bus at the same time without data corruption by using
arbitration.
Arbitration starts as soon as two or more masters place information on the bus at the same time,
and stops (arbitration is lost) for the master that intends to send a logical one while the other
master sends a logical zero.
As soon as arbitration is lost by a master, it stops sending data and listens to the bus in order to
detect a stop. When the stop is detected, the master who has lost arbitration may put its data on
the bus by respecting arbitration.
Arbitration is illustrated in Figure 24-20 on page 236.
24.12.2
Different Multi-master Modes
Two multi-master modes may be distinguished:
1. TWI is considered as a Master only and will never be addressed.
2. TWI may be either a Master or a Slave and may be addressed.
Note:
24.12.2.1
Arbitration is supported in both Multi-master modes.
TWI as Master Only
In this mode, TWI is considered as a Master only (MSEN is always at one) and must be driven
like a Master with the ARBLST (ARBitration Lost) flag in addition.
If arbitration is lost (ARBLST = 1), the programmer must reinitiate the data transfer.
If the user starts a transfer (ex.: DADR + START + W + Write in THR) and if the bus is busy, the
TWI automatically waits for a STOP condition on the bus to initiate the transfer (see Figure 2419 on page 236).
Note:
24.12.2.2
The state of the bus (busy or free) is not indicated in the user interface.
TWI as Master or Slave
The automatic reversal from Master to Slave is not supported in case of a lost arbitration.
Then, in the case where TWI may be either a Master or a Slave, the programmer must manage
the pseudo Multi-master mode described in the steps below.
1. Program TWI in Slave mode (SADR + MSDIS + SVEN) and perform Slave Access (if
TWI is addressed).
2. If TWI has to be set in Master mode, wait until TXCOMP flag is at 1.
3. Program Master mode (DADR + SVDIS + MSEN) and start the transfer (ex: START +
Write in THR).
4. As soon as the Master mode is enabled, TWI scans the bus in order to detect if it is busy
or free. When the bus is considered as free, TWI initiates the transfer.
5. As soon as the transfer is initiated and until a STOP condition is sent, the arbitration
becomes relevant and the user must monitor the ARBLST flag.
6. If the arbitration is lost (ARBLST is set to 1), the user must program the TWI in Slave
mode in the case where the Master that won the arbitration wanted to access the TWI.
235
32058K AVR32-01/12
AT32UC3A
7. If TWI has to be set in Slave mode, wait until TXCOMP flag is at 1 and then program the
Slave mode.
Note:
In the case where the arbitration is lost and TWI is addressed, TWI will not acknowledge even if it
is programmed in Slave mode as soon as ARBLST is set to 1. Then, the Master must repeat
SADR.
Figure 24-19. Programmer Sends Data While the Bus is Busy
TWCK
START sent by the TWI
STOP sent by the master
DATA sent by a master
TWD
DATA sent by the TWI
Bus is busy
Bus is free
Transfer is kept
TWI DATA transfer
A transfer is programmed
(DADR + W + START + Write THR)
Bus is considered as free
Transfer is initiated
Figure 24-20. Arbitration Cases
TWCK
TWD
TWCK
Data from a Master
S
1
0
0 1 1
Data from TWI
S
1
0
1
TWD
S
1
0
0
P
Arbitration is lost
TWI stops sending data
1 1
Data from the master
P
Arbitration is lost
S
1
0
1
S
1
0
0 1
1
S
1
0
0 1
1
The master stops sending data
Data from the TWI
ARBLST
Bus is busy
Bus is free
Transfer is kept
TWI DATA transfer
A transfer is programmed
(DADR + W + START + Write THR)
Transfer is stopped
Transfer is programmed again
(DADR + W + START + Write THR)
Bus is considered as free
Transfer is initiated
The flowchart shown in Figure 24-21 on page 237 gives an example of read and write operations
in Multi-master mode.
236
32058K AVR32-01/12
AT32UC3A
Figure 24-21. Multi-master Flowchart
START
Programm the SLAVE mode:
SADR + MSDIS + SVEN
Read Status Register
SVACC = 1 ?
Yes
GACC = 1 ?
SVREAD = 0 ?
EOSACC = 1 ?
TXRDY= 1 ?
Yes
Yes
Yes
Write in TWI_THR
TXCOMP = 1 ?
RXRDY= 0 ?
Yes
Yes
Read TWI_RHR
Need to perform
a master access ?
GENERAL CALL TREATMENT
Yes
Decoding of the
programming sequence
Prog seq
OK ?
Change SADR
Program the Master mode
DADR + SVDIS + MSEN + CLK + R / W
Read Status Register
Yes
ARBLST = 1 ?
Yes
Yes
Read TWI_RHR
Yes
MREAD = 1 ?
RXRDY= 0 ?
TXRDY= 0 ?
Data to read?
Data to send ?
Yes
Yes
Write in TWI_THR
Stop transfer
Read Status Register
Yes
TXCOMP = 0 ?
237
32058K AVR32-01/12
AT32UC3A
24.13 Slave Mode
24.13.1
Definition
The Slave Mode is defined as a mode where the device receives the clock and the address from
another device called the master.
In this mode, the device never initiates and never completes the transmission (START,
REPEATED_START and STOP conditions are always provided by the master).
24.13.2
Application Block Diagram
Figure 24-22. Slave Mode Typical Application Block Diagram
VDD
R
Master
TWD
Host with
TWI
Interface
24.13.3
R
TWCK
Host with TWI
Interface
Host with TWI
Interface
LCD Controller
Slave 1
Slave 2
Slave 3
Programming Slave Mode
The following fields must be programmed before entering Slave mode:
1. SADR (SMR): The slave device address is used in order to be accessed by master
devices in read or write mode.
2. MSDIS (CR): Disable the master mode.
3. SVEN (CR): Enable the slave mode.
As the device receives the clock, values written in CWGR are not taken into account.
24.13.4
Receiving Data
After a Start or Repeated Start condition is detected and if the address sent by the Master
matches with the Slave address programmed in the SADR (Slave ADdress) field, SVACC (Slave
ACCess) flag is set and SVREAD (Slave READ) indicates the direction of the transfer.
SVACC remains high until a STOP condition or a repeated START is detected. When such a
condition is detected, EOSACC (End Of Slave ACCess) flag is set.
24.13.4.1
Read Sequence
In the case of a Read sequence (SVREAD is high), TWI transfers data written in the THR (TWI
Transmit Holding Register) until a STOP condition or a REPEATED_START + an address different from SADR is detected. Note that at the end of the read sequence TXCOMP (Transmission
Complete) flag is set and SVACC reset.
238
32058K AVR32-01/12
AT32UC3A
As soon as data is written in the THR, TXRDY (Transmit Holding Register Ready) flag is reset,
and it is set when the shift register is empty and the sent data acknowledged or not. If the data is
not acknowledged, the NACK flag is set.
Note that a STOP or a repeated START always follows a NACK.
See Figure 24-23 on page 240.
24.13.4.2
Write Sequence
In the case of a Write sequence (SVREAD is low), the RXRDY (Receive Holding Register
Ready) flag is set as soon as a character has been received in the RHR (TWI Receive Holding
Register). RXRDY is reset when reading the RHR.
TWI continues receiving data until a STOP condition or a REPEATED_START + an address different from SADR is detected. Note that at the end of the write sequence TXCOMP flag is set
and SVACC reset.
See Figure 24-24 on page 241.
24.13.4.3
Clock Synchronization Sequence
In the case where THR or RHR is not written/read in time, TWI performs a clock synchronization.
Clock stretching information is given by the SCLWS (Clock Wait state) bit.
See Figure 24-26 on page 242 and Figure 24-27 on page 243.
24.13.4.4
General Call
In the case where a GENERAL CALL is performed, GACC (General Call ACCess) flag is set.
After GACC is set, it is up to the programmer to interpret the meaning of the GENERAL CALL
and to decode the new address programming sequence.
See Figure 24-25 on page 241.
24.13.4.5
PDC
As it is impossible to know the exact number of data to receive/send, the use of PDC is NOT recommended in SLAVE mode.
239
32058K AVR32-01/12
AT32UC3A
As it is impossible to know the exact number of data to receive/send, the use of PDC is NOT recommended in SLAVE mode.
24.13.5
Data Transfer
24.13.5.1
Read Operation
The read mode is defined as a data requirement from the master.
After a START or a REPEATED START condition is detected, the decoding of the address
starts. If the slave address (SADR) is decoded, SVACC is set and SVREAD indicates the direction of the transfer.
Until a STOP or REPEATED START condition is detected, TWI continues sending data loaded
in the THR register.
If a STOP condition or a REPEATED START + an address different from SADR is detected,
SVACC is reset.
Figure 24-23 on page 240 describes the write operation.
Figure 24-23. Read Access Ordered by a MASTER
SADR matches,
TWI answers with an ACK
SADR does not match,
TWI answers with a NACK
TWD
S
ADR
R
NA
DATA
NA
P/S/Sr
SADR R
A
DATA
A
ACK/NACK from the Master
A
DATA
NA
S/Sr
TXRDY
Read RHR
Write THR
NACK
SVACC
SVREAD
SVREAD has to be taken into account only while SVACC is active
EOSVACC
Notes:
24.13.5.2
1. When SVACC is low, the state of SVREAD becomes irrelevant.
2. TXRDY is reset when data has been transmitted from THR to the shift register and set when
this data has been acknowledged or non acknowledged.
Write Operation
The write mode is defined as a data transmission from the master.
After a START or a REPEATED START, the decoding of the address starts. If the slave address
is decoded, SVACC is set and SVREAD indicates the direction of the transfer (SVREAD is low in
this case).
Until a STOP or REPEATED START condition is detected, TWI stores the received data in the
RHR register.
If a STOP condition or a REPEATED START + an address different from SADR is detected,
SVACC is reset.
Figure 24-24 on page 241 describes the Write operation.
240
32058K AVR32-01/12
AT32UC3A
Figure 24-24. Write Access Ordered by a Master
SADR does not match,
TWI answers with a NACK
S
TWD
ADR
W
NA
DATA
NA
SADR matches,
TWI answers with an ACK
P/S/Sr
SADR W
A
DATA
Read RHR
A
A
DATA
NA
S/Sr
RXRDY
SVACC
SVREAD
SVREAD has to be taken into account only while SVACC is active
EOSVACC
Notes:
24.13.5.3
1. When SVACC is low, the state of SVREAD becomes irrelevant.
2. RXRDY is set when data has been transmitted from the shift register to the RHR and reset
when this data is read.
General Call
The general call is performed in order to change the address of the slave.
If a GENERAL CALL is detected, GACC is set.
After the detection of General Call, it is up to the programmer to decode the commands which
come afterwards.
In case of a WRITE command, the programmer has to decode the programming sequence and
program a new SADR if the programming sequence matches.
Figure 24-25 on page 241 describes the General Call access.
Figure 24-25. Master Performs a General Call
0000000 + W
TXD
S
GENERAL CALL
RESET command = 00000110X
WRITE command = 00000100X
A
Reset or write DADD
A
DATA1
A
DATA2
A
New SADR
A
P
New SADR
Programming sequence
GCACC
Reset after read
SVACC
Note:
1. This method allows the user to create an own programming sequence by choosing the programming bytes and the number of them. The programming sequence has to be provided to
the master.
241
32058K AVR32-01/12
AT32UC3A
24.13.6
Clock Synchronization
In both read and write modes, it may happen that THR/RHR buffer is not filled /emptied before
the emission/reception of a new character. In this case, to avoid sending/receiving undesired
data, a clock stretching mechanism is implemented.
24.13.6.1
Clock Synchronization in Read Mode
The clock is tied low if the shift register is empty and if a STOP or REPEATED START condition
was not detected. It is tied low until the shift register is loaded.
Figure 24-26 on page 242 describes the clock synchronization in Read mode.
Figure 24-26. Clock Synchronization in Read Mode
TWI_THR
S
SADR
R
DATA1
1
DATA0
A
DATA0
A
DATA1
DATA2
A
XXXXXXX
DATA2
NA
S
2
TWCK
Write THR
SCLWS
TXRDY
CLOCK is tied low by the TWI
as long as THR is empty
SVACC
SVREAD
As soon as a START is detected
TXCOMP
TWI_THR is transmitted to the shift register
Notes:
Ack or Nack from the master
1
The data is memorized in TWI_THR until a new value is written
2
The clock is stretched after the ACK, the state of TWD is undefined during clock stretching
1. TXRDY is reset when data has been written in the TH to the shift register and set when this data has been acknowledged or
non acknowledged.
2. At the end of the read sequence, TXCOMP is set after a STOP or after a REPEATED_START + an address different from
SADR.
3. SCLWS is automatically set when the clock synchronization mechanism is started.
242
32058K AVR32-01/12
AT32UC3A
24.13.6.2
Clock Synchronization in Write Mode
The clock is tied low if the shift register and the RHR is full. If a STOP or REPEATED_START
condition was not detected, it is tied low until RHR is read.
Figure 24-27 on page 243 describes the clock synchronization in Read mode.
Figure 24-27. Clock Synchronization in Write Mode
TWCK
CLOCK is tied low by the TWI as long as RHR is full
TWD
S
SADR
W
A
DATA0
TWI_RHR
A
DATA1
A
DATA0 is not read in the RHR
SCLWS
DATA2
DATA1
NA
S
ADR
DATA2
SCL is stretched on the last bit of DATA1
RXRDY
Rd DATA0
Rd DATA1
Rd DATA2
SVACC
SVREAD
TXCOMP
Notes:
As soon as a START is detected
1. At the end of the read sequence, TXCOMP is set after a STOP or after a REPEATED_START + an address different from
SADR.
2. SCLWS is automatically set when the clock synchronization mechanism is started and automatically reset when the mechanism is finished.
243
32058K AVR32-01/12
AT32UC3A
24.13.7
Reversal after a Repeated Start
24.13.7.1
Reversal of Read to Write
The master initiates the communication by a read command and finishes it by a write command.
Figure 24-28 on page 244 describes the repeated start + reversal from Read to Write mode.
Figure 24-28. Repeated Start + Reversal from Read to Write Mode
TWI_THR
DATA0
TWD
S
SADR
R
A
DATA0
DATA1
A
DATA1
NA
Sr
SADR
W
A
DATA2
TWI_RHR
A
DATA3
DATA2
A
P
DATA3
SVACC
SVREAD
TXRDY
RXRDY
EOSACC
Cleared after read
As soon as a START is detected
TXCOMP
1. TXCOMP is only set at the end of the transmission because after the repeated start, SADR is detected again.
24.13.7.2
Reversal of Write to Read
The master initiates the communication by a write command and finishes it by a read command.Figure 24-29 on page 244 describes the repeated start + reversal from Write to Read
mode.
Figure 24-29. Repeated Start + Reversal from Write to Read Mode
DATA2
TWI_THR
TWD
S
SADR
W
A
DATA0
TWI_RHR
A
DATA1
DATA0
A
Sr
SADR
R
A
DATA3
DATA2
A
DATA3
NA
P
DATA1
SVACC
SVREAD
TXRDY
RXRDY
EOSACC
TXCOMP
Notes:
Read TWI_RHR
Cleared after read
As soon as a START is detected
1. In this case, if THR has not been written at the end of the read command, the clock is automatically stretched before the
ACK.
2. TXCOMP is only set at the end of the transmission because after the repeated start, SADR is detected again.
244
32058K AVR32-01/12
AT32UC3A
24.13.8
Read Write Flowcharts
The flowchart shown in Figure 24-30 on page 245 gives an example of read and write operations
in Slave mode. A polling or interrupt method can be used to check the status bits. The interrupt
method requires that the interrupt enable register (IER) be configured first.
Figure 24-30. Read Write Flowchart in Slave Mode
Set the SLAVE mode:
SADR + MSDIS + SVEN
Read Status Register
SVACC = 1 ?
GACC = 1 ?
SVREAD = 0 ?
TXRDY= 1 ?
EOSACC = 1 ?
Write in TWI_THR
TXCOMP = 1 ?
RXRDY= 0 ?
END
Read TWI_RHR
GENERAL CALL TREATMENT
Decoding of the
programming sequence
Prog seq
OK ?
Change SADR
245
32058K AVR32-01/12
AT32UC3A
24.14 Two-wire Interface (TWI) User Interface
24.14.1
Register Mapping
Table 24-4.
TWI User Interface
Offset
Register
Name
Access
Reset
0x00
Control Register
CR
Write-only
N/A
0x04
Master Mode Register
MMR
Read-write
0x00000000
0x08
Slave Mode Register
SMR
Read-write
0x00000000
0x0C
Internal Address Register
IADR
Read-write
0x00000000
0x10
Clock Waveform Generator Register
CWGR
Read-write
0x00000000
0x20
Status Register
SR
Read-only
0x0000F009
0x24
Interrupt Enable Register
IER
Write-only
N/A
0x28
Interrupt Disable Register
IDR
Write-only
N/A
0x2C
Interrupt Mask Register
IMR
Read-only
0x00000000
0x30
Receive Holding Register
RHR
Read-only
0x00000000
0x34
Transmit Holding Register
THR
Write-only
0x00000000
0x38 - 0xF8
Reserved
–
–
–
0xFC
Version Register
TWI_VER
Read-only
0x00000000(1)
0x38 - 0xFC
Reserved
–
–
–
Note:
1. Values in the Version Register vary with the version of the IP block implementation.
24.14.2
TWI Control Register
Name:
CR
Access:
Write-only
Reset Value: 0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
SWRST
6
–
5
SVDIS
4
SVEN
3
MSDIS
2
MSEN
1
STOP
0
START
• START: Send a START Condition
0 = No effect.
1 = A frame beginning with a START bit is transmitted according to the features defined in the mode register.
246
32058K AVR32-01/12
AT32UC3A
This action is necessary when the TWI peripheral wants to read data from a slave. When configured in Master Mode with a
write operation, a frame is sent as soon as the user writes a character in the Transmit Holding Register (THR).
• STOP: Send a STOP Condition
0 = No effect.
1 = STOP Condition is sent just after completing the current byte transmission in master read mode.
- In single data byte master read, the START and STOP must both be set.
- In multiple data bytes master read, the STOP must be set after the last data received but one.
- In master read mode, if a NACK bit is received, the STOP is automatically performed.
- In multiple data write operation, when both THR and shift register are empty, a STOP condition is automatically sent.
• MSEN: TWI Master Mode Enabled
0 = No effect.
1 = If MSDIS = 0, the master mode is enabled.
Note:
Switching from Slave to Master mode is only permitted when TXCOMP = 1.
• MSDIS: TWI Master Mode Disabled
0 = No effect.
1 = The master mode is disabled, all pending data is transmitted. The shifter and holding characters (if it contains data) are
transmitted in case of write operation. In read operation, the character being transferred must be completely received
before disabling.
• SVEN: TWI Slave Mode Enabled
0 = No effect.
1 = If SVDIS = 0, the slave mode is enabled.
Note:
Switching from Master to Slave mode is only permitted when TXCOMP = 1.
• SVDIS: TWI Slave Mode Disabled
0 = No effect.
1 = The slave mode is disabled. The shifter and holding characters (if it contains data) are transmitted in case of read operation. In write operation, the character being transferred must be completely received before disabling.
• SWRST: Software Reset
0 = No effect.
1 = Equivalent to a system reset.
247
32058K AVR32-01/12
AT32UC3A
24.14.3
TWI Master Mode Register
Name:
MMR
Access:
Read-write
Reset Value: 0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
21
20
19
DADR
18
17
16
15
–
14
–
13
–
12
MREAD
11
–
10
–
9
7
–
6
–
5
–
4
–
3
–
2
–
1
–
8
IADRSZ
0
–
• IADRSZ: Internal Device Address Size
IADRSZ[9:8]
Description
0
0
No internal device address
0
1
One-byte internal device address
1
0
Two-byte internal device address
1
1
Three-byte internal device address
• MREAD: Master Read Direction
0 = Master write direction.
1 = Master read direction.
• DADR: Device Address
The device address is used to access slave devices in read or write mode. Those bits are only used in Master mode.
248
32058K AVR32-01/12
AT32UC3A
24.14.4
TWI Slave Mode Register
Name:
SMR
Access:
Read-write
Reset Value: 0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
21
20
19
SADR
18
17
16
15
–
14
–
13
–
12
–
11
–
10
–
9
8
7
–
6
–
5
–
4
–
3
–
2
–
1
–
0
–
• SADR: Slave Address
The slave device address is used in Slave mode in order to be accessed by master devices in read or write mode.
SADR must be programmed before enabling the Slave mode or after a general call. Writes at other times have no effect.
249
32058K AVR32-01/12
AT32UC3A
24.14.5
TWI Internal Address Register
Name:
IADR
Access:
Read-write
Reset Value: 0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
22
21
20
19
18
17
16
11
10
9
8
3
2
1
0
IADR
15
14
13
12
IADR
7
6
5
4
IADR
• IADR: Internal Address
0, 1, 2 or 3 bytes depending on IADRSZ.
250
32058K AVR32-01/12
AT32UC3A
24.14.6
TWI Clock Waveform Generator Register
Name:
CWGR
Access:
Read-write
Reset Value: 0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
22
21
20
19
18
17
CKDIV
16
15
14
13
12
11
10
9
8
3
2
1
0
CHDIV
7
6
5
4
CLDIV
CWGR is only used in Master mode.
• CLDIV: Clock Low Divider
The SCL low period is defined as follows:
T low = ( ( CLDIV × 2
CKDIV
) + 4 ) × T MCK
• CHDIV: Clock High Divider
The SCL high period is defined as follows:
T high = ( ( CHDIV × 2
CKDIV
) + 4 ) × T MCK
• CKDIV: Clock Divider
The CKDIV is used to increase both SCL high and low periods.
251
32058K AVR32-01/12
AT32UC3A
24.14.7
TWI Status Register
Name:
SR
Access:
Read-only
Reset Value: 0x0000F009
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
TXBUFE
14
RXBUFF
13
ENDTX
12
ENDRX
11
EOSACC
10
SCLWS
9
ARBLST
8
NACK
7
–
6
OVRE
5
GACC
4
SVACC
3
SVREAD
2
TXRDY
1
RXRDY
0
TXCOMP
• TXCOMP: Transmission Completed (automatically set / reset)
TXCOMP used in Master mode:
0 = During the length of the current frame.
1 = When both holding and shifter registers are empty and STOP condition has been sent.
TXCOMP behavior in Master mode can be seen in Figure 24-8 on page 225 and in Figure 24-10 on page 226.
TXCOMP used in Slave mode:
0 = As soon as a Start is detected.
1 = After a Stop or a Repeated Start + an address different from SADR is detected.
TXCOMP behavior in Slave mode can be seen in Figure 24-26 on page 242, Figure 24-27 on page 243, Figure 24-28 on
page 244 and Figure 24-29 on page 244.
• RXRDY: Receive Holding Register Ready (automatically set / reset)
0 = No character has been received since the last RHR read operation.
1 = A byte has been received in the RHR since the last read.
RXRDY behavior in Master mode can be seen in Figure 24-10 on page 226.
RXRDY behavior in Slave mode can be seen in Figure 24-24 on page 241, Figure 24-27 on page 243, Figure 24-28 on
page 244 and Figure 24-29 on page 244.
• TXRDY: Transmit Holding Register Ready (automatically set / reset)
TXRDY used in Master mode:
0 = The transmit holding register has not been transferred into shift register. Set to 0 when writing into THR register.
1 = As soon as a data byte is transferred from THR to internal shifter or if a NACK error is detected, TXRDY is set at the
same time as TXCOMP and NACK. TXRDY is also set when MSEN is set (enable TWI).
TXRDY behavior in Master mode can be seen in Figure 24-8 on page 225.
252
32058K AVR32-01/12
AT32UC3A
TXRDY used in Slave mode:
0 = As soon as data is written in the THR, until this data has been transmitted and acknowledged (ACK or NACK).
1 = It indicates that the THR is empty and that data has been transmitted and acknowledged.
If TXRDY is high and if a NACK has been detected, the transmission will be stopped. Thus when TRDY = NACK = 1, the
programmer must not fill THR to avoid losing it.
TXRDY behavior in Slave mode can be seen in Figure 24-23 on page 240, Figure 24-26 on page 242, Figure 24-28 on
page 244 and Figure 24-29 on page 244.
• SVREAD: Slave Read (automatically set / reset)
This bit is only used in Slave mode. When SVACC is low (no Slave access has been detected) SVREAD is irrelevant.
0 = Indicates that a write access is performed by a Master.
1 = Indicates that a read access is performed by a Master.
SVREAD behavior can be seen in Figure 24-23 on page 240, Figure 24-24 on page 241, Figure 24-28 on page 244 and
Figure 24-29 on page 244.
• SVACC: Slave Access (automatically set / reset)
This bit is only used in Slave mode.
0 = TWI is not addressed. SVACC is automatically cleared after a NACK or a STOP condition is detected.
1 = Indicates that the address decoding sequence has matched (A Master has sent SADR). SVACC remains high until a
NACK or a STOP condition is detected.
SVACC behavior can be seen in Figure 24-23 on page 240, Figure 24-24 on page 241, Figure 24-28 on page 244 and Figure 24-29 on page 244.
• GACC: General Call Access (clear on read)
This bit is only used in Slave mode.
0 = No General Call has been detected.
1 = A General Call has been detected. After the detection of General Call, the programmer decoded the commands that follow and the programming sequence.
GACC behavior can be seen in Figure 24-25 on page 241.
• OVRE: Overrun Error (clear on read)
This bit is only used in Master mode.
0 = RHR has not been loaded while RXRDY was set
1 = RHR has been loaded while RXRDY was set. Reset by read in SR when TXCOMP is set.
• NACK: Not Acknowledged (clear on read)
NACK used in Master mode:
0 = Each data byte has been correctly received by the far-end side TWI slave component.
1 = A data byte has not been acknowledged by the slave component. Set at the same time as TXCOMP.
253
32058K AVR32-01/12
AT32UC3A
NACK used in Slave Read mode:
0 = Each data byte has been correctly received by the Master.
1 = In read mode, a data byte has not been acknowledged by the Master. When NACK is set the programmer must not fill
THR even if TXRDY is set, because it means that the Master will stop the data transfer or re initiate it.
Note that in Slave Write mode all data are acknowledged by the TWI.
• ARBLST: Arbitration Lost (clear on read)
This bit is only used in Master mode.
0 = Arbitration won.
1 = Arbitration lost. Another master of the TWI bus has won the multi-master arbitration. TXCOMP is set at the same time.
• SCLWS: Clock Wait State (automatically set / reset)
This bit is only used in Slave mode.
0 = The clock is not stretched.
1 = The clock is stretched. THR / RHR buffer is not filled / emptied before the emission / reception of a new character.
SCLWS behavior can be seen in Figure 24-26 on page 242 and Figure 24-27 on page 243.
• EOSACC: End Of Slave Access (clear on read)
This bit is only used in Slave mode.
0 = A slave access is being performing.
1 = The Slave Access is finished. End Of Slave Access is automatically set as soon as SVACC is reset.
EOSACC behavior can be seen in Figure 24-28 on page 244 and Figure 24-29 on page 244
• ENDRX: End of RX buffer
This bit is only used in Master mode.
0 = The Receive Counter Register has not reached 0 since the last write in RCR or RNCR.
1 = The Receive Counter Register has reached 0 since the last write in RCR or RNCR.
• ENDTX: End of TX buffer
This bit is only used in Master mode.
0 = The Transmit Counter Register has not reached 0 since the last write in TCR or TNCR.
1 = The Transmit Counter Register has reached 0 since the last write in TCR or TNCR.
• RXBUFF: RX Buffer Full
This bit is only used in Master mode.
0 = RCR or RNCR have a value other than 0.
1 = Both RCR and RNCR have a value of 0.
• TXBUFE: TX Buffer Empty
This bit is only used in Master mode.
0 = TCR or TNCR have a value other than 0.
254
32058K AVR32-01/12
AT32UC3A
1 = Both TCR and TNCR have a value of 0.
24.14.8
TWI Interrupt Enable Register
Name:
IER
Access:
Write-only
Reset Value: 0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
TXBUFE
14
RXBUFF
13
ENDTX
12
ENDRX
11
EOSACC
10
SCL_WS
9
ARBLST
8
NACK
7
–
6
OVRE
5
GACC
4
SVACC
3
–
2
TXRDY
1
RXRDY
0
TXCOMP
• TXCOMP: Transmission Completed Interrupt Enable
• RXRDY: Receive Holding Register Ready Interrupt Enable
• TXRDY: Transmit Holding Register Ready Interrupt Enable
• SVACC: Slave Access Interrupt Enable
• GACC: General Call Access Interrupt Enable
• OVRE: Overrun Error Interrupt Enable
• NACK: Not Acknowledge Interrupt Enable
• ARBLST: Arbitration Lost Interrupt Enable
• SCL_WS: Clock Wait State Interrupt Enable
• EOSACC: End Of Slave Access Interrupt Enable
• ENDRX: End of Receive Buffer Interrupt Enable
• ENDTX: End of Transmit Buffer Interrupt Enable
• RXBUFF: Receive Buffer Full Interrupt Enable
• TXBUFE: Transmit Buffer Empty Interrupt Enable
0 = No effect.
255
32058K AVR32-01/12
AT32UC3A
1 = Enables the corresponding interrupt.
24.14.9
TWI Interrupt Disable Register
Name:
IDR
Access:
Write-only
Reset Value: 0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
TXBUFE
14
RXBUFF
13
ENDTX
12
ENDRX
11
EOSACC
10
SCL_WS
9
ARBLST
8
NACK
7
–
6
OVRE
5
GACC
4
SVACC
3
–
2
TXRDY
1
RXRDY
0
TXCOMP
• TXCOMP: Transmission Completed Interrupt Disable
• RXRDY: Receive Holding Register Ready Interrupt Disable
• TXRDY: Transmit Holding Register Ready Interrupt Disable
• SVACC: Slave Access Interrupt Disable
• GACC: General Call Access Interrupt Disable
• OVRE: Overrun Error Interrupt Disable
• NACK: Not Acknowledge Interrupt Disable
• ARBLST: Arbitration Lost Interrupt Disable
• SCL_WS: Clock Wait State Interrupt Disable
• EOSACC: End Of Slave Access Interrupt Disable
• ENDRX: End of Receive Buffer Interrupt Disable
• ENDTX: End of Transmit Buffer Interrupt Disable
• RXBUFF: Receive Buffer Full Interrupt Disable
• TXBUFE: Transmit Buffer Empty Interrupt Disable
0 = No effect.
1 = Disables the corresponding interrupt.
256
32058K AVR32-01/12
AT32UC3A
24.14.10 TWI Interrupt Mask Register
Name:
IMR
Access:
Read-only
Reset Value: 0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
TXBUFE
14
RXBUFF
13
ENDTX
12
ENDRX
11
EOSACC
10
SCL_WS
9
ARBLST
8
NACK
7
–
6
OVRE
5
GACC
4
SVACC
3
–
2
TXRDY
1
RXRDY
0
TXCOMP
• TXCOMP: Transmission Completed Interrupt Mask
• RXRDY: Receive Holding Register Ready Interrupt Mask
• TXRDY: Transmit Holding Register Ready Interrupt Mask
• SVACC: Slave Access Interrupt Mask
• GACC: General Call Access Interrupt Mask
• OVRE: Overrun Error Interrupt Mask
• NACK: Not Acknowledge Interrupt Mask
• ARBLST: Arbitration Lost Interrupt Mask
• SCL_WS: Clock Wait State Interrupt Mask
• EOSACC: End Of Slave Access Interrupt Mask
• ENDRX: End of Receive Buffer Interrupt Mask
• ENDTX: End of Transmit Buffer Interrupt Mask
• RXBUFF: Receive Buffer Full Interrupt Mask
• TXBUFE: Transmit Buffer Empty Interrupt Mask
0 = The corresponding interrupt is disabled.
1 = The corresponding interrupt is enabled.
24.14.11 TWI Receive Holding Register
Name:
RHR
257
32058K AVR32-01/12
AT32UC3A
Access:
Read-only
Reset Value: 0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
RXDATA
• RXDATA: Master or Slave Receive Holding Data
24.14.12 TWI Transmit Holding Register
Name:
THR
Access:
Read-write
Reset Value: 0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
TXDATA
• TXDATA: Master or Slave Transmit Holding Data
258
32058K AVR32-01/12
AT32UC3A
25. Synchronous Serial Controller (SSC)
Rev: 3.0.0.2
25.1
Features
•
•
•
•
•
Provides Serial Synchronous Communication Links Used in Audio and Telecom Applications
Contains an Independent Receiver and Transmitter and a Common Clock Divider
Interfaced with Two PDCA Channels (DMA Access) to Reduce Processor Overhead
Offers a Configurable Frame Sync and Data Length
Receiver and Transmitter Can be Programmed to Start Automatically or on Detection of Different
Events on the Frame Sync Signal
• Receiver and Transmitter Include a Data Signal, a Clock Signal and a Frame Synchronization
Signal
25.2
Overview
The Atmel Synchronous Serial Controller (SSC) provides a synchronous communication link
with external devices. It supports many serial synchronous communication protocols generally
used in audio and telecom applications such as I2S, Short Frame Sync, Long Frame Sync, etc.
The SSC contains an independent receiver and transmitter and a common clock divider. The
receiver and the transmitter each interface with three signals: the TX_DATA/RX_DATA signal
for data, the TX_CLOCK/RX_CLOCK signal for the clock and the
TX_FRAME_SYNC/RX_FRAME_SYNC signal for the Frame Sync. The transfers can be programmed to start automatically or on different events detected on the Frame Sync signal.
The SSC’s high-level of programmability and its two dedicated PDCA channels of up to 32 bits
permit a continuous high bit rate data transfer without processor intervention.
Featuring connection to two PDCA channels, the SSC permits interfacing with low processor
overhead to the following:
• CODEC’s in master or slave mode
• DAC through dedicated serial interface, particularly I2S
• Magnetic card reader
259
32058K AVR32-01/12
AT32UC3A
25.3
Block Diagram
Figure 25-1. Block Diagram
High
Speed
Bus
Peripheral Bus
Bridge
PDCA
Peripheral
Bus
TX_FRAME_SYNC
TX_CLOCK
TX_DATA
Power CLK_SSC
Manager
PIO
SSC Interface
RX_FRAME_SYNC
RX_CLOCK
Interrupt Control
RX_DATA
SSC Interrupt
25.4
Application Block Diagram
Figure 25-2. Application Block Diagram
OS or RTOS Driver
Power
Management
Interrupt
Management
Test
Management
SSC
Serial AUDIO
Codec
Time Slot
Frame
Management Management
Line Interface
260
32058K AVR32-01/12
AT32UC3A
25.5
I/O Lines Description
Table 25-1.
I/O Lines Description
Pin Name
Pin Description
RX_FRAME_SYNC
Receiver Frame Synchro
Input/Output
RX_CLOCK
Receiver Clock
Input/Output
RX_DATA
Receiver Data
Input
TX_FRAME_SYNC
Transmitter Frame Synchro
Input/Output
TX_CLOCK
Transmitter Clock
Input/Output
TX_DATA
Transmitter Data
Output
25.6
25.6.1
Type
Product Dependencies
I/O Lines
The pins used for interfacing the compliant external devices may be multiplexed with PIO lines.
Before using the SSC receiver, the PIO controller must be configured to dedicate the SSC
receiver I/O lines to the SSC peripheral mode.
Before using the SSC transmitter, the PIO controller must be configured to dedicate the SSC
transmitter I/O lines to the SSC peripheral mode.
25.6.2
Power Management
The SSC clock is generated by the power manager. Before using the SSC, the programmer
must ensure that the SSC clock is enabled in the power manager.
In the SSC description, Master Clock (CLK_SSC) is the bus clock of the peripheral bus to which
the SSC is connected.
25.6.3
Interrupt
The SSC interface has an interrupt line connected to the interrupt controller. Handling interrupts
requires programming the interrupt controller before configuring the SSC.
All SSC interrupts can be enabled/disabled configuring the SSC Interrupt mask register. Each
pending and unmasked SSC interrupt will assert the SSC interrupt line. The SSC interrupt service routine can get the interrupt origin by reading the SSC interrupt status register.
25.7
Functional Description
This chapter contains the functional description of the following: SSC Functional Block, Clock
Management, Data format, Start, Transmitter, Receiver and Frame Sync.
The receiver and transmitter operate separately. However, they can work synchronously by programming the receiver to use the transmit clock and/or to start a data transfer when transmission
starts. Alternatively, this can be done by programming the transmitter to use the receive clock
and/or to start a data transfer when reception starts. The transmitter and the receiver can be programmed to operate with the clock signals provided on either the TX_CLOCK or RX_CLOCK
pins. This allows the SSC to support many slave-mode data transfers. The maximum clock
speed allowed on the TX_CLOCK and RX_CLOCK pins is the master clock divided by 2.
261
32058K AVR32-01/12
AT32UC3A
Figure 25-3. SSC Functional Block Diagram
Transmitter
Clock Output
Controller
TX_CLOCK
Frame Sync
Controller
TX_FRAME_SYNC
TX_CLOCK Input
CLK_SSC Clock
Divider
Transmit Clock TX clock
Controller
RX clock
TX_FRAME_SYNC
RX_FRAME_SYNC
Start
Selector
Transmit Shift Register
Transmit Holding
Register
TX_PDCA
Peripheral
Bus
TX_DATA
Transmit Sync
Holding Register
Load Shift
User
Interface
Receiver
RX_CLOCK
Input
TX clock
TX_FRAME_SYNC
RX_FRAME_SYNC
Receive Clock RX clock
Controller
Start
Selector
Interrupt Control
RX_CLOCK
Frame Sync
Controller
RX_FRAME_SYNC
Receive Shift Register
RX_PDCA
PDCA
Clock Output
Controller
Receive Holding
Register
RX_DATA
Receive Sync
Holding Register
Load Shift
Interrupt Controller
25.7.1
Clock Management
The transmitter clock can be generated by:
• an external clock received on the TX_CLOCK I/O pad
• the receiver clock
• the internal clock divider
The receiver clock can be generated by:
• an external clock received on the RX_CLOCK I/O pad
• the transmitter clock
• the internal clock divider
Furthermore, the transmitter block can generate an external clock on the TX_CLOCK I/O pad,
and the receiver block can generate an external clock on the RX_CLOCK I/O pad.
This allows the SSC to support many Master and Slave Mode data transfers.
262
32058K AVR32-01/12
AT32UC3A
25.7.1.1
Clock Divider
Figure 25-4. Divided Clock Block Diagram
Clock Divider
CMR
CLK_SSC
/2
12-bit Counter
Divided Clock
The Master Clock divider is determined by the 12-bit field DIV counter and comparator (so its
maximal value is 4095) in the Clock Mode Register CMR, allowing a Master Clock division by up
to 8190. The Divided Clock is provided to both the Receiver and Transmitter. When this field is
programmed to 0, the Clock Divider is not used and remains inactive.
When DIV is set to a value equal to or greater than 1, the Divided Clock has a frequency of Master Clock divided by 2 times DIV. Each level of the Divided Clock has a duration of the Master
Clock multiplied by DIV. This ensures a 50% duty cycle for the Divided Clock regardless of
whether the DIV value is even or odd.
Divided Clock Generation
Figure 25-5.
Master Clock
Divided Clock
DIV = 1
Divided Clock Frequency = CLK_SSC/2
Master Clock
Divided Clock
DIV = 3
Divided Clock Frequency = CLK_SSC/6
Table 25-2.
25.7.1.2
Maximum
Minimum
CLK_SSC / 2
CLK_SSC / 8190
Transmitter Clock Management
The transmitter clock is generated from the receiver clock or the divider clock or an external
clock scanned on the TX_CLOCK I/O pad. The transmitter clock is selected by the CKS field in
TCMR (Transmit Clock Mode Register). Transmit Clock can be inverted independently by the
CKI bits in TCMR.
263
32058K AVR32-01/12
AT32UC3A
The transmitter can also drive the TX_CLOCK I/O pad continuously or be limited to the actual
data transfer. The clock output is configured by the TCMR register. The Transmit Clock Inversion (CKI) bits have no effect on the clock outputs. Programming the TCMR register to select
TX_CLOCK pin (CKS field) and at the same time Continuous Transmit Clock (CKO field) might
lead to unpredictable results.
Figure 25-6. Transmitter Clock Management
TX_CLOCK(pin)
Clock
Output
Tri-state
Controller
MUX
Receiver
Clock
Divider
Clock
Data Transfer
CKO
CKS
25.7.1.3
INV
MUX
Tri-state
Controller
CKI
CKG
Transmitter
Clock
Receiver Clock Management
The receiver clock is generated from the transmitter clock or the divider clock or an external
clock scanned on the RX_CLOCK I/O pad. The Receive Clock is selected by the CKS field in
RCMR (Receive Clock Mode Register). Receive Clocks can be inverted independently by the
CKI bits in RCMR.
The receiver can also drive the RX_CLOCK I/O pad continuously or be limited to the actual data
transfer. The clock output is configured by the RCMR register. The Receive Clock Inversion
(CKI) bits have no effect on the clock outputs. Programming the RCMR register to select
RX_CLOCK pin (CKS field) and at the same time Continuous Receive Clock (CKO field) can
lead to unpredictable results.
264
32058K AVR32-01/12
AT32UC3A
Figure 25-7. Receiver Clock Management
RX_CLOCK (pin)
Tri-state
Controller
MUX
Clock
Output
Transmitter
Clock
Divider
Clock
Data Transfer
CKO
CKS
25.7.1.4
INV
MUX
Tri-state
Controller
CKI
CKG
Receiver
Clock
Serial Clock Ratio Considerations
The Transmitter and the Receiver can be programmed to operate with the clock signals provided
on either the TX_CLOCK or RX_CLOCK pins. This allows the SSC to support many slave-mode
data transfers. In this case, the maximum clock speed allowed on the RX_CLOCK pin is:
– Master Clock divided by 2 if Receiver Frame Synchro is input
– Master Clock divided by 3 if Receiver Frame Synchro is output
In addition, the maximum clock speed allowed on the TX_CLOCK pin is:
– Master Clock divided by 6 if Transmit Frame Synchro is input
– Master Clock divided by 2 if Transmit Frame Synchro is output
25.7.2
Transmitter Operations
A transmitted frame is triggered by a start event and can be followed by synchronization data
before data transmission.
The start event is configured by setting the Transmit Clock Mode Register (TCMR). See Section
“25.7.4” on page 267.
The frame synchronization is configured setting the Transmit Frame Mode Register (TFMR).
See Section “25.7.5” on page 269.
To transmit data, the transmitter uses a shift register clocked by the transmitter clock signal and
the start mode selected in the TCMR. Data is written by the application to the THR register then
transferred to the shift register according to the data format selected.
When both the THR and the transmit shift register are empty, the status flag TXEMPTY is set in
SR. When the Transmit Holding register is transferred in the Transmit shift register, the status
flag TXRDY is set in SR and additional data can be loaded in the holding register.
265
32058K AVR32-01/12
AT32UC3A
Figure 25-8. Transmitter Block Diagram
CR.TXEN
SR.TXEN
CR.TXDIS
TFMR.DATDEF
1
TX_FRAME_SYNC
RX_FRAME_SYNC
Transmitter Clock
Start
Selector
TX_DATA
0
TFMR.MSBF
Transmit Shift Register
0
TFMR.FSDEN
TCMR.STTDLY
TFMR.DATLEN
25.7.3
TCMR.STTDLY
TFMR.FSDEN
TFMR.DATNB
THR
1
TSHR
TFMR.FSLEN
Receiver Operations
A received frame is triggered by a start event and can be followed by synchronization data
before data transmission.
The start event is configured setting the Receive Clock Mode Register (RCMR). See Section
“25.7.4” on page 267.
The frame synchronization is configured setting the Receive Frame Mode Register (RFMR). See
Section “25.7.5” on page 269.
The receiver uses a shift register clocked by the receiver clock signal and the start mode
selected in the RCMR. The data is transferred from the shift register depending on the data format selected.
When the receiver shift register is full, the SSC transfers this data in the holding register, the status flag RXRDY is set in SR and the data can be read in the receiver holding register. If another
transfer occurs before read of the RHR register, the status flag OVERUN is set in SR and the
receiver shift register is transferred in the RHR register.
266
32058K AVR32-01/12
AT32UC3A
Figure 25-9. Receiver Block Diagram
R X _ C L O C K (p in )
T ri-sta te
C o n tro lle r
MUX
C lo ck
O u tp u t
T ra n sm itte r
C lo ck
D ivid e r
C lo ck
D a ta Tra n sfe r
CKO
CKS
25.7.4
IN V
MUX
T ri-sta te
C o n tro lle r
CKI
CKG
R e ce ive r
C lo ck
Start
The transmitter and receiver can both be programmed to start their operations when an event
occurs, respectively in the Transmit Start Selection (START) field of TCMR and in the Receive
Start Selection (START) field of RCMR.
Under the following conditions the start event is independently programmable:
• Continuous. In this case, the transmission starts as soon as a word is written in THR and the
reception starts as soon as the Receiver is enabled.
• Synchronously with the transmitter/receiver
• On detection of a falling/rising edge on TX_FRAME_SYNC/RX_FRAME_SYNC
• On detection of a low level/high level on TX_FRAME_SYNC/RX_FRAME_SYNC
• On detection of a level change or an edge on TX_FRAME_SYNC/RX_FRAME_SYNC
A start can be programmed in the same manner on either side of the Transmit/Receive Clock
Register (RCMR/TCMR). Thus, the start could be on TX_FRAME_SYNC (Transmit) or
RX_FRAME_SYNC (Receive).
Moreover, the Receiver can start when data is detected in the bit stream with the Compare
Functions.
Detection on TX_FRAME_SYNC/RX_FRAME_SYNC input/output is done by the field FSOS of
the Transmit/Receive Frame Mode Register (TFMR/RFMR).
267
32058K AVR32-01/12
AT32UC3A
Figure 25-10. Transmit Start Mode
TX_CLOCK (Input)
TX_FRAME_SYNC (Input)
TX_DATA (Output)
Start= Low Level on TX_FRAME_SYNC
TX_DATA (Output)
Start= Falling Edge on TX_FRAME_SYNC
X
B0
B0
X
TX_DATA (Output)
Start= High Level on TX_FRAME_SYNC
STTDLY
B0
B0
B1
B0
B1
B0
B1
B0
B1
X
TX_DATA (Output)
Start= Level Change on TX_FRAME_SYNC
X
X
STTDLY
B1
X
TX_DATA (Output)
Start= Rising Edge on TX_FRAME_SYNC
TX_DATA (Output)
Start= Any Edge on TX_FRAME_SYNC
B1
B0
B1
STTDLY
STTDLY
B1
STTDLY
STTDLY
Figure 25-11. Receive Pulse/Edge Start Modes
RX_CLOCK
RX_FRAME_SYNC (Input)
RX_DATA (Input)
X
Start = Low Level on RX_FRAME_SYNC
RX_DATA (Input)
Start = Falling Edge on RX_FRAME_SYNC
STTDLY
B0
X
RX_DATA (Input)
RX_DATA (Input)
B0
B0
B1
B0
B1
B0
B1
B0
B1
X
Start = Rising Edge on RX_FRAME_SYNC
RX_DATA (Input)
X
Start = Level Change on RX_FRAME_SYNC
RX_DATA (Input)
STTDLY
X
Start = High Level on RX_FRAME_SYNC
Start = Any Edge on RX_FRAME_SYNC
B1
X
B0
B1
STTDLY
STTDLY
B1
STTDLY
STTDLY
268
32058K AVR32-01/12
AT32UC3A
25.7.5
Frame Sync
The Transmitter and Receiver Frame Sync pins, TX_FRAME_SYNC and RX_FRAME_SYNC,
can be programmed to generate different kinds of frame synchronization signals. The Frame
Sync Output Selection (FSOS) field in the Receive Frame Mode Register (RFMR) and in the
Transmit Frame Mode Register (TFMR) are used to select the required waveform.
• Programmable low or high levels during data transfer are supported.
• Programmable high levels before the start of data transfers or toggling are also supported.
If a pulse waveform is selected, the Frame Sync Length (FSLEN) field in RFMR and TFMR programs the length of the pulse, from 1 bit time up to 16 bit time.
The periodicity of the Receive and Transmit Frame Sync pulse output can be programmed
through the Period Divider Selection (PERIOD) field in RCMR and TCMR.
25.7.5.1
Frame Sync Data
Frame Sync Data transmits or receives a specific tag during the Frame Sync signal.
During the Frame Sync signal, the Receiver can sample the RX_DATA line and store the data in
the Receive Sync Holding Register and the transmitter can transfer Transmit Sync Holding Register in the Shifter Register. The data length to be sampled/shifted out during the Frame Sync
signal is programmed by the FSLEN field in RFMR/TFMR.
Concerning the Receive Frame Sync Data operation, if the Frame Sync Length is equal to or
lower than the delay between the start event and the actual data reception, the data sampling
operation is performed in the Receive Sync Holding Register through the Receive Shift Register.
The Transmit Frame Sync Operation is performed by the transmitter only if the bit Frame Sync
Data Enable (FSDEN) in TFMR is set. If the Frame Sync length is equal to or lower than the
delay between the start event and the actual data transmission, the normal transmission has priority and the data contained in the Transmit Sync Holding Register is transferred in the Transmit
Register, then shifted out.
25.7.5.2
Frame Sync Edge Detection
The Frame Sync Edge detection is programmed by the FSEDGE field in RFMR/TFMR. This sets
the corresponding flags RXSYN/TXSYN in the SSC Status Register (SR) on frame synchro
edge detection (signals RX_FRAME_SYNC/TX_FRAME_SYNC).
269
32058K AVR32-01/12
AT32UC3A
25.7.6
Receive Compare Modes
Figure 25-12. Receive Compare Modes
RX_CLOCK
RX_DATA
(Input)
CMP0
CMP1
CMP2
CMP3
Ignored
B0
B1
B2
Start
FSLEN
Up to 16 Bits
(4 in This Example)
25.7.6.1
25.7.7
STTDLY
DATLEN
Compare Functions
Compare 0 can be one start event of the Receiver. In this case, the receiver compares at each
new sample the last FSLEN bits received at the FSLEN lower bit of the data contained in the
Compare 0 Register (RC0R). When this start event is selected, the user can program the
Receiver to start a new data transfer either by writing a new Compare 0, or by receiving continuously until Compare 1 occurs. This selection is done with the bit (STOP) in RCMR.
Data Format
The data framing format of both the transmitter and the receiver are programmable through the
Transmitter Frame Mode Register (TFMR) and the Receiver Frame Mode Register (RFMR). In
either case, the user can independently select:
• the event that starts the data transfer (START)
• the delay in number of bit periods between the start event and the first data bit (STTDLY)
• the length of the data (DATLEN)
• the number of data to be transferred for each start event (DATNB).
• the length of synchronization transferred for each start event (FSLEN)
• the bit sense: most or lowest significant bit first (MSBF).
Additionally, the transmitter can be used to transfer synchronization and select the level driven
on the TX_DATA pin while not in data transfer operation. This is done respectively by the Frame
Sync Data Enable (FSDEN) and by the Data Default Value (DATDEF) bits in TFMR.
270
32058K AVR32-01/12
AT32UC3A
Data Frame Registers
Table 25-3.
Transmitter
Receiver
Field
Length
Comment
TFMR
RFMR
DATLEN
Up to 32
Size of word
TFMR
RFMR
DATNB
Up to 16
Number of words transmitted in frame
TFMR
RFMR
MSBF
TFMR
RFMR
FSLEN
Up to 16
Size of Synchro data register
TFMR
DATDEF
0 or 1
Data default value ended
TFMR
FSDEN
Most significant bit first
Enable send TSHR
TCMR
RCMR
PERIOD
Up to 512
Frame size
TCMR
RCMR
STTDLY
Up to 255
Size of transmit start delay
Figure 25-13. Transmit and Receive Frame Format in Edge/Pulse Start Modes
Start
Start
PERIOD
TX_FRAME_SYNC
/
(1)
RX_FRAME_SYNC
FSLEN
TX_DATA
(If FSDEN = 1)
TX_DATA
(If FSDEN = 0)
Sync Data
From TSHR
Default
From DATDEF
Default
From DATDEF
RX_DATA
Sync Data
Ignored
To RSHR
STTDLY
Data
Data
From THR
From THR
Data
Data
From THR
From THR
Data
Data
To RHR
To RHR
Default
Sync Data
From DATDEF
Default
From DATDEF
Ignored
Sync Data
DATLEN
DATLEN
DATNB
Note:
1. Example of input on falling edge of TX_FRAME_SYNC/RX_FRAME_SYNC.
271
32058K AVR32-01/12
AT32UC3A
Figure 25-14. Transmit Frame Format in Continuous Mode
Start
TX_DATA
Data
Data
From THR
From THR
Default
DATLEN
DATLEN
Start: 1. TXEMPTY set to 1
2. Write into the THR
Note:
1. STTDLY is set to 0. In this example, THR is loaded twice. FSDEN value has no effect on the
transmission. SyncData cannot be output in continuous mode.
Figure 25-15. Receive Frame Format in Continuous Mode
Start = Enable Receiver
RX_DATA
Note:
25.7.8
Data
Data
To RHR
To RHR
DATLEN
DATLEN
1. STTDLY is set to 0.
Loop Mode
The receiver can be programmed to receive transmissions from the transmitter. This is done by
setting the Loop Mode (LOOP) bit in RFMR. In this case, RX_DATA is connected to TX_DATA,
RX_FRAME_SYNC is connected to TX_FRAME_SYNC and RX_CLOCK is connected to
TX_CLOCK.
25.7.9
Interrupt
Most bits in SR have a corresponding bit in interrupt management registers.
The SSC can be programmed to generate an interrupt when it detects an event. The interrupt is
controlled by writing IER (Interrupt Enable Register) and IDR (Interrupt Disable Register) These
registers enable and disable, respectively, the corresponding interrupt by setting and clearing
the corresponding bit in IMR (Interrupt Mask Register), which controls the generation of interrupts by asserting the SSC interrupt line connected to the interrupt controller.
272
32058K AVR32-01/12
AT32UC3A
Figure 25-16. Interrupt Block Diagram
IMR
PDCA
IER
IDR
Set
Clear
TXBUFE
ENDTX
Transmitter
TXRDY
TXEMPTY
TXSYNC
Interrupt
Control
RXBUFF
ENDRX
SSC Interrupt
Receiver
RXRDY
OVRUN
RXSYNC
25.8
SSC Application Examples
The SSC can support several serial communication modes used in audio or high speed serial
links. Some standard applications are shown in the following figures. All serial link applications
supported by the SSC are not listed here.
Figure 25-17. Audio Application Block Diagram
TX_CLOCK
Clock SCK
Word Select WS
TX_FRAME_SYNC
SSC
TX_DATA
RX_DATA
RX_FRAME_SYNC
RX_CLOCK
I2S
RECEIVER
Data SD
Clock SCK
Word Select WS
Data SD
MSB
LSB
Left Channel
MSB
Right Channel
273
32058K AVR32-01/12
AT32UC3A
Figure 25-18. Codec Application Block Diagram
TX_CLOCK
TX_FRAME_SYNC
SSC
Frame sync (FSYNC)
CODEC
Serial Data Out
TX_DATA
Serial Data In
RX_DATA
RX_FRAME_SYNC
RX_CLOCK
Serial Data Clock (SCLK)
Serial Data Clock (SCLK)
Frame sync (FSYNC)
First Time Slot
Dstart
Dend
Serial Data Out
Serial Data In
Figure 25-19. Time Slot Application Block Diagram
TX_CLOCK
TX_FRAME_SYNC
SSC
TX_DATA
RX_DATA
SCLK
FSYNC
CODEC
First
Time Slot
Data Out
Data in
RX_FRAME_SYNC
RX_CLOCK
CODEC
Second
Time Slot
Serial Data Clock (SCLK)
Frame sync (FSYNC)
First Time Slot
Dstart
Second Time Slot
Dend
Serial Data Out
Serial Data In
274
32058K AVR32-01/12
AT32UC3A
25.9
User Interface
Table 25-4.
Register Mapping
Offset
Register
Register Name
Access
Reset
CR
Write
–
CMR
Read/Write
0x0
0x0
Control Register
0x4
Clock Mode Register
0x8
Reserved
–
–
–
0xC
Reserved
–
–
–
0x10
Receive Clock Mode Register
RCMR
Read/Write
0x0
0x14
Receive Frame Mode Register
RFMR
Read/Write
0x0
0x18
Transmit Clock Mode Register
TCMR
Read/Write
0x0
0x1C
Transmit Frame Mode Register
TFMR
Read/Write
0x0
0x20
Receive Holding Register
RHR
Read
0x0
0x24
Transmit Holding Register
THR
Write
–
0x28
Reserved
–
–
–
0x2C
Reserved
–
–
–
0x30
Receive Sync. Holding Register
RSHR
Read
0x0
0x34
Transmit Sync. Holding Register
TSHR
Read/Write
0x0
0x38
Receive Compare 0 Register
RC0R
Read/Write
0x0
0x3C
Receive Compare 1 Register
RC1R
Read/Write
0x0
0x40
Status Register
SR
Read
0x000000CC
0x44
Interrupt Enable Register
IER
Write
–
0x48
Interrupt Disable Register
IDR
Write
–
0x4C
Interrupt Mask Register
IMR
Read
0x0
–
–
–
0x50-0xFC
Reserved
275
32058K AVR32-01/12
AT32UC3A
25.9.1
Name:
Control Register
CR
Access Type:
Write-only
Offset:
0x00
Reset value:
-
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
SWRST
14
–
13
–
12
–
11
–
10
–
9
TXDIS
8
TXEN
7
–
6
–
5
–
4
–
3
–
2
–
1
RXDIS
0
RXEN
• SWRST: Software Reset
0: No effect.
1: Performs a software reset. Has priority on any other bit in CR.
• TXDIS: Transmit Disable
0: No effect.
1: Disables Transmit. If a character is currently being transmitted, disables at end of current character transmission.
• TXEN: Transmit Enable
0: No effect.
1: Enables Transmit if TXDIS is not set.
• RXDIS: Receive Disable
0: No effect.
1: Disables Receive. If a character is currently being received, disables at end of current character reception.
• RXEN: Receive Enable
0: No effect.
1: Enables Receive if RXDIS is not set.
276
32058K AVR32-01/12
AT32UC3A
25.9.2
Name:
Clock Mode Register
CMR
Access Type:
Read/Write
Offset:
0x04
Reset value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
10
9
8
7
6
5
4
3
1
0
DIV
2
DIV
• DIV: Clock Divider
0: The Clock Divider is not active.
Any Other Value: The Divided Clock equals the Master Clock divided by 2 times DIV. The maximum bit rate is CLK_SSC/2.
The minimum bit rate is CLK_SSC/2 x 4095 = CLK_SSC/8190.
277
32058K AVR32-01/12
AT32UC3A
25.9.3
Name:
Receive Clock Mode Register
RCMR
Access Type:
Read/Write
Offset:
0x10
Reset value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
10
9
8
1
0
PERIOD
23
22
21
20
STTDLY
15
–
14
–
13
–
12
STOP
11
7
6
5
CKI
4
3
CKO
CKG
START
2
CKS
• PERIOD: Receive Period Divider Selection
This field selects the divider to apply to the selected Receive Clock in order to generate a new Frame Sync Signal. If 0, no
PERIOD signal is generated. If not 0, a PERIOD signal is generated each 2 x (PERIOD+1) Receive Clock.
• STTDLY: Receive Start Delay
If STTDLY is not 0, a delay of STTDLY clock cycles is inserted between the start event and the actual start of reception.
When the Receiver is programmed to start synchronously with the Transmitter, the delay is also applied.
Note: It is very important that STTDLY be set carefully. If STTDLY must be set, it should be done in relation to TAG
(Receive Sync Data) reception.
• STOP: Receive Stop Selection
0: After completion of a data transfer when starting with a Compare 0, the receiver stops the data transfer and waits for a
new compare 0.
1: After starting a receive with a Compare 0, the receiver operates in a continuous mode until a Compare 1 is detected.
• START: Receive Start Selection
START
Receive Start
0x0
Continuous, as soon as the receiver is enabled, and immediately after the end of
transfer of the previous data.
0x1
Transmit start
0x2
Detection of a low level on RX_FRAME_SYNC signal
0x3
Detection of a high level on RX_FRAME_SYNC signal
0x4
Detection of a falling edge on RX_FRAME_SYNC signal
0x5
Detection of a rising edge on RX_FRAME_SYNC signal
0x6
Detection of any level change on RX_FRAME_SYNC signal
0x7
Detection of any edge on RX_FRAME_SYNC signal
0x8
Compare 0
0x9-0xF
Reserved
278
32058K AVR32-01/12
AT32UC3A
• CKG: Receive Clock Gating Selection
CKG
Receive Clock Gating
0x0
None, continuous clock
0x1
Receive Clock enabled only if RX_FRAME_SYNC Low
0x2
Receive Clock enabled only if RX_FRAME_SYNC High
0x3
Reserved
• CKI: Receive Clock Inversion
0: The data inputs (Data and Frame Sync signals) are sampled on Receive Clock falling edge. The Frame Sync signal output is shifted out on Receive Clock rising edge.
1: The data inputs (Data and Frame Sync signals) are sampled on Receive Clock rising edge. The Frame Sync signal output is shifted out on Receive Clock falling edge.
CKI affects only the Receive Clock and not the output clock signal.
• CKO: Receive Clock Output Mode Selection
CKO
Receive Clock Output Mode
0x0
None
0x1
Continuous Receive Clock
Output
0x2
Receive Clock only during data transfers
Output
0x3-0x7
RX_CLOCK pin
Input-only
Reserved
• CKS: Receive Clock Selection
CKS
Selected Receive Clock
0x0
Divided Clock
0x1
TX_CLOCK Clock signal
0x2
RX_CLOCK pin
0x3
Reserved
279
32058K AVR32-01/12
AT32UC3A
25.9.4
Name:
Receive Frame Mode Register
RFMR
Access Type:
Read/Write
Offset:
0x14
Reset value:
0x00000000
31
30
29
28
27
–
26
–
21
FSOS
20
19
18
FSLENHI
23
–
22
15
–
14
–
13
–
12
–
11
7
MSBF
6
–
5
LOOP
4
3
25
–
24
FSEDGE
17
16
9
8
1
0
FSLEN
10
DATNB
2
DATLEN
• FSLENHI: Receive Frame Sync Length High part
The four MSB of the FSLEN bitfield.
• FSEDGE: Frame Sync Edge Detection
Determines which edge on Frame Sync will generate the interrupt RXSYN in the SSC Status Register.
FSEDGE
Frame Sync Edge Detection
0x0
Positive Edge Detection
0x1
Negative Edge Detection
• FSOS: Receive Frame Sync Output Selection
FSOS
Selected Receive Frame Sync Signal
RX_FRAME_SYNC Pin
0x0
None
0x1
Negative Pulse
Output
0x2
Positive Pulse
Output
0x3
Driven Low during data transfer
Output
0x4
Driven High during data transfer
Output
0x5
Toggling at each start of data transfer
Output
0x6-0x7
Reserved
Input-only
Undefined
• FSLEN: Receive Frame Sync Length
This field defines the length of the Receive Frame Sync Signal and the number of bits sampled and stored in the Receive
Sync Data Register. When this mode is selected by the START field in the Receive Clock Mode Register, it also determines the length of the sampled data to be compared to the Compare 0 or Compare 1 register. Note: The four most
significant bits fo this bitfield are in the FSLENHI bitfield.
Pulse length is equal to ({FSLENHI,FSLEN} + 1) Receive Clock periods. Thus, if {FSLENHI,FSLEN} is 0, the Receive
Frame Sync signal is generated during one Receive Clock period.
• DATNB: Data Number per Frame
280
32058K AVR32-01/12
AT32UC3A
This field defines the number of data words to be received after each transfer start, which is equal to (DATNB + 1).
• MSBF: Most Significant Bit First
0: The lowest significant bit of the data register is sampled first in the bit stream.
1: The most significant bit of the data register is sampled first in the bit stream.
• LOOP: Loop Mode
0: Normal operating mode.
1: RX_DATA is driven by TX_DATA, RX_FRAME_SYNC is driven by TX_FRAME_SYNC and TX_CLOCK drives
RX_CLOCK.
• DATLEN: Data Length
0: Forbidden value (1-bit data length not supported).
Any other value: The bit stream contains DATLEN + 1 data bits. Moreover, it defines the transfer size performed by the
PDCA assigned to the Receiver. If DATLEN is lower or equal to 7, data transfers are in bytes. If DATLEN is between 8 and
15 (included), half-words are transferred, and for any other value, 32-bit words are transferred.
281
32058K AVR32-01/12
AT32UC3A
25.9.5
Name:
Transmit Clock Mode Register
TCMR
Access Type:
Read/Write
Offset:
0x18
Reset value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
10
9
8
1
0
PERIOD
23
22
21
20
STTDLY
15
–
14
–
13
–
12
–
11
7
6
5
CKI
4
3
CKO
CKG
START
2
CKS
• PERIOD: Transmit Period Divider Selection
This field selects the divider to apply to the selected Transmit Clock to generate a new Frame Sync Signal. If 0, no period
signal is generated. If not 0, a period signal is generated at each 2 x (PERIOD+1) Transmit Clock.
• STTDLY: Transmit Start Delay
If STTDLY is not 0, a delay of STTDLY clock cycles is inserted between the start event and the actual start of transmission
of data. When the Transmitter is programmed to start synchronously with the Receiver, the delay is also applied.
Note: STTDLY must be set carefully. If STTDLY is too short in respect to TAG (Transmit Sync Data) emission, data is emitted instead of the end of TAG.
• START: Transmit Start Selection
START
Transmit Start
0x0
Continuous, as soon as a word is written in the THR Register (if Transmit is enabled), and immediately
after the end of transfer of the previous data.
0x1
Receive start
0x2
Detection of a low level on TX_FRAME_SYNC signal
0x3
Detection of a high level on TX_FRAME_SYNC signal
0x4
Detection of a falling edge on TX_FRAME_SYNC signal
0x5
Detection of a rising edge on TX_FRAME_SYNC signal
0x6
Detection of any level change on TX_FRAME_SYNC signal
0x7
Detection of any edge on TX_FRAME_SYNC signal
0x8 - 0xF
Reserved
282
32058K AVR32-01/12
AT32UC3A
• CKG: Transmit Clock Gating Selection
CKG
Transmit Clock Gating
0x0
None, continuous clock
0x1
Transmit Clock enabled only if TX_FRAME_SYNC Low
0x2
Transmit Clock enabled only if TX_FRAME_SYNC High
0x3
Reserved
• CKI: Transmit Clock Inversion
0: The data outputs (Data and Frame Sync signals) are shifted out on Transmit Clock falling edge. The Frame sync signal
input is sampled on Transmit clock rising edge.
1: The data outputs (Data and Frame Sync signals) are shifted out on Transmit Clock rising edge. The Frame sync signal
input is sampled on Transmit clock falling edge.
CKI affects only the Transmit Clock and not the output clock signal.
• CKO: Transmit Clock Output Mode Selection
CKO
Transmit Clock Output Mode
0x0
None
0x1
Continuous Transmit Clock
Output
0x2
Transmit Clock only during data transfers
Output
0x3-0x7
TX_CLOCK pin
Input-only
Reserved
• CKS: Transmit Clock Selection
CKS
Selected Transmit Clock
0x0
Divided Clock
0x1
RX_CLOCK Clock signal
0x2
TX_CLOCK Pin
0x3
Reserved
283
32058K AVR32-01/12
AT32UC3A
25.9.6
Name:
Transmit Frame Mode Register
TFMR
Access Type:
Read/Write
Offset:
0x1C
Reset value:
0x00000000
31
30
29
28
27
–
26
–
21
FSOS
20
19
18
FSLENHI
23
FSDEN
22
15
–
14
–
13
–
12
–
11
7
MSBF
6
–
5
DATDEF
4
3
25
–
24
FSEDGE
17
16
9
8
1
0
FSLEN
10
DATNB
2
DATLEN
• FSLENHI: Transmit Frame Sync Length High part
The four MSB of the FSLEN bitfield.
• FSEDGE: Frame Sync Edge Detection
Determines which edge on frame sync will generate the interrupt TXSYN (Status Register).
FSEDGE
Frame Sync Edge Detection
0x0
Positive Edge Detection
0x1
Negative Edge Detection
• FSDEN: Frame Sync Data Enable
0: The TX_DATA line is driven with the default value during the Transmit Frame Sync signal.
1: TSHR value is shifted out during the transmission of the Transmit Frame Sync signal.
• FSOS: Transmit Frame Sync Output Selection
FSOS
Selected Transmit Frame Sync Signal
TX_FRAME_SYNC Pin
0x0
None
0x1
Negative Pulse
Output
0x2
Positive Pulse
Output
0x3
Driven Low during data transfer
Output
0x4
Driven High during data transfer
Output
0x5
Toggling at each start of data transfer
Output
0x6-0x7
Reserved
Input-only
Undefined
• FSLEN: Transmit Frame Sync Length
This field defines the length of the Transmit Frame Sync signal and the number of bits shifted out from the Transmit Sync
Data Register if FSDEN is 1. Note: The four most significant bits fo this bitfield are in the FSLENHI bitfield.
284
32058K AVR32-01/12
AT32UC3A
Pulse length is equal to ({FSLENHI,FSLEN} + 1) Transmit Clock periods, i.e., the pulse length can range from 1 to 16
Transmit Clock periods. If {FSLENHI,FSLEN} is 0, the Transmit Frame Sync signal is generated during one Transmit Clock
period.
• DATNB: Data Number per frame
This field defines the number of data words to be transferred after each transfer start, which is equal to (DATNB +1).
• MSBF: Most Significant Bit First
0: The lowest significant bit of the data register is shifted out first in the bit stream.
1: The most significant bit of the data register is shifted out first in the bit stream.
• DATDEF: Data Default Value
This bit defines the level driven on the TX_DATA pin while out of transmission. Note that if the pin is defined as multi-drive
by the PIO Controller, the pin is enabled only if the SCC TX_DATA output is 1.
• DATLEN: Data Length
0: Forbidden value (1-bit data length not supported).
Any other value: The bit stream contains DATLEN + 1 data bits. Moreover, it defines the transfer size performed by the
PDCA assigned to the Transmit. If DATLEN is lower or equal to 7, data transfers are bytes, if DATLEN is between 8 and 15
(included), half-words are transferred, and for any other value, 32-bit words are transferred.
285
32058K AVR32-01/12
AT32UC3A
25.9.7
Name:
SSC Receive Holding Register
RHR
Access Type:
Read-only
Offset:
0x20
Reset value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
RDAT
23
22
21
20
RDAT
15
14
13
12
RDAT
7
6
5
4
RDAT
• RDAT: Receive Data
Right aligned regardless of the number of data bits defined by DATLEN in RFMR.
286
32058K AVR32-01/12
AT32UC3A
25.9.8
Name:
Transmit Holding Register
THR
Access Type:
Write-only
Offset:
0x24
Reset value:
-
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
TDAT
23
22
21
20
TDAT
15
14
13
12
TDAT
7
6
5
4
TDAT
• TDAT: Transmit Data
Right aligned regardless of the number of data bits defined by DATLEN in TFMR.
287
32058K AVR32-01/12
AT32UC3A
25.9.9
Name:
Receive Synchronization Holding Register
RSHR
Access Type:
Read-only
Offset:
0x30
Reset value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
RSDAT
7
6
5
4
RSDAT
• RSDAT: Receive Synchronization Data
288
32058K AVR32-01/12
AT32UC3A
25.9.10
Name:
Transmit Synchronization Holding Register
TSHR
Access Type:
Read/Write
Offset:
0x34
Reset value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
TSDAT
7
6
5
4
TSDAT
• TSDAT: Transmit Synchronization Data
289
32058K AVR32-01/12
AT32UC3A
25.9.11
Name:
Receive Compare 0 Register
RC0R
Access Type:
Read/Write
Offset:
0x38
Reset value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
CP0
7
6
5
4
CP0
• CP0: Receive Compare Data 0
290
32058K AVR32-01/12
AT32UC3A
25.9.12
Name:
Receive Compare 1 Register
RC1R
Access Type:
Read/Write
Offset:
0x3C
Reset value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
CP1
7
6
5
4
CP1
• CP1: Receive Compare Data 1
291
32058K AVR32-01/12
AT32UC3A
25.9.13
Name:
Status Register
SR
Access Type:
Read-only
Offset:
0x40
Reset value:
0x000000CC
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
RXEN
16
TXEN
15
–
14
–
13
–
12
–
11
RXSYN
10
TXSYN
9
CP1
8
CP0
7
RXBUFF
6
ENDRX
5
OVRUN
4
RXRDY
3
TXBUFE
2
ENDTX
1
TXEMPTY
0
TXRDY
• RXEN: Receive Enable
0: Receive is disabled.
1: Receive is enabled.
• TXEN: Transmit Enable
0: Transmit is disabled.
1: Transmit is enabled.
• RXSYN: Receive Sync
0: An Rx Sync has not occurred since the last read of the Status Register.
1: An Rx Sync has occurred since the last read of the Status Register.
• TXSYN: Transmit Sync
0: A Tx Sync has not occurred since the last read of the Status Register.
1: A Tx Sync has occurred since the last read of the Status Register.
• CP1: Compare 1
0: A compare 1 has not occurred since the last read of the Status Register.
1: A compare 1 has occurred since the last read of the Status Register.
• CP0: Compare 0
0: A compare 0 has not occurred since the last read of the Status Register.
1: A compare 0 has occurred since the last read of the Status Register.
• RXBUFF: Receive Buffer Full
0: RCR or RNCR have a value other than 0.
1: Both RCR and RNCR have a value of 0.
• ENDRX: End of Reception
0: Data is written on the Receive Counter Register or Receive Next Counter Register.
292
32058K AVR32-01/12
AT32UC3A
1: End of PDCA transfer when Receive Counter Register has arrived at zero.
• OVRUN: Receive Overrun
0: No data has been loaded in RHR while previous data has not been read since the last read of the Status Register.
1: Data has been loaded in RHR while previous data has not yet been read since the last read of the Status Register.
• RXRDY: Receive Ready
0: RHR is empty.
1: Data has been received and loaded in RHR.
• TXBUFE: Transmit Buffer Empty
0: TCR or TNCR have a value other than 0.
1: Both TCR and TNCR have a value of 0.
• ENDTX: End of Transmission
0: The register TCR has not reached 0 since the last write in TCR or TNCR.
1: The register TCR has reached 0 since the last write in TCR or TNCR.
• TXEMPTY: Transmit Empty
0: Data remains in THR or is currently transmitted from TSR.
1: Last data written in THR has been loaded in TSR and last data loaded in TSR has been transmitted.
• TXRDY: Transmit Ready
0: Data has been loaded in THR and is waiting to be loaded in the Transmit Shift Register (TSR).
1: THR is empty.
293
32058K AVR32-01/12
AT32UC3A
25.9.14
Name:
Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x44
Reset value:
-
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
RXSYN
10
TXSYN
9
CP1
8
CP0
7
RXBUFF
6
ENDRX
5
OVRUN
4
RXRDY
3
TXBUFE
2
ENDTX
1
TXEMPTY
0
TXRDY
• RXSYN: Rx Sync Interrupt Enable
0: No effect.
1: Enables the Rx Sync Interrupt.
• TXSYN: Tx Sync Interrupt Enable
0: No effect.
1: Enables the Tx Sync Interrupt.
• CP1: Compare 1 Interrupt Enable
0: No effect.
1: Enables the Compare 1 Interrupt.
• CP0: Compare 0 Interrupt Enable
0: No effect.
1: Enables the Compare 0 Interrupt.
• RXBUFF: Receive Buffer Full Interrupt Enable
0: No effect.
1: Enables the Receive Buffer Full Interrupt.
• ENDRX: End of Reception Interrupt Enable
0: No effect.
1: Enables the End of Reception Interrupt.
• OVRUN: Receive Overrun Interrupt Enable
0: No effect.
1: Enables the Receive Overrun Interrupt.
• RXRDY: Receive Ready Interrupt Enable
0: No effect.
294
32058K AVR32-01/12
AT32UC3A
1: Enables the Receive Ready Interrupt.
• TXBUFE: Transmit Buffer Empty Interrupt Enable
0: No effect.
1: Enables the Transmit Buffer Empty Interrupt
• ENDTX: End of Transmission Interrupt Enable
0: No effect.
1: Enables the End of Transmission Interrupt.
• TXEMPTY: Transmit Empty Interrupt Enable
0: No effect.
1: Enables the Transmit Empty Interrupt.
• TXRDY: Transmit Ready Interrupt Enable
0: No effect.
1: Enables the Transmit Ready Interrupt.
295
32058K AVR32-01/12
AT32UC3A
25.9.15
Name:
Interrupt Disable Register
IDR
Access Type:
Write-only
Offset:
0x48
Reset value:
-
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
RXSYN
10
TXSYN
9
CP1
8
CP0
7
RXBUFF
6
ENDRX
5
OVRUN
4
RXRDY
3
TXBUFE
2
ENDTX
1
TXEMPTY
0
TXRDY
• RXSYN: Rx Sync Interrupt Enable
0: No effect.
1: Disables the Rx Sync Interrupt.
• TXSYN: Tx Sync Interrupt Enable
0: No effect.
1: Disables the Tx Sync Interrupt.
• CP1: Compare 1 Interrupt Disable
0: No effect.
1: Disables the Compare 1 Interrupt.
• CP0: Compare 0 Interrupt Disable
0: No effect.
1: Disables the Compare 0 Interrupt.
• RXBUFF: Receive Buffer Full Interrupt Disable
0: No effect.
1: Disables the Receive Buffer Full Interrupt.
• ENDRX: End of Reception Interrupt Disable
0: No effect.
1: Disables the End of Reception Interrupt.
• OVRUN: Receive Overrun Interrupt Disable
0: No effect.
1: Disables the Receive Overrun Interrupt.
• RXRDY: Receive Ready Interrupt Disable
0: No effect.
296
32058K AVR32-01/12
AT32UC3A
1: Disables the Receive Ready Interrupt.
• TXBUFE: Transmit Buffer Empty Interrupt Disable
0: No effect.
1: Disables the Transmit Buffer Empty Interrupt.
• ENDTX: End of Transmission Interrupt Disable
0: No effect.
1: Disables the End of Transmission Interrupt.
• TXEMPTY: Transmit Empty Interrupt Disable
0: No effect.
1: Disables the Transmit Empty Interrupt.
• TXRDY: Transmit Ready Interrupt Disable
0: No effect.
1: Disables the Transmit Ready Interrupt.
297
32058K AVR32-01/12
AT32UC3A
25.9.16
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
Offset:
0x4C
Reset value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
RXSYN
10
TXSYN
9
CP1
8
CP0
7
RXBUFF
6
ENDRX
5
OVRUN
4
RXRDY
3
TXBUFE
2
ENDTX
1
TXEMPTY
0
TXRDY
• RXSYN: Rx Sync Interrupt Mask
0: The Rx Sync Interrupt is disabled.
1: The Rx Sync Interrupt is enabled.
• TXSYN: Tx Sync Interrupt Mask
0: The Tx Sync Interrupt is disabled.
1: The Tx Sync Interrupt is enabled.
• CP1: Compare 1 Interrupt Mask
0: The Compare 1 Interrupt is disabled.
1: The Compare 1 Interrupt is enabled.
• CP0: Compare 0 Interrupt Mask
0: The Compare 0 Interrupt is disabled.
1: The Compare 0 Interrupt is enabled.
• RXBUFF: Receive Buffer Full Interrupt Mask
0: The Receive Buffer Full Interrupt is disabled.
1: The Receive Buffer Full Interrupt is enabled.
• ENDRX: End of Reception Interrupt Mask
0: The End of Reception Interrupt is disabled.
1: The End of Reception Interrupt is enabled.
• OVRUN: Receive Overrun Interrupt Mask
0: The Receive Overrun Interrupt is disabled.
1: The Receive Overrun Interrupt is enabled.
• RXRDY: Receive Ready Interrupt Mask
0: The Receive Ready Interrupt is disabled.
298
32058K AVR32-01/12
AT32UC3A
26. Universal Synchronous/Asynchronous Receiver/Transmitter (USART)
Rev. 4.0.0.2
26.1
Features
• Programmable Baud Rate Generator
• 5- to 9-bit Full-duplex Synchronous or Asynchronous Serial Communications
•
•
•
•
•
•
26.2
– 1, 1.5 or 2 Stop Bits in Asynchronous Mode or 1 or 2 Stop Bits in Synchronous Mode
– Parity Generation and Error Detection
– Framing Error Detection, Overrun Error Detection
– MSB- or LSB-first
– Optional Break Generation and Detection
– By 8 or by 16 Over-sampling Receiver Frequency
– Optional Hardware Handshaking RTS-CTS
– Receiver Time-out and Transmitter Timeguard
– Optional Multidrop Mode with Address Generation and Detection
RS485 with Driver Control Signal
ISO7816, T = 0 or T = 1 Protocols for Interfacing with Smart Cards
– NACK Handling, Error Counter with Repetition and Iteration Limit
IrDA Modulation and Demodulation
– Communication at up to 115.2 Kbps
SPI Mode
– Master or Slave
– Serial Clock Programmable Phase and Polarity
– SPI Serial Clock (CLK) Frequency up to Internal Clock Frequency CLK_USART/4
Test Modes
– Remote Loopback, Local Loopback, Automatic Echo
Supports Connection of Two Peripheral DMA Controller Channels (PDC)
– Offers Buffer Transfer without Processor Intervention
Overview
The Universal Synchronous Asynchronous Receiver Transceiver (USART) provides one full
duplex universal synchronous asynchronous serial link. Data frame format is widely programmable (data length, parity, number of stop bits) to support a maximum of standards. The receiver
implements parity error, framing error and overrun error detection. The receiver time-out enables
handling variable-length frames and the transmitter timeguard facilitates communications with
slow remote devices. Multidrop communications are also supported through address bit handling in reception and transmission.
The USART features three test modes: remote loopback, local loopback and automatic echo.
The USART supports specific operating modes providing interfaces on RS485 and SPI buses,
with ISO7816 T = 0 or T = 1 smart card slots and infrared transceivers. The hardware handshaking feature enables an out-of-band flow control by automatic management of the pins RTS and
CTS.
The USART supports the connection to the Peripheral DMA Controller, which enables data
transfers to the transmitter and from the receiver. The PDC provides chained buffer management without any intervention of the processor.
299
32058K AVR32-01/12
AT32UC3A
26.3
Block Diagram
Figure 26-1. USART Block Diagram
Peripheral DMA
Controller
Channel
Channel
USART
PIO
Controller
RXD
Receiver
INTC
USART
Interrupt
RTS
TXD
Transmitter
CTS
CLK_USART
Power
Manager
DIV
BaudRate
Generator
CLK_USART/DIV
CLK
User
Interface
Peripheral bus
300
32058K AVR32-01/12
AT32UC3A
26.4
Application Block Diagram
Figure 26-2. Application Block Diagram
IrLAP
PPP
Modem
Driver
Serial
Driver
Field Bus
Driver
EMV
Driver
IrDA
Driver
SPI
Driver
USART
RS232
Drivers
RS232
Drivers
RS485
Drivers
Serial
Port
Differential
Bus
Smart
Card
Slot
IrDA
Transceivers
SPI
Transceiver
Modem
PSTN
301
32058K
AVR32-01/12
AT32UC3A
26.5
I/O Lines Description
Table 26-1.
I/O Line Description
Name
Description
Type
Active Level
CLK
Serial Clock
I/O
TXD
Transmit Serial Data
or Master Out Slave In (MOSI) in SPI Master Mode
or Master In Slave Out (MISO) in SPI Slave Mode
I/O
RXD
Receive Serial Data
or Master In Slave Out (MISO) in SPI Master Mode
or Master Out Slave In (MOSI) in SPI Slave Mode
Input
CTS
Clear to Send
or Slave Select (NSS) in SPI Slave Mode
Input
Low
RTS
Request to Send
or Slave Select (NSS) in SPI Master Mode
Output
Low
302
32058K
AVR32-01/12
AT32UC3A
26.6
Product Dependencies
26.6.1
I/O Lines
The pins used for interfacing the USART may be multiplexed with the PIO lines. The programmer must first program the PIO controller to assign the desired USART pins to their peripheral
function. If I/O lines of the USART are not used by the application, they can be used for other
purposes by the PIO Controller.
To prevent the TXD line from falling when the USART is disabled, the use of an internal pull up
is mandatory. If the hardware handshaking feature or Modem mode is used, the internal pull up
on TXD must also be enabled.
26.6.2
Power Manager (PM)
The USART is not continuously clocked. The programmer must first enable the USART Clock in
the Power Manager (PM) before using the USART. However, if the application does not require
USART operations, the USART clock can be stopped when not needed and be restarted later.
In this case, the USART will resume its operations where it left off.
Configuring the USART does not require the USART clock to be enabled.
26.6.3
Interrupt
The USART interrupt line is connected on one of the internal sources of the Advanced Interrupt
Controller. Using the USART interrupt requires the INTC to be programmed first. Note that it is
not recommended to use the USART interrupt line in edge sensitive mode.
303
32058K
AVR32-01/12
AT32UC3A
26.7
Functional Description
The USART is capable of managing several types of serial synchronous or asynchronous
communications.
It supports the following communication modes:
•5- to 9-bit full-duplex asynchronous serial communication
–MSB- or LSB-first
–1, 1.5 or 2 stop bits
–Parity even, odd, marked, space or none
–By 8 or by 16 over-sampling receiver frequency
–Optional hardware handshaking
–Optional break management
–Optional multidrop serial communication
•High-speed 5- to 9-bit full-duplex synchronous serial communication
–MSB- or LSB-first
–1 or 2 stop bits
–Parity even, odd, marked, space or none
–By 8 or by 16 over-sampling frequency
–Optional hardware handshaking
–Optional break management
–Optional multidrop serial communication
•RS485 with driver control signal
•ISO7816, T0 or T1 protocols for interfacing with smart cards
–NACK handling, error counter with repetition and iteration limit
•InfraRed IrDA Modulation and Demodulation
• SPI Mode
– Master or Slave
– Serial Clock Programmable Phase and Polarity
– SPI Serial Clock (CLK) Frequency up to Internal Clock Frequency CLK_USART/4
•Test modes
–Remote loopback, local loopback, automatic echo
304
32058K
AVR32-01/12
AT32UC3A
26.7.1
Baud Rate Generator
The Baud Rate Generator provides the bit period clock named the Baud Rate Clock to both the
receiver and the transmitter.
The Baud Rate Generator clock source can be selected by setting the USCLKS field in the Mode
Register (MR) between:
•the CLK_USART
•a division of the CLK_USART, the divider being product dependent, but generally set to 8
•the external clock, available on the CLK pin
The Baud Rate Generator is based upon a 16-bit divider, which is programmed with the CD field
of the Baud Rate Generator Register (BRGR). If CD is programmed at 0, the Baud Rate Generator does not generate any clock. If CD is programmed at 1, the divider is bypassed and
becomes inactive.
If the external CLK clock is selected, the duration of the low and high levels of the signal provided on the CLK pin must be longer than a CLK_USART period. The frequency of the signal
provided on CLK must be at least 4.5 times lower than CLK_USART.
Figure 26-3. Baud Rate Generator
USCLKS
CLK_USART
CLK_USART/DIV
CLK
Reserved
CD
CD
0
1
2
16-bit Counter
CLK
FIDI
>1
3
1
0
0
Sampling
Divider
0
SYNC
OVER
0
BaudRate
Clock
1
1
SYNC
USCLKS= 3
26.7.1.1
Sampling
Clock
Baud Rate in Asynchronous Mode
If the USART is programmed to operate in asynchronous mode, the selected clock is first
divided by CD, which is field programmed in the Baud Rate Generator Register (BRGR). The
resulting clock is provided to the receiver as a sampling clock and then divided by 16 or 8,
depending on the programming of the OVER bit in MR.
If OVER is set to 1, the receiver sampling is 8 times higher than the baud rate clock. If OVER is
cleared, the sampling is performed at 16 times the baud rate clock.
The following formula performs the calculation of the Baud Rate.
SelectedClock
Baudrate = -------------------------------------------( 8 ( 2 – Over )CD )
305
32058K
AVR32-01/12
AT32UC3A
This gives a maximum baud rate of CLK_USART divided by 8, assuming that CLK_USART is
the highest possible clock and that OVER is programmed at 1.
26.7.1.2
Baud Rate Calculation Example
Table 26-2 on page 306 shows calculations of CD to obtain a baud rate at 38400 bauds for different source clock frequencies. This table also shows the actual resulting baud rate and the
error.
Table 26-2.
Baud Rate Example (OVER = 0)
Source Clock
Expected Baud
Rate
MHz
Bit/s
3 686 400
38 400
6.00
6
38 400.00
0.00%
4 915 200
38 400
8.00
8
38 400.00
0.00%
5 000 000
38 400
8.14
8
39 062.50
1.70%
7 372 800
38 400
12.00
12
38 400.00
0.00%
8 000 000
38 400
13.02
13
38 461.54
0.16%
12 000 000
38 400
19.53
20
37 500.00
2.40%
12 288 000
38 400
20.00
20
38 400.00
0.00%
14 318 180
38 400
23.30
23
38 908.10
1.31%
14 745 600
38 400
24.00
24
38 400.00
0.00%
18 432 000
38 400
30.00
30
38 400.00
0.00%
24 000 000
38 400
39.06
39
38 461.54
0.16%
24 576 000
38 400
40.00
40
38 400.00
0.00%
25 000 000
38 400
40.69
40
38 109.76
0.76%
32 000 000
38 400
52.08
52
38 461.54
0.16%
32 768 000
38 400
53.33
53
38 641.51
0.63%
33 000 000
38 400
53.71
54
38 194.44
0.54%
40 000 000
38 400
65.10
65
38 461.54
0.16%
50 000 000
38 400
81.38
81
38 580.25
0.47%
60 000 000
38 400
97.66
98
38 265.31
0.35%
70 000 000
38 400
113.93
114
38 377.19
0.06%
Calculation Result
CD
Actual Baud Rate
Error
Bit/s
The baud rate is calculated with the following formula:
BaudRate = ( CLKUSART ) ⁄ CD × 16
The baud rate error is calculated with the following formula. It is not recommended to work with
an error higher than 5%.
ExpectedBaudRate
Error = 1 – ⎛ ---------------------------------------------------⎞
⎝ ActualBaudRate ⎠
306
32058K
AVR32-01/12
AT32UC3A
26.7.1.3
Fractional Baud Rate in Asynchronous Mode
The Baud Rate generator previously defined is subject to the following limitation: the output frequency changes by only integer multiples of the reference frequency. An approach to this
problem is to integrate a fractional N clock generator that has a high resolution. The generator
architecture is modified to obtain Baud Rate changes by a fraction of the reference source clock.
This fractional part is programmed with the FP field in the Baud Rate Generator Register
(BRGR). If FP is not 0, the fractional part is activated. The resolution is one eighth of the clock
divider. This feature is only available when using USART normal mode. The fractional Baud
Rate is calculated using the following formula:
SelectedClock
Baudrate = ---------------------------------------------------------------FP-⎞ ⎞
⎛ 8 ( 2 – Over ) ⎛ CD + -----⎝
⎝
8 ⎠⎠
The modified architecture is presented below:
Figure 26-4. Fractional Baud Rate Generator
FP
USCLKS
CLK_USART
CLK_USART/DIV
CLK
Reserved
0
1
2
3
CD
Modulus
Control
16-bit Counter
FP
CD
glitch-free
logic
FIDI
>1
1
0
CLK
0
OVER
Sampling
Divider
0
SYNC
0
BaudRate
Clock
1
1
SYNC
USCLKS = 3
26.7.1.4
Sampling
Clock
Baud Rate in Synchronous Mode or SPI Mode
If the USART is programmed to operate in synchronous mode, the selected clock is simply
divided by the field CD in BRGR.
SelectedClockBaudRate = ------------------------------------CD
In synchronous mode, if the external clock is selected (USCLKS = 3), the clock is provided
directly by the signal on the USART CLK pin. No division is active. The value written in BRGR
307
32058K
AVR32-01/12
AT32UC3A
has no effect. The external clock frequency must be at least 4.5 times lower than the system
clock.
When either the external clock CLK or the internal clock divided (CLK_USART/DIV) is selected,
the value programmed in CD must be even if the user has to ensure a 50:50 mark/space ratio on
the CLK pin. If the internal clock CLK_USART is selected, the Baud Rate Generator ensures a
50:50 duty cycle on the CLK pin, even if the value programmed in CD is odd.
26.7.1.5
Baud Rate in ISO 7816 Mode
The ISO7816 specification defines the bit rate with the following formula:
Di
B = ------ × f
Fi
where:
•B is the bit rate
•Di is the bit-rate adjustment factor
•Fi is the clock frequency division factor
•f is the ISO7816 clock frequency (Hz)
Di is a binary value encoded on a 4-bit field, named DI, as represented in Table 26-3 on page
308.
Binary and Decimal Values for Di
Table 26-3.
DI field
0001
0010
0011
0100
0101
0110
1000
1001
1
2
4
8
16
32
12
20
Di (decimal)
Fi is a binary value encoded on a 4-bit field, named FI, as represented in Table 26-4 on page
308.
Binary and Decimal Values for Fi
Table 26-4.
FI field
0000
0001
0010
0011
0100
0101
0110
1001
1010
1011
1100
1101
Fi (decimal
372
372
558
744
1116
1488
1860
512
768
1024
1536
2048
Table 26-5 on page 308 shows the resulting Fi/Di Ratio, which is the ratio between the ISO7816
clock and the baud rate clock.
Possible Values for the Fi/Di Ratio
Table 26-5.
Fi/Di
372
558
774
1116
1488
1806
512
768
1024
1536
2048
1
372
558
744
1116
1488
1860
512
768
1024
1536
2048
2
186
279
372
558
744
930
256
384
512
768
1024
4
93
139.5
186
279
372
465
128
192
256
384
512
8
46.5
69.75
93
139.5
186
232.5
64
96
128
192
256
16
23.25
34.87
46.5
69.75
93
116.2
32
48
64
96
128
32
11.62
17.43
23.25
34.87
46.5
58.13
16
24
32
48
64
12
31
46.5
62
93
124
155
42.66
64
85.33
128
170.6
20
18.6
27.9
37.2
55.8
74.4
93
25.6
38.4
51.2
76.8
102.4
If the USART is configured in ISO7816 Mode, the clock selected by the USCLKS field in the
Mode Register (MR) is first divided by the value programmed in the field CD in the Baud Rate
308
32058K
AVR32-01/12
AT32UC3A
Generator Register (BRGR). The resulting clock can be provided to the CLK pin to feed the
smart card clock inputs. This means that the CLKO bit can be set in MR.
This clock is then divided by the value programmed in the FI_DI_RATIO field in the FI_DI_Ratio
register (FIDI). This is performed by the Sampling Divider, which performs a division by up to
2047 in ISO7816 Mode. The non-integer values of the Fi/Di Ratio are not supported and the user
must program the FI_DI_RATIO field to a value as close as possible to the expected value.
The FI_DI_RATIO field resets to the value 0x174 (372 in decimal) and is the most common
divider between the ISO7816 clock and the bit rate (Fi = 372, Di = 1).
Figure 26-5 on page 309 shows the relation between the Elementary Time Unit, corresponding
to a bit time, and the ISO 7816 clock.
Figure 26-5. Elementary Time Unit (ETU)
FI_DI_RATIO
ISO7816 Clock Cycles
ISO7816 Clock
on CLK
ISO7816 I/O Line
on TXD
1 ETU
26.7.2
Receiver and Transmitter Control
After reset, the receiver is disabled. The user must enable the receiver by setting the RXEN bit
in the Control Register (CR). However, the receiver registers can be programmed before the
receiver clock is enabled.
After reset, the transmitter is disabled. The user must enable it by setting the TXEN bit in the
Control Register (CR). However, the transmitter registers can be programmed before being
enabled.
The Receiver and the Transmitter can be enabled together or independently.
At any time, the software can perform a reset on the receiver or the transmitter of the USART by
setting the corresponding bit, RSTRX and RSTTX respectively, in the Control Register (CR).
The software resets clear the status flag and reset internal state machines but the user interface
configuration registers hold the value configured prior to software reset. Regardless of what the
receiver or the transmitter is performing, the communication is immediately stopped.
The user can also independently disable the receiver or the transmitter by setting RXDIS and
TXDIS respectively in CR. If the receiver is disabled during a character reception, the USART
waits until the end of reception of the current character, then the reception is stopped. If the
transmitter is disabled while it is operating, the USART waits the end of transmission of both the
current character and character being stored in the Transmit Holding Register (THR). If a timeguard is programmed, it is handled normally.
309
32058K
AVR32-01/12
AT32UC3A
26.7.3
26.7.3.1
Synchronous and Asynchronous Modes
Transmitter Operations
The transmitter performs the same in both synchronous and asynchronous operating modes
(SYNC = 0 or SYNC = 1). One start bit, up to 9 data bits, one optional parity bit and up to two
stop bits are successively shifted out on the TXD pin at each falling edge of the programmed
serial clock.
The number of data bits is selected by the CHRL field and the MODE 9 bit in the Mode Register
(MR). Nine bits are selected by setting the MODE 9 bit regardless of the CHRL field. The parity
bit is set according to the PAR field in MR. The even, odd, space, marked or none parity bit can
be configured. The MSBF field in MR configures which data bit is sent first. If written at 1, the
most significant bit is sent first. At 0, the less significant bit is sent first. The number of stop bits is
selected by the NBSTOP field in MR. The 1.5 stop bit is supported in asynchronous mode only.
Figure 26-6. Character Transmit
Example: 8-bit, Parity Enabled One Stop
Baud Rate
Clock
TXD
D0
Start
Bit
D1
D2
D3
D4
D5
D6
D7
Parity
Bit
Stop
Bit
The characters are sent by writing in the Transmit Holding Register (THR). The transmitter
reports two status bits in the Channel Status Register (CSR): TXRDY (Transmitter Ready),
which indicates that THR is empty and TXEMPTY, which indicates that all the characters written
in THR have been processed. When the current character processing is completed, the last
character written in THR is transferred into the Shift Register of the transmitter and THR
becomes empty, thus TXRDY rises.
Both TXRDY and TXEMPTY bits are low when the transmitter is disabled. Writing a character in
THR while TXRDY is low has no effect and the written character is lost.
Figure 26-7. Transmitter Status
Baud Rate
Clock
TXD
Start
D0
Bit
D1
D2
D3
D4
D5
D6
D7
Parity Stop Start
D0
Bit Bit Bit
D1
D2
D3
D4
D5
D6
D7
Parity Stop
Bit Bit
Write
US_THR
TXRDY
TXEMPTY
310
32058K
AVR32-01/12
AT32UC3A
26.7.3.2
Manchester Encoder
When the Manchester encoder is in use, characters transmitted through the USART are
encoded based on biphase Manchester II format. To enable this mode, set the MAN field in the
MR register to 1. Depending on polarity configuration, a logic level (zero or one), is transmitted
as a coded signal one-to-zero or zero-to-one. Thus, a transition always occurs at the midpoint of
each bit time. It consumes more bandwidth than the original NRZ signal (2x) but the receiver has
more error control since the expected input must show a change at the center of a bit cell. An
example of Manchester encoded sequence is: the byte 0xB1 or 10110001 encodes to 10 01 10
10 01 01 01 10, assuming the default polarity of the encoder. Figure 26-8 on page 311 illustrates
this coding scheme.
Figure 26-8. NRZ to Manchester Encoding
NRZ
encoded
data
Manchester
encoded
data
1
0
1
1
0
0
0
1
Txd
The Manchester encoded character can also be encapsulated by adding both a configurable
preamble and a start frame delimiter pattern. Depending on the configuration, the preamble is a
training sequence, composed of a pre-defined pattern with a programmable length from 1 to 15
bit times. If the preamble length is set to 0, the preamble waveform is not generated prior to any
character. The preamble pattern is chosen among the following sequences: ALL_ONE,
ALL_ZERO, ONE_ZERO or ZERO_ONE, writing the field TX_PP in the MAN register, the field
TX_PL is used to configure the preamble length. Figure 26-9 on page 312 illustrates and defines
the valid patterns. To improve flexibility, the encoding scheme can be configured using the
TX_MPOL field in the MAN register. If the TX_MPOL field is set to zero (default), a logic zero is
encoded with a zero-to-one transition and a logic one is encoded with a one-to-zero transition. If
the TX_MPOL field is set to one, a logic one is encoded with a one-to-zero transition and a logic
zero is encoded with a zero-to-one transition.
311
32058K
AVR32-01/12
AT32UC3A
Figure 26-9. Preamble Patterns, Default Polarity Assumed
Manchester
encoded
data
Txd
SFD
DATA
SFD
DATA
SFD
DATA
SFD
DATA
8 bit width "ALL_ONE" Preamble
Manchester
encoded
data
Txd
8 bit width "ALL_ZERO" Preamble
Manchester
encoded
data
Txd
8 bit width "ZERO_ONE" Preamble
Manchester
encoded
data
Txd
8 bit width "ONE_ZERO" Preamble
A start frame delimiter is to be configured using the ONEBIT field in the MR register. It consists
of a user-defined pattern that indicates the beginning of a valid data. Figure 26-10 on page 313
illustrates these patterns. If the start frame delimiter, also known as start bit, is one bit, (ONEBIT
at 1), a logic zero is Manchester encoded and indicates that a new character is being sent serially on the line. If the start frame delimiter is a synchronization pattern also referred to as sync
(ONEBIT at 0), a sequence of 3 bit times is sent serially on the line to indicate the start of a new
character. The sync waveform is in itself an invalid Manchester waveform as the transition
occurs at the middle of the second bit time. Two distinct sync patterns are used: the command
sync and the data sync. The command sync has a logic one level for one and a half bit times,
then a transition to logic zero for the second one and a half bit times. If the MODSYNC field in
the MR register is set to 1, the next character is a command. If it is set to 0, the next character is
a data. When direct memory access is used, the MODSYNC field can be immediately updated
with a modified character located in memory. To enable this mode, VAR_SYNC field in MR register must be set to 1. In this case, the MODSYNC field in MR is bypassed and the sync
configuration is held in the TXSYNH in the THR register. The USART character format is modified and includes sync information.
312
32058K
AVR32-01/12
AT32UC3A
Figure 26-10. Start Frame Delimiter
Preamble Length
is set to 0
Manchester
encoded
data
Manchester
encoded
data
Manchester
encoded
data
SFD
DATA
Txd
SFD
One bit start frame delimiter
DATA
Txd
Command Sync
start frame delimiter
SFD
DATA
Txd
Data Sync
start frame delimiter
26.7.3.3
Drift Compensation
Drift compensation is available only in 16X oversampling mode. An hardware recovery system
allows a larger clock drift. To enable the hardware system, the bit in the MAN register must be
set. If the RXD edge is one 16X clock cycle from the expected edge, this is considered as normal jitter and no corrective actions is taken. If the RXD event is between 4 and 2 clock cycles
before the expected edge, then the current period is shortened by one clock cycle. If the RXD
event is between 2 and 3 clock cycles after the expected edge, then the current period is lengthened by one clock cycle. These intervals are considered to be drift and so corrective actions are
automatically taken.
Figure 26-11. Bit Resynchronization
Oversampling
16x Clock
RXD
Sampling
point
Expected edge
Synchro.
Error
26.7.3.4
Synchro.
Jump
Tolerance
Sync
Jump
Synchro.
Error
Asynchronous Receiver
If the USART is programmed in asynchronous operating mode (SYNC = 0), the receiver oversamples the RXD input line. The oversampling is either 16 or 8 times the Baud Rate clock,
depending on the OVER bit in the Mode Register (MR).
313
32058K
AVR32-01/12
AT32UC3A
The receiver samples the RXD line. If the line is sampled during one half of a bit time at 0, a start
bit is detected and data, parity and stop bits are successively sampled on the bit rate clock.
If the oversampling is 16, (OVER at 0), a start is detected at the eighth sample at 0. Then, data
bits, parity bit and stop bit are sampled on each 16 sampling clock cycle. If the oversampling is 8
(OVER at 1), a start bit is detected at the fourth sample at 0. Then, data bits, parity bit and stop
bit are sampled on each 8 sampling clock cycle.
The number of data bits, first bit sent and parity mode are selected by the same fields and bits
as the transmitter, i.e. respectively CHRL, MODE9, MSBF and PAR. For the synchronization
mechanism only, the number of stop bits has no effect on the receiver as it considers only one
stop bit, regardless of the field NBSTOP, so that resynchronization between the receiver and the
transmitter can occur. Moreover, as soon as the stop bit is sampled, the receiver starts looking
for a new start bit so that resynchronization can also be accomplished when the transmitter is
operating with one stop bit.
Figure 26-12 on page 314 and Figure 26-13 on page 314 illustrate start detection and character
reception when USART operates in asynchronous mode.
Figure 26-12. Asynchronous Start Detection
Baud Rate
Clock
Sampling
Clock (x16)
RXD
Sampling
1
2
3
4
5
6
7
8
1
2
3
4
2
3
4
5
6
7
8
9 10 11 12 13 14 15 16
D0
Sampling
Start
Detection
RXD
Sampling
1
2
3
4
5
6
7
0 1
Start
Rejection
Figure 26-13. Asynchronous Character Reception
Example: 8-bit, Parity Enabled
Baud Rate
Clock
RXD
Start
Detection
16
16
16
16
16
16
16
16
16
16
samples samples samples samples samples samples samples samples samples samples
D0
D1
D2
D3
D4
D5
D6
D7
Parity
Bit
Stop
Bit
314
32058K
AVR32-01/12
AT32UC3A
26.7.3.5
Manchester Decoder
When the MAN field in MR register is set to 1, the Manchester decoder is enabled. The decoder
performs both preamble and start frame delimiter detection. One input line is dedicated to Manchester encoded input data.
An optional preamble sequence can be defined, its length is user-defined and totally independent of the emitter side. Use RX_PL in MAN register to configure the length of the preamble
sequence. If the length is set to 0, no preamble is detected and the function is disabled. In addition, the polarity of the input stream is programmable with RX_MPOL field in MAN register.
Depending on the desired application the preamble pattern matching is to be defined via the
RX_PP field in MAN. See Figure 26-9 on page 312 for available preamble patterns.
Unlike preamble, the start frame delimiter is shared between Manchester Encoder and Decoder.
So, if ONEBIT field is set to 1, only a zero encoded Manchester can be detected as a valid start
frame delimiter. If ONEBIT is set to 0, only a sync pattern is detected as a valid start frame
delimiter. Decoder operates by detecting transition on incoming stream. If RXD is sampled during one quarter of a bit time at zero, a start bit is detected. See Figure 26-14 on page 315.. The
sample pulse rejection mechanism applies.
Figure 26-14. Asynchronous Start Bit Detection
Sampling
Clock
(16 x)
Manchester
encoded
data
Txd
1
2
3
4
Start
Detection
The receiver is activated and starts Preamble and Frame Delimiter detection, sampling the data
at one quarter and then three quarters. If a valid preamble pattern or start frame delimiter is
detected, the receiver continues decoding with the same synchronization. If the stream does not
match a valid pattern or a valid start frame delimiter, the receiver re-synchronizes on the next
valid edge.The minimum time threshold to estimate the bit value is three quarters of a bit time.
If a valid preamble (if used) followed with a valid start frame delimiter is detected, the incoming
stream is decoded into NRZ data and passed to USART for processing. Figure 26-15 on page
316 illustrates Manchester pattern mismatch. When incoming data stream is passed to the
USART, the receiver is also able to detect Manchester code violation. A code violation is a lack
of transition in the middle of a bit cell. In this case, MANE flag in CSR register is raised. It is
cleared by writing the Control Register (CR) with the RSTSTA bit at 1. See Figure 26-16 on page
316 for an example of Manchester error detection during data phase.
315
32058K
AVR32-01/12
AT32UC3A
Figure 26-15. Preamble Pattern Mismatch
Preamble Mismatch
Manchester coding error
Manchester
encoded
data
Preamble Mismatch
invalid pattern
SFD
Txd
DATA
Preamble Length is set to 8
Figure 26-16. Manchester Error Flag
Preamble Length
is set to 4
Elementary character bit time
SFD
Manchester
encoded
data
Txd
Entering USART character area
sampling points
Preamble subpacket
and Start Frame Delimiter
were successfully
decoded
Manchester
Coding Error
detected
When the start frame delimiter is a sync pattern (ONEBIT field at 0), both command and data
delimiter are supported. If a valid sync is detected, the received character is written as RXCHR
field in the RHR register and the RXSYNH is updated. RXCHR is set to 1 when the received
character is a command, and it is set to 0 if the received character is a data. This mechanism
alleviates and simplifies the direct memory access as the character contains its own sync field in
the same register.
As the decoder is setup to be used in unipolar mode, the first bit of the frame has to be a zero-toone transition.
26.7.3.6
Radio Interface: Manchester Encoded USART Application
This section describes low data rate RF transmission systems and their integration with a Manchester encoded USART. These systems are based on transmitter and receiver ICs that support
ASK and FSK modulation schemes.
The goal is to perform full duplex radio transmission of characters using two different frequency
carriers. See the configuration in Figure 26-17 on page 317.
316
32058K
AVR32-01/12
AT32UC3A
Figure 26-17. Manchester Encoded Characters RF Transmission
Fup frequency Carrier
ASK/FSK
Upstream Receiver
Upstream
Emitter
LNA
VCO
RF filter
Demod
control
Fdown frequency Carrier
Serial
Configuration
Interface
bi-dir
line
ASK/FSK
downstream transmitter
Downstream
Receiver
Manchester
decoder
USART
Receiver
Manchester
encoder
USART
Emitter
PA
RF filter
Mod
VCO
control
The USART module is configured as a Manchester encoder/decoder. Looking at the downstream communication channel, Manchester encoded characters are serially sent to the RF
emitter. This may also include a user defined preamble and a start frame delimiter. Mostly, preamble is used in the RF receiver to distinguish between a valid data from a transmitter and
signals due to noise. The Manchester stream is then modulated. See Figure 26-18 on page 317
for an example of ASK modulation scheme. When a logic one is sent to the ASK modulator, the
power amplifier, referred to as PA, is enabled and transmits an RF signal at downstream frequency. When a logic zero is transmitted, the RF signal is turned off. If the FSK modulator is
activated, two different frequencies are used to transmit data. When a logic 1 is sent, the modulator outputs an RF signal at frequency F0 and switches to F1 if the data sent is a 0. See Figure
26-19 on page 318.
From the receiver side, another carrier frequency is used. The RF receiver performs a bit check
operation examining demodulated data stream. If a valid pattern is detected, the receiver
switches to receiving mode. The demodulated stream is sent to the Manchester decoder.
Because of bit checking inside RF IC, the data transferred to the microcontroller is reduced by a
user-defined number of bits. The Manchester preamble length is to be defined in accordance
with the RF IC configuration.
Figure 26-18. ASK Modulator Output
1
0
0
1
NRZ stream
Manchester
encoded
data
default polarity
unipolar output
Txd
ASK Modulator
Output
Uptstream Frequency F0
317
32058K
AVR32-01/12
AT32UC3A
Figure 26-19. FSK Modulator Output
1
0
0
1
NRZ stream
Manchester
encoded
data
default polarity
unipolar output
Txd
FSK Modulator
Output
Uptstream Frequencies
[F0, F0+offset]
26.7.4
Synchronous Receiver
In synchronous mode (SYNC = 1), the receiver samples the RXD signal on each rising edge of
the Baud Rate Clock. If a low level is detected, it is considered as a start. All data bits, the parity
bit and the stop bits are sampled and the receiver waits for the next start bit. Synchronous mode
operations provide a high speed transfer capability.
Configuration fields and bits are the same as in asynchronous mode.
Figure 26-20 on page 318 illustrates a character reception in synchronous mode.
Figure 26-20. Synchronous Mode Character Reception
Example: 8-bit, Parity Enabled 1 Stop
Baud Rate
Clock
RXD
Sampling
Start
D0
D1
D2
D3
D4
D5
D6
D7
Stop Bit
Parity Bit
26.7.4.1
Receiver Operations
When a character reception is completed, it is transferred to the Receive Holding Register
(RHR) and the RXRDY bit in the Status Register (CSR) rises. If a character is completed while
the RXRDY is set, the OVRE (Overrun Error) bit is set. The last character is transferred into
RHR and overwrites the previous one. The OVRE bit is cleared by writing the Control Register
(CR) with the RSTSTA (Reset Status) bit at 1.
318
32058K
AVR32-01/12
AT32UC3A
Figure 26-21. Receiver Status
Baud Rate
Clock
RXD
Start
D0
Bit
D1
D2
D3
D4
D5
D6
D7
Parity Stop Start
D0
Bit Bit Bit
D1
D2
D3
D4
D5
D6
D7
Parity Stop
Bit Bit
RSTSTA = 1
Write
US_CR
Read
US_RHR
RXRDY
OVRE
319
32058K
AVR32-01/12
AT32UC3A
26.7.4.2
Parity
The USART supports five parity modes selected by programming the PAR field in the Mode
Register (MR). The PAR field also enables the Multidrop mode, see ”Multidrop Mode” on page
321. Even and odd parity bit generation and error detection are supported.
If even parity is selected, the parity generator of the transmitter drives the parity bit at 0 if a number of 1s in the character data bit is even, and at 1 if the number of 1s is odd. Accordingly, the
receiver parity checker counts the number of received 1s and reports a parity error if the sampled parity bit does not correspond. If odd parity is selected, the parity generator of the
transmitter drives the parity bit at 1 if a number of 1s in the character data bit is even, and at 0 if
the number of 1s is odd. Accordingly, the receiver parity checker counts the number of received
1s and reports a parity error if the sampled parity bit does not correspond. If the mark parity is
used, the parity generator of the transmitter drives the parity bit at 1 for all characters. The
receiver parity checker reports an error if the parity bit is sampled at 0. If the space parity is
used, the parity generator of the transmitter drives the parity bit at 0 for all characters. The
receiver parity checker reports an error if the parity bit is sampled at 1. If parity is disabled, the
transmitter does not generate any parity bit and the receiver does not report any parity error.
Table 26-6 on page 320 shows an example of the parity bit for the character 0x41 (character
ASCII “A”) depending on the configuration of the USART. Because there are two bits at 1, 1 bit is
added when a parity is odd, or 0 is added when a parity is even.
Table 26-6.
Parity Bit Examples
Character
Hexa
Binary
Parity Bit
Parity Mode
A
0x41
0100 0001
1
Odd
A
0x41
0100 0001
0
Even
A
0x41
0100 0001
1
Mark
A
0x41
0100 0001
0
Space
A
0x41
0100 0001
None
None
When the receiver detects a parity error, it sets the PARE (Parity Error) bit in the Channel Status
Register (CSR). The PARE bit can be cleared by writing the Control Register (CR) with the RSTSTA bit at 1. Figure 26-22 on page 321 illustrates the parity bit status setting and clearing.
320
32058K
AVR32-01/12
AT32UC3A
Figure 26-22. Parity Error
Baud Rate
Clock
RXD
Start
D0
Bit
Write
US_CR
D1
D2
D3
D4
D5
D6
D7
Bad Stop
Parity Bit
Bit
RSTSTA = 1
PARE
RXRDY
26.7.4.3
Multidrop Mode
If the PAR field in the Mode Register (MR) is programmed to the value 0x6 or 0x07, the USART
runs in Multidrop Mode. This mode differentiates the data characters and the address characters. Data is transmitted with the parity bit at 0 and addresses are transmitted with the parity bit
at 1.
If the USART is configured in multidrop mode, the receiver sets the PARE parity error bit when
the parity bit is high and the transmitter is able to send a character with the parity bit high when
the Control Register is written with the SENDA bit at 1.
To handle parity error, the PARE bit is cleared when the Control Register is written with the bit
RSTSTA at 1.
The transmitter sends an address byte (parity bit set) when SENDA is written to CR. In this case,
the next byte written to THR is transmitted as an address. Any character written in THR without
having written the command SENDA is transmitted normally with the parity at 0.
26.7.4.4
Transmitter Timeguard
The timeguard feature enables the USART interface with slow remote devices.
The timeguard function enables the transmitter to insert an idle state on the TXD line between
two characters. This idle state actually acts as a long stop bit.
The duration of the idle state is programmed in the TG field of the Transmitter Timeguard Register (TTGR). When this field is programmed at zero no timeguard is generated. Otherwise, the
transmitter holds a high level on TXD after each transmitted byte during the number of bit periods programmed in TG in addition to the number of stop bits.
As illustrated in Figure 26-23 on page 322, the behavior of TXRDY and TXEMPTY status bits is
modified by the programming of a timeguard. TXRDY rises only when the start bit of the next
character is sent, and thus remains at 0 during the timeguard transmission if a character has
been written in THR. TXEMPTY remains low until the timeguard transmission is completed as
the timeguard is part of the current character being transmitted.
321
32058K
AVR32-01/12
AT32UC3A
Figure 26-23. Timeguard Operations
TG = 4
TG = 4
Baud Rate
Clock
TXD
Start
D0
Bit
D1
D2
D3
D4
D5
D6
D7
Parity Stop
Bit Bit
Start
D0
Bit
D1
D2
D3
D4
D5
D6
D7
Parity Stop
Bit Bit
Write
US_THR
TXRDY
TXEMPTY
Table 26-7 on page 322 indicates the maximum length of a timeguard period that the transmitter
can handle in relation to the function of the Baud Rate.
Table 26-7.
26.7.4.5
Maximum Timeguard Length Depending on Baud Rate
Baud Rate
Bit time
Timeguard
Bit/sec
µs
ms
1 200
833
212.50
9 600
104
26.56
14400
69.4
17.71
19200
52.1
13.28
28800
34.7
8.85
33400
29.9
7.63
56000
17.9
4.55
57600
17.4
4.43
115200
8.7
2.21
Receiver Time-out
The Receiver Time-out provides support in handling variable-length frames. This feature detects
an idle condition on the RXD line. When a time-out is detected, the bit TIMEOUT in the Channel
Status Register (CSR) rises and can generate an interrupt, thus indicating to the driver an end of
frame.
The time-out delay period (during which the receiver waits for a new character) is programmed
in the TO field of the Receiver Time-out Register (RTOR). If the TO field is programmed at 0, the
Receiver Time-out is disabled and no time-out is detected. The TIMEOUT bit in CSR remains at
0. Otherwise, the receiver loads a 16-bit counter with the value programmed in TO. This counter
is decremented at each bit period and reloaded each time a new character is received. If the
counter reaches 0, the TIMEOUT bit in the Status Register rises. Then, the user can either:
• Stop the counter clock until a new character is received. This is performed by writing the
Control Register (CR) with the STTTO (Start Time-out) bit at 1. In this case, the idle state on
RXD before a new character is received will not provide a time-out. This prevents having to
322
32058K
AVR32-01/12
AT32UC3A
handle an interrupt before a character is received and allows waiting for the next idle state on
RXD after a frame is received.
• Obtain an interrupt while no character is received. This is performed by writing CR with the
RETTO (Reload and Start Time-out) bit at 1. If RETTO is performed, the counter starts
counting down immediately from the value TO. This enables generation of a periodic interrupt
so that a user time-out can be handled, for example when no key is pressed on a keyboard.
If STTTO is performed, the counter clock is stopped until a first character is received. The idle
state on RXD before the start of the frame does not provide a time-out. This prevents having to
obtain a periodic interrupt and enables a wait of the end of frame when the idle state on RXD is
detected.
If RETTO is performed, the counter starts counting down immediately from the value TO. This
enables generation of a periodic interrupt so that a user time-out can be handled, for example
when no key is pressed on a keyboard.
Figure 26-24 on page 323 shows the block diagram of the Receiver Time-out feature.
Figure 26-24. Receiver Time-out Block Diagram
TO
Baud Rate
Clock
1
D
Q
Clock
16-bit Time-out
Counter
16-bit
Value
=
STTTO
Clear
Character
Received
Load
TIMEOUT
0
RETTO
Table 26-8 on page 323 gives the maximum time-out period for some standard baud rates.
Table 26-8.
Maximum Time-out Period
Baud Rate
Bit Time
Time-out
bit/sec
µs
ms
600
1 667
109 225
1 200
833
54 613
2 400
417
27 306
4 800
208
13 653
9 600
104
6 827
14400
69
4 551
19200
52
3 413
28800
35
2 276
33400
30
1 962
56000
18
1 170
57600
17
1 138
200000
5
328
323
32058K
AVR32-01/12
AT32UC3A
26.7.4.6
Framing Error
The receiver is capable of detecting framing errors. A framing error happens when the stop bit of
a received character is detected at level 0. This can occur if the receiver and the transmitter are
fully desynchronized.
A framing error is reported on the FRAME bit of the Channel Status Register (CSR). The
FRAME bit is asserted in the middle of the stop bit as soon as the framing error is detected. It is
cleared by writing the Control Register (CR) with the RSTSTA bit at 1.
Figure 26-25. Framing Error Status
Baud Rate
Clock
RXD
Start
D0
Bit
D1
D2
D3
D4
D5
D6
D7
Parity Stop
Bit Bit
RSTSTA = 1
Write
US_CR
FRAME
RXRDY
26.7.4.7
Transmit Break
The user can request the transmitter to generate a break condition on the TXD line. A break condition drives the TXD line low during at least one complete character. It appears the same as a
0x00 character sent with the parity and the stop bits at 0. However, the transmitter holds the
TXD line at least during one character until the user requests the break condition to be removed.
A break is transmitted by writing the Control Register (CR) with the STTBRK bit at 1. This can be
performed at any time, either while the transmitter is empty (no character in either the Shift Register or in THR) or when a character is being transmitted. If a break is requested while a
character is being shifted out, the character is first completed before the TXD line is held low.
Once STTBRK command is requested further STTBRK commands are ignored until the end of
the break is completed.
The break condition is removed by writing CR with the STPBRK bit at 1. If the STPBRK is
requested before the end of the minimum break duration (one character, including start, data,
parity and stop bits), the transmitter ensures that the break condition completes.
The transmitter considers the break as though it is a character, i.e. the STTBRK and STPBRK
commands are taken into account only if the TXRDY bit in CSR is at 1 and the start of the break
condition clears the TXRDY and TXEMPTY bits as if a character is processed.
Writing CR with the both STTBRK and STPBRK bits at 1 can lead to an unpredictable result. All
STPBRK commands requested without a previous STTBRK command are ignored. A byte written into the Transmit Holding Register while a break is pending, but not started, is ignored.
324
32058K
AVR32-01/12
AT32UC3A
After the break condition, the transmitter returns the TXD line to 1 for a minimum of 12 bit times.
Thus, the transmitter ensures that the remote receiver detects correctly the end of break and the
start of the next character. If the timeguard is programmed with a value higher than 12, the TXD
line is held high for the timeguard period.
After holding the TXD line for this period, the transmitter resumes normal operations.
Figure 26-26 on page 325 illustrates the effect of both the Start Break (STTBRK) and Stop Break
(STPBRK) commands on the TXD line.
Figure 26-26. Break Transmission
Baud Rate
Clock
TXD
Start
D0
Bit
D1
D2
D3
D4
D5
D6
D7
Parity Stop
Bit Bit
Break Transmission
STTBRK = 1
End of Break
STPBRK = 1
Write
US_CR
TXRDY
TXEMPTY
26.7.4.8
Receive Break
The receiver detects a break condition when all data, parity and stop bits are low. This corresponds to detecting a framing error with data at 0x00, but FRAME remains low.
When the low stop bit is detected, the receiver asserts the RXBRK bit in CSR. This bit may be
cleared by writing the Control Register (CR) with the bit RSTSTA at 1.
An end of receive break is detected by a high level for at least 2/16 of a bit period in asynchronous operating mode or one sample at high level in synchronous operating mode. The end of
break detection also asserts the RXBRK bit.
26.7.4.9
Hardware Handshaking
The USART features a hardware handshaking out-of-band flow control. The RTS and CTS pins
are used to connect with the remote device, as shown in Figure 26-27 on page 325.
Figure 26-27. Connection with a Remote Device for Hardware Handshaking
USART
Remote
Device
TXD
RXD
RXD
TXD
CTS
RTS
RTS
CTS
325
32058K
AVR32-01/12
AT32UC3A
Setting the USART to operate with hardware handshaking is performed by writing the MODE
field in the Mode Register (MR) to the value 0x2.
The USART behavior when hardware handshaking is enabled is the same as the behavior in
standard synchronous or asynchronous mode, except that the receiver drives the RTS pin as
described below and the level on the CTS pin modifies the behavior of the transmitter as
described below. Using this mode requires using the PDC channel for reception. The transmitter
can handle hardware handshaking in any case.
Figure 26-28 on page 326 shows how the receiver operates if hardware handshaking is enabled.
The RTS pin is driven high if the receiver is disabled and if the status RXBUFF (Receive Buffer
Full) coming from the PDC channel is high. Normally, the remote device does not start transmitting while its CTS pin (driven by RTS) is high. As soon as the Receiver is enabled, the RTS falls,
indicating to the remote device that it can start transmitting. Defining a new buffer to the PDC
clears the status bit RXBUFF and, as a result, asserts the pin RTS low.
Figure 26-28. Receiver Behavior when Operating with Hardware Handshaking
RXD
Write
US_CR
RXEN = 1
RXDIS = 1
RTS
RXBUFF
Figure 26-29 on page 326 shows how the transmitter operates if hardware handshaking is
enabled. The CTS pin disables the transmitter. If a character is being processing, the transmitter
is disabled only after the completion of the current character and transmission of the next character happens as soon as the pin CTS falls.
Figure 26-29. Transmitter Behavior when Operating with Hardware Handshaking
CTS
TXD
26.7.5
ISO7816 Mode
The USART features an ISO7816-compatible operating mode. This mode permits interfacing
with smart cards and Security Access Modules (SAM) communicating through an ISO7816 link.
Both T = 0 and T = 1 protocols defined by the ISO7816 specification are supported.
Setting the USART in ISO7816 mode is performed by writing the MODE field in the Mode Register (MR) to the value 0x4 for protocol T = 0 and to the value 0x5 for protocol T = 1.
26.7.5.1
ISO7816 Mode Overview
The ISO7816 is a half duplex communication on only one bidirectional line. The baud rate is
determined by a division of the clock provided to the remote device (see ”Baud Rate Generator”
on page 305).
The USART connects to a smart card as shown in Figure 26-30 on page 327. The TXD line
becomes bidirectional and the Baud Rate Generator feeds the ISO7816 clock on the CLK pin.
326
32058K
AVR32-01/12
AT32UC3A
As the TXD pin becomes bidirectional, its output remains driven by the output of the transmitter
but only when the transmitter is active while its input is directed to the input of the receiver. The
USART is considered as the master of the communication as it generates the clock.
Figure 26-30. Connection of a Smart Card to the USART
USART
CLK
TXD
CLK
I/O
Smart
Card
When operating in ISO7816, either in T = 0 or T = 1 modes, the character format is fixed. The
configuration is 8 data bits, even parity and 1 or 2 stop bits, regardless of the values programmed in the CHRL, MODE9, PAR and CHMODE fields. MSBF can be used to transmit LSB
or MSB first. Parity Bit (PAR) can be used to transmit in normal or inverse mode. Refer to
”USART Mode Register” on page 343 and ”PAR: Parity Type” on page 345.
The USART cannot operate concurrently in both receiver and transmitter modes as the communication is unidirectional at a time. It has to be configured according to the required mode by
enabling or disabling either the receiver or the transmitter as desired. Enabling both the receiver
and the transmitter at the same time in ISO7816 mode may lead to unpredictable results.
The ISO7816 specification defines an inverse transmission format. Data bits of the character
must be transmitted on the I/O line at their negative value. The USART does not support this format and the user has to perform an exclusive OR on the data before writing it in the Transmit
Holding Register (THR) or after reading it in the Receive Holding Register (RHR).
26.7.5.2
Protocol T = 0
In T = 0 protocol, a character is made up of one start bit, eight data bits, one parity bit and one
guard time, which lasts two bit times. The transmitter shifts out the bits and does not drive the
I/O line during the guard time.
If no parity error is detected, the I/O line remains at 1 during the guard time and the transmitter
can continue with the transmission of the next character, as shown in Figure 26-31 on page 328.
If a parity error is detected by the receiver, it drives the I/O line at 0 during the guard time, as
shown in Figure 26-32 on page 328. This error bit is also named NACK, for Non Acknowledge.
In this case, the character lasts 1 bit time more, as the guard time length is the same and is
added to the error bit time which lasts 1 bit time.
When the USART is the receiver and it detects an error, it does not load the erroneous character
in the Receive Holding Register (RHR). It appropriately sets the PARE bit in the Status Register
(SR) so that the software can handle the error.
327
32058K
AVR32-01/12
AT32UC3A
Figure 26-31. T = 0 Protocol without Parity Error
Baud Rate
Clock
RXD
Start
Bit
D0
D2
D1
D4
D3
D5
D6
D7
Parity Guard Guard Next
Bit Time 1 Time 2 Start
Bit
Figure 26-32. T = 0 Protocol with Parity Error
Baud Rate
Clock
Error
I/O
Start
Bit
D0
D1
D2
D3
D4
D5
D6
D7
Parity Guard
Bit Time 1
Guard Start
Time 2 Bit
D0
D1
Repetition
26.7.5.3
Receive Error Counter
The USART receiver also records the total number of errors. This can be read in the Number of
Error (NER) register. The NB_ERRORS field can record up to 255 errors. Reading NER automatically clears the NB_ERRORS field.
26.7.5.4
Receive NACK Inhibit
The USART can also be configured to inhibit an error. This can be achieved by setting the
INACK bit in the Mode Register (MR). If INACK is at 1, no error signal is driven on the I/O line
even if a parity bit is detected, but the INACK bit is set in the Status Register (SR). The INACK
bit can be cleared by writing the Control Register (CR) with the RSTNACK bit at 1.
Moreover, if INACK is set, the erroneous received character is stored in the Receive Holding
Register, as if no error occurred. However, the RXRDY bit does not raise.
26.7.5.5
Transmit Character Repetition
When the USART is transmitting a character and gets a NACK, it can automatically repeat the
character before moving on to the next one. Repetition is enabled by writing the
MAX_ITERATION field in the Mode Register (MR) at a value higher than 0. Each character can
be transmitted up to eight times; the first transmission plus seven repetitions.
If MAX_ITERATION does not equal zero, the USART repeats the character as many times as
the value loaded in MAX_ITERATION.
When the USART repetition number reaches MAX_ITERATION, the ITERATION bit is set in the
Channel Status Register (CSR). If the repetition of the character is acknowledged by the
receiver, the repetitions are stopped and the iteration counter is cleared.
The ITERATION bit in CSR can be cleared by writing the Control Register with the RSIT bit at 1.
26.7.5.6
Disable Successive Receive NACK
The receiver can limit the number of successive NACKs sent back to the remote transmitter.
This is programmed by setting the bit DSNACK in the Mode Register (MR). The maximum number of NACK transmitted is programmed in the MAX_ITERATION field. As soon as
328
32058K
AVR32-01/12
AT32UC3A
MAX_ITERATION is reached, the character is considered as correct, an acknowledge is sent on
the line and the ITERATION bit in the Channel Status Register is set.
26.7.5.7
Protocol T = 1
When operating in ISO7816 protocol T = 1, the transmission is similar to an asynchronous format with only one stop bit. The parity is generated when transmitting and checked when
receiving. Parity error detection sets the PARE bit in the Channel Status Register (CSR).
26.7.6
IrDA Mode
The USART features an IrDA mode supplying half-duplex point-to-point wireless communication. It embeds the modulator and demodulator which allows a glueless connection to the
infrared transceivers, as shown in Figure 26-33 on page 329. The modulator and demodulator
are compliant with the IrDA specification version 1.1 and support data transfer speeds ranging
from 2.4 Kb/s to 115.2 Kb/s.
The USART IrDA mode is enabled by setting the MODE field in the Mode Register (MR) to the
value 0x8. The IrDA Filter Register (IFR) allows configuring the demodulator filter. The USART
transmitter and receiver operate in a normal asynchronous mode and all parameters are accessible. Note that the modulator and the demodulator are activated.
Figure 26-33. Connection to IrDA Transceivers
USART
IrDA
Transceivers
Receiver
Demodulator
RXD
Transmitter
Modulator
TXD
RX
TX
The receiver and the transmitter must be enabled or disabled according to the direction of the
transmission to be managed.
26.7.6.1
IrDA Modulation
For baud rates up to and including 115.2 Kbits/sec, the RZI modulation scheme is used. “0” is
represented by a light pulse of 3/16th of a bit time. Some examples of signal pulse duration are
shown in Table 26-9 on page 329.
Table 26-9.
IrDA Pulse Duration
Baud Rate
Pulse Duration (3/16)
2.4 Kb/s
78.13 µs
9.6 Kb/s
19.53 µs
19.2 Kb/s
9.77 µs
329
32058K
AVR32-01/12
AT32UC3A
Table 26-9.
IrDA Pulse Duration
Baud Rate
Pulse Duration (3/16)
38.4 Kb/s
4.88 µs
57.6 Kb/s
3.26 µs
115.2 Kb/s
1.63 µs
Figure 26-34 on page 330 shows an example of character transmission.
Figure 26-34. IrDA Modulation
Start
Bit
Transmitter
Output
0
Stop
Bit
Data Bits
1
0
1
0
1
0
1
0
1
TXD
3
16 Bit Period
Bit Period
26.7.6.2
IrDA Baud Rate
Table 26-10 on page 330 gives some examples of CD values, baud rate error and pulse duration. Note that the requirement on the maximum acceptable error of ±1.87% must be met.
Table 26-10. IrDA Baud Rate Error
Peripheral Clock
Baud Rate
CD
Baud Rate Error
Pulse Time
3 686 400
115 200
2
0.00%
1.63
20 000 000
115 200
11
1.38%
1.63
32 768 000
115 200
18
1.25%
1.63
40 000 000
115 200
22
1.38%
1.63
3 686 400
57 600
4
0.00%
3.26
20 000 000
57 600
22
1.38%
3.26
32 768 000
57 600
36
1.25%
3.26
40 000 000
57 600
43
0.93%
3.26
3 686 400
38 400
6
0.00%
4.88
20 000 000
38 400
33
1.38%
4.88
32 768 000
38 400
53
0.63%
4.88
40 000 000
38 400
65
0.16%
4.88
3 686 400
19 200
12
0.00%
9.77
20 000 000
19 200
65
0.16%
9.77
32 768 000
19 200
107
0.31%
9.77
40 000 000
19 200
130
0.16%
9.77
330
32058K
AVR32-01/12
AT32UC3A
Table 26-10. IrDA Baud Rate Error (Continued)
Peripheral Clock
26.7.6.3
Baud Rate
CD
Baud Rate Error
Pulse Time
3 686 400
9 600
24
0.00%
19.53
20 000 000
9 600
130
0.16%
19.53
32 768 000
9 600
213
0.16%
19.53
40 000 000
9 600
260
0.16%
19.53
3 686 400
2 400
96
0.00%
78.13
20 000 000
2 400
521
0.03%
78.13
32 768 000
2 400
853
0.04%
78.13
IrDA Demodulator
The demodulator is based on the IrDA Receive filter comprised of an 8-bit down counter which is
loaded with the value programmed in IFR. When a falling edge is detected on the RXD pin, the
Filter Counter starts counting down at the CLK_USART speed. If a rising edge is detected on the
RXD pin, the counter stops and is reloaded with IFR. If no rising edge is detected when the
counter reaches 0, the input of the receiver is driven low during one bit time.
Figure 26-35 on page 331 illustrates the operations of the IrDA demodulator.
Figure 26-35. IrDA Demodulator Operations
CLK_USART
RXD
Counter
Value
Receiver
Input
6
5
4
3
2
6
6
5
4
3
2
1
0
Pulse
Accepted
Pulse
Rejected
Driven Low During 16 Baud Rate Clock Cycles
As the IrDA mode uses the same logic as the ISO7816, note that the FI_DI_RATIO field in FIDI
must be set to a value higher than 0 in order to assure IrDA communications operate correctly.
331
32058K
AVR32-01/12
AT32UC3A
26.7.7
RS485 Mode
The USART features the RS485 mode to enable line driver control. While operating in RS485
mode, the USART behaves as though in asynchronous or synchronous mode and configuration
of all the parameters is possible. The difference is that the RTS pin is driven high when the
transmitter is operating. The behavior of the RTS pin is controlled by the TXEMPTY bit. A typical
connection of the USART to a RS485 bus is shown in Figure 26-36 on page 332.
Figure 26-36. Typical Connection to a RS485 Bus
USART
RXD
Differential
Bus
TXD
RTS
The USART is set in RS485 mode by programming the MODE field in the Mode Register (MR)
to the value 0x1.
The RTS pin is at a level inverse to the TXEMPTY bit. Significantly, the RTS pin remains high
when a timeguard is programmed so that the line can remain driven after the last character completion. Figure 26-37 on page 332 gives an example of the RTS waveform during a character
transmission when the timeguard is enabled.
Figure 26-37. Example of RTS Drive with Timeguard
TG = 4
Baud Rate
Clock
TXD
Start
D0
Bit
D1
D2
D3
D4
D5
D6
D7
Parity Stop
Bit Bit
Write
US_THR
TXRDY
TXEMPTY
RTS
332
32058K
AVR32-01/12
AT32UC3A
26.7.8
SPI Mode
The Serial Peripheral Interface (SPI) Mode is a synchronous serial data link that provides communication with external devices in Master or Slave Mode. It also enables communication
between processors if an external processor is connected to the system.
The Serial Peripheral Interface is essentially a shift register that serially transmits data bits to
other SPIs. During a data transfer, one SPI system acts as the “master” which controls the data
flow, while the other devices act as “slaves'' which have data shifted into and out by the master.
Different CPUs can take turns being masters and one master may simultaneously shift data into
multiple slaves. (Multiple Master Protocol is the opposite of Single Master Protocol, where one
CPU is always the master while all of the others are always slaves.) However, only one slave
may drive its output to write data back to the master at any given time.
A slave device is selected when its NSS signal is asserted by the master. The USART in SPI
Master mode can address only one SPI Slave because it can generate only one NSS signal.
The SPI system consists of two data lines and two control lines:
• Master Out Slave In (MOSI): This data line supplies the output data from the master shifted
into the input of the slave.
• Master In Slave Out (MISO): This data line supplies the output data from a slave to the input of
the master.
• Serial Clock (CLK): This control line is driven by the master and regulates the flow of the data
bits. The master may transmit data at a variety of baud rates. The CLK line cycles once for
each bit that is transmitted.
• Slave Select (NSS): This control line allows the master to select or deselect the slave.
26.7.8.1
Modes of Operation
The USART can operate in Master Mode or in Slave Mode.
Operation in SPI Master Mode is programmed by writing at 0xE the MODE field in the Mode
Register. In this case the SPI lines must be connected as described below:
• the MOSI line is driven by the output pin TXD
• the MISO line drives the input pin RXD
• the CLK line is driven by the output pin CLK
• the NSS line is driven by the output pin RTS
Operation in SPI Slave Mode is programmed by writing at 0xF the MODE field in the Mode Register. In this case the SPI lines must be connected as described below:
• the MOSI line drives the input pin RXD
• the MISO line is driven by the output pin TXD
• the CLK line drives the input pin CLK
• the NSS line drives the input pin CTS
In order to avoid unpredicted behavior, any change of the SPI Mode must be followed by a software reset of the transmitter and of the receiver (except the initial configuration after a hardware
reset).
333
32058K
AVR32-01/12
AT32UC3A
26.7.8.2
Baud Rate
In SPI Mode, the baudrate generator operates in the same way as in USART synchronous
mode: See Section “26.7.1.4” on page 307. However, there are some restrictions:
In SPI Master Mode:
• the external clock CLK must not be selected (USCLKS … 0x3), and the bit CLKO must be set
to “1” in the Mode Register (MR), in order to generate correctly the serial clock on the CLK pin.
• to obtain correct behavior of the receiver and the transmitter, the value programmed in CD of
must be superior or equal to 4.
• if the internal clock divided (CLK_USART/DIV) is selected, the value programmed in CD must
be even to ensure a 50:50 mark/space ratio on the CLK pin, this value can be odd if the
internal clock is selected (CLK_USART).
In SPI Slave Mode:
• the external clock (CLK) selection is forced regardless of the value of the USCLKS field in the
Mode Register (MR). Likewise, the value written in BRGR has no effect, because the clock is
provided directly by the signal on the USART CLK pin.
• to obtain correct behavior of the receiver and the transmitter, the external clock (CLK)
frequency must be at least 4 times lower than the system clock.
26.7.8.3
Data Transfer
Up to 9 data bits are successively shifted out on the TXD pin at each rising or falling edge
(depending of CPOL and CPHA) of the programmed serial clock. There is no Start bit, no Parity
bit and no Stop bit.
The number of data bits is selected by the CHRL field and the MODE 9 bit in the Mode Register
(MR). The 9 bits are selected by setting the MODE 9 bit regardless of the CHRL field. The MSB
data bit is always sent first in SPI Mode (Master or Slave).
Four combinations of polarity and phase are available for data transfers. The clock polarity is
programmed with the CPOL bit in the Mode Register. The clock phase is programmed with the
CPHA bit. These two parameters determine the edges of the clock signal upon which data is
driven and sampled. Each of the two parameters has two possible states, resulting in four possible combinations that are incompatible with one another. Thus, a master/slave pair must use the
same parameter pair values to communicate. If multiple slaves are used and fixed in different
configurations, the master must reconfigure itself each time it needs to communicate with a different slave.
Table 26-11. SPI Bus Protocol Mode
SPI Bus Protocol Mode
CPOL
CPHA
0
0
1
1
0
0
2
1
1
3
1
0
334
32058K
AVR32-01/12
AT32UC3A
Figure 26-38. SPI Transfer Format (CPHA=1, 8 bits per transfer)
CLK cycle (for reference)
2
1
4
3
5
7
6
8
CLK
(CPOL= 0)
CLK
(CPOL= 1)
MOSI
SPI Master ->TXD
SPI Slave ->RXD
MISO
SPI Master ->RXD
SPI Slave ->TXD
MSB
MSB
6
5
4
3
2
1
LSB
6
5
4
3
2
1
LSB
NSS
SPI Master ->RTS
SPI Slave ->CTS
Figure 26-39. SPI Transfer Format (CPHA=0, 8 bits per transfer)
CLK cycle (for reference)
1
2
3
4
5
6
7
8
CLK
(CPOL= 0)
CLK
(CPOL= 1)
MOSI
SPI Master -> TXD
SPI Slave -> RXD
MSB
6
5
4
3
2
1
LSB
MISO
SPI Master -> RXD
SPI Slave -> TXD
MSB
6
5
4
3
2
1
LSB
NSS
SPI Master -> RTS
SPI Slave -> CTS
26.7.8.4
Receiver and Transmitter Control
See Section “26.7.2” on page 309.
335
32058K
AVR32-01/12
AT32UC3A
26.7.8.5
Character Transmission
The characters are sent by writing in the Transmit Holding Register (THR). The transmitter
reports two status bits in the Channel Status Register (CSR): TXRDY (Transmitter Ready),
which indicates that THR is empty and TXEMPTY, which indicates that all the characters written
in THR have been processed. When the current character processing is completed, the last
character written in THR is transferred into the Shift Register of the transmitter and THR
becomes empty, thus TXRDY rises.
Both TXRDY and TXEMPTY bits are low when the transmitter is disabled. Writing a character in
THR while TXRDY is low has no effect and the written character is lost.
If the USART is in SPI Slave Mode and if a character must be sent while the Transmit Holding
Register (THR) is empty, the UNRE (Underrun Error) bit is set. The TXD transmission line stays
at high level during all this time. The UNRE bit is cleared by writing the Control Register (CR)
with the RSTSTA (Reset Status) bit at 1.
In SPI Master Mode, the slave select line (NSS) is asserted at low level 1 Tbit before the transmission of the MSB bit and released at high level 1 Tbit after the transmission of the LSB bit. So,
the slave select line (NSS) is always released between each character transmission and a minimum delay of 3 Tbits always inserted. However, in order to address slave devices supporting the
CSAAT mode (Chip Select Active After Transfer), the slave select line (NSS) can be forced at
low level by writing the Control Register (CR) with the RTSEN bit at 1. The slave select line
(NSS) can be released at high level only by writing the Control Register (CR) with the RTSDIS
bit at 1 (for example, when all data have been transferred to the slave device).
In SPI Slave Mode, the transmitter does not require a falling edge of the slave select line (NSS)
to initiate a character transmission but only a low level. However, this low level must be present
on the slave select line (NSS) at least 1 Tbit before the first serial clock cycle corresponding to
the MSB bit.
26.7.8.6
Character Reception
When a character reception is completed, it is transferred to the Receive Holding Register
(RHR) and the RXRDY bit in the Status Register (CSR) rises. If a character is completed while
RXRDY is set, the OVRE (Overrun Error) bit is set. The last character is transferred into RHR
and overwrites the previous one. The OVRE bit is cleared by writing the Control Register (CR)
with the RSTSTA (Reset Status) bit at 1.
To ensure correct behavior of the receiver in SPI Slave Mode, the master device sending the
frame must ensure a minimum delay of 1 Tbit between each character transmission. The
receiver does not require a falling edge of the slave select line (NSS) to initiate a character
reception but only a low level. However, this low level must be present on the slave select line
(NSS) at least 1 Tbit before the first serial clock cycle corresponding to the MSB bit.
26.7.8.7
Receiver Timeout
Because the receiver baudrate clock is active only during data transfers in SPI Mode, a receiver
timeout is impossible in this mode, whatever the Time-out value is (field TO) in the Time-out
Register (RTOR).
336
32058K
AVR32-01/12
AT32UC3A
26.7.9
Test Modes
The USART can be programmed to operate in three different test modes. The internal loopback
capability allows on-board diagnostics. In the loopback mode the USART interface pins are disconnected or not and reconfigured for loopback internally or externally.
26.7.9.1
Normal Mode
Normal mode connects the RXD pin on the receiver input and the transmitter output on the TXD
pin.
Figure 26-40. Normal Mode Configuration
RXD
Receiver
TXD
Transmitter
26.7.9.2
Automatic Echo Mode
Automatic echo mode allows bit-by-bit retransmission. When a bit is received on the RXD pin, it
is sent to the TXD pin, as shown in Figure 26-41 on page 337. Programming the transmitter has
no effect on the TXD pin. The RXD pin is still connected to the receiver input, thus the receiver
remains active.
Figure 26-41. Automatic Echo Mode Configuration
RXD
Receiver
TXD
Transmitter
26.7.9.3
Local Loopback Mode
Local loopback mode connects the output of the transmitter directly to the input of the receiver,
as shown in Figure 26-42 on page 337. The TXD and RXD pins are not used. The RXD pin has
no effect on the receiver and the TXD pin is continuously driven high, as in idle state.
Figure 26-42. Local Loopback Mode Configuration
RXD
Receiver
Transmitter
1
TXD
337
32058K
AVR32-01/12
AT32UC3A
26.7.9.4
Remote Loopback Mode
Remote loopback mode directly connects the RXD pin to the TXD pin, as shown in Figure 26-43
on page 338. The transmitter and the receiver are disabled and have no effect. This mode
allows bit-by-bit retransmission.
Figure 26-43. Remote Loopback Mode Configuration
Receiver
1
RXD
TXD
Transmitter
338
32058K
AVR32-01/12
AT32UC3A
26.8
Universal Synchronous/Asynchronous Receiver/Transmitter (USART) User Interface
26.8.1
Register Mapping
Table 26-12.
Register Mapping
Offset
Register
Name
Access
Reset
0x0000
Control Register
CR
Write-only
–
0x0004
Mode Register
MR
Read-write
–
0x0008
Interrupt Enable Register
IER
Write-only
–
0x000C
Interrupt Disable Register
IDR
Write-only
–
0x0010
Interrupt Mask Register
IMR
Read-only
0x0
0x0014
Channel Status Register
CSR
Read-only
–
0x0018
Receiver Holding Register
RHR
Read-only
0x0
0x001C
Transmitter Holding Register
THR
Write-only
–
0x0020
Baud Rate Generator Register
BRGR
Read-write
0x0
0x0024
Receiver Time-out Register
RTOR
Read-write
0x0
0x0028
Transmitter Timeguard Register
TTGR
Read-write
0x0
–
–
–
0x2C - 0x3C
Reserved
0x0040
FI DI Ratio Register
FIDI
Read-write
0x174
0x0044
Number of Errors Register
NER
Read-only
–
0x0048
Reserved
–
–
–
0x004C
IrDA Filter Register
IFR
Read-write
0x0
0x0050
Manchester Encoder Decoder Register
MAN
Read-write
0x30011004
–
–
–
VERSION
Read-only
0x–(3)
–
–
–
0x5C - 0xF8
0xFC
0x5C - 0xFC
Reserved
Version Register
Reserved
3. Values in the Version Register vary with the version of the IP block implementation.
339
32058K
AVR32-01/12
AT32UC3A
26.8.2
USART Control Register
Name:
CR
Access Type:
Write-only
Offset:
0x0
Reset Value:
-
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
RTSDIS/RCS
18
RTSEN/FCS
17
–
16
–
15
RETTO
14
RSTNACK
13
RSTIT
12
SENDA
11
STTTO
10
STPBRK
9
STTBRK
8
RSTSTA
7
TXDIS
6
TXEN
5
RXDIS
4
RXEN
3
RSTTX
2
RSTRX
1
–
0
–
• RTSDIS/RCS: Request to Send Disable/Release SPI Chip Select
– If USART does not operate in SPI Master Mode (MODE … 0xE):
0: No effect.
1: Drives the pin RTS to 1.
– If USART operates in SPI Master Mode (MODE = 0xE):
RCS = 0: No effect.
RCS = 1: Releases the Slave Select Line NSS (RTS pin).
• RTSEN/FCS: Request to Send Enable/Force SPI Chip Select
– If USART does not operate in SPI Master Mode (MODE … 0xE):
0: No effect.
1: Drives the pin RTS to 0.
– If USART operates in SPI Master Mode (MODE = 0xE):
FCS = 0: No effect.
FCS = 1: Forces the Slave Select Line NSS (RTS pin) to 0, even if USART is no transmitting, in order to address SPI slave
devices supporting the CSAAT Mode (Chip Select Active After Transfer).
• RETTO: Rearm Time-out
0: No effect
1: Restart Time-out
• RSTNACK: Reset Non Acknowledge
0: No effect
1: Resets NACK in CSR.
340
32058K
AVR32-01/12
AT32UC3A
• RSTIT: Reset Iterations
0: No effect.
1: Resets ITERATION in CSR. No effect if the ISO7816 is not enabled.
• SENDA: Send Address
0: No effect.
1: In Multidrop Mode only, the next character written to the THR is sent with the address bit set.
• STTTO: Start Time-out
0: No effect.
1: Starts waiting for a character before clocking the time-out counter. Resets the status bit TIMEOUT in CSR.
• STPBRK: Stop Break
0: No effect.
1: Stops transmission of the break after a minimum of one character length and transmits a high level during 12-bit periods.
No effect if no break is being transmitted.
• STTBRK: Start Break
0: No effect.
1: Starts transmission of a break after the characters present in THR and the Transmit Shift Register have been transmitted. No effect if a break is already being transmitted.
• RSTSTA: Reset Status Bits
0: No effect.
1: Resets the status bits PARE, FRAME, OVRE, MANERR and RXBRK in CSR.
• TXDIS: Transmitter Disable
0: No effect.
1: Disables the transmitter.
• TXEN: Transmitter Enable
0: No effect.
1: Enables the transmitter if TXDIS is 0.
• RXDIS: Receiver Disable
0: No effect.
1: Disables the receiver.
• RXEN: Receiver Enable
0: No effect.
1: Enables the receiver, if RXDIS is 0.
• RSTTX: Reset Transmitter
0: No effect.
341
32058K
AVR32-01/12
AT32UC3A
1: Resets the transmitter.
• RSTRX: Reset Receiver
0: No effect.
1: Resets the receiver.
342
32058K
AVR32-01/12
AT32UC3A
26.8.3
USART Mode Register
Name:
MR
Access Type:
Read-write
Offset:
0x4
Reset Value:
-
31
ONEBIT
30
MODSYNC
29
MAN
28
FILTER
27
–
26
25
MAX_ITERATION
24
23
–
22
VAR_SYNC
21
DSNACK
20
INACK
19
OVER
18
CLKO
17
MODE9
16
MSBF/CPOL
15
14
13
12
11
10
PAR
9
8
SYNC/CPHA
4
3
2
1
0
CHMODE
7
NBSTOP
6
CHRL
5
USCLKS
MODE
• ONEBIT: Start Frame Delimiter Selector
0: Start Frame delimiter is COMMAND or DATA SYNC.
1: Start Frame delimiter is One Bit.
• MODSYNC: Manchester Synchronization Mode
0:The Manchester Start bit is a 0 to 1 transition
1: The Manchester Start bit is a 1 to 0 transition.
• MAN: Manchester Encoder/Decoder Enable
0: Manchester Encoder/Decoder are disabled.
1: Manchester Encoder/Decoder are enabled.
• FILTER: Infrared Receive Line Filter
0: The USART does not filter the receive line.
1: The USART filters the receive line using a three-sample filter (1/16-bit clock) (2 over 3 majority).
• MAX_ITERATION
Defines the maximum number of iterations in mode ISO7816, protocol T= 0.
• VAR_SYNC: Variable Synchronization of Command/Data Sync Start Frame Delimiter
0: User defined configuration of command or data sync field depending on SYNC value.
1: The sync field is updated when a character is written into THR register.
• DSNACK: Disable Successive NACK
0: NACK is sent on the ISO line as soon as a parity error occurs in the received character (unless INACK is set).
343
32058K
AVR32-01/12
AT32UC3A
1: Successive parity errors are counted up to the value specified in the MAX_ITERATION field. These parity errors generate a NACK on the ISO line. As soon as this value is reached, no additional NACK is sent on the ISO line. The flag
ITERATION is asserted.
• INACK: Inhibit Non Acknowledge
0: The NACK is generated.
1: The NACK is not generated.
• OVER: Oversampling Mode
0: 16x Oversampling.
1: 8x Oversampling.
• CLKO: Clock Output Select
0: The USART does not drive the CLK pin.
1: The USART drives the CLK pin if USCLKS does not select the external clock CLK.
• MODE9: 9-bit Character Length
0: CHRL defines character length.
1: 9-bit character length.
• MSBF/CPOL: Bit Order or SPI Clock Polarity
– If USART does not operate in SPI Mode (MODE … 0xE and 0xF):
MSBF = 0: Least Significant Bit is sent/received first.
MSBF = 1: Most Significant Bit is sent/received first.
– If USART operates in SPI Mode (Slave or Master, MODE = 0xE or 0xF):
CPOL = 0: The inactive state value of SPCK is logic level zero.
CPOL = 1: The inactive state value of SPCK is logic level one.
CPOL is used to determine the inactive state value of the serial clock (SPCK). It is used with CPHA to produce the required
clock/data relationship between master and slave devices.
• CHMODE: Channel Mode
CHMODE
Mode Description
0
0
Normal Mode
0
1
Automatic Echo. Receiver input is connected to the TXD pin.
1
0
Local Loopback. Transmitter output is connected to the Receiver Input..
1
1
Remote Loopback. RXD pin is internally connected to the TXD pin.
• NBSTOP: Number of Stop Bits
NBSTOP
0
0
Asynchronous (SYNC = 0)
Synchronous (SYNC = 1)
1 stop bit
1 stop bit
344
32058K
AVR32-01/12
AT32UC3A
0
1
1.5 stop bits
Reserved
1
0
2 stop bits
2 stop bits
1
1
Reserved
Reserved
• PAR: Parity Type
PAR
Parity Type
0
0
0
Even parity
0
0
1
Odd parity
0
1
0
Parity forced to 0 (Space)
0
1
1
Parity forced to 1 (Mark)
1
0
x
No parity
1
1
x
Multidrop mode
• SYNC/CPHA: Synchronous Mode Select or SPI Clock Phase
– If USART does not operate in SPI Mode (MODE is … 0xE and 0xF):
SYNC = 0: USART operates in Asynchronous Mode.
SYNC = 1: USART operates in Synchronous Mode.
– If USART operates in SPI Mode (MODE = 0xE or 0xF):
CPHA = 0: Data is changed on the leading edge of SPCK and captured on the following edge of SPCK.
CPHA = 1: Data is captured on the leading edge of SPCK and changed on the following edge of SPCK.
CPHA determines which edge of SPCK causes data to change and which edge causes data to be captured. CPHA is used
with CPOL to produce the required clock/data relationship between master and slave devices.
• CHRL: Character Length.
CHRL
Character Length
0
0
5 bits
0
1
6 bits
1
0
7 bits
1
1
8 bits
• USCLKS: Clock Selection
USCLKS
Selected Clock
0
0
CLK_USART
0
1
CLK_USART/DIV (DIV = xx)
1
0
Reserved
1
1
CLK
345
32058K
AVR32-01/12
AT32UC3A
• MODE
MODE
Mode of the USART
0
0
0
0
Normal
0
0
0
1
RS485
0
0
1
0
Hardware Handshaking
0
1
0
0
IS07816 Protocol: T = 0
0
1
1
0
IS07816 Protocol: T = 1
1
0
0
0
IrDA
1
1
1
0
SPI Master
1
1
1
1
SPI Slave
Others
Reserved
346
32058K
AVR32-01/12
AT32UC3A
26.8.4
USART Interrupt Enable Register
Name:
IER
Access Type:
Write-only
Offset:
0x8
Reset Value:
-
31
–
30
–
29
–
28
–
27
–
26
25
24
MANEA
23
–
22
–
21
–
20
MANE
19
CTSIC
18
–
17
–
16
–
15
14
13
NACK
12
RXBUFF
11
TXBUFE
10
ITER/UNRE
9
TXEMPTY
8
TIMEOUT
7
PARE
6
FRAME
5
OVRE
4
ENDTX
3
ENDRX
2
RXBRK
1
TXRDY
0
RXRDY
• MANEA: Manchester Error Interrupt Enable
• MANE: Manchester Error Interrupt Enable
• CTSIC: Clear to Send Input Change Interrupt Enable
• NACK: Non Acknowledge Interrupt Enable
• RXBUFF: Buffer Full Interrupt Enable
• TXBUFE: Buffer Empty Interrupt Enable
• ITER/UNRE: Iteration or SPI Underrun Error Interrupt Enable
• TXEMPTY: TXEMPTY Interrupt Enable
• TIMEOUT: Time-out Interrupt Enable
• PARE: Parity Error Interrupt Enable
• FRAME: Framing Error Interrupt Enable
• OVRE: Overrun Error Interrupt Enable
• ENDTX: End of Transmit Interrupt Enable
• ENDRX: End of Receive Transfer Interrupt Enable
• RXBRK: Receiver Break Interrupt Enable
• TXRDY: TXRDY Interrupt Enable
• RXRDY: RXRDY Interrupt Enable
347
32058K
AVR32-01/12
AT32UC3A
26.8.5
USART Interrupt Disable Register
Name:
IDR
348
32058K
AVR32-01/12
AT32UC3A
Access Type:
Write-only
Offset:
0xC
Reset Value:
-
31
–
30
–
29
28
27
26
25
24
MANEA
23
–
22
–
21
–
20
MANE
19
CTSIC
18
–
17
–
16
–
15
14
13
NACK
12
RXBUFF
11
TXBUFE
10
ITER/UNRE
9
TXEMPTY
8
TIMEOUT
7
PARE
6
FRAME
5
OVRE
4
ENDTX
3
ENDRX
2
RXBRK
1
TXRDY
0
RXRDY
• MANEA: Manchester Error Interrupt Disable
• MANE: Manchester Error Interrupt Disable
• CTSIC: Clear to Send Input Change Interrupt Disable
• NACK: Non Acknowledge Interrupt Disable
• RXBUFF: Buffer Full Interrupt Disable
• TXBUFE: Buffer Empty Interrupt Disable
• ITER/UNRE: Iteration or SPI Underrun Error Interrupt Enable
• TXEMPTY: TXEMPTY Interrupt Disable
• TIMEOUT: Time-out Interrupt Disable
• PARE: Parity Error Interrupt Disable
• FRAME: Framing Error Interrupt Disable
• OVRE: Overrun Error Interrupt Disable
• ENDTX: End of Transmit Interrupt Disable
• ENDRX: End of Receive Transfer Interrupt Disable
• RXBRK: Receiver Break Interrupt Disable
• TXRDY: TXRDY Interrupt Disable
• RXRDY: RXRDY Interrupt Disable
349
32058K
AVR32-01/12
AT32UC3A
26.8.6
USART Interrupt Mask Register
Name:
IMR
Access Type:
Read-only
Offset:
0x10
Reset Value:
0x00000000
31
–
30
–
29
28
27
26
25
24
MANEA
23
–
22
–
21
–
20
MANE
19
CTSIC
18
–
17
–
16
–
15
14
13
NACK
12
RXBUFF
11
TXBUFE
10
ITER/UNRE
9
TXEMPTY
8
TIMEOUT
7
PARE
6
FRAME
5
OVRE
4
ENDTX
3
ENDRX
2
RXBRK
1
TXRDY
0
RXRDY
• MANEA: Manchester Error Interrupt Mask
• MANE: Manchester Error Interrupt Mask
• CTSIC: Clear to Send Input Change Interrupt Mask
• NACK: Non Acknowledge Interrupt Mask
• RXBUFF: Buffer Full Interrupt Mask
• TXBUFE: Buffer Empty Interrupt Mask
• ITER/UNRE: Iteration or SPI Underrun Error Interrupt Enable
• TXEMPTY: TXEMPTY Interrupt Mask
• TIMEOUT: Time-out Interrupt Mask
• PARE: Parity Error Interrupt Mask
• FRAME: Framing Error Interrupt Mask
• OVRE: Overrun Error Interrupt Mask
• ENDTX: End of Transmit Interrupt Mask
• ENDRX: End of Receive Transfer Interrupt Mask
• RXBRK: Receiver Break Interrupt Mask
• TXRDY: TXRDY Interrupt Mask
• RXRDY: RXRDY Interrupt Mask
350
32058K
AVR32-01/12
AT32UC3A
26.8.7
USART Channel Status Register
Name:
CSR
351
32058K
AVR32-01/12
AT32UC3A
Access Type:
Read-only
Offset:
0x14
Reset Value:
-
31
–
30
–
29
28
27
26
25
24
MANERR
23
CTS
22
–
21
–
20
–
19
CTSIC
18
–
17
–
16
–
15
14
13
NACK
12
RXBUFF
11
TXBUFE
10
ITER/UNRE
9
TXEMPTY
8
TIMEOUT
7
PARE
6
FRAME
5
OVRE
4
ENDTX
3
ENDRX
2
RXBRK
1
TXRDY
0
RXRDY
• MANERR: Manchester Error
0: No Manchester error has been detected since the last RSTSTA.
1: At least one Manchester error has been detected since the last RSTSTA.
• CTS: Image of CTS Input
0: CTS is at 0.
1: CTS is at 1.
• CTSIC: Clear to Send Input Change Flag
0: No input change has been detected on the CTS pin since the last read of CSR.
1: At least one input change has been detected on the CTS pin since the last read of CSR.
• NACK: Non Acknowledge
0: No Non Acknowledge has not been detected since the last RSTNACK.
1: At least one Non Acknowledge has been detected since the last RSTNACK.
• RXBUFF: Reception Buffer Full
0: The signal Buffer Full from the Receive PDC channel is inactive.
1: The signal Buffer Full from the Receive PDC channel is active.
• TXBUFE: Transmission Buffer Empty
0: The signal Buffer Empty from the Transmit PDC channel is inactive.
1: The signal Buffer Empty from the Transmit PDC channel is active.
• ITER/UNRE: Max number of Repetitions Reached or SPI Underrun Error
– If USART does not operate in SPI Slave Mode (MODE … 0xF):
ITER = 0: Maximum number of repetitions has not been reached since the last RSTSTA.
ITER = 1: Maximum number of repetitions has been reached since the last RSTSTA.
352
32058K
AVR32-01/12
AT32UC3A
– If USART operates in SPI Slave Mode (MODE = 0xF):
UNRE = 0: No SPI underrun error has occurred since the last RSTSTA.
UNRE = 1: At least one SPI underrun error has occurred since the last RSTSTA.
• TXEMPTY: Transmitter Empty
0: There are characters in either THR or the Transmit Shift Register, or the transmitter is disabled.
TXEMPTY == 1: Means that the Transmit Shift Register is empty and that there is no data in THR.
• TIMEOUT: Receiver Time-out
0: There has not been a time-out since the last Start Time-out command (STTTO in CR) or the Time-out Register is 0.
1: There has been a time-out since the last Start Time-out command (STTTO in CR).
• PARE: Parity Error
0: No parity error has been detected since the last RSTSTA.
1: At least one parity error has been detected since the last RSTSTA.
• FRAME: Framing Error
0: No stop bit has been detected low since the last RSTSTA.
1: At least one stop bit has been detected low since the last RSTSTA.
• OVRE: Overrun Error
0: No overrun error has occurred since the last RSTSTA.
1: At least one overrun error has occurred since the last RSTSTA.
• ENDTX: End of Transmitter Transfer
0: The End of Transfer signal from the Transmit PDC channel is inactive.
1: The End of Transfer signal from the Transmit PDC channel is active.
• ENDRX: End of Receiver Transfer
0: The End of Transfer signal from the Receive PDC channel is inactive.
1: The End of Transfer signal from the Receive PDC channel is active.
• RXBRK: Break Received/End of Break
0: No Break received or End of Break detected since the last RSTSTA.
1: Break Received or End of Break detected since the last RSTSTA.
• TXRDY: Transmitter Ready
0: A character is in the THR waiting to be transferred to the Transmit Shift Register, or an STTBRK command has been
requested, or the transmitter is disabled. As soon as the transmitter is enabled, TXRDY becomes 1.
1: There is no character in the THR.
• RXRDY: Receiver Ready
0: No complete character has been received since the last read of RHR or the receiver is disabled. If characters were being
received when the receiver was disabled, RXRDY changes to 1 when the receiver is enabled.
353
32058K
AVR32-01/12
AT32UC3A
1: At least one complete character has been received and RHR has not yet been read.
354
32058K
AVR32-01/12
AT32UC3A
26.8.8
USART Receive Holding Register
Name:
RHR
Access Type:
Read-only
Offset:
0x18
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
RXSYNH
14
–
13
–
12
–
11
–
10
–
9
–
8
RXCHR
7
6
5
4
3
2
1
0
RXCHR
• RXSYNH: Received Sync
0: Last Character received is a Data.
1: Last Character received is a Command.
• RXCHR: Received Character
Last character received if RXRDY is set.
355
32058K
AVR32-01/12
AT32UC3A
26.8.9
USART Transmit Holding Register
Name:
THR
Access Type:
Write-only
Offset:
0x1C
Reset Value:
-
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
TXSYNH
14
–
13
–
12
–
11
–
10
–
9
–
8
TXCHR
7
6
5
4
3
2
1
0
TXCHR
• TXSYNH: Sync Field to be transmitted
0: The next character sent is encoded as a data. Start Frame Delimiter is DATA SYNC.
1: The next character sent is encoded as a command. Start Frame Delimiter is COMMAND SYNC.
• TXCHR: Character to be Transmitted
Next character to be transmitted after the current character if TXRDY is not set.
356
32058K
AVR32-01/12
AT32UC3A
26.8.10
USART Baud Rate Generator Register
Name:
BRGR
Access Type:
Read-write
Offset:
0x20
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
17
FP–
16
15
14
13
12
11
10
9
8
3
2
1
0
CD
7
6
5
4
CD
• FP: Fractional Part
0: Fractional divider is disabled.
1 - 7: Baudrate resolution, defined by FP x 1/8.
• CD: Clock Divider
MODE ≠ ISO7816
SYNC = 1
or
MODE = SPI
(Master or Slave)
SYNC = 0
CD
OVER = 0
0
1 to 65535
MODE = ISO7816
OVER = 1
Baud Rate Clock Disabled
Baud Rate =
Selected Clock/16/CD
Baud Rate =
Selected Clock/8/CD
Baud Rate =
Selected Clock /CD
Baud Rate = Selected
Clock/CD/FI_DI_RATIO
357
32058K
AVR32-01/12
AT32UC3A
26.8.11
USART Receiver Time-out Register
Name:
RTOR
Access Type:
Read-write
Offset:
0x24
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
–
–
–
–
–
–
–
–
TO
7
6
5
4
TO
• TO: Time-out Value
0: The Receiver Time-out is disabled.
1 - 65535: The Receiver Time-out is enabled and the Time-out delay is TO x Bit Period.
358
32058K
AVR32-01/12
AT32UC3A
26.8.12
USART Transmitter Timeguard Register
Name:
TTGR
Access Type:
Read-write
Offset:
0x28
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
TG
• TG: Timeguard Value
0: The Transmitter Timeguard is disabled.
1 - 255: The Transmitter timeguard is enabled and the timeguard delay is TG x Bit Period.
359
32058K
AVR32-01/12
AT32UC3A
26.8.13
USART FI DI RATIO Register
Name:
FIDI
Access Type:
Read-write
Offset:
0x40
Reset Value:
0x00000174
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
9
FI_DI_RATIO
8
7
6
5
4
3
2
1
0
FI_DI_RATIO
• FI_DI_RATIO: FI Over DI Ratio Value
0: If ISO7816 mode is selected, the Baud Rate Generator generates no signal.
1 - 2047: If ISO7816 mode is selected, the Baud Rate is the clock provided on CLK divided by FI_DI_RATIO.
360
32058K
AVR32-01/12
AT32UC3A
26.8.14
USART Number of Errors Register
Name:
NER
Access Type:
Read-only
Offset:
0x44
Reset Value:
-
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
NB_ERRORS
• NB_ERRORS: Number of Errors
Total number of errors that occurred during an ISO7816 transfer. This register automatically clears when read.
361
32058K
AVR32-01/12
AT32UC3A
26.8.15
USART IrDA FILTER Register
Name:
IFR
Access Type:
Read-write
Offset:
0x4C
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
IRDA_FILTER
• IRDA_FILTER: IrDA Filter
Sets the filter of the IrDA demodulator.
362
32058K
AVR32-01/12
AT32UC3A
26.8.16
USART Manchester Configuration Register
Name:
MAN
Access Type:
Read-write
Offset:
0x50
Reset Value:
0x30011004
31
–
30
DRIFT
29
1
28
RX_MPOL
27
–
26
–
23
–
22
–
21
–
20
–
19
18
15
–
14
–
13
–
12
TX_MPOL
11
–
10
–
7
–
6
–
5
–
4
–
3
2
25
24
RX_PP
17
16
RX_PL
9
8
TX_PP
1
0
TX_PL
• DRIFT: Drift compensation
0: The USART can not recover from an important clock drift
1: The USART can recover from clock drift. The 16X clock mode must be enabled.
• RX_MPOL: Receiver Manchester Polarity
0: Logic Zero is coded as a zero-to-one transition, Logic One is coded as a one-to-zero transition.
1: Logic Zero is coded as a one-to-zero transition, Logic One is coded as a zero-to-one transition.
• RX_PP: Receiver Preamble Pattern detected
RX_PP
Preamble Pattern default polarity assumed (RX_MPOL field not set)
0
0
ALL_ONE
0
1
ALL_ZERO
1
0
ZERO_ONE
1
1
ONE_ZERO
• RX_PL: Receiver Preamble Length
0: The receiver preamble pattern detection is disabled
1 - 15: The detected preamble length is RX_PL x Bit Period
• TX_MPOL: Transmitter Manchester Polarity
0: Logic Zero is coded as a zero-to-one transition, Logic One is coded as a one-to-zero transition.
1: Logic Zero is coded as a one-to-zero transition, Logic One is coded as a zero-to-one transition.
363
32058K
AVR32-01/12
AT32UC3A
• TX_PP: Transmitter Preamble Pattern
TX_PP
Preamble Pattern default polarity assumed (TX_MPOL field not set)
0
0
ALL_ONE
0
1
ALL_ZERO
1
0
ZERO_ONE
1
1
ONE_ZERO
• TX_PL: Transmitter Preamble Length
0: The Transmitter Preamble pattern generation is disabled
1 - 15: The Preamble Length is TX_PL x Bit Period
364
32058K
AVR32-01/12
AT32UC3A
26.8.17
USART Version Register
Name:
VERSION
Access Type:
Read-only
Offset:
0xFC
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
17
VARIANT
16
15
–
14
–
13
–
12
–
11
10
9
8
7
6
5
4
1
0
VERSION
3
2
VERSION
• VARIANT
Reserved. No functionality associated.
• VERSION
Version of the module. No functionality associated.
365
32058K
AVR32-01/12
AT32UC3A
27. Static Memory Controller (SMC)
Rev. 1.0.0.0
27.1
Features
•
•
•
•
•
•
•
•
•
•
•
•
27.2
4 Chip Selects Available
64-Mbyte Address Space per Chip Select
8-, 16- or 32-bit Data Bus
Word, Halfword, Byte Transfers
Byte Write or Byte Select Lines
Programmable Setup, Pulse And Hold Time for Read Signals per Chip Select
Programmable Setup, Pulse And Hold Time for Write Signals per Chip Select
Programmable Data Float Time per Chip Select
Compliant with LCD Module
External Wait Request
Automatic Switch to Slow Clock Mode
Asynchronous Read in Page Mode Supported: Page Size Ranges from 4 to 32 Bytes
Overview
The Static Memory Controller (SMC) generates the signals that control the access to the external memory devices or peripheral devices. It has 4 Chip Selects and a 26-bit address bus. The
32-bit data bus can be configured to interface with 8-, or16-, or 32-bit external devices. Separate
read and write control signals allow for direct memory and peripheral interfacing. Read and write
signal waveforms are fully parametrizable.
The SMC can manage wait requests from external devices to extend the current access. The
SMC is provided with an automatic slow clock mode. In slow clock mode, it switches from userprogrammed waveforms to slow-rate specific waveforms on read and write signals. The SMC
supports asynchronous burst read in page mode access for page size up to 32 bytes.
366
32058K
AVR32-01/12
AT32UC3A
27.3
Block Diagram
Figure 27-1. Block Diagram
NCS[5:0]
Bus
Matrix
NRD
SMC
Chip Select
NWR0/NWE
A0/NBS0
SMC
NWR1/NBS1
A1/NWR2/NBS2
PM
CLK_SMC
NWR3/NBS3
GPIO
Controller
NCS[5:0]
NRD
NWR0/NWE
A0/NBS0
NWR1/NBS1
A1/NWR2/NBS2
NWR3/NBS3
A[25:2]
D[31:0]
NWAIT
A[25:2]
D[31:0]
NWAIT
User Interface
Peripheral Bus
27.4
I/O Lines Description
Table 27-1.
I/O Line Description
Name
Description
Type
Active Level
NCS[3:0]
Static Memory Controller Chip Select Lines
Output
Low
NRD
Read Signal
Output
Low
NWR0/NWE
Write 0/Write Enable Signal
Output
Low
A0/NBS0
Address Bit 0/Byte 0 Select Signal
Output
Low
NWR1/NBS1
Write 1/Byte 1 Select Signal
Output
Low
A1/NWR2/NBS2
Address Bit 1/Write 2/Byte 2 Select Signal
Output
Low
NWR3/NBS3
Write 3/Byte 3 Select Signal
Output
Low
A[25:2]
Address Bus
Output
D[31:0]
Data Bus
NWAIT
External Wait Signal
I/O
Input
Low
367
32058K
AVR32-01/12
AT32UC3A
27.5
Product Dependencies
27.5.1
EBI I/O Lines
The Static Memory Controller signals pass througth the EBI module where they are multiplexed.
The programmer must first configure the GPIO controller to assign the EBI pins corresponding to
SMC signals to their peripheral function. If I/O lines of the EBI corresponding to SMC
signals are not used by the application, they can be used for other purposes by the GPIO
Controller.
27.6
Functionnal Description
27.6.1
27.6.1.1
Application Example
Hardware Interface
Figure 27-2. SMC Connections to Static Memory Devices
D0-D31
A0/NBS0
NWR0/NWE
NWR1/NBS1
A1/NWR2/NBS2
NWR3/NBS3
D0-D7
128K x 8
SRAM
D0-D7
NRD
NWR0/NWE
D16-D23
A2-A25
NWR1/NBS1
WE
128K x 8
SRAM
D24-D31
D0-D7
Static Memory
Controller
27.6.2
OE
WE
A0-A16
NRD
OE
A2-A18
OE
WE
128K x 8
SRAM
D0-D7
CS
A0-A16
NRD
D0-D7
A2-A18
CS
A1/NWR2/NBS2
128K x 8
SRAM
CS
CS
A0-A16
NCS0
NCS1
NCS2
NCS3
NCS4
NCS5
D8-D15
A2-A18
A0-A16
A2-A18
NRD
OE
NWR3/NBS3 WE
External Memory Mapping
The SMC provides up to 26 address lines, A[25:0]. This allows each chip select line to address
up to 64 Mbytes of memory.
368
32058K
AVR32-01/12
AT32UC3A
If the physical memory device connected on one chip select is smaller than 64 Mbytes, it wraps
around and appears to be repeated within this space. The SMC correctly handles any valid
access to the memory device within the page (see Figure 27-3).
A[25:0] is only significant for 8-bit memory, A[25:1] is used for 16-bit memory, A[25:2] is used for
32-bit memory.
Memory Connections for 6 External Devices
Figure 27-3.
NCS[0] - NCS[5]
NRD
SMC
NWE
NCS5
A[25:0]
NCS4
D[31:0]
NCS3
NCS2
NCS1
NCS0
Memory Enable
Memory Enable
Memory Enable
Memory Enable
Memory Enable
Memory Enable
Output Enable
Write Enable
8 or 16 or 32
27.6.3
27.6.3.1
A[25:0]
D[31:0] or D[15:0] or
D[7:0]
Connection to External Devices
Data Bus Width
A data bus width of 8, 16 or 32 bits can be selected for each chip select. This option is controlled
by the field DBW in MODE (Mode Register) for the corresponding chip select.
Figure 27-4 shows how to connect a 512K x 8-bit memory on NCS2. Figure 27-5 shows how to
connect a 512K x 16-bit memory on NCS2. Figure 27-6 shows two 16-bit memories connected
as a single 32-bit memory.
27.6.3.2
Byte Write or Byte Select Access
Each chip select with a 16-bit or 32-bit data bus can operate with one of two different types of
write access: byte write or byte select access. This is controlled by the BAT field of the MODE
register for the corresponding chip select.
369
32058K
AVR32-01/12
AT32UC3A
Figure 27-4.
Memory Connection for an 8-bit Data Bus
D[7:0]
D[7:0]
A[18:2]
A[18:2]
SMC
A0
A0
A1
A1
NWE
Write Enable
NRD
Output Enable
NCS[2]
Figure 27-5.
Memory Enable
Memory Connection for a 16-bit Data Bus
D[15:0]
D[15:0]
A[19:2]
A[18:1]
A1
SMC
A[0]
NBS0
Low Byte Enable
NBS1
High Byte Enable
NWE
Write Enable
NRD
Output Enable
NCS[2]
Memory Enable
Figure 27-6. Memory Connection for a 32-bit Data Bus
D[31:16]
SMC
D[31:16]
D[15:0]
D[15:0]
A[20:2]
A[18:0]
NBS0
Byte 0 Enable
NBS1
Byte 1 Enable
NBS2
Byte 2 Enable
NBS3
Byte 3 Enable
NWE
Write Enable
NRD
Output Enable
NCS[2]
Memory Enable
370
32058K
AVR32-01/12
AT32UC3A
– Byte Write Access
Byte write access supports one byte write signal per byte of the data bus and a single read
signal.
Note that the SMC does not allow boot in Byte Write Access mode.
• For 16-bit devices: the SMC provides NWR0 and NWR1 write signals for respectively byte0
(lower byte) and byte1 (upper byte) of a 16-bit bus. One single read signal (NRD) is provided.
Byte Write Access is used to connect 2 x 8-bit devices as a 16-bit memory.
• For 32-bit devices: NWR0, NWR1, NWR2 and NWR3, are the write signals of byte0 (lower
byte), byte1, byte2 and byte 3 (upper byte) respectively. One single read signal (NRD) is
provided.
Byte Write Access is used to connect 4 x 8-bit devices as a 32-bit memory.
Byte Write option is illustrated on Figure 27-7.
– Byte Select Access
In this mode, read/write operations can be enabled/disabled at a byte level. One byte-select line
per byte of the data bus is provided. One NRD and one NWE signal control read and write.
• For 16-bit devices: the SMC provides NBS0 and NBS1 selection signals for respectively byte0
(lower byte) and byte1 (upper byte) of a 16-bit bus.
Byte Select Access is used to connect one 16-bit device.
• For 32-bit devices: NBS0, NBS1, NBS2 and NBS3, are the selection signals of byte0 (lower
byte), byte1, byte2 and byte 3 (upper byte) respectively. Byte Select Access is used to connect
two 16-bit devices.
Figure 27-8 shows how to connect two 16-bit devices on a 32-bit data bus in Byte Select Access
mode, on NCS3 (BAT = Byte Select Access).
371
32058K
AVR32-01/12
AT32UC3A
Figure 27-7.
Connection of 2 x 8-bit Devices on a 16-bit Bus: Byte Write Option
D[7:0]
D[7:0]
D[15:8]
A[24:2]
SMC
A1
NWR0
A[23:1]
A[0]
Write Enable
NWR1
NRD
NCS[3]
Read Enable
Memory Enable
D[15:8]
A[23:1]
A[0]
Write Enable
Read Enable
Memory Enable
– Signal Multiplexing
Depending on the BAT, only the write signals or the byte select signals are used. To save IOs at
the external bus interface, control signals at the SMC interface are multiplexed.
For 32-bit devices, bits A0 and A1 are unused. For 16-bit devices, bit A0 of address is unused.
When Byte Select Option is selected, NWR1 to NWR3 are unused. When Byte Write option is
selected, NBS0 to NBS3 are unused.
372
32058K
AVR32-01/12
AT32UC3A
Figure 27-8. Connection of 2x16-bit Data Bus on a 32-bit Data Bus (Byte Select Option)
D[15:0]
D[15:0]
D[31:16]
A[25:2]
A[23:0]
NWE
Write Enable
NBS0
Low Byte Enable
NBS1
High Byte Enable
NBS2
SMC
NBS3
Read Enable
NRD
Memory Enable
NCS[3]
D[31:16]
A[23:0]
Write Enable
Low Byte Enable
High Byte Enable
Read Enable
Memory Enable
Table 27-2.
SMC Multiplexed Signal Translation
Signal Name
Device Type
32-bit Bus
16-bit Bus
8-bit Bus
1x32-bit
2x16-bit
4 x 8-bit
1x16-bit
2 x 8-bit
Byte Select
Byte Select
Byte Write
Byte Select
Byte Write
NBS0_A0
NBS0
NBS0
NWE_NWR0
NWE
NWE
NWR0
NWE
NWR0
NBS1_NWR1
NBS1
NBS1
NWR1
NBS1
NWR1
NBS2_NWR2_A1
NBS2
NBS2
NWR2
A1
A1
NBS3_NWR3
NBS3
NBS3
NWR3
Byte Access Type (BAT)
27.6.4
NBS0
1 x 8-bit
A0
NWE
A1
Standard Read and Write Protocols
In the following sections, the byte access type is not considered. Byte select lines (NBS0 to
NBS3) always have the same timing as the A address bus. NWE represents either the NWE signal in byte select access type or one of the byte write lines (NWR0 to NWR3) in byte write
access type. NWR0 to NWR3 have the same timings and protocol as NWE. In the same way,
NCS represents one of the NCS[0..3] chip select lines.
373
32058K
AVR32-01/12
AT32UC3A
27.6.4.1
Read Waveforms
The read cycle is shown on Figure 27-9.
The read cycle starts with the address setting on the memory address bus, i.e.:
{A[25:2], A1, A0} for 8-bit devices
{A[25:2], A1} for 16-bit devices
A[25:2] for 32-bit devices.
Figure 27-9. Standard Read Cycle
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NRD
NCS
D[15:0]
NRD_SETUP
NCS_RD_SETUP
NRD_PULSE
NCS_RD_PULSE
NRD_HOLD
NCS_RD_HOLD
NRD_CYCLE
– NRD Waveform
The NRD signal is characterized by a setup timing, a pulse width and a hold timing.
1. NRD_SETUP: the NRD setup time is defined as the setup of address before the NRD
falling edge;
2. NRD_PULSE: the NRD pulse length is the time between NRD falling edge and NRD rising edge;
3. NRD_HOLD: the NRD hold time is defined as the hold time of address after the NRD rising edge.
374
32058K
AVR32-01/12
AT32UC3A
– NCS Waveform
Similarly, the NCS signal can be divided into a setup time, pulse length and hold time:
1. NCS_RD_SETUP: the NCS setup time is defined as the setup time of address before the
NCS falling edge.
2. NCS_RD_PULSE: the NCS pulse length is the time between NCS falling edge and NCS
rising edge;
3. NCS_RD_HOLD: the NCS hold time is defined as the hold time of address after the NCS
rising edge.
– Read Cycle
The NRD_CYCLE time is defined as the total duration of the read cycle, i.e., from the time where
address is set on the address bus to the point where address may change. The total read cycle
time is equal to:
NRD_CYCLE = NRD_SETUP + NRD_PULSE + NRD_HOLD
= NCS_RD_SETUP + NCS_RD_PULSE + NCS_RD_HOLD
All NRD and NCS timings are defined separately for each chip select as an integer number of
Master Clock cycles. To ensure that the NRD and NCS timings are coherent, user must define
the total read cycle instead of the hold timing. NRD_CYCLE implicitly defines the NRD hold time
and NCS hold time as:
NRD_HOLD = NRD_CYCLE - NRD SETUP - NRD PULSE
NCS_RD_HOLD = NRD_CYCLE - NCS_RD_SETUP - NCS_RD_PULSE
– Null Delay Setup and Hold
If null setup and hold parameters are programmed for NRD and/or NCS, NRD and NCS remain
active continuously in case of consecutive read cycles in the same memory (see Figure 27-10).
375
32058K
AVR32-01/12
AT32UC3A
Figure 27-10. No Setup, No Hold On NRD and NCS Read Signals
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NRD
NCS
D[15:0]
NRD_SETUP
NCS_RD_PULSE
NRD_CYCLE
NRD_PULSE
NCS_RD_PULSE
NRD_CYCLE
NRD_PULSE
NCS_RD_PULSE
NRD_CYCLE
– Null Pulse
Programming null pulse is not permitted. Pulse must be at least set to 1. A null value leads to
unpredictable behavior.
27.6.4.2
Read Mode
As NCS and NRD waveforms are defined independently of one other, the SMC needs to know
when the read data is available on the data bus. The SMC does not compare NCS and NRD timings to know which signal rises first. The READ_MODE parameter in the MODE register of the
corresponding chip select indicates which signal of NRD and NCS controls the read operation.
– Read is Controlled by NRD (READ_MODE = 1):
Figure 27-11 shows the waveforms of a read operation of a typical asynchronous RAM. The
read data is available tPACC after the falling edge of NRD, and turns to ‘Z’ after the rising edge of
NRD. In this case, the READ_MODE must be set to 1 (read is controlled by NRD), to indicate
that data is available with the rising edge of NRD. The SMC samples the read data internally on
the rising edge of Master Clock that generates the rising edge of NRD, whatever the programmed waveform of NCS may be.
376
32058K
AVR32-01/12
AT32UC3A
Figure 27-11. READ_MODE = 1: Data is sampled by SMC before the rising edge of NRD
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NRD
NCS
tPACC
D[15:0]
Data Sampling
– Read is Controlled by NCS (READ_MODE = 0)
Figure 27-12 shows the typical read cycle of an LCD module. The read data is valid tPACC after
the falling edge of the NCS signal and remains valid until the rising edge of NCS. Data must be
sampled when NCS is raised. In that case, the READ_MODE must be set to 0 (read is controlled
by NCS): the SMC internally samples the data on the rising edge of Master Clock that generates
the rising edge of NCS, whatever the programmed waveform of NRD may be.
377
32058K
AVR32-01/12
AT32UC3A
Figure 27-12. READ_MODE = 0: Data is sampled by SMC before the rising edge of NCS
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NRD
NCS
tPACC
D[15:0]
Data Sampling
27.6.4.3
Write Waveforms
The write protocol is similar to the read protocol. It is depicted in Figure 27-13. The write cycle
starts with the address setting on the memory address bus.
– NWE Waveforms
The NWE signal is characterized by a setup timing, a pulse width and a hold timing.
1. NWE_SETUP: the NWE setup time is defined as the setup of address and data before
the NWE falling edge;
2. NWE_PULSE: The NWE pulse length is the time between NWE falling edge and NWE
rising edge;
3. NWE_HOLD: The NWE hold time is defined as the hold time of address and data after
the NWE rising edge.
The NWE waveforms apply to all byte-write lines in Byte Write access mode: NWR0 to NWR3.
27.6.4.4
NCS Waveforms
The NCS signal waveforms in write operation are not the same that those applied in read operations, but are separately defined:
1. NCS_WR_SETUP: the NCS setup time is defined as the setup time of address before
the NCS falling edge.
2. NCS_WR_PULSE: the NCS pulse length is the time between NCS falling edge and NCS
rising edge;
3. NCS_WR_HOLD: the NCS hold time is defined as the hold time of address after the
NCS rising edge.
378
32058K
AVR32-01/12
AT32UC3A
Figure 27-13. Write Cycle
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NWE
NCS
NWE_SETUP
NCS_WR_SETUP
NWE_PULSE
NCS_WR_PULSE
NWE_HOLD
NCS_WR_HOLD
NWE_CYCLE
– Write Cycle
The write_cycle time is defined as the total duration of the write cycle, that is, from the time
where address is set on the address bus to the point where address may change. The total write
cycle time is equal to:
NWE_CYCLE = NWE_SETUP + NWE_PULSE + NWE_HOLD
= NCS_WR_SETUP + NCS_WR_PULSE + NCS_WR_HOLD
All NWE and NCS (write) timings are defined separately for each chip select as an integer number of Master Clock cycles. To ensure that the NWE and NCS timings are coherent, the user
must define the total write cycle instead of the hold timing. This implicitly defines the NWE hold
time and NCS (write) hold times as:
NWE_HOLD = NWE_CYCLE - NWE_SETUP - NWE_PULSE
NCS_WR_HOLD = NWE_CYCLE - NCS_WR_SETUP - NCS_WR_PULSE
– Null Delay Setup and Hold
If null setup parameters are programmed for NWE and/or NCS, NWE and/or NCS remain active
continuously in case of consecutive write cycles in the same memory (see Figure 27-14). However, for devices that perform write operations on the rising edge of NWE or NCS, such as
SRAM, either a setup or a hold must be programmed.
379
32058K
AVR32-01/12
AT32UC3A
Figure 27-14. Null Setup and Hold Values of NCS and NWE in Write Cycle
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NWE,
NWE0, NWE1
NCS
D[15:0]
NWE_SETUP
NCS_WR_SETUP
NWE_CYCLE
NWE_PULSE
NWE_PULSE
NCS_WR_PULSE
NCS_WR_PULSE
NWE_CYCLE
NWE_CYCLE
– Null Pulse
Programming null pulse is not permitted. Pulse must be at least set to 1. A null value leads to
unpredictable behavior.
27.6.4.5
Write Mode
The WRITE_MODE parameter in the MODE register of the corresponding chip select indicates
which signal controls the write operation.
– Write is Controlled by NWE (WRITE_MODE = 1):
Figure 27-15 shows the waveforms of a write operation with WRITE_MODE set to 1. The data is
put on the bus during the pulse and hold steps of the NWE signal. The internal data buffers are
turned out after the NWE_SETUP time, and until the end of the write cycle, regardless of the
programmed waveform on NCS.
380
32058K
AVR32-01/12
AT32UC3A
Figure 27-15. WRITE_MODE = 1. The write operation is controlled by NWE
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NWE,
NWR0, NWR1
NCS
D[15:0]
– Write is Controlled by NCS (WRITE_MODE = 0)
Figure 27-16 shows the waveforms of a write operation with WRITE_MODE set to 0. The data is
put on the bus during the pulse and hold steps of the NCS signal. The internal data buffers are
turned out after the NCS_WR_SETUP time, and until the end of the write cycle, regardless of
the programmed waveform on NWE.
Figure 27-16. WRITE_MODE = 0. The write operation is controlled by NCS
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NWE,
NWR0, NWR1
NCS
D[15:0]
27.6.4.6
Coding Timing Parameters
All timing parameters are defined for one chip select and are grouped together in one REGISTER according to their type.
381
32058K
AVR32-01/12
AT32UC3A
The SETUP register groups the definition of all setup parameters:
• NRD_SETUP, NCS_RD_SETUP, NWE_SETUP, NCS_WR_SETUP
The PULSE register groups the definition of all pulse parameters:
• NRD_PULSE, ncs_rd_pULSE, nwe_pULSE, ncs_wr_pULSE
The CYCLE register groups the definition of all cycle parameters:
• NRD_CYCLE, NWE_CYCLEe
Table 27-3 shows how the timing parameters are coded and their permitted range.
Coding and Range of Timing Parameters
Table 27-3.
Permitted Range
Coded Value
Number of Bits
Effective Value
Coded Value
Effective Value
setup [5:0]
6
128 x setup[5] + setup[4:0]
0 ≤ ≤ 31
128 ≤ ≤ 128+31
pulse [6:0]
7
256 x pulse[6] + pulse[5:0]
0 ≤ ≤ 63
256 ≤ ≤ 256+63
0 ≤ ≤ 127
256 ≤ ≤ 256+127
512 ≤ ≤ 512+127
768 ≤ ≤ 768+127
cycle [8:0]
27.6.4.7
9
256 x cycle[8:7] + cycle[6:0]
Usage Restriction
The SMC does not check the validity of the user-programmed parameters. If the sum of
SETUP and PULSE parameters is larger than the corresponding CYCLE parameter, this
leads to unpredictable behavior of the SMC.
For read operations:
Null but positive setup and hold of address and NRD and/or NCS can not be guaranteed at the
memory interface because of the propagation delay of theses signals through external logic and
pads. If positive setup and hold values must be verified, then it is strictly recommended to program non-null values so as to cover possible skews between address, NCS and NRD signals.
For write operations:
If a null hold value is programmed on NWE, the SMC can guarantee a positive hold of address,
byte select lines, and NCS signal after the rising edge of NWE. This is true for WRITE_MODE =
1 only. See ”Early Read Wait State” on page 385.
For read and write operations: a null value for pulse parameters is forbidden and may lead to
unpredictable behavior.
In read and write cycles, the setup and hold time parameters are defined in reference to the
address bus. For external devices that require setup and hold time between NCS and NRD signals (read), or between NCS and NWE signals (write), these setup and hold times must be
converted into setup and hold times in reference to the address bus.
27.6.5
Automatic Wait States
Under certain circumstances, the SMC automatically inserts idle cycles between accesses to
avoid bus contention or operation conflict.
382
32058K
AVR32-01/12
AT32UC3A
27.6.5.1
Chip Select Wait States
The SMC always inserts an idle cycle between 2 transfers on separate chip selects. This idle
cycle ensures that there is no bus contention between the de-activation of one device and the
activation of the next one.
During chip select wait state, all control lines are turned inactive: NBS0 to NBS3, NWR0 to
NWR3, NCS[0..5], NRD lines are all set to 1.
Figure 27-17 illustrates a chip select wait state between access on Chip Select 0 and Chip
Select 2.
383
32058K
AVR32-01/12
AT32UC3A
Figure 27-17. Chip Select Wait State between a Read Access on NCS0 and a Write Access on
NCS2
CLK_SMC
A[25:2]
S0, NBS1,
A0, A1
NRD
NWE
NCS0
NCS2
NWE_CYCLE
NRD_CYCLE
D[15:0]
Read to Write
Wait State
Chip Select
Wait State
384
32058K
AVR32-01/12
AT32UC3A
27.6.5.2
Early Read Wait State
In some cases, the SMC inserts a wait state cycle between a write access and a read access to
allow time for the write cycle to end before the subsequent read cycle begins. This wait state is
not generated in addition to a chip select wait state. The early read cycle thus only occurs
between a write and read access to the same memory device (same chip select).
• An early read wait state is automatically inserted if at least one of the following conditions is
valid:
• if the write controlling signal has no hold time and the read controlling signal has no setup time
(Figure 27-18).
• in NCS write controlled mode (WRITE_MODE = 0), if there is no hold timing on the NCS signal
and the NCS_RD_SETUP parameter is set to 0, regardless of the read mode (Figure 27-19).
The write operation must end with a NCS rising edge. Without an Early Read Wait State, the
write operation could not complete properly.
• in NWE controlled mode (WRITE_MODE = 1) and if there is no hold timing (NWE_HOLD = 0),
the feedback of the write control signal is used to control address, data, chip select and byte
select lines. If the external write control signal is not inactivated as expected due to load
capacitances, an Early Read Wait State is inserted and address, data and control signals are
maintained one more cycle. See Figure 27-20.
Figure 27-18. Early Read Wait State: Write with No Hold Followed by Read with No Setup.
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NWE
NRD
No hold
No setup
D[15:0]
Write cycle
Early Read
Wait state
Read cycle
385
32058K
AVR32-01/12
AT32UC3A
Figure 27-19. Early Read Wait State: NCS Controlled Write with No Hold Followed by a Read
with No Setup.
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NWE
NRD
No hold
No setup
D[15:0]
Write cycle
(WRITE_MODE=0)
Read cycle
Early Read
Wait state (READ_MODE=0 or READ_MODE=1)
386
32058K
AVR32-01/12
AT32UC3A
Figure 27-20. Early Read Wait State: NWE-controlled Write with No Hold Followed by a Read
with one Set-up Cycle.
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
Internal write controlling signal
external write controlling
signal(NWE)
No hold
Read setup=1
NRD
D[15:0]
Write cycle
(WRITE_MODE = 1)
27.6.5.3
Early Read
Wait state
Read cycle
(READ_MODE=0 or READ_MODE=1)
Reload User Configuration Wait State
The user may change any of the configuration parameters by writing the SMC user interface.
When detecting that a new user configuration has been written in the user interface, the SMC
inserts a wait state before starting the next access. The so called “Reload User Configuration
Wait State” is used by the SMC to load the new set of parameters to apply to next accesses.
The Reload Configuration Wait State is not applied in addition to the Chip Select Wait State. If
accesses before and after re-programming the user interface are made to different devices
(Chip Selects), then one single Chip Select Wait State is applied.
On the other hand, if accesses before and after writing the user interface are made to the same
device, a Reload Configuration Wait State is inserted, even if the change does not concern the
current Chip Select.
– User Procedure
To insert a Reload Configuration Wait State, the SMC detects a write access to any MODE register of the user interface. If the user only modifies timing registers (SETUP, PULSE, CYCLE
registers) in the user interface, he must validate the modification by writing the MODE, even if no
change was made on the mode parameters.
– Slow Clock Mode Transition
A Reload Configuration Wait State is also inserted when the Slow Clock Mode is entered or
exited, after the end of the current transfer (see ”Slow Clock Mode” on page 398).
387
32058K
AVR32-01/12
AT32UC3A
27.6.5.4
Read to Write Wait State
Due to an internal mechanism, a wait cycle is always inserted between consecutive read and
write SMC accesses.
This wait cycle is referred to as a read to write wait state in this document.
This wait cycle is applied in addition to chip select and reload user configuration wait states
when they are to be inserted. See Figure 27-17 on page 384.
27.6.6
Data Float Wait States
Some memory devices are slow to release the external bus. For such devices, it is necessary to
add wait states (data float wait states) after a read access:
• before starting a read access to a different external memory
• before starting a write access to the same device or to a different external one.
The Data Float Output Time (t DF ) for each external memory device is programmed in the
TDF_CYCLES field of the MODE register for the corresponding chip select. The value of
TDF_CYCLES indicates the number of data float wait cycles (between 0 and 15) before the
external device releases the bus, and represents the time allowed for the data output to go to
high impedance after the memory is disabled.
Data float wait states do not delay internal memory accesses. Hence, a single access to an
external memory with long t DF will not slow down the execution of a program from internal
memory.
The data float wait states management depends on the READ_MODE and the TDF_MODE
fields of the MODE register for the corresponding chip select.
27.6.6.1
READ_MODE
Setting the READ_MODE to 1 indicates to the SMC that the NRD signal is responsible for turning off the tri-state buffers of the external memory device. The Data Float Period then begins
after the rising edge of the NRD signal and lasts TDF_CYCLES MCK cycles.
When the read operation is controlled by the NCS signal (READ_MODE = 0), the TDF field gives
the number of MCK cycles during which the data bus remains busy after the rising edge of NCS.
Figure 27-21 illustrates the Data Float Period in NRD-controlled mode (READ_MODE =1),
assuming a data float period of 2 cycles (TDF_CYCLES = 2). Figure 27-22 shows the read operation when controlled by NCS (READ_MODE = 0) and the TDF_CYCLES parameter equals 3.
388
32058K
AVR32-01/12
AT32UC3A
Figure 27-21. TDF Period in NRD Controlled Read Access (TDF = 2)
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NRD
NCS
D[15:0]
tPACC
TDF = 2 clock cycles
NRD controlled read operation
Figure 27-22. TDF Period in NCS Controlled Read Operation (TDF = 3)
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NRD
NCS
D[15:0]
tPACC
TDF = 3 clock cycles
NCS controlled read operation
389
32058K
AVR32-01/12
AT32UC3A
27.6.6.2
TDF Optimization Enabled (TDF_MODE = 1)
When the TDF_MODE of the MODE register is set to 1 (TDF optimization is enabled), the SMC
takes advantage of the setup period of the next access to optimize the number of wait states
cycle to insert.
Figure 27-23 shows a read access controlled by NRD, followed by a write access controlled by
NWE, on Chip Select 0. Chip Select 0 has been programmed with:
NRD_HOLD = 4; READ_MODE = 1 (NRD controlled)
NWE_SETUP = 3; WRITE_MODE = 1 (NWE controlled)
TDF_CYCLES = 6; TDF_MODE = 1 (optimization enabled).
Figure 27-23. TDF Optimization: No TDF wait states are inserted if the TDF period is over when the next access begins
CLK_SMC
A[25:2]
NRD
NRD_HOLD = 4
NWE
NWE_SETUP = 3
NCS0
TDF_CYCLES = 6
D[15:0]
Read access on NCS0 (NRD controlled)
27.6.6.3
Read to Write
Wait State
write access on NCS0 (NWE controlled)
TDF Optimization Disabled (TDF_MODE = 0)
When optimization is disabled, tdf wait states are inserted at the end of the read transfer, so that
the data float period is ended when the second access begins. If the hold period of the read1
controlling signal overlaps the data float period, no additional tdf wait states will be inserted.
Figure 27-24, Figure 27-25 and Figure 27-26 illustrate the cases:
• read access followed by a read access on another chip select,
• read access followed by a write access on another chip select,
390
32058K
AVR32-01/12
AT32UC3A
• read access followed by a write access on the same chip select,
with no TDF optimization.
Figure 27-24. TDF Optimization Disabled (TDF Mode = 0). TDF wait states between 2 read accesses on different chip
selects.
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
Read1 controlling
signal(NRD)
Read1 hold = 1
Read2 controlling
signal(NRD)
Read2 setup = 1
TDF_CYCLES = 6
D[15:0]
5 TDF WAIT STATES
Read 2 cycle
TDF_MODE=0
(optimization disabled)
Read1 cycle
TDF_CYCLES = 6
Chip Select Wait State
Figure 27-25. TDF Mode = 0: TDF wait states between a read and a write access on different chip selects.
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
Read1 controlling
signal(NRD)
Read1 hold = 1
Write2 controlling
signal(NWE)
Write2 setup = 1
TDF_CYCLES = 4
D[15:0]
Read1 cycle
TDF_CYCLES = 4
2 TDF WAIT STATES
Read to Write Chip Select
Wait State Wait State
Write 2 cycle
TDF_MODE=0
(optimization disabled)
391
32058K
AVR32-01/12
AT32UC3A
Figure 27-26. TDF Mode = 0: TDF wait states between read and write accesses on the same chip select.
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
Read1 controlling
signal(NRD)
Write2 setup = 1
Read1 hold = 1
Write2 controlling
signal(NWE)
TDF_CYCLES = 5
D[15:0]
4 TDF WAIT STATES
Read1 cycle
TDF_CYCLES = 5
27.6.7
Read to Write
Wait State
Write 2 cycle
TDF_MODE=0
(optimization disabled)
External Wait
Any access can be extended by an external device using the NWAIT input signal of the SMC.
The EXNW_MODE field of the MODE register on the corresponding chip select must be set to
either to “10” (frozen mode) or “11” (ready mode). When the EXNW_MODE is set to “00” (disabled), the NWAIT signal is simply ignored on the corresponding chip select. The NWAIT signal
delays the read or write operation in regards to the read or write controlling signal, depending on
the read and write modes of the corresponding chip select.
27.6.7.1
Restriction
When one of the EXNW_MODE is enabled, it is mandatory to program at least one hold
cycle for the read/write controlling signal. For that reason, the NWAIT signal cannot be
used in Page Mode (”Asynchronous Page Mode” on page 400), or in Slow Clock Mode
(”Slow Clock Mode” on page 398).
The NWAIT signal is assumed to be a response of the external device to the read/write
request of the SMC. Then NWAIT is examined by the SMC only in the pulse state of the
read or write controlling signal. The assertion of the NWAIT signal outside the expected
period has no impact on SMC behavior.
392
32058K
AVR32-01/12
AT32UC3A
27.6.7.2
Frozen Mode
When the external device asserts the NWAIT signal (active low), and after internal synchronization of this signal, the SMC state is frozen, i.e., SMC internal counters are frozen, and all control
signals remain unchanged. When the resynchronized NWAIT signal is deasserted, the SMC
completes the access, resuming the access from the point where it was stopped. See Figure 2727. This mode must be selected when the external device uses the NWAIT signal to delay the
access and to freeze the SMC.
The assertion of the NWAIT signal outside the expected period is ignored as illustrated in Figure
27-28.
Figure 27-27. Write Access with NWAIT Assertion in Frozen Mode (EXNW_MODE = 10).
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
FROZEN STATE
4
3
2
1
1
1
1
3
2
2
2
2
0
NWE
NCS
6
5
4
0
D[15:0]
NWAIT
Internally synchronized
NWAIT signal
Write cycle
EXNW_MODE = 10 (Frozen)
WRITE_MODE = 1 (NWE_controlled)
NWE_PULSE = 5
NCS_WR_PULSE = 7
393
32058K
AVR32-01/12
AT32UC3A
Figure 27-28. Read Access with NWAIT Assertion in Frozen Mode (EXNW_MODE = 10).
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
FROZEN STATE
NCS
NRD
4
3
1
0
2
2
5
5
2
5
1
0
4
3
2
1
0
2
1
0
NWAIT
Internally synchronized
NWAIT signal
Read cycle
EXNW_MODE = 10 (Frozen)
READ_MODE = 0 (NCS_controlled)
NRD_PULSE = 2, NRD_HOLD = 6
NCS_RD_PULSE = 5, NCS_RD_HOLD = 3
Assertion is ignored
394
32058K
AVR32-01/12
AT32UC3A
27.6.7.3
Ready Mode
In Ready mode (EXNW_MODE = 11), the SMC behaves differently. Normally, the SMC begins
the access by down counting the setup and pulse counters of the read/write controlling signal. In
the last cycle of the pulse phase, the resynchronized NWAIT signal is examined.
If asserted, the SMC suspends the access as shown in Figure 27-29 and Figure 27-30. After
deassertion, the access is completed: the hold step of the access is performed.
This mode must be selected when the external device uses deassertion of the NWAIT signal to
indicate its ability to complete the read or write operation.
If the NWAIT signal is deasserted before the end of the pulse, or asserted after the end of the
pulse of the controlling read/write signal, it has no impact on the access length as shown in Figure 27-30.
Figure 27-29. NWAIT Assertion in Write Access: Ready Mode (EXNW_MODE = 11).
CLK _SM C
A [2 5 :2 ]
NBS0, NBS1,
A0, A1
FRO ZEN STATE
4
3
2
1
0
2
1
0
0
NW E
NCS
6
5
4
3
1
1
0
D [1 5 :0 ]
N W A IT
In te rn a lly s y n c h ro n iz e d
N W A IT s ig n a l
W rite c y c le
E X N W _ M O D E = 1 1 (R e a d y m o d e )
W R IT E _ M O D E = 1 (N W E _ c o n tro lle d )
N W E_PU LSE = 5
N C S _W R _PU LSE = 7
395
32058K
AVR32-01/12
AT32UC3A
Figure 27-30. NWAIT Assertion in Read Access: Ready Mode (EXNW_MODE = 11).
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
Wait STATE
NCS
NRD
6
5
4
3
2
6
5
4
3
1
2
0
0
1
1
0
NWAIT
Internally synchronized
NWAIT signal
Read cycle
Assertion is ignored
EXNW_MODE = 11 (Ready mode)
READ_MODE = 0 (NCS_controlled)
NRD_PULSE = 7
NCS_RD_PULSE = 7
Assertion is ignored
396
32058K
AVR32-01/12
AT32UC3A
27.6.7.4
NWAIT Latency and Read/write Timings
There may be a latency between the assertion of the read/write controlling signal and the assertion of the NWAIT signal by the device. The programmed pulse length of the read/write
controlling signal must be at least equal to this latency plus the 2 cycles of resynchronization + 1
cycle. Otherwise, the SMC may enter the hold state of the access without detecting the NWAIT
signal assertion. This is true in frozen mode as well as in ready mode. This is illustrated on Figure 27-31.
When EXNW_MODE is enabled (ready or frozen), the user must program a pulse length of the
read and write controlling signal of at least:
minimal pulse length = NWAIT latency + 2 resynchronization cycles + 1 cycle
Figure 27-31. NWAIT Latency
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
Wait STATE
4
3
2
1
0
0
0
NRD
Minimal pulse length
NWAIT
nternally synchronized
NWAIT signal
NWAIT latency 2 cycle resynchronization
Read cycle
EXNW_MODE = 10 or 11
READ_MODE = 1 (NRD_controlled)
NRD_PULSE = 5
397
32058K
AVR32-01/12
AT32UC3A
27.6.8
Slow Clock Mode
The SMC is able to automatically apply a set of “slow clock mode” read/write waveforms when
an internal signal driven by the Power Management Controller is asserted because CLK_SMC
has been turned to a very slow clock rate (typically 32kHz clock rate). In this mode, the user-programmed waveforms are ignored and the slow clock mode waveforms are applied. This mode is
provided so as to avoid reprogramming the User Interface with appropriate waveforms at very
slow clock rate. When activated, the slow mode is active on all chip selects.
27.6.8.1
Slow Clock Mode Waveforms
Figure 27-32 illustrates the read and write operations in slow clock mode. They are valid on all
chip selects. indicates the value of read and write parameters in slow clock mode.
Figure 27-32. Read/write Cycles in Slow Clock Mode
CLK_SMC
CLK_SMC
A[25:2]
A[25:2]
NBS0, NBS1,
A0, A1
NBS0, NBS1,
A0, A1
NWE
1
NRD
1
1
1
1
NCS
NCS
NRD_CYCLES = 2
NWE_CYCLES = 3
SLOW CLOCK MODE READ
SLOW CLOCK MODE WRITE
Table 27-4.
Read and Write Timing Parameters in Slow Clock Mode
Read Parameters
Duration (cycles)
Write Parameters
Duration (cycles)
NRD_SETUP
1
NWE_SETUP
1
NRD_PULSE
1
NWE_PULSE
1
NCS_RD_SETUP
0
NCS_WR_SETUP
0
NCS_RD_PULSE
2
NCS_WR_PULSE
3
NRD_CYCLE
2
NWE_CYCLE
3
398
32058K
AVR32-01/12
AT32UC3A
27.6.8.2
Switching from (to) Slow Clock Mode to (from) Normal Mode
When switching from slow clock mode to the normal mode, the current slow clock mode transfer
is completed at high clock rate, with the set of slow clock mode parameters.See Figure 27-33 on
page 399. The external device may not be fast enough to support such timings.
Figure 27-34 illustrates the recommended procedure to properly switch from one mode to the
other.
Figure 27-33. Clock Rate Transition Occurs while the SMC is Performing a Write Operation
Slow Clock Mode
Internal signal from PM
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NWE
1
1
1
1
1
1
2
3
2
NCS
NWE_CYCLE = 3
NWE_CYCLE = 7
SLOW CLOCK MODE WRITE SLOW CLOCK MODE WRITE
This write cycle finishes with the slow clock mode set
Of parameters after the clock rate transition
NORMAL MODE WRITE
Slow clock mode transition is detected:
Reload Configuration Wait State
399
32058K
AVR32-01/12
AT32UC3A
Figure 27-34. Recommended Procedure to Switch from Slow Clock Mode to Normal Mode or from Normal Mode to Slow
Clock Mode
Slow Clock Mode
Internal signal from PM
CLK_SMC
A[25:2]
NBS0, NBS1,
A0, A1
NWE
1
1
1
2
3
2
NCS
SLOW CLOCK MODE WRITE
IDLE STATE
NORMAL MODE WRITE
Reload Configuration
Wait State
27.6.9
Asynchronous Page Mode
The SMC supports asynchronous burst reads in page mode, providing that the page mode is
enabled in the MODE register (PMEN field). The page size must be configured in the MODE
register (PS field) to 4, 8, 16 or 32 bytes.
The page defines a set of consecutive bytes into memory. A 4-byte page (resp. 8-, 16-, 32-byte
page) is always aligned to 4-byte boundaries (resp. 8-, 16-, 32-byte boundaries) of memory. The
MSB of data address defines the address of the page in memory, the LSB of address define the
address of the data in the page as detailed in Table 27-5.
With page mode memory devices, the first access to one page (tpa) takes longer than the subsequent accesses to the page (t sa ) as shown in Figure 27-35. When in page mode, the SMC
enables the user to define different read timings for the first access within one page, and next
accesses within the page.
Table 27-5.
Page Size
Page Address(1)
Data Address in the Page(2)
4 bytes
A[25:2]
A[1:0]
8 bytes
A[25:3]
A[2:0]
16 bytes
A[25:4]
A[3:0]
32 bytes
A[25:5]
A[4:0]
Notes:
27.6.9.1
Page Address and Data Address within a Page
1. A denotes the address bus of the memory device
2. For 16-bit devices, the bit 0 of address is ignored. For 32-bit devices, bits [1:0] are ignored.
Protocol and Timings in Page Mode
Figure 27-35 shows the NRD and NCS timings in page mode access.
400
32058K
AVR32-01/12
AT32UC3A
Figure 27-35. Page Mode Read Protocol (Address MSB and LSB are defined in Table 27-5)
CLK_SMC
A[MSB]
A[LSB]
NRD
tpa
tsa
NCS
tsa
D[15:0]
NCS_RD_PULSE
NRD_PULSE
NRD_PULSE
The NRD and NCS signals are held low during all read transfers, whatever the programmed values of the setup and hold timings in the User Interface may be. Moreover, the NRD and NCS
timings are identical. The pulse length of the first access to the page is defined with the
NCS_RD_PULSE field of the PULSE register. The pulse length of subsequent accesses within
the page are defined using the NRD_PULSE parameter.
In page mode, the programming of the read timings is described in Table 27-6:
Table 27-6.
Programming of Read Timings in Page Mode
Parameter
Value
Definition
READ_MODE
‘x’
No impact
NCS_RD_SETUP
‘x’
No impact
NCS_RD_PULSE
tpa
Access time of first access to the page
NRD_SETUP
‘x’
No impact
NRD_PULSE
tsa
Access time of subsequent accesses in the page
NRD_CYCLE
‘x’
No impact
The SMC does not check the coherency of timings. It will always apply the NCS_RD_PULSE
timings as page access timing (tpa) and the NRD_PULSE for accesses to the page (tsa), even if
the programmed value for tpa is shorter than the programmed value for tsa.
27.6.9.2
Byte Access Type in Page Mode
The Byte Access Type configuration remains active in page mode. For 16-bit or 32-bit page
mode devices that require byte selection signals, configure the BAT field of the REGISTER to 0
(byte select access type).
401
32058K
AVR32-01/12
AT32UC3A
27.6.9.3
Page Mode Restriction
The page mode is not compatible with the use of the NWAIT signal. Using the page mode and
the NWAIT signal may lead to unpredictable behavior.
27.6.9.4
Sequential and Non-sequential Accesses
If the chip select and the MSB of addresses as defined in Table 27-5 are identical, then the current access lies in the same page as the previous one, and no page break occurs.
Using this information, all data within the same page, sequential or not sequential, are accessed
with a minimum access time (tsa). Figure 27-36 illustrates access to an 8-bit memory device in
page mode, with 8-byte pages. Access to D1 causes a page access with a long access time
(tpa). Accesses to D3 and D7, though they are not sequential accesses, only require a short
access time (tsa).
If the MSB of addresses are different, the SMC performs the access of a new page. In the same
way, if the chip select is different from the previous access, a page break occurs. If two sequential accesses are made to the page mode memory, but separated by an other internal or external
peripheral access, a page break occurs on the second access because the chip select of the
device was deasserted between both accesses.
Figure 27-36. Access to Non-sequential Data within the Same Page
CLK_SMC
Page address
A[25:3]
A[2], A1, A0
A1
A3
A7
NRD
NCS
D[7:0]
D1
NCS_RD_PULSE
D3
NRD_PULSE
D7
NRD_PULSE
402
32058K
AVR32-01/12
AT32UC3A
27.7
User Interface
The SMC is programmed using the registers listed in Table 27-7. For each chip select, a set of 4 registers is used to program the parameters of the external device connected on it. In Table 27-7, “CS_number” denotes the chip select number.
16 bytes (0x10) are required per chip select.
The user must complete writing the configuration by writing any one of the MODE registers.
SMC Register Mapping
Table 27-7.
Offset
Register
Name
Access
Reset State
0x10 x CS_number + 0x00
SMC Setup Register
SETUP
Read/Write
–
0x10 x CS_number + 0x04
SMC Pulse Register
PULSE
Read/Write
–
0x10 x CS_number + 0x08
SMC Cycle Register
CYCLE
Read/Write
–
0x10 x CS_number + 0x0C
SMC Mode Register
MODE
Read/Write
–
403
32058K
AVR32-01/12
AT32UC3A
27.7.1
Setup Register
Register Name:
SETUP[0 ..3]
Access Type:
Read/Write
Offset:
0x10 x CS_number + 0x00
Reset Value:
–
31
30
–
–
23
22
–
–
15
14
–
–
7
6
–
–
29
28
27
26
25
24
18
17
16
10
9
8
1
0
NCS_RD_SETUP
21
20
19
NRD_SETUP
13
12
11
NCS_WR_SETUP
5
4
3
2
NWE_SETUP
• NCS_RD_SETUP: NCS Setup Length in READ Access
In read access, the NCS signal setup length is defined as:
NCS setup length = (128* NCS_RD_SETUP[5] + NCS_RD_SETUP[4:0]) clock cycles
• NRD_SETUP: NRD Setup Length
The NRD signal setup length is defined in clock cycles as:
NRD setup length = (128* NRD_SETUP[5] + NRD_SETUP[4:0]) clock cycles
• NCS_WR_SETUP: NCS Setup Length in WRITE Access
In write access, the NCS signal setup length is defined as:
NCS setup length = (128* NCS_WR_SETUP[5] + NCS_WR_SETUP[4:0]) clock cycles
• NWE_SETUP: NWE Setup Length
The NWE signal setup length is defined as:
NWE setup length = (128* NWE_SETUP[5] + NWE_SETUP[4:0]) clock cycles
404
32058K
AVR32-01/12
AT32UC3A
27.7.2
Pulse Register
Register Name:
PULSE[0..3]
Access Type:
Read/Write
Offset:
0x10 x CS_number + 0x04
Reset Value:
–
31
30
29
28
–
23
22
21
20
–
15
26
25
24
19
18
17
16
10
9
8
2
1
0
NRD_PULSE
14
13
12
–
7
27
NCS_RD_PULSE
11
NCS_WR_PULSE
6
5
4
–
3
NWE_PULSE
• NCS_RD_PULSE: NCS Pulse Length in READ Access
In standard read access, the NCS signal pulse length is defined as:
NCS pulse length = (256* NCS_RD_PULSE[6] + NCS_RD_PULSE[5:0]) clock cycles
The NCS pulse length must be at least 1 clock cycle.
In page mode read access, the NCS_RD_PULSE parameter defines the duration of the first access to one page.
• NRD_PULSE: NRD Pulse Length
In standard read access, the NRD signal pulse length is defined in clock cycles as:
NRD pulse length = (256* NRD_PULSE[6] + NRD_PULSE[5:0]) clock cycles
The NRD pulse length must be at least 1 clock cycle.
In page mode read access, the NRD_PULSE parameter defines the duration of the subsequent accesses in the page.
• NCS_WR_PULSE: NCS Pulse Length in WRITE Access
In write access, the NCS signal pulse length is defined as:
NCS pulse length = (256* NCS_WR_PULSE[6] + NCS_WR_PULSE[5:0]) clock cycles
The NCS pulse length must be at least 1 clock cycle.
• NWE_PULSE: NWE Pulse Length
The NWE signal pulse length is defined as:
NWE pulse length = (256* NWE_PULSE[6] + NWE_PULSE[5:0]) clock cycles
The NWE pulse length must be at least 1 clock cycle.
405
32058K
AVR32-01/12
AT32UC3A
27.7.3
Cycle Register
Register Name:
CYCLE[0..3]
Access Type:
Read/Write
Offset:
0x10 x CS_number + 0x08
Reset Value:
–
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
NRD_CYCLE
23
22
21
20
19
18
17
16
NRD_CYCLE
15
14
13
12
11
10
9
8
–
–
–
–
–
–
–
NWE_CYCLE
7
6
5
4
3
2
1
0
NWE_CYCLE
• NRD_CYCLE: Total Read Cycle Length
The total read cycle length is the total duration in clock cycles of the read cycle. It is equal to the sum of the setup, pulse
and hold steps of the NRD and NCS signals. It is defined as:
Read cycle length = (NRD_CYCLE[8:7]*256 + NRD_CYCLE[6:0]) clock cycles
• NWE_CYCLE: Total Write Cycle Length
The total write cycle length is the total duration in clock cycles of the write cycle. It is equal to the sum of the setup, pulse
and hold steps of the NWE and NCS signals. It is defined as:
Write cycle length = (NWE_CYCLE[8:7]*256 + NWE_CYCLE[6:0]) clock cycles
406
32058K
AVR32-01/12
AT32UC3A
27.7.4
MODE Register
Register Name:
MODE[0..3]
Access Type:
Read/Write
Offset:
0x10 x CS_number + 0x0C
Reset Value:
–
31
30
–
–
23
22
21
20
–
–
–
TDF_MODE
15
14
13
–
–
7
6
–
29
28
PS
12
DBW
5
–
4
EXNW_MODE
27
26
25
24
–
–
–
PMEN
19
18
17
16
TDF_CYCLES
11
10
9
8
–
–
–
BAT
3
2
1
0
–
WRITE_MOD
E
READ_MODE
–
• PS: Page Size
If page mode is enabled, this field indicates the size of the page in bytes.
Page size settings.
Table 27-8.
PS
Page Size
0
0
4-byte page
0
1
8-byte page
1
0
16-byte page
1
1
32-byte page
• PMEN: Page Mode Enabled
1: Asynchronous burst read in page mode is applied on the corresponding chip select.
0: Standard read is applied.
• TDF_MODE: TDF Optimization
1: TDF optimization is enabled.
– The number of TDF wait states is optimized using the setup period of the next read/write access.
0: TDF optimization is disabled.
– The number of TDF wait states is inserted before the next access begins.
• TDF_CYCLES: Data Float Time
This field gives the integer number of clock cycles required by the external device to release the data after the rising edge
of the read controlling signal. The SMC always provide one full cycle of bus turnaround after the TDF_CYCLES period. The
407
32058K
AVR32-01/12
AT32UC3A
external bus cannot be used by another chip select during TDF_CYCLES + 1 cycles. From 0 up to 15 TDF_CYCLES can
be set.
• Data Bus Width (DBW)
DBW
Data Bus Width
0
0
8-bit bus
0
1
16-bit bus
1
0
32-bit bus
1
1
Reserved
• BAT: Byte Access Type
This field is used only if DBW defines a 16- or 32-bit data bus.
1: Byte write access type:
– Write operation is controlled using NCS, NWR0, NWR1, NWR2, NWR3.
– Read operation is controlled using NCS and NRD.
0: Byte select access type:
– Write operation is controlled using NCS, NWE, NBS0, NBS1, NBS2 and NBS3
– Read operation is controlled using NCS, NRD, NBS0, NBS1, NBS2 and NBS3
• EXNW_MODE: NWAIT Mode
The NWAIT signal is used to extend the current read or write signal. It is only taken into account during the pulse phase of
the read and write controlling signal. When the use of NWAIT is enabled, at least one cycle hold duration must be programmed for the read and write controlling signal.
EXNW_MODE
Table 27-9.
EXNW_MODE
NWAIT Mode
0
0
Disabled
0
1
Reserved
1
0
Frozen Mode
1
1
Ready Mode
• Disabled Mode: The NWAIT input signal is ignored on the corresponding Chip Select.
• Frozen Mode: If asserted, the NWAIT signal freezes the current read or write cycle. After deassertion, the read/write
cycle is resumed from the point where it was stopped.
• Ready Mode: The NWAIT signal indicates the availability of the external device at the end of the pulse of the controlling
read or write signal, to complete the access. If high, the access normally completes. If low, the access is extended until
NWAIT returns high.
• WRITE_MODE
1: The write operation is controlled by the NWE signal.
– If TDF optimization is enabled (TDF_MODE =1), TDF wait states will be inserted after the setup of NWE.
0: The write operation is controlled by the NCS signal.
– If TDF optimization is enabled (TDF_MODE =1), TDF wait states will be inserted after the setup of NCS.
• READ_MODE:
1: The read operation is controlled by the NRD signal.
408
32058K
AVR32-01/12
AT32UC3A
– If TDF cycles are programmed, the external bus is marked busy after the rising edge of NRD.
– If TDF optimization is enabled (TDF_MODE =1), TDF wait states are inserted after the setup of NRD.
0: The read operation is controlled by the NCS signal.
– If TDF cycles are programmed, the external bus is marked busy after the rising edge of NCS.
– If TDF optimization is enabled (TDF_MODE =1), TDF wait states are inserted after the setup of NCS.
409
32058K
AVR32-01/12
AT32UC3A
28. SDRAM Controller (SDRAMC)
Rev: 2.0.1.1
28.1
Features
• Numerous Configurations Supported
•
•
•
•
•
•
28.2
– 2K, 4K, 8K Row Address Memory Parts
– SDRAM with Two or Four Internal Banks
– SDRAM with 16- or 32-bit Data Path
Programming Facilities
– Word, Half-word, Byte Access
– Automatic Page Break When Memory Boundary Has Been Reached
– Multibank Ping-pong Access
– Timing Parameters Specified by Software
– Automatic Refresh Operation, Refresh Rate is Programmable
– Automatic Update of DS, TCR and PASR Parameters (Mobile SDRAM Devices)
Energy-saving Capabilities
– Self-refresh, Power-down and Deep Power Modes Supported
– Supports Mobile SDRAM Devices
Error Detection
– Refresh Error Interrupt
SDRAM Power-up Initialization by Software
CAS Latency of 1, 2, 3 Supported
Auto Precharge Command Not Used
Description
The SDRAM Controller (SDRAMC) extends the memory capabilities of a chip by providing the
interface to an external 16-bit or 32-bit SDRAM device. The page size supports ranges from
2048 to 8192 and the number of columns from 256 to 2048. It supports byte (8-bit), half-word
(16-bit) and word (32-bit) accesses.
The SDRAM Controller supports a read or write burst length of one location. It keeps track of the
active row in each bank, thus maximizing SDRAM performance, e.g., the application may be
placed in one bank and data in the other banks. So as to optimize performance, it is advisable to
avoid accessing different rows in the same bank.
The SDRAM controller supports a CAS latency of 1, 2 or 3 and optimizes the read access
depending on the frequency.
The different modes available - self-refresh, power-down and deep power-down modes - minimize power consumption on the SDRAM device.
410
32058K
AVR32-01/12
AT32UC3A
28.3
Block Diagram
Figure 28-1. SDRAM Controller Block Diagram
PIO
Controller
SDRAMC
Chip Select
SDCK
Memory
Controller
SDCKE
SDCS
SDRAMC
Interrupt
BA[1:0]
RAS
PMC
SDRAMC
MCK
CAS
SDWE
NBS[3:0]
SDRAMC_A[12:0]
D[31:0]
User Interface
Peripheral Bus
28.4
I/O Lines Description
Table 28-1.
I/O Line Description
Name
Description
Type
Active Level
SDCK
SDRAM Clock
Output
SDCKE
SDRAM Clock Enable
Output
High
SDCS
SDRAM Controller Chip Select
Output
Low
BA[1:0]
Bank Select Signals
Output
RAS
Row Signal
Output
Low
CAS
Column Signal
Output
Low
SDWE
SDRAM Write Enable
Output
Low
NBS[3:0]
Data Mask Enable Signals
Output
Low
SDRAMC_A[12:0]
Address Bus
Output
D[31:0]
Data Bus
I/O
411
32058K
AVR32-01/12
AT32UC3A
28.5
Application Example
28.5.1
Hardware Interface
Figure 28-2 shows an example of SDRAM device connection to the SDRAM Controller using a
32-bit data bus width. Figure 28-3 shows an example of SDRAM device connection using a 16bit data bus width. It is important to note that these examples are given for a direct connection of
the devices to the SDRAM Controller, without External Bus Interface or PIO Controller
multiplexing.
Figure 28-2. SDRAM Controller Connections to SDRAM Devices: 32-bit Data Bus Width
D0-D31
RAS
CAS
SDCK
SDCKE
SDWE
NBS0
NBS1
NBS2
NBS3
D0-D7
2M x 8
SDRAM
D8-D15
D0-D7
CS
CLK
CKE
SDWE WE
RAS
CAS
DQM
NBS0
A0-A9, A11
A10
BA0
BA1
SDRAMC_A[0-9], SDRAMC_A11
SDRAMC_A10
BA0
BA1
2M x 8
SDRAM
D0-D7
CS
CLK
CKE
SDWE
WE
RAS
CAS
DQM
NBS1
A0-A9, A11
A10
BA0
BA1
SDRAMC_A[0-9], SDRAMC_A11
SDRAMC_A10
BA0
BA1
SDRAMC_A[0-12]
BA0
BA1
2M x 8
SDRAM
SDCS
D16-D23 D0-D7
CS
CLK
CKE
SDWE WE
RAS
CAS
DQM
NBS2
SDRAM
Controller
A0-A9, A11
A10
BA0
BA1
D24-D31
SDRAMC_A[0-9], SDRAMC_ A11
SDRAMC_A10
BA0
BA1
2M x 8
SDRAM
D0-D7
CS
CLK
CKE
SDWE
WE
RAS
CAS
DQM
NBS3
SDRAMC_A[0-9], SDRAMC_A11
A0-A9, A11
SDRAMC_A10
A10
BA0
BA0
BA1 BA1
Figure 28-3. SDRAM Controller Connections to SDRAM Devices: 16-bit Data Bus Width
D0-D31
RAS
CAS
SDCK
SDCKE
SDWE
NBS0
NBS1
D0-D7
2M x 8
SDRAM
D8-D15
D0-D7
CS
CLK
CKE
SDWE WE
RAS
CAS
DQM
NBS0
A0-A9, A11
A10
BA0
BA1
SDRAMC_A[0-9], SDRAMC_A11
SDRAMC_A10
BA0
BA1
2M x 8
SDRAM
D0-D7
CS
CLK
CKE
SDWE
WE
RAS
CAS
DQM
NBS1
A0-A9, A11
A10
BA0
BA1
SDRAMC_A[0-9], SDRAMC_A11
SDRAMC_A10
BA0
BA1
SDRAMC_A[0-12]
BA0
BA1
SDRAM
Controller
SDCS
412
32058K
AVR32-01/12
AT32UC3A
28.5.2
Software Interface
The SDRAM address space is organized into banks, rows, and columns. The SDRAM controller
allows mapping different memory types according to the values set in the SDRAMC configuration register.
The SDRAM Controller’s function is to make the SDRAM device access protocol transparent to
the user. Table 28-2 to Table 28-7 illustrate the SDRAM device memory mapping seen by the
user in correlation with the device structure. Various configurations are illustrated.
28.5.2.1
32-bit Memory Data Bus Width
Table 28-2.
SDRAM Configuration Mapping: 2K Rows, 256/512/1024/2048 Columns
CPU Address Line
2
7
2
6
2
5
2
4
2
3
2
2
2
1
2
0
1
9
1
8
1
7
1
6
Bk[1:0]
1
4
1
3
1
2
1
1
1
0
9
8
7
Row[10:0]
Bk[1:0]
Bk[1:0]
6
5
4
3
2
Column[7:0]
Row[10:0]
0
M[1:0]
Column[9:0]
Row[10:0]
1
M[1:0]
Column[8:0]
Row[10:0]
Bk[1:0]
Table 28-3.
1
5
M[1:0]
Column[10:0]
M[1:0]
SDRAM Configuration Mapping: 4K Rows, 256/512/1024/2048 Columns
CPU Address Line
2
7
2
6
2
5
2
4
2
3
2
2
2
1
2
0
1
9
1
8
1
7
1
6
Bk[1:0]
1
4
1
3
1
2
1
1
1
0
9
8
7
Row[11:0]
Bk[1:0]
Bk[1:0]
6
5
4
3
2
Column[7:0]
Row[11:0]
0
M[1:0]
Column[9:0]
Row[11:0]
1
M[1:0]
Column[8:0]
Row[11:0]
Bk[1:0]
Table 28-4.
1
5
M[1:0]
Column[10:0]
M[1:0]
SDRAM Configuration Mapping: 8K Rows, 256/512/1024/2048 Columns
CPU Address Line
2
7
2
6
2
5
2
4
2
3
2
2
2
1
2
0
1
9
1
8
1
7
Bk[1:0]
Row[12:0]
Bk[1:0]
Bk[1:0]
Bk[1:0]
Notes:
1
6
Row[12:0]
Row[12:0]
Row[12:0]
1
5
1
4
1
3
1
2
1
1
1
0
9
8
7
6
5
Column[7:0]
Column[8:0]
Column[9:0]
Column[10:0]
4
3
2
1
0
M[1:0]
M[1:0]
M[1:0]
M[1:0]
1. M[1:0] is the byte address inside a 32-bit word.
2. Bk[1] = BA1, Bk[0] = BA0.
413
32058K
AVR32-01/12
AT32UC3A
28.5.2.2
16-bit Memory Data Bus Width
Table 28-5.
SDRAM Configuration Mapping: 2K Rows, 256/512/1024/2048 Columns
CPU Address Line
2
7
2
6
2
5
2
4
2
3
2
2
2
1
2
0
1
9
1
8
1
7
1
6
1
5
Bk[1:0]
1
3
1
2
1
1
1
0
9
8
7
6
Row[10:0]
Bk[1:0]
4
3
2
1
M
0
M
0
Column[9:0]
Row[10:0]
0
M
0
Column[8:0]
Row[10:0]
Bk[1:0]
5
Column[7:0]
Row[10:0]
Bk[1:0]
Table 28-6.
1
4
M
0
Column[10:0]
SDRAM Configuration Mapping: 4K Rows, 256/512/1024/2048 Columns
CPU Address Line
2
7
2
6
2
5
2
4
2
3
2
2
2
1
2
0
1
9
1
8
1
7
1
6
1
5
Bk[1:0]
1
3
1
2
1
1
1
0
9
8
7
6
Row[11:0]
Bk[1:0]
4
3
2
1
M
0
M
0
Column[9:0]
Row[11:0]
0
M
0
Column[8:0]
Row[11:0]
Bk[1:0]
5
Column[7:0]
Row[11:0]
Bk[1:0]
Table 28-7.
1
4
M
0
Column[10:0]
SDRAM Configuration Mapping: 8K Rows, 256/512/1024/2048 Columns
CPU Address Line
2
7
2
6
2
5
2
4
2
3
2
2
2
1
2
0
1
9
1
8
1
7
1
6
Bk[1:0]
Row[12:0]
Bk[1:0]
Bk[1:0]
Bk[1:0]
Notes:
1
5
Row[12:0]
Row[12:0]
Row[12:0]
1
4
1
3
1
2
1
1
1
0
9
8
7
6
5
4
Column[7:0]
Column[8:0]
Column[9:0]
Column[10:0]
3
2
1
0
M
0
M
0
M
0
M
0
1. M0 is the byte address inside a 16-bit half-word.
2. Bk[1] = BA1, Bk[0] = BA0.
414
32058K
AVR32-01/12
AT32UC3A
28.6
Product Dependencies
28.6.1
SDRAM Device Initialization
The initialization sequence is generated by software. The SDRAM devices are initialized by the
following sequence:
1. SDRAM features must be set in the configuration register: asynchronous timings (TRC,
TRAS, ...), number of column, rows, CAS latency, and the data bus width.
2. For mobile SDRAM, temperature-compensated self refresh (TCSR), drive strength (DS)
and partial array self refresh (PASR) must be set in the Low Power Register.
3. The SDRAM memory type must be set in the Memory Device Register.
4. An No Operation (NOP)command must be issued to the SDRAM devices to start the
SDRAM clock. The application must set Mode to 1 in the and perform a write access to
any SDRAM address.
5. A minimum pause of 200 µs is provided to precede any signal toggle.
6. An All Banks Precharge command must be issued to the SDRAM devices. The application must set Mode to 2 in the Mode Register and perform a write access to any SDRAM
address.
7. Eight auto-refresh (CBR) cycles are provided. The application must set the Mode to 4 in
the Mode Register and performs a write access to any SDRAM location eight times.
8. A Mode Register set (MRS) cycle must be issued to program the parameters of the
SDRAM devices, in particular CAS latency and burst length. The application must set
Mode to 3 in the Mode Register and perform a write access to the SDRAM. The write
address must be chosen so that BA[1:0] are set to 0. For example, with a 16-bit 128 MB
SDRAM (12 rows, 9 columns, 4 banks) bank address, the SDRAM write access should
be done at the address 0x20000000.
9. For mobile SDRAM initialization, an Extended Mode Register set (EMRS) cycle must be
issued to program the SDRAM parameters (TCSR, PASR, DS). The application must set
Mode to 5 in the Mode Register and perform a write access to the SDRAM. The write
address must be chosen so that BA[1] or BA[0] are set to 1. For example, with a 16-bit
128 MB SDRAM, (12 rows, 9 columns, 4 banks) bank address the SDRAM write access
should be done at the address 0x20800000 or 0x20400000.
10. The application must go into Normal Mode, setting Mode to 0 in the Mode Register and
performing a write access at any location in the SDRAM.
11. Write the refresh rate into the count field in the SDRAMC Refresh Timer register.
(Refresh rate = delay between refresh cycles). The SDRAM device requires a refresh
every 15.625 us or 7.81 us. With a 100 MHz frequency, the Refresh Timer Counter Register must be set with the value 1562 (15.625 µs x 100 MHz) or 781 (7.81 µs x 100 MHz).
After initialization, the SDRAM devices are fully functional.
415
32058K
AVR32-01/12
AT32UC3A
Figure 28-4. SDRAM Device Initialization Sequence
SDCKE
tRP
tRC
tMRD
SDCK
SDRAMC_A[9:0]
A10
SDRAMC_A[12:11]
SDCS
RAS
CAS
SDWE
NBS
Inputs Stable for
200 μsec
28.6.2
Precharge All Banks
1st Auto-refresh
8th Auto-refresh
MRS Command
Valid Command
I/O Lines
The pins used for interfacing the SDRAM Controller may be multiplexed with the PIO lines. The
programmer must first program the PIO controller to assign the SDRAM Controller pins to their
peripheral function. If I/O lines of the SDRAM Controller are not used by the application, they
can be used for other purposes by the PIO Controller.
28.6.3
Interrupt
The SDRAM Controller has an interrupt line connected to the interrupt controller. In order to handle interrupts, the interrupt controller must be programmed before configuring the SDRAM
Controller.
Using the SDRAM Controller interrupt requires the IC to be programmed first.)
416
32058K
AVR32-01/12
AT32UC3A
28.7
Functional Description
28.7.1
SDRAM Controller Write Cycle
The SDRAM Controller allows burst access or single access. In both cases, the SDRAM controller keeps track of the active row in each bank, thus maximizing performance. To initiate a burst
access, the SDRAM Controller uses the transfer type signal provided by the master requesting
the access. If the next access is a sequential write access, writing to the SDRAM device is carried out. If the next access is a write-sequential access, but the current access is to a boundary
page, or if the next access is in another row, then the SDRAM Controller generates a precharge
command, activates the new row and initiates a write command. To comply with SDRAM timing
parameters, additional clock cycles are inserted between precharge/active (tRP) commands and
active/write (tRCD) commands. For definition of these timing parameters, refer to the ”SDRAMC
Configuration Register” on page 427. This is described in Figure 28-5 below.
Figure 28-5. Write Burst, 32-bit SDRAM Access
tRCD = 3
SDCS
SDCK
SDRAMC_A[12:0]
Row n
col a
col b
col c
col d
col e
col f
col g
col h
col i
col j
col k
col l
Dnb
Dnc
Dnd
Dne
Dnf
Dng
Dnh
Dni
Dnj
Dnk
Dnl
RAS
CAS
SDWE
D[31:0]
28.7.2
Dna
SDRAM Controller Read Cycle
The SDRAM Controller allows burst access, incremental burst of unspecified length or single
access. In all cases, the SDRAM Controller keeps track of the active row in each bank, thus
maximizing performance of the SDRAM. If row and bank addresses do not match the previous
row/bank address, then the SDRAM controller automatically generates a precharge command,
activates the new row and starts the read command. To comply with the SDRAM timing parameters, additional clock cycles on SDCK are inserted between precharge and active commands
(tRP) and between active and read command (tRCD). These two parameters are set in the configuration register of the SDRAM Controller. After a read command, additional wait states are
generated to comply with the CAS latency (1, 2 or 3 clock delays specified in the configuration
register).
417
32058K
AVR32-01/12
AT32UC3A
For a single access or an incremented burst of unspecified length, the SDRAM Controller anticipates the next access. While the last value of the column is returned by the SDRAM Controller
on the bus, the SDRAM Controller anticipates the read to the next column and thus anticipates
the CAS latency. This reduces the effect of the CAS latency on the internal bus.
For burst access of specified length (4, 8, 16 words), access is not anticipated. This case leads
to the best performance. If the burst is broken (border, busy mode, etc.), the next access is handled as an incrementing burst of unspecified length.
Figure 28-6. Read Burst, 32-bit SDRAM Access
tRCD = 3
CAS = 2
SDCS
SDCK
SDRAMC_A[12:0]
Row n
col a
col b
col c
col d
col e
col f
RAS
CAS
SDWE
D[31:0]
(Input)
28.7.3
Dna
Dnb
Dnc
Dnd
Dne
Dnf
Border Management
When the memory row boundary has been reached, an automatic page break is inserted. In this
case, the SDRAM controller generates a precharge command, activates the new row and initiates a read or write command. To comply with SDRAM timing parameters, an additional clock
cycle is inserted between the precharge/active (tRP) command and the active/read (tRCD) command. This is described in Figure 28-7 below.
418
32058K
AVR32-01/12
AT32UC3A
Figure 28-7. Read Burst with Boundary Row Access
TRP = 3
TRCD = 3
CAS = 2
SDCS
SDCK
Row n
SDRAMC_A[12:0]
col a
col b
col c
col d
Row m
col a
col b
col c
col d
col e
RAS
CAS
SDWE
D[31:0]
28.7.4
Dna
Dnb
Dnc
Dnd
Dma
Dmb
Dmc
Dmd
Dme
SDRAM Controller Refresh Cycles
An auto-refresh command is used to refresh the SDRAM device. Refresh addresses are generated internally by the SDRAM device and incremented after each auto-refresh automatically.
The SDRAM Controller generates these auto-refresh commands periodically. An internal timer is
loaded with the value in the register TR that indicates the number of clock cycles between
refresh cycles.
A refresh error interrupt is generated when the previous auto-refresh command did not perform.
It is acknowledged by reading the Interrupt Status Register (ISR).
When the SDRAM Controller initiates a refresh of the SDRAM device, internal memory accesses
are not delayed. However, if the CPU tries to access the SDRAM, the slave indicates that the
device is busy and the master is held by a wait signal. See Figure 28-8.
419
32058K
AVR32-01/12
AT32UC3A
Figure 28-8. Refresh Cycle Followed by a Read Access
tRP = 3
tRC = 8
tRCD = 3
CAS = 2
SDCS
SDCK
Row n
Row m
col c col d
SDRAMC_A[12:0]
col a
RAS
CAS
SDWE
D[31:0]
(input)
28.7.5
Dnb
Dnc
Dnd
Dma
Power Management
Three low-power modes are available:
• Self-refresh Mode: The SDRAM executes its own Auto-refresh cycle without control of the
SDRAM Controller. Current drained by the SDRAM is very low.
• Power-down Mode: Auto-refresh cycles are controlled by the SDRAM Controller. Between
auto-refresh cycles, the SDRAM is in power-down. Current drained in Power-down mode is
higher than in Self-refresh Mode.
• Deep Power-down Mode: (Only available with Mobile SDRAM) The SDRAM contents are lost,
but the SDRAM does not drain any current.
The SDRAM Controller activates one low-power mode as soon as the SDRAM device is not
selected. It is possible to delay the entry in self-refresh and power-down mode after the last
access by programming a timeout value in the Low Power Register.
28.7.5.1
Self-refresh Mode
This mode is selected by programming the LPCB field to 1 in the SDRAMC Low Power Register.
In self-refresh mode, the SDRAM device retains data without external clocking and provides its
own internal clocking, thus performing its own auto-refresh cycles. All the inputs to the SDRAM
device become “don’t care” except SDCKE, which remains low. As soon as the SDRAM device
is selected, the SDRAM Controller provides a sequence of commands and exits self-refresh
mode.
Some low-power SDRAMs (e.g., mobile SDRAM) can refresh only one quarter or a half quarter
or all banks of the SDRAM array. This feature reduces the self-refresh current. To configure this
feature, Temperature Compensated Self Refresh (TCSR), Partial Array Self Refresh (PASR)
and Drive Strength (DS) parameters must be set in the Low Power Register and transmitted to
the low-power SDRAM during initialization.
420
32058K
AVR32-01/12
AT32UC3A
After initialization, as soon as PASR/DS/TCSR fields are modified and self-refresh mode is activated, the Extended Mode Register is accessed automatically and PASR/DS/TCSR bits are
updated before entry into self-refresh mode.
The SDRAM device must remain in self-refresh mode for a minimum period of tRAS and may
remain in self-refresh mode for an indefinite period. This is described in Figure 28-9.
Figure 28-9. Self-refresh Mode Behavior
Self Refresh Mode
SRCB = 1
TXSR = 3
Write
SDRAMC_SRR
Row
SDRAMC_A[12:0]
SDCK
SDCKE
SDCS
RAS
CAS
SDWE
Access Request
to the SDRAM Controller
28.7.5.2
Low-power Mode
This mode is selected by programming the LPCB field to 2 in the SDRAMC Low Power Register.
Power consumption is greater than in self-refresh mode. All the input and output buffers of the
SDRAM device are deactivated except SDCKE, which remains low. In contrast to self-refresh
mode, the SDRAM device cannot remain in low-power mode longer than the refresh period (64
ms for a whole device refresh operation). As no auto-refresh operations are performed by the
SDRAM itself, the SDRAM Controller carries out the refresh operation. The exit procedure is
faster than in self-refresh mode.
This is described in Figure 28-10.
421
32058K
AVR32-01/12
AT32UC3A
Figure 28-10. Low-power Mode Behavior
TRCD = 3
CAS = 2
Low Power Mode
SDCS
SDCK
SDRAMC_A[12:0]
Row n
col a
col b
col c
col d
col e
col f
RAS
CAS
SDCKE
D[31:0]
(input)
28.7.5.3
Dna
Dnb
Dnc
Dnd
Dne
Dnf
Deep Power-down Mode
This mode is selected by programming the LPCB field to 3 in the SDRAMC Low Power Register.
When this mode is activated, all internal voltage generators inside the SDRAM are stopped and
all data is lost.
When this mode is enabled, the application must not access to the SDRAM until a new initialization sequence is done (See ”SDRAM Device Initialization” on page 415).
This is described in Figure 28-11.
422
32058K
AVR32-01/12
AT32UC3A
Figure 28-11. Deep Power-down Mode Behavior
tRP = 3
SDCS
SDCK
Row n
SDRAMC_A[12:0]
col c
col d
RAS
CAS
SDWE
CKE
D[31:0]
(input)
Dnb
Dnc
Dnd
423
32058K
AVR32-01/12
AT32UC3A
28.8
SDRAM Controller User Interface
Table 28-8.
Offset
SDRAM Controller Memory Map
Register
Name
Access
Reset State
0x00
SDRAMC Mode Register
MR
Read/Write
0x00000000
0x04
SDRAMC Refresh Timer Register
TR
Read/Write
0x00000000
0x08
SDRAMC Configuration Register
CR
Read/Write
0x852372C0
0x0C
SDRAMC High Speed Register
HSR
Read/Write
0x00
0x10
SDRAMC Low Power Register
LPR
Read/Write
0x0
0x14
SDRAMC Interrupt Enable Register
IER
Write-only
–
0x18
SDRAMC Interrupt Disable Register
IDR
Write-only
–
0x1C
SDRAMC Interrupt Mask Register
IMR
Read-only
0x0
0x20
SDRAMC Interrupt Status Register
ISR
Read-only
0x0
0x24
SDRAMC Memory Device Register
MDR
Read/Write
0x0
–
–
–
0x28 - 0xFC
Reserved
424
32058K
AVR32-01/12
AT32UC3A
28.8.1
SDRAMC Mode Register
Register Name:
MR
Access Type:
Read/Write
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
–
5
–
4
–
3
–
2
1
0
MODE
• MODE: SDRAMC Command Mode
This field defines the command issued by the SDRAM Controller when the SDRAM device is accessed.
Table 28-9.
MODE
Description
0
0
0
Normal mode. Any access to the SDRAM is decoded normally.
0
0
1
The SDRAM Controller issues a NOP command when the SDRAM device is accessed regardless of the cycle.
0
1
0
The SDRAM Controller issues an “All Banks Precharge” command when the SDRAM device is accessed
regardless of the cycle.
0
1
1
The SDRAM Controller issues a “Load Mode Register” command when the SDRAM device is accessed
regardless of the cycle. The command will load the CAS latency from the Configuration Register and every other
value set to 0 into the Mode Register.
1
0
0
The SDRAM Controller issues an “Auto-Refresh” Command when the SDRAM device is accessed regardless of
the cycle. Previously, an “All Banks Precharge” command must be issued.
1
0
1
The SDRAM Controller issues an extended load mode register command when the SDRAM device is accessed
regardless of the cycle. The command will load the PASR, DS and TCR from the Low Power Register and every
other value set to 0 into the Extended Mode Register.
1
1
0
Deep power-down mode. Enters deep power-down mode.
425
32058K
AVR32-01/12
AT32UC3A
28.8.2
SDRAMC Refresh Timer Register
Register Name:
TR
Access Type:
Read/Write
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
10
9
8
7
6
5
4
1
0
COUNT
3
2
COUNT
• COUNT: SDRAMC Refresh Timer Count
This 12-bit field is loaded into a timer that generates the refresh pulse. Each time the refresh pulse is generated, a refresh
burst is initiated. The value to be loaded depends on the SDRAMC clock frequency (MCK: Master Clock), the refresh rate
of the SDRAM device and the refresh burst length where 15.6 µs per row is a typical value for a burst of length one.
To refresh the SDRAM device, this 12-bit field must be written. If this condition is not satisfied, no refresh command is
issued and no refresh of the SDRAM device is carried out.
426
32058K
AVR32-01/12
AT32UC3A
28.8.3
SDRAMC Configuration Register
Register Name:
CR
Access Type:
Read/Write
Reset Value:
0x852372C0
31
30
29
28
27
26
TXSR
23
22
21
20
19
18
TRCD
15
24
17
16
9
8
TRP
14
13
12
11
10
TRC
7
DBW
25
TRAS
TWR
6
5
CAS
4
NB
3
2
NR
1
0
NC
• NC: Number of Column Bits
Reset value is 8 column bits.
NC
Column Bits
0
0
8
0
1
9
1
0
10
1
1
11
• NR: Number of Row Bits
Reset value is 11 row bits.
NR
Row Bits
0
0
11
0
1
12
1
0
13
1
1
Reserved
• NB: Number of Banks
Reset value is two banks.
NB
Number of Banks
0
2
1
4
427
32058K
AVR32-01/12
AT32UC3A
• CAS: CAS Latency
Reset value is two cycles.
In the SDRAMC, only a CAS latency of one, two and three cycles is managed.
CAS
CAS Latency (Cycles)
0
0
Reserved
0
1
1
1
0
2
1
1
3
• DBW: Data Bus Width
Reset value is 16 bits
0: Data bus width is 32 bits.
1: Data bus width is 16 bits.
• TWR: Write Recovery Delay
Reset value is two cycles.
This field defines the Write Recovery Time in number of cycles. Number of cycles is between 0 and 15.
• TRC: Row Cycle Delay
Reset value is seven cycles.
This field defines the delay between a Refresh and an Activate Command in number of cycles. Number of cycles is
between 0 and 15.
• TRP: Row Precharge Delay
Reset value is three cycles.
This field defines the delay between a Precharge Command and another Command in number of cycles. Number of cycles
is between 0 and 15.
• TRCD: Row to Column Delay
Reset value is two cycles.
This field defines the delay between an Activate Command and a Read/Write Command in number of cycles. Number of
cycles is between 0 and 15.
• TRAS: Active to Precharge Delay
Reset value is five cycles.
This field defines the delay between an Activate Command and a Precharge Command in number of cycles. Number of
cycles is between 0 and 15.
• TXSR: Exit Self Refresh to Active Delay
Reset value is height cycles.
This field defines the delay between SCKE set high and an Activate Command in number of cycles. Number of cycles is
between 0 and 15.
428
32058K
AVR32-01/12
AT32UC3A
28.8.4
SDRAMC High Speed Register
Register Name:
HSR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
–
5
–
4
–
3
–
2
–
1
–
0
DA
• DA: Decode Cycle Enable
A decode cycle can be added on the addresses as soon as a non-sequential access is performed on the HSB bus.
The addition of the decode cycle allows the SDRAMC to gain time to access the SDRAM memory.
0: Decode cycle is disabled.
1: Decode cycle is enabled.
429
32058K
AVR32-01/12
AT32UC3A
28.8.5
SDRAMC Low Power Register
Register Name:
LPR
Access Type:
Read/Write
Reset Value:
0x0
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
12
11
10
9
7
–
6
5
PASR
TIMEOUT
DS
4
3
–
8
TCSR
2
–
1
0
LPCB
• LPCB: Low-power Configuration Bits
00
Low Power Feature is inhibited: no Power-down, Self-refresh or Deep Power-down command is issued to
the SDRAM device.
01
The SDRAM Controller issues a Self-refresh command to the SDRAM device, the SDCLK clock is
deactivated and the SDCKE signal is set low. The SDRAM device leaves the Self Refresh Mode when
accessed and enters it after the access.
10
The SDRAM Controller issues a Power-down Command to the SDRAM device after each access, the
SDCKE signal is set to low. The SDRAM device leaves the Power-down Mode when accessed and
enters it after the access.
11
The SDRAM Controller issues a Deep Power-down command to the SDRAM device. This mode is
unique to low-power SDRAM.
• PASR: Partial Array Self-refresh (only for low-power SDRAM)
PASR parameter is transmitted to the SDRAM during initialization to specify whether only one quarter, one half or all banks
of the SDRAM array are enabled. Disabled banks are not refreshed in self-refresh mode. This parameter must be set
according to the SDRAM device specification.
After initialization, as soon as PASR field is modified and self-refresh mode is activated, the Extended Mode Register is
accessed automatically and PASR bits are updated before entry in self-refresh mode.
• TCSR: Temperature Compensated Self-Refresh (only for low-power SDRAM)
TCSR parameter is transmitted to the SDRAM during initialization to set the refresh interval during self-refresh mode
depending on the temperature of the low-power SDRAM. This parameter must be set according to the SDRAM device
specification.
After initialization, as soon as TCSR field is modified and self-refresh mode is activated, the Extended Mode Register is
accessed automatically and TCSR bits are updated before entry in self-refresh mode.
• DS: Drive Strength (only for low-power SDRAM)
DS parameter is transmitted to the SDRAM during initialization to select the SDRAM strength of data output. This parameter must be set according to the SDRAM device specification.
430
32058K
AVR32-01/12
AT32UC3A
After initialization, as soon as DS field is modified and self-refresh mode is activated, the Extended Mode Register is
accessed automatically and DS bits are updated before entry in self-refresh mode.
• TIMEOUT: Time to define when low-power mode is enabled
00
The SDRAM controller activates the SDRAM low-power mode immediately after the end of the last transfer.
01
The SDRAM controller activates the SDRAM low-power mode 64 clock cycles after the end of the last
transfer.
10
The SDRAM controller activates the SDRAM low-power mode 128 clock cycles after the end of the last
transfer.
11
Reserved.
431
32058K
AVR32-01/12
AT32UC3A
28.8.6
SDRAMC Interrupt Enable Register
Register Name:
IER
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
–
5
–
4
–
3
–
2
–
1
–
0
RES
• RES: Refresh Error Status
0: No effect.
1: Enables the refresh error interrupt.
432
32058K
AVR32-01/12
AT32UC3A
28.8.7
SDRAMC Interrupt Disable Register
Register Name:
IDR
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
–
5
–
4
–
3
–
2
–
1
–
0
RES
• RES: Refresh Error Status
0: No effect.
1: Disables the refresh error interrupt.
433
32058K
AVR32-01/12
AT32UC3A
28.8.8
SDRAMC Interrupt Mask Register
Register Name:
IMR
Access Type:
Read-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
–
5
–
4
–
3
–
2
–
1
–
0
RES
• RES: Refresh Error Status
0: The refresh error interrupt is disabled.
1: The refresh error interrupt is enabled.
434
32058K
AVR32-01/12
AT32UC3A
28.8.9
SDRAMC Interrupt Status Register
Register Name:
ISR
Access Type:
Read-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
–
5
–
4
–
3
–
2
–
1
–
0
RES
• RES: Refresh Error Status
0: No refresh error has been detected since the register was last read.
1: A refresh error has been detected since the register was last read.
435
32058K
AVR32-01/12
AT32UC3A
28.8.10
SDRAMC Memory Device Register
Register Name:
MDR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
–
5
–
4
–
3
–
2
–
1
0
MD
• MD: Memory Device Type
00
SDRAM
01
Low-power SDRAM
10
Reserved
11
Reserved.
436
32058K
AVR32-01/12
AT32UC3A
29. Ethernet MAC (MACB)
Rev: 1.1.2.5
29.1
Features
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
29.2
Compatible with IEEE Standard 802.3
10 and 100 Mbit/s Operation
Full- and Half-duplex Operation
Statistics Counter Registers
MII/RMII Interface to the Physical Layer
Interrupt Generation to Signal Receive and Transmit Completion
DMA Master on Receive and Transmit Channels
Transmit and Receive FIFOs
Automatic Pad and CRC Generation on Transmitted Frames
Automatic Discard of Frames Received with Errors
Address Checking Logic Supports Up to Four Specific 48-bit Addresses
Supports Promiscuous Mode Where All Valid Received Frames are Copied to Memory
Hash Matching of Unicast and Multicast Destination Addresses
External Address Matching of Received Frames
Physical Layer Management through MDIO Interface
Half-duplex Flow Control by Forcing Collisions on Incoming Frames
Full-duplex Flow Control with Recognition of Incoming Pause Frames and Hardware Generation
of Transmitted Pause Frames
Support for 802.1Q VLAN Tagging with Recognition of Incoming VLAN and Priority Tagged
Frames
Multiple Buffers per Receive and Transmit Frame
Wake-on-LAN Support
Jumbo Frames Up to 10240 bytes Supported
Description
The MACB module implements a 10/100 Ethernet MAC compatible with the IEEE 802.3 standard using an address checker, statistics and control registers, receive and transmit blocks, and
a DMA interface.
The address checker recognizes four specific 48-bit addresses and contains a 64-bit hash register for matching multicast and unicast addresses. It can recognize the broadcast address of all
ones, copy all frames, and act on an external address match signal.
The statistics register block contains registers for counting various types of events associated
with transmit and receive operations. These registers, along with the status words stored in the
receive buffer list, enable software to generate network management statistics compatible with
IEEE 802.3.
437
32058K
AVR32-01/12
AT32UC3A
29.3
Block Diagram
Figure 29-1. MACB Block Diagram
Address Checker
Peripheral Bus
Slave
Register Interface
Statistics Registers
Control Registers
MDIO
DMA Interface
RX FIFO
TX FIFO
Ethernet Receive
MII/RMII
High Speed Bus
Master
Ethernet Transmit
29.4
Product Dependencies
29.4.1
I/O Lines
The pins used for interfacing the compliant external devices may be multiplexed with PIO lines.
The programmer must first program the PIO controllers to assign the MACB pins to their peripheral functions.
29.4.2
Power Management
The MACB clock is generated by the Power Manager. Before using the MACB, the programmer
must ensure that the MACB clock is enabled in the Power Manager.
In the MACB description, Master Clock (MCK) is the clock of the peripheral bus to which the
MACB is connected.
The synchronization module in the MACB requires that the bus clock (hclk) runs on at least the
speed of the macb_tx/rx_clk, which is 25MHz in 100Mbps, and 2.5MHZ in 10Mbps in MII mode
and 50MHz in 100Mbps, and 5MHZ in 10Mbps in RMII mode.
To prevent bus errors the MACB operation must be terminated before entering sleep mode.
438
32058K
AVR32-01/12
AT32UC3A
29.4.3
Interrupt
The MACB interface has an interrupt line connected to the Interrupt Controller. Handling the
MACB interrupt requires programming the interrupt controller before configuring the MACB.
29.5
Functional Description
Figure 29-1 on page 438 illustrates the different blocks of the MACB module.
The control registers drive the MDIO interface, setup DMA activity, start frame transmission and
select modes of operation such as full- or half-duplex.
The receive block checks for valid preamble, FCS, alignment and length, and presents received
frames to the address checking block and DMA interface.
The transmit block takes data from the DMA interface, adds preamble and, if necessary, pad
and FCS, and transmits data according to the CSMA/CD (carrier sense multiple access with collision detect) protocol. The start of transmission is deferred if CRS (carrier sense) is active.
If COL (collision) becomes active during transmission, a jam sequence is asserted and the
transmission is retried after a random back off. CRS and COL have no effect in full duplex mode.
The DMA block connects to external memory through its high speed bus (HSB) interface. It contains receive and transmit FIFOs for buffering frame data. It loads the transmit FIFO and empties
the receive FIFO using HSB bus master operations. Receive data is not sent to memory until the
address checking logic has determined that the frame should be copied. Receive or transmit
frames are stored in one or more buffers. Receive buffers have a fixed length of 128 bytes.
Transmit buffers range in length between 0 and 2047 bytes, and up to 128 buffers are permitted
per frame. The DMA block manages the transmit and receive framebuffer queues. These
queues can hold multiple frames.
29.5.1
Memory Interface
Frame data is transferred to and from the MACB through the DMA interface. All transfers are 32bit words and may be single accesses or bursts of 2, 3 or 4 words. Burst accesses do not cross
sixteen-byte boundaries. Bursts of 4 words are the default data transfer; single accesses or
bursts of less than four words may be used to transfer data at the beginning or the end of a
buffer.
The DMA controller performs six types of operation on the bus. In order of priority, these are:
1. Receive buffer manager write
2. Receive buffer manager read
3. Transmit data DMA read
4. Receive data DMA write
5. Transmit buffer manager read
6. Transmit buffer manager write
29.5.1.1
FIFO
The FIFO depths are 124 bytes.
Data is typically transferred into and out of the FIFOs in bursts of four words. For receive, a bus
request is asserted when the FIFO contains four words and has space for three more. For transmit, a bus request is generated when there is space for four words, or when there is space for
two words if the next transfer is to be only one or two words.
439
32058K
AVR32-01/12
AT32UC3A
Thus the bus latency must be less than the time it takes to load the FIFO and transmit or receive
three words (12 bytes) of data.
At 100 Mbit/s, it takes 960 ns to transmit or receive 12 bytes of data. In addition, six master clock
cycles should be allowed for data to be loaded from the bus and to propagate through the
FIFOs. For a 60 MHz master clock this takes 100 ns, making the bus latency requirement 860
ns.
29.5.1.2
Receive Buffers
Received frames, optionally including CRC/FCS, are written to receive buffers stored in memory. Each receive buffer is 128 bytes long. The start location for each receive buffer is stored in
memory in a list of receive buffer descriptors at a location pointed to by the receive buffer queue
pointer register. The receive buffer start location is a word address. For the first buffer of a
frame, the start location can be offset by up to three bytes depending on the value written to bits
14 and 15 of the network configuration register. If the start location of the buffer is offset the
available length of the first buffer of a frame is reduced by the corresponding number of bytes.
Each list entry consists of two words, the first being the address of the receive buffer and the
second being the receive status. If the length of a receive frame exceeds the buffer length, the
status word for the used buffer is written with zeroes except for the “start of frame” bit and the
offset bits, if appropriate. Bit zero of the address field is written to one to show the buffer has
been used. The receive buffer manager then reads the location of the next receive buffer and
fills that with receive frame data. The final buffer descriptor status word contains the complete
frame status. Refer to Table 29-1 for details of the receive buffer descriptor list.
Receive Buffer Descriptor Entry
Table 29-1.
Bit
Function
Word 0
31:2
Address of beginning of buffer
1
Wrap - marks last descriptor in receive buffer descriptor list.
0
Ownership - needs to be zero for the MACB to write data to the receive buffer. The MACB sets this to one once it has
successfully written a frame to memory.
Software has to clear this bit before the buffer can be used again.
Word 1
31
Global all ones broadcast address detected
30
Multicast hash match
29
Unicast hash match
28
External address match
27
Reserved for future use
26
Specific address register 1 match
25
Specific address register 2 match
24
Specific address register 3 match
23
Specific address register 4 match
22
Type ID match
21
VLAN tag detected (i.e., type id of 0x8100)
440
32058K
AVR32-01/12
AT32UC3A
Receive Buffer Descriptor Entry (Continued)
Table 29-1.
Bit
Function
20
19:17
Priority tag detected (i.e., type id of 0x8100 and null VLAN identifier)
VLAN priority (only valid if bit 21 is set)
16
Concatenation format indicator (CFI) bit (only valid if bit 21 is set)
15
End of frame - when set the buffer contains the end of a frame. If end of frame is not set, then the only other valid status
are bits 12, 13 and 14.
14
Start of frame - when set the buffer contains the start of a frame. If both bits 15 and 14 are set, then the buffer contains a
whole frame.
13:12
Receive buffer offset - indicates the number of bytes by which the data in the first buffer is offset from the word address.
Updated with the current values of the network configuration register. If jumbo frame mode is enabled through bit 3 of the
network configuration register, then bits 13:12 of the receive buffer descriptor entry are used to indicate bits 13:12 of the
frame length.
11:0
Length of frame including FCS (if selected). Bits 13:12 are also used if jumbo frame mode is selected.
To receive frames, the buffer descriptors must be initialized by writing an appropriate address to
bits 31 to 2 in the first word of each list entry. Bit zero must be written with zero. Bit one is the
wrap bit and indicates the last entry in the list.
The start location of the receive buffer descriptor list must be written to the receive buffer queue
pointer register before setting the receive enable bit in the network control register to enable
receive. As soon as the receive block starts writing received frame data to the receive FIFO, the
receive buffer manager reads the first receive buffer location pointed to by the receive buffer
queue pointer register.
If the filter block then indicates that the frame should be copied to memory, the receive data
DMA operation starts writing data into the receive buffer. If an error occurs, the buffer is recovered. If the current buffer pointer has its wrap bit set or is the 1024th descriptor, the next receive
buffer location is read from the beginning of the receive descriptor list. Otherwise, the next
receive buffer location is read from the next word in memory.
There is an 11-bit counter to count out the 2048 word locations of a maximum length, receive
buffer descriptor list. This is added with the value originally written to the receive buffer queue
pointer register to produce a pointer into the list. A read of the receive buffer queue pointer register returns the pointer value, which is the queue entry currently being accessed. The counter is
reset after receive status is written to a descriptor that has its wrap bit set or rolls over to zero
after 1024 descriptors have been accessed. The value written to the receive buffer pointer register may be any word-aligned address, provided that there are at least 2048 word locations
available between the pointer and the top of the memory.
The System Bus specification states that bursts should not cross 1K boundaries. As receive buffer manager writes are bursts of two words, to ensure that this does not occur, it is best to write
the pointer register with the least three significant bits set to zero. As receive buffers are used,
the receive buffer manager sets bit zero of the first word of the descriptor to indicate used. If a
receive error is detected the receive buffer currently being written is recovered. Previous buffers
are not recovered. Software should search through the used bits in the buffer descriptors to find
out how many frames have been received. It should be checking the start-of-frame and end-offrame bits, and not rely on the value returned by the receive buffer queue pointer register which
changes continuously as more buffers are used.
441
32058K
AVR32-01/12
AT32UC3A
For CRC errored frames, excessive length frames or length field mismatched frames, all of
which are counted in the statistics registers, it is possible that a frame fragment might be stored
in a sequence of receive buffers. Software can detect this by looking for start of frame bit set in a
buffer following a buffer with no end of frame bit set.
For a properly working Ethernet system, there should be no excessively long frames or frames
greater than 128 bytes with CRC/FCS errors. Collision fragments are less than 128 bytes long.
Therefore, it is a rare occurrence to find a frame fragment in a receive buffer.
If bit zero is set when the receive buffer manager reads the location of the receive buffer, then
the buffer has already been used and cannot be used again until software has processed the
frame and cleared bit zero. In this case, the DMA block sets the buffer not available bit in the
receive status register and triggers an interrupt.
If bit zero is set when the receive buffer manager reads the location of the receive buffer and a
frame is being received, the frame is discarded and the receive resource error statistics register
is incremented.
A receive overrun condition occurs when bus was not granted in time or because HRESP was
not OK (bus error). In a receive overrun condition, the receive overrun interrupt is asserted and
the buffer currently being written is recovered. The next frame received with an address that is
recognized reuses the buffer.
If bit 17 of the network configuration register is set, the FCS of received frames shall not be copied to memory. The frame length indicated in the receive status field shall be reduced by four
bytes in this case.
29.5.1.3
Transmit Buffer
Frames to be transmitted are stored in one or more transmit buffers. Transmit buffers can be
between 0 and 2047 bytes long, so it is possible to transmit frames longer than the maximum
length specified in IEEE Standard 802.3. Zero length buffers are allowed. The maximum number
of buffers permitted for each transmit frame is 128.
The start location for each transmit buffer is stored in memory in a list of transmit buffer descriptors at a location pointed to by the transmit buffer queue pointer register. Each list entry consists
of two words, the first being the byte address of the transmit buffer and the second containing
the transmit control and status. Frames can be transmitted with or without automatic CRC generation. If CRC is automatically generated, padding is also automatically generated to take
frames to a minimum length of 64 bytes. Table 29-2 on page 443 defines an entry in the transmit
buffer descriptor list. To transmit frames, the buffer descriptors must be initialized by writing an
appropriate byte address to bits 31 to 0 in the first word of each list entry. The second transmit
buffer descriptor is initialized with control information that indicates the length of the buffer,
whether or not it is to be transmitted with CRC and whether the buffer is the last buffer in the
frame.
After transmission, the control bits are written back to the second word of the first buffer along
with the “used” bit and other status information. Before a transmission, bit 31 is the “used” bit
which must be zero when the control word is read. It is written to one when a frame has been
transmitted. Bits 27, 28 and 29 indicate various transmit error conditions. Bit 30 is the “wrap” bit
which can be set for any buffer within a frame. If no wrap bit is encountered after 1024 descriptors, the queue pointer rolls over to the start.
The transmit buffer queue pointer register must not be written while transmit is active. If a new
value is written to the transmit buffer queue pointer register, the queue pointer resets itself to
442
32058K
AVR32-01/12
AT32UC3A
point to the beginning of the new queue. If transmit is disabled by writing to bit 3 of the network
control, the transmit buffer queue pointer register resets to point to the beginning of the transmit
queue. Note that disabling receive does not have the same effect on the receive queue pointer.
Once the transmit queue is initialized, transmit is activated by writing to bit 9, the Transmit Start
bit of the network control register. Transmit is halted when a buffer descriptor with its used bit set
is read, or if a transmit error occurs, or by writing to the transmit halt bit of the network control
register. (Transmission is suspended if a pause frame is received while the pause enable bit is
set in the network configuration register.) Rewriting the start bit while transmission is active is
allowed.
Transmission control is implemented with a Tx_go variable which is readable in the transmit status register at bit location 3. The Tx_go variable is reset when:
– transmit is disabled
– a buffer descriptor with its ownership bit set is read
– a new value is written to the transmit buffer queue pointer register
– bit 10, tx_halt, of the network control register is written
– there is a transmit error such as too many retries or a transmit underrun.
To set tx_go, write to bit 9, tx_start, of the network control register. Transmit halt does not take
effect until any ongoing transmit finishes. If a collision occurs during transmission of a multi-buffer frame, transmission automatically restarts from the first buffer of the frame. If a “used” bit is
read midway through transmission of a multi-buffer frame, this is treated as a transmit error.
Transmission stops, tx_er is asserted and the FCS is bad.
If transmission stops due to a transmit error, the transmit queue pointer resets to point to the
beginning of the transmit queue. Software needs to re-initialize the transmit queue after a transmit error.
If transmission stops due to a “used” bit being read at the start of the frame, the transmission
queue pointer is not reset and transmit starts from the same transmit buffer descriptor when the
transmit start bit is written
Transmit Buffer Descriptor Entry
Table 29-2.
Bit
Function
Word 0
31:0
Byte Address of buffer
Word 1
31
Used. Needs to be zero for the MACB to read data from the transmit buffer. The MACB sets this to one for the first buffer
of a frame once it has been successfully transmitted.
Software has to clear this bit before the buffer can be used again.
Note:
This bit is only set for the first buffer in a frame unlike receive where all buffers have the Used bit set once used.
30
Wrap. Marks last descriptor in transmit buffer descriptor list.
29
Retry limit exceeded, transmit error detected
28
Transmit underrun, occurs either when hresp is not OK (bus error) or the transmit data could not be fetched in time or
when buffers are exhausted in mid frame.
27
Buffers exhausted in mid frame
26:17
Reserved
443
32058K
AVR32-01/12
AT32UC3A
Transmit Buffer Descriptor Entry (Continued)
Table 29-2.
Bit
Function
16
No CRC. When set, no CRC is appended to the current frame. This bit only needs to be set for the last buffer of a frame.
15
Last buffer. When set, this bit indicates the last buffer in the current frame has been reached.
14:11
Reserved
10:0
Length of buffer
29.5.2
Transmit Block
This block transmits frames in accordance with the Ethernet IEEE 802.3 CSMA/CD protocol.
Frame assembly starts by adding preamble and the start frame delimiter. Data is taken from the
transmit FIFO a word at a time. Data is transmitted least significant nibble first. If necessary,
padding is added to increase the frame length to 60 bytes. CRC is calculated as a 32-bit polynomial. This is inverted and appended to the end of the frame, taking the frame length to a
minimum of 64 bytes. If the No CRC bit is set in the second word of the last buffer descriptor of a
transmit frame, neither pad nor CRC are appended.
In full-duplex mode, frames are transmitted immediately. Back-to-back frames are transmitted at
least 96 bit times apart to guarantee the interframe gap.
In half-duplex mode, the transmitter checks carrier sense. If asserted, it waits for it to de-assert
and then starts transmission after the interframe gap of 96 bit times. If the collision signal is
asserted during transmission, the transmitter transmits a jam sequence of 32 bits taken from the
data register and retries transmission after the back off time has elapsed.
The back-off time is based on an XOR of the 10 least significant bits of the data coming from the
transmit FIFO and a 10-bit pseudo random number. The number of bits used depends on the
number of collisions seen. After the first collision, 1 bit is used, after the second 2, and so on up
to 10. Above 10, all 10 bits are used. An error is indicated and no further attempts are made if 16
attempts cause collisions.
If transmit DMA underruns, bad CRC is automatically appended using the same mechanism as
jam insertion and the TX_ER signal is asserted. In a properly configured system, this should
never happen.
If the back pressure bit is set in the network control register in half duplex mode, the transmit
block transmits 64 bits of data, which can consist of 16 nibbles of 1011 or in bit-rate mode 64 1s,
whenever it sees an incoming frame to force a collision. This provides a way of implementing
flow control in half-duplex mode.
29.5.3
Pause Frame Support
The start of an 802.3 pause frame is as follows:
Table 29-3.
Start of an 802.3 Pause Frame
Destination Address
Source
Address
Type
(Mac Control Frame)
Pause
Opcode
Pause Time
0x0180C2000001
6 bytes
0x8808
0x0001
2 bytes
The network configuration register contains a receive pause enable bit (13). If a valid pause
frame is received, the pause time register is updated with the frame’s pause time, regardless of
444
32058K
AVR32-01/12
AT32UC3A
its current contents and regardless of the state of the configuration register bit 13. An interrupt
(12) is triggered when a pause frame is received, assuming it is enabled in the interrupt mask
register. If bit 13 is set in the network configuration register and the value of the pause time register is non-zero, no new frame is transmitted until the pause time register has decremented to
zero.
The loading of a new pause time, and hence the pausing of transmission, only occurs when the
MACB is configured for full-duplex operation. If the MACB is configured for half-duplex, there is
no transmission pause, but the pause frame received interrupt is still triggered.
A valid pause frame is defined as having a destination address that matches either the address
stored in specific address register 1 or matches 0x0180C2000001 and has the MAC control
frame type ID of 0x8808 and the pause opcode of 0x0001. Pause frames that have FCS or other
errors are treated as invalid and are discarded. Valid pause frames received increment the
Pause Frame Received statistic register.
The pause time register decrements every 512 bit times (i.e., 128 rx_clks in nibble mode)
once transmission has stopped. For test purposes, the register decrements every rx_clk cycle
once transmission has stopped if bit 12 (retry test) is set in the network configuration register. If
the pause enable bit (13) is not set in the network configuration register, then the decrementing
occurs regardless of whether transmission has stopped or not.
An interrupt (13) is asserted whenever the pause time register decrements to zero (assuming it
is enabled in the interrupt mask register). Automatic transmission of pause frames is supported
through the transmit pause frame bits of the network control register and the tx_pause and
tx_pause_zero inputs. If either bit 11 or bit 12 of the network control register is written to with
a 1, or if the input signal tx_pause is toggled, a pause frame is transmitted only if full duplex is
selected in the network configuration register and transmit is enabled in the network control
register.
Pause frame transmission occurs immediately if transmit is inactive or if transmit is active
between the current frame and the next frame due to be transmitted. The transmitted pause
frame is comprised of the items in the following list:
• a destination address of 01-80-C2-00-00-01
• a source address taken from the specific address 1 register
• a type ID of 88-08 (MAC control frame)
• a pause opcode of 00-01
• a pause quantum
• fill of 00 to take the frame to minimum frame length
• valid FCS
The pause quantum used in the generated frame depends on the trigger source for the frame as
follows:
1. If bit 11 is written with a one, the pause quantum comes from the transmit pause quantum register. The Transmit Pause Quantum register resets to a value of 0xFFFF giving
a maximum pause quantum as a default.
2. If bit 12 is written with a one, the pause quantum is zero.
3. If the tx_pause input is toggled and the tx_pause_zero input is held low until the
next toggle, the pause quantum comes from the transmit pause quantum register.
4. If the tx_pause input is toggled and the tx_pause_zero input is held high until the
next toggle, the pause quantum is zero.
445
32058K
AVR32-01/12
AT32UC3A
After transmission, no interrupts are generated and the only statistics register that is incremented is the pause frames transmitted register.
29.5.4
Receive Block
The receive block checks for valid preamble, FCS, alignment and length, presents received
frames to the DMA block and stores the frames destination address for use by the address
checking block. If, during frame reception, the frame is found to be too long or rx_er is asserted,
a bad frame indication is sent to the DMA block. The DMA block then ceases sending data to
memory. At the end of frame reception, the receive block indicates to the DMA block whether the
frame is good or bad. The DMA block recovers the current receive buffer if the frame was bad.
The receive block signals the register block to increment the alignment error, the CRC (FCS)
error, the short frame, long frame, jabber error, the receive symbol error statistics and the length
field mismatch statistics.
The enable bit for jumbo frames in the network configuration register allows the MACB to receive
jumbo frames of up to 10240 bytes in size. This operation does not form part of the IEEE802.3
specification and is disabled by default. When jumbo frames are enabled, frames received with a
frame size greater than 10240 bytes are discarded.
29.5.5
Address Checking Block
The address checking (or filter) block indicates to the DMA block which receive frames should
be copied to memory. Whether a frame is copied depends on what is enabled in the network
configuration register, the state of the external match pin, the contents of the specific address
and hash registers and the frame’s destination address. In this implementation of the MACB, the
frame’s source address is not checked. Provided that bit 18 of the Network Configuration register is not set, a frame is not copied to memory if the MACB is transmitting in half duplex mode at
the time a destination address is received. If bit 18 of the Network Configuration register is set,
frames can be received while transmitting in half-duplex mode.
Ethernet frames are transmitted a byte at a time, least significant bit first. The first six bytes (48
bits) of an Ethernet frame make up the destination address. The first bit of the destination
address, the LSB of the first byte of the frame, is the group/individual bit: this is One for multicast
addresses and Zero for unicast. The All Ones address is the broadcast address, and a special
case of multicast.
The MACB supports recognition of four specific addresses. Each specific address requires two
registers, specific address register bottom and specific address register top. Specific address
register bottom stores the first four bytes of the destination address and specific address register
top contains the last two bytes. The addresses stored can be specific, group, local or universal.
The destination address of received frames is compared against the data stored in the specific
address registers once they have been activated. The addresses are deactivated at reset or
when their corresponding specific address register bottom is written. They are activated when
specific address register top is written. If a receive frame address matches an active address,
the frame is copied to memory.
446
32058K
AVR32-01/12
AT32UC3A
The following example illustrates the use of the address match registers for a MAC address of
21:43:65:87:A9:CB.
Preamble 55
SFD D5
DA (Octet0 - LSB) 21
DA(Octet 1) 43
DA(Octet 2) 65
DA(Octet 3) 87
DA(Octet 4) A9
DA (Octet5 - MSB) CB
SA (LSB) 00
SA 00
SA 00
SA 00
SA 00
SA (MSB) 43
SA (LSB) 21
The sequence above shows the beginning of an Ethernet frame. Byte order of transmission is
from top to bottom as shown. For a successful match to specific address 1, the following
address matching registers must be set up:
• Base address + 0x98 0x87654321 (Bottom)
• Base address + 0x9C 0x0000CBA9 (Top)
And for a successful match to the Type ID register, the following should be set up:
• Base address + 0xB8 0x00004321
29.5.6
Broadcast Address
The broadcast address of 0xFFFFFFFFFFFF is recognized unless the ‘no broadcast’ bit in the
network configuration register is set.
29.5.7
Hash Addressing
The hash address register is 64 bits long and takes up two locations in the memory map. The
least significant bits are stored in hash register bottom and the most significant bits in hash register top.
The unicast hash enable and the multicast hash enable bits in the network configuration register
enable the reception of hash matched frames. The destination address is reduced to a 6-bit
index into the 64-bit hash register using the following hash function. The hash function is an
exclusive or of every sixth bit of the destination address.
447
32058K
AVR32-01/12
AT32UC3A
hash_index[5] = da[5] ^ da[11] ^ da[17] ^ da[23] ^ da[29] ^ da[35] ^ da[41] ^ da[47]
hash_index[4] = da[4] ^ da[10] ^ da[16] ^ da[22] ^ da[28] ^ da[34] ^ da[40] ^ da[46]
hash_index[3] = da[3] ^ da[09] ^ da[15] ^ da[21] ^ da[27] ^ da[33] ^ da[39] ^ da[45]
hash_index[2] = da[2] ^ da[08] ^ da[14] ^ da[20] ^ da[26] ^ da[32] ^ da[38] ^ da[44]
hash_index[1] = da[1] ^ da[07] ^ da[13] ^ da[19] ^ da[25] ^ da[31] ^ da[37] ^ da[43]
hash_index[0] = da[0] ^ da[06] ^ da[12] ^ da[18] ^ da[24] ^ da[30] ^ da[36] ^ da[42]
da[0] represents the least significant bit of the first byte received, that is, the multicast/unicast
indicator, and da[47] represents the most significant bit of the last byte received.
If the hash index points to a bit that is set in the hash register, then the frame is matched according to whether the frame is multicast or unicast.
A multicast match is signalled if the multicast hash enable bit is set. da[0] is 1 and the hash index
points to a bit set in the hash register.
A unicast match is signalled if the unicast hash enable bit is set. da[0] is 0 and the hash index
points to a bit set in the hash register.
To receive all multicast frames, the hash register should be set with all ones and the multicast
hash enable bit should be set in the network configuration register.
29.5.8
External Address Matching
The external address signal (eam) is enabled by bit 9 in the network configuration register.
When enabled, the filter block sends the store frame and the external address match status signal to the DMA block if the external address match signal is asserted (from a source external to
the MACB) and the destination address has been received and the frame has not completed.
For the DMA block to be able to copy the frame to memory, the external address signal must be
asserted before four words have been loaded into the receive FIFO.
29.5.9
Copy All Frames (or Promiscuous Mode)
If the copy all frames bit is set in the network configuration register, then all non-errored frames
are copied to memory. For example, frames that are too long, too short, or have FCS errors or
rx_er asserted during reception are discarded and all others are received. Frames with FCS
errors are copied to memory if bit 19 in the network configuration register is set.
29.5.10
Type ID Checking
The contents of the type_id register are compared against the length/type ID of received frames
(i.e., bytes 13 and 14). Bit 22 in the receive buffer descriptor status is set if there is a match. The
reset state of this register is zero which is unlikely to match the length/type ID of any valid Ethernet frame.
Note:
A type ID match does not affect whether a frame is copied to memory.
448
32058K
AVR32-01/12
AT32UC3A
29.5.11
VLAN Support
An Ethernet encoded 802.1Q VLAN tag looks like this:
Table 29-4.
802.1Q VLAN Tag
TPID (Tag Protocol Identifier) 16 bits
TCI (Tag Control Information) 16 bits
0x8100
First 3 bits priority, then CFI bit, last 12 bits VID
The VLAN tag is inserted at the 13th byte of the frame, adding an extra four bytes to the frame. If
the VID (VLAN identifier) is null (0x000), this indicates a priority-tagged frame. The MAC can
support frame lengths up to 1536 bytes, 18 bytes more than the original Ethernet maximum
frame length of 1518 bytes. This is achieved by setting bit 8 in the network configuration register.
The following bits in the receive buffer descriptor status word give information about VLAN
tagged frames:
• Bit 21 set if receive frame is VLAN tagged (i.e. type id of 0x8100)
• Bit 20 set if receive frame is priority tagged (i.e. type id of 0x8100 and null VID). (If bit 20 is
set bit 21 is set also.)
• Bit 19, 18 and 17 set to priority if bit 21 is set
• Bit 16 set to CFI if bit 21 is set
29.5.12
PHY Maintenance
The register MAN enables the MACB to communicate with a PHY by means of the MDIO interface. It is used during auto-negotiation to ensure that the MACB and the PHY are configured for
the same speed and duplex configuration.
The PHY maintenance register is implemented as a shift register. Writing to the register starts a
shift operation which is signalled as complete when bit two is set in the network status register
(about 2000 MCK cycles later when bit ten is set to zero, and bit eleven is set to one in the network configuration register). An interrupt is generated as this bit is set. During this time, the MSB
of the register is output on the MDIO pin and the LSB updated from the MDIO pin with each
MDC cycle. This causes transmission of a PHY management frame on MDIO.
Reading during the shift operation returns the current contents of the shift register. At the end of
management operation, the bits have shifted back to their original locations. For a read operation, the data bits are updated with data read from the PHY. It is important to write the correct
values to the register to ensure a valid PHY management frame is produced.
The MDIO interface can read IEEE 802.3 clause 45 PHYs as well as clause 22 PHYs. To read
clause 45 PHYs, bits[31:28] should be written as 0x0011. For a description of MDC generation,
see the network configuration register in the ”Network Control Register” on page 456.
29.5.13
Media Independent Interface
The Ethernet MAC is capable of interfacing to both RMII and MII Interfaces. The RMII bit in the
USRIO register controls the interface that is selected. When this bit is set, the RMII interface is
selected, else the MII interface is selected.
449
32058K
AVR32-01/12
AT32UC3A
The MII and RMII interface are capable of both 10Mb/s and 100Mb/s data rates as described in
the IEEE 802.3u standard. The signals used by the MII and RMII interfaces are described in
Table 29-5.
Pin Configuration
Table 29-5.
Pin Name
MII
RMII
ETXCK_EREFCK
ETXCK: Transmit Clock
EREFCK: Reference Clock
ECRS
ECRS: Carrier Sense
ECOL
ECOL: Collision Detect
ERXDV
ERXDV: Data Valid
ECRSDV: Carrier Sense/Data Valid
ERX0 - ERX3
ERX0 - ERX3: 4-bit Receive Data
ERX0 - ERX1: 2-bit Receive Data
ERXER
ERXER: Receive Error
ERXER: Receive Error
ERXCK
ERXCK: Receive Clock
ETXEN
ETXEN: Transmit Enable
ETXEN: Transmit Enable
ETX0-ETX3
ETX0 - ETX3: 4-bit Transmit Data
ETX0 - ETX1: 2-bit Transmit Data
ETXER
ETXER: Transmit Error
The intent of the RMII is to provide a reduced pin count alternative to the IEEE 802.3u MII. It
uses 2 bits for transmit (ETX0 and ETX1) and two bits for receive (ERX0 and ERX1). There is a
Transmit Enable (ETXEN), a Receive Error (ERXER), a Carrier Sense (ECRS_DV), and a 50
MHz Reference Clock (ETXCK_EREFCK) for 100Mb/s data rate.
29.5.13.1
RMII Transmit and Receive Operation
The same signals are used internally for both the RMII and the MII operations. The RMII maps
these signals in a more pin-efficient manner. The transmit and receive bits are converted from a
4-bit parallel format to a 2-bit parallel scheme that is clocked at twice the rate. The carrier sense
and data valid signals are combined into the ECRSDV signal. This signal contains information
on carrier sense, FIFO status, and validity of the data. Transmit error bit (ETXER) and collision
detect (ECOL) are not used in RMII mode.
450
32058K
AVR32-01/12
AT32UC3A
29.6
Programming Interface
29.6.1
Initialization
29.6.1.1
Configuration
Initialization of the MACB configuration (e.g. frequency ratios) must be done while the transmit
and receive circuits are disabled. See the description of the network control register and network
configuration register later in this document.
29.6.1.2
Receive Buffer List
Receive data is written to areas of data (i.e., buffers) in system memory. These buffers are listed
in another data structure that also resides in main memory. This data structure (receive buffer
queue) is a sequence of descriptor entries as defined in ”Receive Buffer Descriptor Entry” on
page 440. It points to this data structure.
Figure 29-2. Receive Buffer List
Receive Buffer 0
Receive Buffer Queue Pointer
(MAC Register)
Receive Buffer 1
Receive Buffer N
Receive Buffer Descriptor List
(In memory)
(In memory)
To create the list of buffers:
1. Allocate a number (n) of buffers of 128 bytes in system memory.
2. Allocate an area 2n words for the receive buffer descriptor entry in system memory and
create n entries in this list. Mark all entries in this list as owned by MACB, i.e., bit 0 of
word 0 set to 0.
3. If less than 1024 buffers are defined, the last descriptor must be marked with the wrap bit
(bit 1 in word 0 set to 1).
4. Write address of receive buffer descriptor entry to MACB register receive_buffer queue
pointer.
5. The receive circuits can then be enabled by writing to the address recognition registers
and then to the network control register.
451
32058K
AVR32-01/12
AT32UC3A
29.6.1.3
Transmit Buffer List
Transmit data is read from the system memory These buffers are listed in another data structure
that also resides in main memory. This data structure (Transmit Buffer Queue) is a sequence of
descriptor entries (as defined in Table 29-2 on page 443) that points to this data structure.
To create this list of buffers:
1. Allocate a number (n) of buffers of between 1 and 2047 bytes of data to be transmitted
in system memory. Up to 128 buffers per frame are allowed.
2. Allocate an area 2n words for the transmit buffer descriptor entry in system memory
and create N entries in this list. Mark all entries in this list as owned by MACB, i.e. bit 31
of word 1 set to 0.
3. If fewer than 1024 buffers are defined, the last descriptor must be marked with the wrap
bit — bit 30 in word 1 set to 1.
4. Write address of transmit buffer descriptor entry to MACB register transmit_buffer
queue pointer.
5. The transmit circuits can then be enabled by writing to the network control register.
29.6.1.4
Address Matching
The MACB register-pair hash address and the four specific address register-pairs must be written with the required values. Each register-pair comprises a bottom register and top register,
with the bottom register being written first. The address matching is disabled for a particular register-pair after the bottom-register has been written and re-enabled when the top register is
written. See Section “29.5.5” on page 446. for details of address matching. Each register-pair
may be written at any time, regardless of whether the receive circuits are enabled or disabled.
29.6.1.5
Interrupts
There are 14 interrupt conditions that are detected within the MACB. These are ORed to make a
single interrupt. This interrupt is passed to the interrupt controller. On receipt of the interrupt signal, the CPU enters the interrupt handler. To ascertain which interrupt has been generated, read
the interrupt status register. Note that this register clears itself when read. At reset, all interrupts
are disabled. To enable an interrupt, write to interrupt enable register with the pertinent interrupt
bit set to 1. To disable an interrupt, write to interrupt disable register with the pertinent interrupt
bit set to 1. To check whether an interrupt is enabled or disabled, read interrupt mask register: if
the bit is set to 1, the interrupt is disabled.
29.6.1.6
Transmitting Frames
To set up a frame for transmission:
1. Enable transmit in the network control register.
2. Allocate an area of system memory for transmit data. This does not have to be contiguous, varying byte lengths can be used as long as they conclude on byte borders.
3. Set-up the transmit buffer list.
4. Set the network control register to enable transmission and enable interrupts.
5. Write data for transmission into these buffers.
6. Write the address to transmit buffer descriptor queue pointer.
7. Write control and length to word one of the transmit buffer descriptor entry.
8. Write to the transmit start bit in the network control register.
452
32058K
AVR32-01/12
AT32UC3A
29.6.1.7
Receiving Frames
When a frame is received and the receive circuits are enabled, the MACB checks the address
and, in the following cases, the frame is written to system memory:
• if it matches one of the four specific address registers.
• if it matches the hash address function.
• if it is a broadcast address (0xFFFFFFFFFFFF) and broadcasts are allowed.
• if the MACB is configured to copy all frames.
• if the EAM is asserted before four words have been loaded into the receive FIFO.
The register receive buffer queue pointer points to the next entry (see Table 29-1 on page 440)
and the MACB uses this as the address in system memory to write the frame to. Once the frame
has been completely and successfully received and written to system memory, the MACB then
updates the receive buffer descriptor entry with the reason for the address match and marks the
area as being owned by software. Once this is complete an interrupt receive complete is set.
Software is then responsible for handling the data in the buffer and then releasing the buffer by
writing the ownership bit back to 0.
If the MACB is unable to write the data at a rate to match the incoming frame, then an interrupt
receive overrun is set. If there is no receive buffer available, i.e., the next buffer is still owned by
software, the interrupt receive buffer not available is set. If the frame is not successfully
received, a statistic register is incremented and the frame is discarded without informing
software.
453
32058K
AVR32-01/12
AT32UC3A
29.7
Ethernet MAC (MACB) User Interface
Table 29-6.
Ethernet MAC (MACB) Register Mapping
Offset
Register
Name
Access
Reset Value
0x00
Network Control Register
NCR
Read/Write
0
0x04
Network Configuration Register
NCFG
Read/Write
0x800
0x08
Network Status Register
NSR
Read-only
-
0x0C
Reserved
0x10
Reserved
0x14
Transmit Status Register
TSR
Read/Write
0x0000_0000
0x18
Receive Buffer Queue Pointer Register
RBQP
Read/Write
0x0000_0000
0x1C
Transmit Buffer Queue Pointer Register
TBQP
Read/Write
0x0000_0000
0x20
Receive Status Register
RSR
Read/Write
0x0000_0000
0x24
Interrupt Status Register
ISR
Read/Write
0x0000_0000
0x28
Interrupt Enable Register
IER
Write-only
-
0x2C
Interrupt Disable Register
IDR
Write-only
-
0x30
Interrupt Mask Register
IMR
Read-only
0x0000_3FFF
0x34
Phy Maintenance Register
MAN
Read/Write
0x0000_0000
0x38
Pause Time Register
PTR
Read/Write
0x0000_0000
0x3C
Pause Frames Received Register
PFR
Read/Write
0x0000_0000
0x40
Frames Transmitted Ok Register
FTO
Read/Write
0x0000_0000
0x44
Single Collision Frames Register
SCF
Read/Write
0x0000_0000
0x48
Multiple Collision Frames Register
MCF
Read/Write
0x0000_0000
0x4C
Frames Received Ok Register
FRO
Read/Write
0x0000_0000
0x50
Frame Check Sequence Errors Register
FCSE
Read/Write
0x0000_0000
0x54
Alignment Errors Register
ALE
Read/Write
0x0000_0000
0x58
Deferred Transmission Frames Register
DTF
Read/Write
0x0000_0000
0x5C
Late Collisions Register
LCOL
Read/Write
0x0000_0000
0x60
Excessive Collisions Register
EXCOL
Read/Write
0x0000_0000
0x64
Transmit Underrun Errors Register
TUND
Read/Write
0x0000_0000
0x68
Carrier Sense Errors Register
CSE
Read/Write
0x0000_0000
0x6C
Receive Resource Errors Register
RRE
Read/Write
0x0000_0000
0x70
Receive Overrun Errors Register
ROV
Read/Write
0x0000_0000
0x74
Receive Symbol Errors Register
RSE
Read/Write
0x0000_0000
0x78
Excessive Length Errors Register
ELE
Read/Write
0x0000_0000
0x7C
Receive Jabbers Register
RJA
Read/Write
0x0000_0000
0x80
Undersize Frames Register
USF
Read/Write
0x0000_0000
0x84
SQE Test Errors Register
STE
Read/Write
0x0000_0000
0x88
Received Length Field Mismatch Register
RLE
Read/Write
0x0000_0000
454
32058K
AVR32-01/12
AT32UC3A
Table 29-6.
Ethernet MAC (MACB) Register Mapping (Continued)
Offset
Register
Name
Access
Reset Value
0x8C
Transmitted Pause Frames Register
TPF
Read/Write
0x0000_0000
0x90
Hash Register Bottom [31:0] Register
HRB
Read/Write
0x0000_0000
0x94
Hash Register Top [63:32] Register
HRT
Read/Write
0x0000_0000
0x98
Specific Address 1 Bottom Register
SA1B
Read/Write
0x0000_0000
0x9C
Specific Address 1 Top Register
SA1T
Read/Write
0x0000_0000
0xA0
Specific Address 2 Bottom Register
SA2B
Read/Write
0x0000_0000
0xA4
Specific Address 2 Top Register
SA2T
Read/Write
0x0000_0000
0xA8
Specific Address 3 Bottom Register
SA3B
Read/Write
0x0000_0000
0xAC
Specific Address 3 Top Register
SA3T
Read/Write
0x0000_0000
0xB0
Specific Address 4 Bottom Register
SA4B
Read/Write
0x0000_0000
0xB4
Specific Address 4 Top Register
SA4T
Read/Write
0x0000_0000
0xB8
Type ID Checking Register
TID
Read/Write
0x0000_0000
0xBC
Transmit Pause Quantum Register
TPQ
Read/Write
0x0000_FFFF
0xC0
User Input/output Register
USRIO
Read/Write
0x0000_0000
0xC4
Wake on LAN Register
WOL
Read/Write
0x0000_0000
0xC8 - 0xFC
Reserved
–
–
–
455
32058K
AVR32-01/12
AT32UC3A
29.7.1
Network Control Register
Register Name:
NCR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
TZQ
11
TPF
10
THALT
9
TSTART
8
BP
7
WESTAT
6
INCSTAT
5
CLRSTAT
4
MPE
3
TE
2
RE
1
LLB
0
LB
• LB: LoopBack
Asserts the loopback signal to the PHY.
• LLB: LoopBack Local
connects txd to rxd, tx_en to rx_dv, forces full duplex and drives rx_clk and tx_clk with pclk divided by 4. rx_clk and tx_clk
may glitch as the MACB is switched into and out of internal loop back. It is important that receive and transmit circuits have
already been disabled when making the switch into and out of internal loop back. This function may not be supported by
some instantiations of the MACB.
• RE: Receive enable
When set, enables the MACB to receive data. When reset, frame reception stops immediately and the receive FIFO is
cleared. The receive queue pointer register is unaffected.
• TE: Transmit enable
When set, enables the Ethernet transmitter to send data. When reset, transmission stops immediately, the transmit FIFO
and control registers are cleared and the transmit queue pointer register resets to point to the start of the transmit descriptor list.
• MPE: Management port enable
Set to one to enable the management port. When zero, forces MDIO to high impedance state and MDC low.
• CLRSTAT: Clear statistics registers
This bit is write only. Writing a one clears the statistics registers.
• INCSTAT: Increment statistics registers
This bit is write only. Writing a one increments all the statistics registers by one for test purposes.
• WESTAT: Write enable for statistics registers
Setting this bit to one makes the statistics registers writable for functional test purposes.
• BP: Back pressure
If set in half duplex mode, forces collisions on all received frames.
456
32058K
AVR32-01/12
AT32UC3A
• TSTART: Start transmission
Writing one to this bit starts transmission.
• THALT: Transmit halt
Writing one to this bit halts transmission as soon as any ongoing frame transmission ends.
• TPF: Transmit pause frame
Writing one to this bit transmits a pause frame with the pause quantum from the transmit pause quantum register at the
next available transmitter idle time.
• TZQ: Transmit zero quantum pause frame
Writing a one to this bit transmits a pause frame with zero pause quantum at the next available transmitter idle time.
457
32058K
AVR32-01/12
AT32UC3A
29.7.2
Network Configuration Register
Register Name:
NCFGR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
IRXFCS
18
EFRHD
17
DRFCS
16
RLCE
15
14
13
PAE
12
RTY
11
10
9
EAE
8
BIG
5
NBC
4
CAF
3
JFRAME
2
Bit rate
1
FD
0
SPD
RBOF
7
UNI
6
MTI
CLK
• SPD: Speed
Set to 1 to indicate 100 Mbit/s operation, 0 for 10 Mbit/s. The value of this pin is reflected on the speed pin.
• FD: Full Duplex
If set to 1, the transmit block ignores the state of collision and carrier sense and allows receive while transmitting. Also controls the half_duplex pin.
• Bit rate:
If set to 1 to configure the interface for serial operation. Must be set before receive and transmit enable in the network control register. If set a serial interface is configured with transmit and receive data being driven out on txd[0] and received on
rxd[0] serially. Also the crs and rx_dv are logically ORed together so either may be used as the data valid signal.
• CAF: Copy All Frames
When set to 1, all valid frames are received.
• JFRAME: Jumbo Frames
Set to one to enable jumbo frames of up to 10240 bytes to be accepted.
• NBC: No Broadcast
When set to 1, frames addressed to the broadcast address of all ones are not received.
• MTI: Multicast Hash Enable
When set, multicast frames are received when the 6-bit hash function of the destination address points to a bit that is set in
the hash register.
• UNI: Unicast Hash Enable
When set, unicast frames are received when the 6-bit hash function of the destination address points to a bit that is set in
the hash register.
• BIG: Receive 1536 bytes frames
Setting this bit means the MACB receives frames up to 1536 bytes in length. Normally, the MACB would reject any frame
above 1518 bytes.
• EAE: External address match enable
When set, the eam pin can be used to copy frames to memory.
458
32058K
AVR32-01/12
AT32UC3A
• CLK: MDC clock divider
Set according to system clock speed. This determines by what number system clock is divided to generate MDC.
For conformance with 802.3, MDC must not exceed 2.5MHz (MDC is only active during MDIO read and write operations).
CLK
MDC
00
MCK divided by 8 (MCK up to 20 MHz)
01
MCK divided by 16 (MCK up to 40 MHz)
10
MCK divided by 32 (MCK up to 80 MHz)
11
MCK divided by 64 (MCK up to 160 MHz)
• RTY: Retry test
Must be set to zero for normal operation. If set to one, the back off between collisions is always one slot time. Setting this
bit to one helps testing the too many retries condition. Also used in the pause frame tests to reduce the pause counters
decrement time from 512 bit times, to every rx_clk cycle.
• PAE: Pause Enable
When set, transmission pauses when a valid pause frame is received.
• RBOF: Receive Buffer Offset
Indicates the number of bytes by which the received data is offset from the start of the first receive buffer.
RBOF
Offset
00
No offset from start of receive buffer
01
One-byte offset from start of receive buffer
10
Two-byte offset from start of receive buffer
11
Three-byte offset from start of receive buffer
• RLCE: Receive Length field Checking Enable
When set, frames with measured lengths shorter than their length fields are discarded. Frames containing a type ID in
bytes 13 and 14 — length/type ID = 0600 — are not be counted as length errors.
• DRFCS: Discard Receive FCS
When set, the FCS field of received frames will not be copied to memory.
• EFRHD:
Enable Frames to be received in half-duplex mode while transmitting.
• IRXFCS: Ignore RX FCS
When set, frames with FCS/CRC errors are not rejected and no FCS error statistics are counted. For normal operation, this
bit must be set to 0.
459
32058K
AVR32-01/12
AT32UC3A
29.7.3
Network Status Register
Register Name:
NSR
Access Type:
Read-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
–
5
–
4
–
3
–
2
IDLE
1
MDIO
0
-
• MDIO
Returns status of the mdio_in pin. Use the PHY maintenance register for reading managed frames rather than this bit.
• IDLE
0 = The PHY logic is running.
1 = The PHY management logic is idle (i.e., has completed).
460
32058K
AVR32-01/12
AT32UC3A
29.7.4
Transmit Status Register
Register Name:
TSR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
UND
5
COMP
4
BEX
3
TGO
2
RLE
1
COL
0
UBR
This register, when read, provides details of the status of a transmit. Once read, individual bits may be cleared by writing 1
to them. It is not possible to set a bit to 1 by writing to the register.
• UBR: Used Bit Read
Set when a transmit buffer descriptor is read with its used bit set. Cleared by writing a one to this bit.
• COL: Collision Occurred
Set by the assertion of collision. Cleared by writing a one to this bit.
• RLE: Retry Limit exceeded
Cleared by writing a one to this bit.
• TGO: Transmit Go
If high transmit is active.
• BEX: Buffers exhausted mid frame
If the buffers run out during transmission of a frame, then transmission stops, FCS shall be bad and tx_er asserted. Cleared
by writing a one to this bit.
• COMP: Transmit Complete
Set when a frame has been transmitted. Cleared by writing a one to this bit.
• UND: Transmit Underrun
Set when transmit DMA was not able to read data from memory, either because the bus was not granted in time, because
a not OK hresp(bus error) was returned or because a used bit was read midway through frame transmission. If this
occurs, the transmitter forces bad CRC. Cleared by writing a one to this bit.
461
32058K
AVR32-01/12
AT32UC3A
29.7.5
Receive Buffer Queue Pointer Register
Register Name:
RBQP
Access Type:
31
Read/Write
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
–
0
–
ADDR
23
22
21
20
ADDR
15
14
13
12
ADDR
7
6
5
4
ADDR
This register points to the entry in the receive buffer queue (descriptor list) currently being used. It is written with the start
location of the receive buffer descriptor list. The lower order bits increment as buffers are used up and wrap to their original
values after either 1024 buffers or when the wrap bit of the entry is set.
Reading this register returns the location of the descriptor currently being accessed. This value increments as buffers are
used. Software should not use this register for determining where to remove received frames from the queue as it constantly changes as new frames are received. Software should instead work its way through the buffer descriptor queue
checking the used bits.
Receive buffer writes also comprise bursts of two words and, as with transmit buffer reads, it is recommended that bit 2 is
always written with zero to prevent a burst crossing a 1K boundary, in violation of the System Bus specification.
• ADDR: Receive buffer queue pointer address
Written with the address of the start of the receive queue, reads as a pointer to the current buffer being used.
462
32058K
AVR32-01/12
AT32UC3A
29.7.6
Transmit Buffer Queue Pointer Register
Register Name:
TBQP
Access Type:
31
Read/Write
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
–
0
–
ADDR
23
22
21
20
ADDR
15
14
13
12
ADDR
7
6
5
4
ADDR
This register points to the entry in the transmit buffer queue (descriptor list) currently being used. It is written with the start
location of the transmit buffer descriptor list. The lower order bits increment as buffers are used up and wrap to their original
values after either 1024 buffers or when the wrap bit of the entry is set. This register can only be written when bit 3 in the
transmit status register is low.
As transmit buffer reads consist of bursts of two words, it is recommended that bit 2 is always written with zero to prevent a
burst crossing a 1K boundary, in violation of the System Bus specification.
• ADDR: Transmit buffer queue pointer address
Written with the address of the start of the transmit queue, reads as a pointer to the first buffer of the frame being transmitted or about to be transmitted.
463
32058K
AVR32-01/12
AT32UC3A
29.7.7
Receive Status Register
Register Name:
RSR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
–
5
–
4
–
3
–
2
OVR
1
REC
0
BNA
This register, when read, provides details of the status of a receive. Once read, individual bits may be cleared by writing 1
to them. It is not possible to set a bit to 1 by writing to the register.
• BNA: Buffer Not Available
An attempt was made to get a new buffer and the pointer indicated that it was owned by the processor. The DMA rereads
the pointer each time a new frame starts until a valid pointer is found. This bit is set at each attempt that fails even if it has
not had a successful pointer read since it has been cleared.
Cleared by writing a one to this bit.
• REC: Frame Received
One or more frames have been received and placed in memory. Cleared by writing a one to this bit.
• OVR: Receive Overrun
The DMA block was unable to store the receive frame to memory, either because the bus was not granted in time or
because a not OK hresp(bus error) was returned. The buffer is recovered if this happens.
Cleared by writing a one to this bit.
464
32058K
AVR32-01/12
AT32UC3A
29.7.8
Interrupt Status Register
Register Name:
ISR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
PTZ
12
PFR
11
HRESP
10
ROVR
9
-
8
–
7
TCOMP
6
TXERR
5
RLE
4
TUND
3
TXUBR
2
RXUBR
1
RCOMP
0
MFD
• MFD: Management Frame Done
The PHY maintenance register has completed its operation. Cleared on read.
• RCOMP: Receive Complete
A frame has been stored in memory. Cleared on read.
• RXUBR: Receive Used Bit Read
Set when a receive buffer descriptor is read with its used bit set. Cleared on read.
• TXUBR: Transmit Used Bit Read
Set when a transmit buffer descriptor is read with its used bit set. Cleared on read.
• TUND: Ethernet Transmit Buffer Underrun
The transmit DMA did not fetch frame data in time for it to be transmitted or hresp returned not OK. Also set if a used bit
is read mid-frame or when a new transmit queue pointer is written. Cleared on read.
• RLE: Retry Limit Exceeded
Cleared on read.
• TXERR: Transmit Error
Transmit buffers exhausted in mid-frame - transmit error. Cleared on read.
• TCOMP: Transmit Complete
Set when a frame has been transmitted. Cleared on read.
• ROVR: Receive Overrun
Set when the receive overrun status bit gets set. Cleared on read.
• HRESP: Hresp not OK
Set when the DMA block sees a bus error. Cleared on read.
• PFR: Pause Frame Received
Indicates a valid pause has been received. Cleared on a read.
• PTZ: Pause Time Zero
Set when the pause time register, 0x38 decrements to zero. Cleared on a read.
465
32058K
AVR32-01/12
AT32UC3A
29.7.9
Interrupt Enable Register
Register Name:
IER
466
32058K
AVR32-01/12
AT32UC3A
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
PTZ
12
PFR
11
HRESP
10
ROVR
9
8
–
7
TCOMP
6
TXERR
5
RLE
4
TUND
3
TXUBR
2
RXUBR
1
RCOMP
0
MFD
• MFD: Management Frame sent
Enable management done interrupt.
• RCOMP: Receive Complete
Enable receive complete interrupt.
• RXUBR: Receive Used Bit Read
Enable receive used bit read interrupt.
• TXUBR: Transmit Used Bit Read
Enable transmit used bit read interrupt.
• TUND: Ethernet Transmit Buffer Underrun
Enable transmit underrun interrupt.
• RLE: Retry Limit Exceeded
Enable retry limit exceeded interrupt.
• TXERR: Transmit Error
Enable transmit buffers exhausted in mid-frame interrupt.
• TCOMP: Transmit Complete
Enable transmit complete interrupt.
• ROVR: Receive Overrun
Enable receive overrun interrupt.
• HRESP: Hresp not OK
Enable Hresp not OK interrupt.
• PFR: Pause Frame Received
Enable pause frame received interrupt.
• PTZ: Pause Time Zero
Enable pause time zero interrupt.
467
32058K
AVR32-01/12
AT32UC3A
29.7.10 Interrupt Disable Register
Register Name:
IDR
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
PTZ
12
PFR
11
HRESP
10
ROVR
9
-
8
–
7
TCOMP
6
TXERR
5
RLE
4
TUND
3
TXUBR
2
RXUBR
1
RCOMP
0
MFD
• MFD: Management Frame sent
Disable management done interrupt.
• RCOMP: Receive Complete
Disable receive complete interrupt.
• RXUBR: Receive Used Bit Read
Disable receive used bit read interrupt.
• TXUBR: Transmit Used Bit Read
Disable transmit used bit read interrupt.
• TUND: Ethernet Transmit Buffer Underrun
Disable transmit underrun interrupt.
• RLE: Retry Limit Exceeded
Disable retry limit exceeded interrupt.
• TXERR: Transmit Error
Disable transmit buffers exhausted in mid-frame interrupt.
• TCOMP: Transmit Complete
Disable transmit complete interrupt.
• ROVR: Receive Overrun
Disable receive overrun interrupt.
• HRESP: Hresp not OK
Disable Hresp not OK interrupt.
• PFR: Pause Frame Received
Disable pause frame received interrupt.
• PTZ: Pause Time Zero
Disable pause time zero interrupt.
468
32058K
AVR32-01/12
AT32UC3A
29.7.11 Interrupt Mask Register
Register Name:
IMR
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
PTZ
12
PFR
11
HRESP
10
ROVR
9
-
8
–
7
TCOMP
6
TXERR
5
RLE
4
TUND
3
TXUBR
2
RXUBR
1
RCOMP
0
MFD
• MFD: Management Frame sent
Management done interrupt masked.
• RCOMP: Receive Complete
Receive complete interrupt masked.
• RXUBR: Receive Used Bit Read
Receive used bit read interrupt masked.
• TXUBR: Transmit Used Bit Read
Transmit used bit read interrupt masked.
• TUND: Ethernet Transmit Buffer Underrun
Transmit underrun interrupt masked.
• RLE: Retry Limit Exceeded
Retry limit exceeded interrupt masked.
• TXERR: Transmit Error
Transmit buffers exhausted in mid-frame interrupt masked.
• TCOMP: Transmit Complete
Transmit complete interrupt masked.
• ROVR: Receive Overrun
Receive overrun interrupt masked.
• HRESP: Hresp not OK
Hresp not OK interrupt masked.
• PFR: Pause Frame Received
Pause frame received interrupt masked.
• PTZ: Pause Time Zero
Pause time zero interrupt masked.
469
32058K
AVR32-01/12
AT32UC3A
29.7.12 PHY Maintenance Register
Register Name:
MAN
Access Type:
31
Read/Write
30
29
SOF
28
27
26
RW
23
PHYA
22
15
14
21
13
25
24
17
16
PHYA
20
REGA
19
18
12
11
10
9
8
3
2
1
0
CODE
DATA
7
6
5
4
DATA
• DATA
For a write operation this is written with the data to be written to the PHY.
After a read operation this contains the data read from the PHY.
• CODE:
Must be written to 10. Reads as written.
• REGA: Register Address
Specifies the register in the PHY to access.
• PHYA: PHY Address
• RW: Read/Write
10 is read; 01 is write. Any other value is an invalid PHY management frame
• SOF: Start of frame
Must be written 01 for a valid frame.
470
32058K
AVR32-01/12
AT32UC3A
29.7.13 Pause Time Register
Register Name:
PTR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
PTIME
7
6
5
4
PTIME
• PTIME: Pause Time
Stores the current value of the pause time register which is decremented every 512 bit times.
471
32058K
AVR32-01/12
AT32UC3A
29.7.14 Hash Register Bottom
Register Name:
HRB
Access Type:
31
Read/Write
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
ADDR
23
22
21
20
ADDR
15
14
13
12
ADDR
7
6
5
4
ADDR
• ADDR:
Bits 31:0 of the hash address register. See ”Hash Addressing” on page 447.
472
32058K
AVR32-01/12
AT32UC3A
29.7.15 Hash Register Top
Register Name:
HRT
Access Type:
31
Read/Write
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
ADDR
23
22
21
20
ADDR
15
14
13
12
ADDR
7
6
5
4
ADDR
• ADDR:
Bits 63:32 of the hash address register. See ”Hash Addressing” on page 447.
473
32058K
AVR32-01/12
AT32UC3A
29.7.16 Specific Address 1 Bottom Register
Register Name:
SA1B
Access Type:
31
Read/Write
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
ADDR
23
22
21
20
ADDR
15
14
13
12
ADDR
7
6
5
4
ADDR
• ADDR
Least significant bits of the destination address. Bit zero indicates whether the address is multicast or unicast and corresponds to the least significant bit of the first byte received.
474
32058K
AVR32-01/12
AT32UC3A
29.7.17 Specific Address 1 Top Register
Register Name:
SA1T
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
ADDR
7
6
5
4
ADDR
• ADDR
The most significant bits of the destination address, that is bits 47 to 32.
475
32058K
AVR32-01/12
AT32UC3A
29.7.18 Specific Address 2 Bottom Register
Register Name:
SA2B
Access Type:
31
Read/Write
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
ADDR
23
22
21
20
ADDR
15
14
13
12
ADDR
7
6
5
4
ADDR
• ADDR
Least significant bits of the destination address. Bit zero indicates whether the address is multicast or unicast and corresponds to the least significant bit of the first byte received.
476
32058K
AVR32-01/12
AT32UC3A
29.7.19 Specific Address 2 Top Register
Register Name:
SA2T
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
ADDR
7
6
5
4
ADDR
• ADDR
The most significant bits of the destination address, that is bits 47 to 32.
477
32058K
AVR32-01/12
AT32UC3A
29.7.20 Specific Address 3 Bottom Register
Register Name:
SA3B
Access Type:
31
Read/Write
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
ADDR
23
22
21
20
ADDR
15
14
13
12
ADDR
7
6
5
4
ADDR
• ADDR
Least significant bits of the destination address. Bit zero indicates whether the address is multicast or unicast and corresponds to the least significant bit of the first byte received.
478
32058K
AVR32-01/12
AT32UC3A
29.7.21 Specific Address 3 Top Register
Register Name:
SA3T
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
ADDR
7
6
5
4
ADDR
• ADDR
The most significant bits of the destination address, that is bits 47 to 32.
479
32058K
AVR32-01/12
AT32UC3A
29.7.22 Specific Address 4 Bottom Register
Register Name:
SA4B
Access Type:
31
Read/Write
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
ADDR
23
22
21
20
ADDR
15
14
13
12
ADDR
7
6
5
4
ADDR
• ADDR
Least significant bits of the destination address. Bit zero indicates whether the address is multicast or unicast and corresponds to the least significant bit of the first byte received.
480
32058K
AVR32-01/12
AT32UC3A
29.7.23 Specific Address 4 Top Register
Register Name:
SA4T
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
ADDR
7
6
5
4
ADDR
• ADDR
The most significant bits of the destination address, that is bits 47 to 32.
481
32058K
AVR32-01/12
AT32UC3A
29.7.24 Type ID Checking Register
Register Name:
TID
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
TID
7
6
5
4
TID
• TID: Type ID checking
For use in comparisons with received frames TypeID/Length field.
482
32058K
AVR32-01/12
AT32UC3A
29.7.25 Transmit Pause Quantum Register
Register Name:
TPQ
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
TPQ
7
6
5
4
TPQ
• TPQ: Transmit Pause Quantum
Used in hardware generation of transmitted pause frames as value for pause quantum.
483
32058K
AVR32-01/12
AT32UC3A
29.7.26 User Input/Output Register
Register Name:
USRIO
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
2
1
0
–
–
–
–
3
TX_PAUSE_
ZERO
TX_PAUSE
EAM
RMII
• RMII
When set, this bit enables the MII operation mode. When reset, it selects the RMII mode.
• EAM
When set, this bit causes a frame to be copied to memory, if this feature is enabled by the EAE bit in NCFGR. Otherwise,
no frame is copied.
• TX_PAUSE
Toggling this bit causes a PAUSE frame to be transmitted.
• TX_PAUSE_ZERO
Selects either zero or the transmit quantum register as the transmitted pause frame quantum.
484
32058K
AVR32-01/12
AT32UC3A
29.7.27 Wake-on-LAN Register
Register Name:
WOL
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
MTI
18
SA1
17
ARP
16
MAG
15
14
13
12
11
10
9
8
3
2
1
0
IP
7
6
5
4
IP
• IP: ARP request IP address
Written to define the least significant 16 bits of the target IP address that is matched to generate a Wake-on-LAN event. A
value of zero does not generate an event, even if this is matched by the received frame.
• MAG: Magic packet event enable
When set, magic packet events causes the wol output to be asserted.
• ARP: ARP request event enable
When set, ARP request events causes the wol output to be asserted.
• SA1: Specific address register 1 event enable
When set, specific address 1 events causes the wol output to be asserted.
• MTI: Multicast hash event enable
When set, multicast hash events causes the wol output to be asserted.
485
32058K
AVR32-01/12
AT32UC3A
29.7.28 MACB Statistic Registers
These registers reset to zero on a read and stick at all ones when they count to their maximum value. They should be read
frequently enough to prevent loss of data. The receive statistics registers are only incremented when the receive enable bit
is set in the network control register. To write to these registers, bit 7, WESTAT, in the network control register, NCR, must
be set. The statistics register block contains the following registers.
29.7.28.1
Pause Frames Received Register
Register Name:
PFR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
FROK
7
6
5
4
FROK
• FROK: Pause Frames received OK
A 16-bit register counting the number of good pause frames received. A good frame has a length of 64 to 1518 (1536 if bit
8, BIG, in network configuration register, NCFGR, is set, 10240 if bit 3, JFRAME in network configuration register, NCFGR,
is set) and has no FCS, alignment or receive symbol errors.
29.7.28.2
Frames Transmitted OK Register
Register Name:
FTO
Access Type:
Read/Write
31
–
30
–
29
–
28
–
23
22
21
20
27
–
26
–
25
–
24
–
19
18
17
16
11
10
9
8
3
2
1
0
FTOK
15
14
13
12
FTOK
7
6
5
4
FTOK
• FTOK: Frames Transmitted OK
A 24-bit register counting the number of frames successfully transmitted, i.e., no underrun and not too many retries.
486
32058K
AVR32-01/12
AT32UC3A
29.7.28.3
Single Collision Frames Register
Register Name:
SCF
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
SCF
7
6
5
4
SCF
• SCF: Single Collision Frames
A 16-bit register counting the number of frames experiencing a single collision before being successfully transmitted, i.e.,
no underrun.
29.7.28.4
Multicollision Frames Register
Register Name:
MCF
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
MCF
7
6
5
4
MCF
• MCF: Multicollision Frames
A 16-bit register counting the number of frames experiencing between two and fifteen collisions prior to being successfully
transmitted, i.e., no underrun and not too many retries.
487
32058K
AVR32-01/12
AT32UC3A
29.7.28.5
Frames Received OK Register
Register Name:
FRO
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
22
21
20
19
18
17
16
11
10
9
8
3
2
1
0
FROK
15
14
13
12
FROK
7
6
5
4
FROK
• FROK: Frames Received OK
A 24-bit register counting the number of good frames received, i.e., address recognized and successfully copied to memory. A good frame is of length 64 to 1518 bytes (1536 if bit 8, BIG, in network configuration register, NCFGR, is set, 10240
if bit 3, JFRAME in network configuration register, NCFGR, is set) and has no FCS, alignment or receive symbol errors.
29.7.28.6
Frames Check Sequence Errors Register
Register Name:
FCSE
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
FCSE
• FCSE: Frame Check Sequence Errors
An 8-bit register counting frames that are an integral number of bytes, have bad CRC and are between 64 and 1518 bytes
in length (1536 if bit 8, BIG, in network configuration register, NCFGR, is set, 10240 if bit 3, JFRAME in network configuration register, NCFGR, is set). This register is also incremented if a symbol error is detected and the frame is of valid length
and has an integral number of bytes.
488
32058K
AVR32-01/12
AT32UC3A
29.7.28.7
Alignment Errors Register
Register Name:
ALE
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
ALE
• ALE: Alignment Errors
An 8-bit register counting frames that are not an integral number of bytes long and have bad CRC when their length is truncated to an integral number of bytes and are between 64 and 1518 bytes in length (1536 if bit 8, BIG, in network
configuration register, NCFGR, is set, 10240 if bit 3, JFRAME in network configuration register, NCFGR, is set). This register is also incremented if a symbol error is detected and the frame is of valid length and does not have an integral number
of bytes.
29.7.28.8
Deferred Transmission Frames Register
Register Name:
DTF
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
DTF
7
6
5
4
DTF
• DTF: Deferred Transmission Frames
A 16-bit register counting the number of frames experiencing deferral due to carrier sense being active on their first attempt
at transmission. Frames involved in any collision are not counted nor are frames that experienced a transmit underrun.
489
32058K
AVR32-01/12
AT32UC3A
29.7.28.9
Late Collisions Register
Register Name:
LCOL
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
LCOL
• LCOL: Late Collisions
An 8-bit register counting the number of frames that experience a collision after the slot time (512 bits) has expired. A late
collision is counted twice; i.e., both as a collision and a late collision.
29.7.28.10 Excessive Collisions Register
Register Name:
EXCOL
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
EXCOL
• EXCOL: Excessive Collisions
An 8-bit register counting the number of frames that failed to be transmitted because they experienced 16 collisions.
490
32058K
AVR32-01/12
AT32UC3A
29.7.28.11 Transmit Underrun Errors Register
Register Name:
TUND
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
TUND
• TUND: Transmit Underruns
An 8-bit register counting the number of frames not transmitted due to a transmit DMA underrun. If this register is incremented, then no other statistics register is incremented.
29.7.28.12 Carrier Sense Errors Register
Register Name:
CSE
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
CSE
• CSE: Carrier Sense Errors
An 8-bit register counting the number of frames transmitted where carrier sense was not seen during transmission or where
carrier sense was deasserted after being asserted in a transmit frame without collision (no underrun). Only incremented in
half-duplex mode. The only effect of a carrier sense error is to increment this register. The behavior of the other statistics
registers is unaffected by the detection of a carrier sense error.
491
32058K
AVR32-01/12
AT32UC3A
29.7.28.13 Receive Resource Errors Register
Register Name:
RRE
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
RRE
7
6
5
4
RRE
• RRE: Receive Resource Errors
A 16-bit register counting the number of frames that were address matched but could not be copied to memory because no
receive buffer was available.
29.7.28.14 Receive Overrun Errors Register
Register Name:
ROVR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
ROVR
• ROVR: Receive Overrun
An 8-bit register counting the number of frames that are address recognized but were not copied to memory due to a
receive DMA overrun.
492
32058K
AVR32-01/12
AT32UC3A
29.7.28.15 Receive Symbol Errors Register
Register Name:
RSE
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
RSE
• RSE: Receive Symbol Errors
An 8-bit register counting the number of frames that had rx_er asserted during reception. Receive symbol errors are also
counted as an FCS or alignment error if the frame is between 64 and 1518 bytes in length (1536 if bit 8, BIG, in network
configuration register, NCFGR, is set, 10240 if bit 3, JFRAME in network configuration register, NCFGR, is set). If the
frame is larger, it is recorded as a jabber error.
29.7.28.16 Excessive Length Errors Register
Register Name:
ELE
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
EXL
• EXL: Excessive Length Errors
An 8-bit register counting the number of frames received exceeding 1518 bytes (1536 if bit 8, BIG, in network configuration
register, NCFGR, is set, 10240 if bit 3, JFRAME in network configuration register, NCFGR, is set) in length but do not have
either a CRC error, an alignment error nor a receive symbol error.
493
32058K
AVR32-01/12
AT32UC3A
29.7.28.17 Receive Jabbers Register
Register Name:
RJA
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
RJB
• RJB: Receive Jabbers
An 8-bit register counting the number of frames received exceeding 1518 bytes (1536 if bit 8, BIG, in network configuration
register, NCFGR, is set, 10240 if bit 3, JFRAME in network configuration register, NCFGR, is set) in length and have either
a CRC error, an alignment error or a receive symbol error.
29.7.28.18 Undersize Frames Register
Register Name:
USF
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
USF
• USF: Undersize frames
An 8-bit register counting the number of frames received less than 64 bytes in length but do not have either a CRC error, an
alignment error or a receive symbol error.
494
32058K
AVR32-01/12
AT32UC3A
29.7.28.19 SQE Test Errors Register
Register Name:
STE
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
SQER
• SQER: SQE test errors
An 8-bit register counting the number of frames where col was not asserted within 96 bit times (an interframe gap) of
tx_en being deasserted in half duplex mode.
29.7.28.20 Received Length Field Mismatch Register
Register Name:
RLE
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
5
4
3
2
1
0
RLFM
• RLFM: Receive Length Field Mismatch
An 8-bit register counting the number of frames received that have a measured length shorter than that extracted from its
length field. Checking is enabled through bit 16 of the network configuration register. Frames containing a type ID in bytes
13 and 14 (i.e., length/type ID 0x0600) are not counted as length field errors, neither are excessive length frames.
495
32058K
AVR32-01/12
AT32UC3A
29.7.28.21 Transmitted Pause Frames Register
Register Name:
TPF
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
TPF
7
6
5
4
TPF
• TPF: Transmitted Pause Frames
A 16-bit register counting the number of pause frames transmitted.
496
32058K
AVR32-01/12
AT32UC3A
30. USB On-The-Go Interface (USBB)
Rev: 3.1.1.1
30.1
Features
•
•
•
•
•
•
•
30.2
USB 2.0 Compliant, Full-/Low-Speed (FS/LS) and On-The-Go (OTG), 12 Mbit/s
7 Pipes/Endpoints
960 bytes of Embedded Dual-Port RAM (DPRAM) for Pipes/Endpoints
Up to 2 Memory Banks per Pipe/Endpoint (Not for Control Pipe/Endpoint)
Flexible Pipe/Endpoint Configuration and Management with Dedicated DMA Channels
On-Chip Transceivers Including Pull-Ups/Pull-downs.
On-Chip OTG pad including VBUS analog comparator
Description
The Universal Serial Bus (USB) MCU device complies with the Universal Serial Bus (USB) 2.0
specification, but it does NOT feature high-speed USB (480 Mbit/s).
Each pipe/endpoint can be configured in one of several transfer types. It can be associated with
one or more banks of a dual-port RAM used to store the current data payload. If several banks
are used (“ping-pong” mode), then one DPRAM bank is read or written by the CPU or the DMA
while the other is read or written by the USB macro core. This feature is mandatory for isochronous pipes/endpoints.
Table 30-1 describes the hardware configuration of the USB MCU device.
Table 30-1.
Description of USB Pipes/Endpoints
Pipe/Endpoint
Mnemonic
Max. Size
Max. Nb. Banks
DMA
Type
0
PEP0
64 bytes
1
N
Control
1
PEP1
bytes
Y
Isochronous/Bulk/Interrupt
2
PEP2
bytes
Y
Isochronous/Bulk/Interrupt
3
PEP3
64 bytes
Y
Bulk/Interrupt
4
PEP4
64 bytes
Y
Bulk/Interrupt
5
PEP5
bytes
Y
Isochronous/Bulk/Interrupt
6
PEP6
bytes
Y
Isochronous/Bulk/Interrupt
The theoretical maximal pipe/endpoint configuration (1600 bytes) exceeds the real DPRAM size
(960 bytes). The user needs to be aware of this when configuring pipes/endpoints. To fully use
the 960 bytes of DPRAM, the user could for example use the configuration described in Table
30-2.
Table 30-2.
Example of Configuration of Pipes/Endpoints Using the Whole DPRAM
Pipe/Endpoint
Mnemonic
Size
Nb. Banks
0
PEP0
64 bytes
1
1
PEP1
64 bytes
2
2
PEP2
64 bytes
2
3
PEP3
64 bytes
1
497
32058K
AVR32-01/12
AT32UC3A
Table 30-2.
Example of Configuration of Pipes/Endpoints Using the Whole DPRAM
Pipe/Endpoint
Mnemonic
Size
Nb. Banks
4
PEP4
64 bytes
1
5
PEP5
256 bytes
1
6
PEP6
256 bytes
1
498
32058K
AVR32-01/12
AT32UC3A
30.3
Block Diagram
The USB controller provides a hardware device to interface a USB link to a data flow stored in a
dual-port RAM (DPRAM).
The USB controller requires a 48 MHz ± 0.25% reference clock, which is the USB generic clock
generated from one of the power manager oscillators, optionally through one of the power manager PLLs.
The 48 MHz clock is used to generate a 12 MHz full-speed (or 1.5 MHz low-speed) bit clock from
the received USB differential data and to transmit data according to full- or low-speed USB
device tolerance. Clock recovery is achieved by a digital phase-locked loop (a DPLL, not represented), which complies with the USB jitter specifications.
Figure 30-1. Block Diagram
USB
32 bits
DPRAM
Local
HSB
Slave Interface
PEP
Allocation
Slave
HSB MUX
HSB
Master
HSB0
DMA
HSB1
PB
VBUS
D-
User Interface
USB 2.0
Core
USB Interrupts
Interrupt
Controller
Power
Manager
D+
GPIO
Controller
USB_ID
USB_VBOF
USB GCLK @ 48 MHz
System Clock
Domain
USB Clock
Domain
499
32058K
AVR32-01/12
AT32UC3A
30.4
Application Block Diagram
Depending on the USB operating mode (device-only, reduced-host or OTG mode) and the
power source (bus-powered or self-powered), there are different typical hardware
implementations.
30.4.1
30.4.1.1
Device Mode
Bus-Powered Device
Figure 30-2. Bus-Powered Device Application Block Diagram
VDD
3.3 V
Regulator
USB
USB
Connector
USB_VBOF
VBUS
VBUS
DD+
D-
39 Ω ± 1%
D+
39 Ω ± 1%
USB_ID
ID
GND
30.4.1.2
Self-Powered Device
Figure 30-3. Self-Powered Device Application Block Diagram
USB
USB
Connector
USB_VBOF
VBUS
VBUS
DD+
USB_ID
39 Ω ± 1%
39 Ω ± 1%
DD+
ID
GND
500
32058K
AVR32-01/12
AT32UC3A
30.4.2
Host and OTG Modes
Figure 30-4. Host and OTG Application Block Diagram
VDD
5 V DC/DC
Generator
USB
USB
Connector
USB_VBOF
VBUS
VBUS
DD+
39 Ω ± 1%
39 Ω ± 1%
USB_ID
DD+
ID
GND
30.5
I/O Lines Description
Table 30-3.
I/O Lines Description
Name
Description
Type
Active Level
USB_VBOF
USB VBus On/Off: Bus Power Control Port
Output
VBUSPO
VBUS
VBus: Bus Power Measurement Port
Input
High
D-
Data -: Differential Data Line - Port
Input/Output
N/A
D+
Data +: Differential Data Line + Port
Input/Output
N/A
USB_ID
USB Identification: Mini Connector Identification Port
Input
Low: Mini-A plug
High Z: Mini-B plug
501
32058K
AVR32-01/12
AT32UC3A
30.6
Product Dependencies
30.6.1
I/O Lines
The USB_VBOF and USB_ID pins are multiplexed with GPIO lines and may also be multiplexed
with lines of other peripherals. In order to use them with the USB, the programmer must first program the GPIO controller to assign them to their USB peripheral functions. Moreover, if USB_ID
is used, the GPIO controller must be configured to enable the internal pull-up resistor of its pin.
If USB_VBOF or USB_ID is not used by the application, the corresponding pin can be used for
other purposes by the GPIO controller or by other peripherals.
30.6.2
Power Management
The 48 MHz USB clock is generated by a dedicated generic clock from the power manager.
Before using the USB, the programmer must ensure that the USB generic clock (USB GCLK) is
enabled at 48 MHz in the power manager.
30.6.3
Interrupts
The USB interface has an interrupt line connected to the interrupt controller. In order to handle
USB interrupts, the interrupt controller must be programmed first.
502
32058K
AVR32-01/12
AT32UC3A
30.7
Functional Description
30.7.1
USB General Operation
30.7.1.1
Introduction
After a hardware reset, the USB controller is disabled. When enabled, the USB controller runs
either in device mode or in host mode according to the ID detection.
If the USB_ID pin is not connected to ground, the ID bit is set by hardware (the internal pull-up
resistor of the USB_ID pin must be enabled by the GPIO controller) and device mode is
engaged.
The ID bit is cleared by hardware when a low level has been detected on the USB_ID pin. Host
mode is then engaged.
30.7.1.2
Power-On and Reset
Figure 30-5 describes the USB controller main states.
Figure 30-5. General States
Macrooff:
USBE = 0
Clock stopped:
FRZCLK = 1
USBE = 0
Reset
<any
other
state>
HW
RESET
USBE = 1
ID = 1
USBE = 0
USBE = 1
ID = 0
Device
USBE = 0
Host
After a hardware reset, the USB controller is in the Reset state. In this state:
•the macro is disabled (USBE = 0);
•the macro clock is stopped in order to minimize power consumption (FRZCLK = 1);
•the pad is in suspend mode;
•the internal states and registers of the device and host modes are reset;
•the DPRAM is not cleared and is accessible;
•the ID and VBUS read-only bits reflect the states of the USB_ID and VBUS input pins;
•the OTGPADE, VBUSPO, FRZCLK, USBE, UIDE, UIMOD and LS bits can be written by
software, so that the user can program pads and speed before enabling the macro, but their
value is only taken into account once the macro is enabled and unfrozen.
After setting USBE, the USB controller enters the Device or the Host mode (according to the ID
detection) in idle state.
The USB controller can be disabled at any time by clearing USBE. In fact, clearing USBE acts
as a hardware reset, except that the OTGPADE, VBUSPO, FRZCLK, UIDE, UIMOD and LS bits
are not reset.
503
32058K
AVR32-01/12
AT32UC3A
30.7.1.3
Interrupts
One interrupt vector is assigned to the USB interface. Figure 30-6 shows the structure of the
USB interrupt system.
Figure 30-6. Interrupt System
USBSTA.IDTI
USBCON.IDTE
USBSTA.VBUSTI
USBCON.VBUSTE
USBSTA.SRPI
USBCON.SRPE
USBSTA.VBERRI
USBCON.VBERRE
USBSTA.BCERRI
USB General
Interrupt
USBCON.BCERRE
USBSTA.ROLEEXI
USBCON.ROLEEXE
USBSTA.HNPERRI
USBCON.HNPERRE
USBSTA.STOI
USBCON.STOE
UESTAX.TXINI
UECONX.TXINE
UESTAX.RXOUTI
UECONX.RXOUTE
UESTAX.RXSTPI
UECONX.RXSTPE
UESTAX.UNDERFI
UECONX.UNDERFE
UESTAX.NAKOUTI
UECONX.NAKOUTE
UESTAX.NAKINI
UECONX.NAKINE
UDINT.SUSP
UESTAX.OVERFI
UDINTE.SUSPE
UECONX.OVERFE
UDINT.SOF
UESTAX.STALLEDI
UDINTE.SOFE
UECONX.STALLEDE
UESTAX.CRCERRI
UECONX.CRCERRE
UESTAX.SHORTPACKET
USB Device
Endpoint X
Interrupt
UDINT.EORST
UDINTE.EORSTE
UDINT.WAKEUP
UDINTE.WAKEUPE
UECONX.SHORTPACKETE
UDINT.EORSM
UECONX.NBUSYBKE
UDINT.UPRSM
UESTAX.NBUSYBK
USB Device
Interrupt
USB
Interrupt
UDINTE.EORSME
UDINTE.UPRSME
UDDMAX_STATUS.EOT_STA
UDINT.EPXINT
UDDMAX_CONTROL.EOT_IRQ_EN
UDINTE.EPXINTE
UDDMAX_STATUS.EOCH_BUFF_STA
UDDMAX_CONTROL.EOBUFF_IRQ_EN
UDDMAX_STATUS.DESC_LD_STA
UDDMAX_CONTROL.DESC_LD_IRQ_EN
USB Device
DMA Channel X
Interrupt
UDINT.DMAXINT
UDINTE.DMAXINTE
UPSTAX.RXINI
UPCONX.RXINE
UPSTAX.TXOUTI
UPCONX.TXOUTE
UPSTAX.TXSTPI
UPCONX.TXSTPE
UPSTAX.UNDERFI
UPCONX.UNDERFIE
UPSTAX.PERRI
UPCONX.PERRE
UHINT.DCONNI
UPSTAX.NAKEDI
UHINTE.DCONNIE
UPCONX.NAKEDE
UHINT.DDISCI
UPCONX.OVERFIE
UHINT.RSTI
UPSTAX.OVERFI
UHINTE.DDISCIE
UPSTAX.RXSTALLDI
UHINTE.RSTIE
UPCONX.RXSTALLDE
UPSTAX.CRCERRI
UPCONX.CRCERRE
UPSTAX.SHORTPACKETI
USB Host
Pipe X
Interrupt
UHINT.RSMEDI
UHINTE.RSMEDIE
UHINT.RXRSMI
UHINTE.RXRSMIE
UPCONX.SHORTPACKETIE
UHINT.HSOFI
UPCONX.NBUSYBKE
UHINT.HWUPI
UPSTAX.NBUSYBK
USB Host
Interrupt
UHINTE.HSOFIE
UHINTE.HWUPIE
UHDMAX_STATUS.EOT_STA
UHINT.PXINT
UHDMAX_CONTROL.EOT_IRQ_EN
UHINTE.PXINTE
UHDMAX_STATUS.EOCH_BUFF_STA
UHDMAX_CONTROL.EOBUFF_IRQ_EN
UHDMAX_STATUS.DESC_LD_STA
UHDMAX_CONTROL.DESC_LD_IRQ_EN
USB Host
DMA Channel X
Interrupt
UHINT.DMAXINT
UHINTE.DMAXINTE
Asynchronous interrupt source
504
32058K
AVR32-01/12
AT32UC3A
See Section 30.7.2.17 on page 520 and Section 30.7.3.13 on page 528 for further details about
device and host interrupts.
There are two kinds of general interrupts: processing, i.e. their generation is part of the normal
processing, and exception, i.e. errors (not related to CPU exceptions).
The processing general interrupts are:
•the ID Transition interrupt (IDTI);
•the VBus Transition interrupt (VBUSTI);
•the SRP interrupt (SRPI);
•the Role Exchange interrupt (ROLEEXI).
The exception general interrupts are:
•the VBus Error interrupt (VBERRI);
•the B-Connection Error interrupt (BCERRI);
•the HNP Error interrupt (HNPERRI);
•the Suspend Time-Out interrupt (STOI).
30.7.1.4
MCU Power Modes
30.7.1.4.1
Run Mode
In this mode, all MCU clocks can run, including the USB clock.
30.7.1.4.2
Idle Mode
In this mode, the CPU is halted, i.e. the CPU clock is stopped. The Idle mode is entered whatever the state of the USB macro. The MCU wakes up on any USB interrupt.
30.7.1.4.3
Frozen Mode
Same as the Idle mode, except that the HSB module is stopped, so the USB DMA, which is an
HSB master, can not be used. Moreover, the USB DMA must be stopped before entering this
sleep mode in order to avoid erratic behavior. The MCU wakes up on any USB interrupt.
30.7.1.4.4
Standby, Stop, DeepStop and Static Modes
Same as the Frozen mode, except that the USB generic clock and other clocks are stopped, so
the USB macro is frozen.
30.7.1.4.5
USB Clock Frozen
In the Run, Idle and Frozen MCU modes, the USB macro can be frozen when the usb line is in
the suspend mode, by setting the FRZCLK bit, what reduces power consumption.
In this case, it is still possible to access the following elements, but only in Run mode:
•the OTGPADE, VBUSPO, FRZCLK, USBE, UIDE, UIMOD and LS bits;
•the DPRAM (through the USB_FIFOX_DATA registers, but not through USB bus transfers
which are frozen).
Moreover, when FRZCLK is set, only the asynchronous interrupt sources may trigger the USB
interrupt:
•the ID Transition interrupt (IDTI);
505
32058K
AVR32-01/12
AT32UC3A
•the VBus Transition interrupt (VBUSTI);
•the Wake-Up interrupt (WAKEUP);
•the Host Wake-Up interrupt (HWUPI).
30.7.1.4.6
USB Suspend mode :
In peripheral mode, the UDINT.SUSP bit indicates that the usb line is in the suspend mode. In
this case, the USB Data transceiver is automatically set in suspend mode to reduce the
consumption.
30.7.1.5
Speed Control
30.7.1.5.1
Device Mode
When the USB interface is in device mode, the speed selection (full-/low-speed) depends on
which of D+ and D- is pulled up. The LS bit allows to connect an internal pull-up resistor either
on D+ (full-speed mode) or on D- (low-speed mode). The LS bit should be configured before
attaching the device, what can be done by clearing the DETACH bit.
Figure 30-7. Speed Selection in Device Mode
RPU
VBUS
UDCON.DETACH
UDCON.LS
D+
D-
30.7.1.5.2
Host Mode
When the USB interface is in host mode, internal pull-down resistors are connected on both D+
and D- and the interface detects the speed of the connected device, which is reflected by the
SPEED bit-field.
30.7.1.6
DPRAM Management
Pipes and endpoints can only be allocated in ascending order (from the pipe/endpoint 0 to the
last pipe/endpoint to be allocated). The firmware shall therefore configure them in the same
order.
The allocation of a pipe/endpoint ki starts when its ALLOC bit is set. Then, the hardware allocates a memory area in the DPRAM and inserts it between the ki-1 and ki+1 pipes/endpoints. The
ki+1 pipe/endpoint memory window slides up and its data is lost. Note that the following pipe/endpoint memory windows (from ki+2) do not slide.
506
32058K
AVR32-01/12
AT32UC3A
Disabling a pipe (PENX = 0) or an endpoint (EPENX = 0) resets neither its ALLOC bit nor its
configuration (PBK/EPBK, PSIZE/EPSIZE, PTOKEN/EPDIR, PTYPE/EPTYPE, PEPNUM, INTFRQ). To free its memory, the firmware should clear its ALLOC bit. The ki+1 pipe/endpoint
memory window then slides down and its data is lost. Note that the following pipe/endpoint
memory windows (from ki+2) do not slide.
Figure 30-8 illustrates the allocation and reorganization of the DPRAM in a typical example.
Figure 30-8. Allocation and Reorganization of the DPRAM
Free Memory
Free Memory
Free Memory
PEP5
PEP5
PEP5
PEP4
PEP4
PEP4 Lost Memory
PEP3
PEP3
(ALLOC stays at 1)
PEP4
PEP2
PEP2
PEP2
PEP2
PEP1
PEP1
PEP1
PEP1
PEP0
PEP0
PEP0
PEP0
U(P/E)RST.(E)PENX = 1
U(P/E)CFGX.ALLOC = 1
U(P/E)RST.(E)PEN3 = 0
Pipes/Endpoints 0..5
Activated
Pipe/Endpoint 3
Disabled
U(P/E)CFG3.ALLOC = 0
Pipe/Endpoint 3
Memory Freed
Free Memory
PEP5
PEP4
Conflict
PEP3 (larger size)
U(P/E)RST.(E)PEN3 = 1
U(P/E)CFG3.ALLOC = 1
Pipe/Endpoint 3
Activated
• First, the pipes/endpoints 0 to 5 are enabled, configured and allocated in ascending order.
Each pipe/endpoint then owns a memory area in the DPRAM.
• Then, the pipe/endpoint 3 is disabled, but its memory is kept allocated by the controller.
• In order to free its memory, its ALLOC bit is then cleared by the firmware. The pipe/endpoint 4
memory window slides down, but the pipe/endpoint 5 does not move.
• Finally, if the firmware chooses to reconfigure the pipe/endpoint 3 with a larger size, the
controller allocates a memory area after the pipe/endpoint 2 memory area and automatically
slides up the pipe/endpoint 4 memory window. The pipe/endpoint 5 does not move and a
memory conflict appears as the memory windows of the pipes/endpoints 4 and 5 overlap. The
data of these pipes/endpoints is potentially lost.
Note that:
•there is no way the data of the pipe/endpoint 0 can be lost (except if it is de-allocated) as
memory allocation and de-allocation may affect only higher pipes/endpoints;
•deactivating then reactivating a same pipe/endpoint with the same configuration only modifies
temporarily the controller DPRAM pointer and size for this pipe/endpoint, but nothing
changes in the DPRAM, so higher endpoints seem to not have been moved and their data is
preserved as far as nothing has been written or received into them while changing the
allocation state of the first pipe/endpoint;
•when the firmware sets the ALLOC bit, the CFGOK bit is set by hardware only if the
configured size and number of banks are correct compared to their maximal allowed values
for the endpoint and to the maximal FIFO size (i.e. the DPRAM size), so the value of CFGOK
does not consider memory allocation conflicts.
507
32058K
AVR32-01/12
AT32UC3A
30.7.1.7
Pad Suspend
Figure 30-9 shows the pad behavior.
Figure 30-9. Pad Behavior
USBE = 1
& DETACH = 0
& Suspend
Idle
USBE = 0
| DETACH = 1
| Suspend
Active
• In the Idle state, the pad is put in low power consumption mode.
• In the Active state, the pad is working.
Figure 30-10 illustrates the pad events leading to a PAD state change.
Figure 30-10. Pad Events
SUSP
Suspend detected
WAKEUP
Cleared by hardware on wake-up
Wake-up detected
Cleared by software to acknowledge the interrupt
PAD State
Active
Idle
Active
The Suspend interrupt flag (SUSP) is set and the Wake-Up interrupt flag (WAKEUP) is cleared
when a USB “Suspend” state has been detected on the USB bus. This event automatically puts
the USB pad in the Idle state. The detection of a non-idle event sets WAKEUP, clears SUSP and
wakes up the USB pad.
Moreover, the pad goes to the Idle state if the macro is disabled or if the DETACH bit is set. It
returns to the Active state when USBE = 1 and DETACH = 0.
508
32058K
AVR32-01/12
AT32UC3A
30.7.1.8
Customizing of OTG Timers
It is possible to refine some OTG timers thanks to the TIMPAGE and TIMVALUE bit-fields, as
shown by Figure 30-4.
Customizing of OTG Timers
Table 30-4.
TIMVALUE
TIMPAGE
00b:
AWaitVrise Time-Out
([OTG] Chapter 6.6.5.1)
01b:
VbBusPulsing Time-Out
([OTG] Chapter 5.3.4)
10b:
PdTmOutCnt Time-Out
([OTG] Chapter 5.3.2)
11b:
SRPDetTmOut Time-Out
([OTG] Chapter 5.3.3)
00b
20 ms
15 ms
93 ms
10 µs
01b
50 ms
23 ms
105 ms
100 µs
10b
70 ms
31 ms
118 ms
1 ms
11b
100 ms
40 ms
131 ms
11 ms
TIMPAGE is used to select the OTG timer to access while TIMVALUE indicates the time-out
value of the selected timer.
TIMPAGE and TIMVALUE can be read or written. Before writing them, the firmware should
unlock write accesses by setting the UNLOCK bit. This is not required for read accesses, except
before accessing TIMPAGE if it has to be written in order to read the TIMVALUE bit-field of
another OTG timer.
30.7.1.9
Plug-In Detection
The USB connection is detected from the VBUS pad. Figure 30-11 shows the architecture of the
plug-in detector.
Figure 30-11. Plug-In Detection Input Block Diagram
VDD
RPU
VBus_pulsing
VBUS
Session_valid
RPD
Va_Vbus_valid
Logic
VBUS
USBSTA
VBUSTI
USBSTA
VBus_discharge
GND
Pad Logic
The control logic of the VBUS pad outputs two signals:
•the Session_valid signal is high when the voltage on the VBUS pad is higher than or equal to
1.4 V;
•the Va_Vbus_valid signal is high when the voltage on the VBUS pad is higher than or equal to
4.4 V.
In device mode, the VBUS bit follows the Session_valid comparator output:
•it is set when the voltage on the VBUS pad is higher than or equal to 1.4 V;
509
32058K
AVR32-01/12
AT32UC3A
•it is cleared when the voltage on the VBUS pad is lower than 1.4 V.
In host mode, the VBUS bit follows an hysteresis based on Session_valid and Va_Vbus_valid:
•it is set when the voltage on the VBUS pad is higher than or equal to 4.4 V;
•it is cleared when the voltage on the VBUS pad is lower than 1.4 V.
The VBus Transition interrupt (VBUSTI) is raised on each transition of the VBUS bit.
The VBUS bit is effective whether the USB macro is enabled or not.
30.7.1.10
ID Detection
Figure 30-12 shows how the ID transitions are detected.
Figure 30-12. ID Detection Input Block Diagram
RPU
VDD
1
USB_ID
UIMOD
0
ID
USBSTA
IDTI
USBSTA
USBCON
UIDE
USBCON
GPIO Controller
The USB mode (device or host) can be either detected from the USB_ID pin or software
selected from the UIMOD bit, according to the UIDE bit. This allows the USB_ID pin to be used
as a general purpose I/O pin even when the USB interface is enabled.
By default, the USB_ID pin is selected (UIDE = 1) and the USB macro is in device mode (ID = 1),
what corresponds to the case where no Mini-A plug is connected, i.e. no plug or a Mini-B plug is
connected and the USB_ID pin is kept high by the internal pull-up resistor from the GPIO controller (which must be enabled if USB_ID is used).
The ID Transition interrupt (IDTI) is raised on each transition of the ID bit, i.e. when a Mini-A plug
(host mode) is connected or disconnected. This does not occur when a Mini-B plug (device
mode) is connected or disconnected.
The ID bit is effective whether the USB macro is enabled or not.
510
32058K
AVR32-01/12
AT32UC3A
30.7.2
30.7.2.1
USB Device Operation
Introduction
In device mode, the USB controller supports full- and low-speed data transfers.
In addition to the default control endpoint, six endpoints are provided, which can be configured
with the types isochronous, bulk or interrupt, as described in Table 30-1 on page 497.
The device mode starts in the Idle state, so the pad consumption is reduced to the minimum.
30.7.2.2
Power-On and Reset
Figure 30-13 describes the USB controller device mode main states.
Figure 30-13. Device Mode States
USBE = 0
| ID = 0
<any
other
state>
USBE = 0
| ID = 0
Reset
Idle
USBE = 1
& ID = 1
HW
RESET
After a hardware reset, the USB controller device mode is in the Reset state. In this state:
•the macro clock is stopped in order to minimize power consumption (FRZCLK = 1);
•the internal registers of the device mode are reset;
•the endpoint banks are de-allocated;
•neither D+ nor D- is pulled up (DETACH = 1).
D+ or D- will be pulled up according to the selected speed as soon as the DETACH bit is cleared
and VBus is present. See Section 30.7.1.5.1 on page 506 for further details.
When the USB macro is enabled (USBE = 1) in device mode (ID = 1), its device mode state
goes to the Idle state with minimal power consumption. This does not require the USB clock to
be activated.
The USB controller device mode can be disabled and reset at any time by disabling the USB
macro (USBE = 0) or when host mode is engaged (ID = 0).
30.7.2.3
USB Reset
The USB bus reset is managed by hardware. It is initiated by a connected host.
When a USB reset is detected on the USB line, the following operations are performed by the
controller:
•all the endpoints are disabled, except the default control endpoint;
•the default control endpoint is reset (see Section 30.7.2.4 on page 512 for more details);
•the data toggle sequence of the default control endpoint is cleared;
•at the end of the reset process, the End of Reset interrupt (EORST) is raised.
511
32058K
AVR32-01/12
AT32UC3A
30.7.2.4
Endpoint Reset
An endpoint can be reset at any time by setting its EPRSTX bit in the UERST register. This is
recommended before using an endpoint upon hardware reset or when a USB bus reset has
been received. This resets:
•the internal state machine of this endpoint;
•the receive and transmit bank FIFO counters;
•all the registers of this endpoint (UECFGX, UESTAX, UECONX), except its configuration
(ALLOC, EPBK, EPSIZE, EPDIR, EPTYPE) and its Data Toggle Sequence bit-field (DTSEQ).
Note that the interrupt sources located in the UESTAX register are not cleared when a USB bus
reset has been received.
The endpoint configuration remains active and the endpoint is still enabled.
The endpoint reset may be associated with a clear of the data toggle sequence as an answer to
the CLEAR_FEATURE USB request. This can be achieved by setting the RSTDT bit (by setting
the RSTDTS bit).
In the end, the firmware has to clear the EPRSTX bit to complete the reset operation and to start
using the FIFO.
30.7.2.5
Endpoint Activation
The endpoint is maintained inactive and reset (see Section 30.7.2.4 on page 512 for more
details) as long as it is disabled (EPENX = 0). The Data Toggle Sequence bit-field (DTSEQ) is
also reset.
The algorithm represented on Figure 30-14 must be followed in order to activate an endpoint.
Figure 30-14. Endpoint Activation Algorithm
Endpoint
Activation
EPENX = 1
Enable the endpoint.
UECFGX
Configure the endpoint:
- type;
- direction;
- size;
- number of banks.
Allocate the configured DPRAM
banks.
EPTYPE
EPDIR
EPSIZE
EPBK
ALLOC
CFGOK ==
1?
Yes
Endpoint
Activated
Test if the endpoint configuration
is correct.
No
ERROR
As long as the endpoint is not correctly configured (CFGOK = 0), the controller does not
acknowledge the packets sent by the host to this endpoint.
The CFGOK bit is set by hardware only if the configured size and number of banks are correct
compared to their maximal allowed values for the endpoint (see Table 30-1 on page 497) and to
the maximal FIFO size (i.e. the DPRAM size).
512
32058K
AVR32-01/12
AT32UC3A
See Section 30.7.1.6 on page 506 for more details about DPRAM management.
30.7.2.6
Address Setup
The USB device address is set up according to the USB protocol:
•after all kinds of resets, the USB device address is 0;
•the host starts a SETUP transaction with a SET_ADDRESS(addr) request;
•the firmware records this address into the UADD bit-field, leaving the ADDEN bit cleared, so
the actual address is still 0;
•the firmware sends a zero-length IN packet from the control endpoint;
•the firmware enables the recorded USB device address by setting ADDEN.
Once the USB device address is configured, the controller filters the packets to only accept
those targeting the address stored in UADD.
UADD and ADDEN shall not be written all at once.
UADD and ADDEN are cleared by hardware:
•on a hardware reset;
•when the USB macro is disabled (USBE = 0);
•when a USB reset is detected.
When UADD or ADDEN is cleared, the default device address 0 is used.
30.7.2.7
Suspend and Wake-Up
When an idle USB bus state has been detected for 3 ms, the controller raises the Suspend interrupt (SUSP). The firmware may then set the FRZCLK bit to reduce power consumption. The
MCU can also enter the Idle or Frozen sleep mode to lower again power consumption.
To recover from the Suspend mode, the firmware should wait for the Wake-Up interrupt
(WAKEUP), which is raised when a non-idle event is detected, then clear FRZCLK.
As the WAKEUP interrupt is raised when a non-idle event is detected, it can occur whether the
controller is in the Suspend mode or not. The SUSP and WAKEUP interrupts are thus independent of each other except that one’s flag is cleared by hardware when the other is raised.
30.7.2.8
Detach
The reset value of the DETACH bit is 1.
It is possible to initiate a device re-enumeration simply by setting then clearing DETACH.
DETACH acts on the pull-up connections of the D+ and D- pads. See Section 30.7.1.5.1 on
page 506 for further details.
30.7.2.9
Remote Wake-Up
The Remote Wake-Up request (also known as Upstream Resume) is the only one the device
may send on its own initiative, but the device should have beforehand been allowed to by a
DEVICE_REMOTE_WAKEUP request from the host.
• First, the USB controller must have detected a “Suspend” state on the bus, i.e. the Remote
Wake-Up request can only be sent after a SUSP interrupt has been raised.
513
32058K
AVR32-01/12
AT32UC3A
• The firmware may then set the RMWKUP bit to send an upstream resume to the host for a
remote wake-up. This will automatically be done by the controller after 5 ms of inactivity on the
USB bus.
• When the controller sends the upstream resume, the Upstream Resume interrupt (UPRSM) is
raised and SUSP is cleared by hardware.
• RMWKUP is cleared by hardware at the end of the upstream resume.
• If the controller detects a valid “End of Resume” signal from the host, the End of Resume
interrupt (EORSM) is raised.
30.7.2.10
STALL Request
For each endpoint, the STALL management is performed using:
•the STALL Request bit (STALLRQ) to initiate a STALL request;
•the STALLed interrupt (STALLEDI) raised when a STALL handshake has been sent.
To answer the next request with a STALL handshake, STALLRQ has to be set by setting the
STALLRQS bit. All following requests will be discarded (RXOUTI, etc. will not be set) and handshaked with a STALL until the STALLRQ bit is cleared, what is done by hardware when a new
SETUP packet is received (for control endpoints) or when the STALLRQC bit is set.
Each time a STALL handshake is sent, the STALLEDI flag is set by the USB controller and the
EPXINT interrupt is raised.
30.7.2.10.1 Special Considerations for Control Endpoints
If a SETUP packet is received into a control endpoint for which a STALL is requested, the
Received SETUP interrupt (RXSTPI) is raised and STALLRQ and STALLEDI are cleared by
hardware. The SETUP has to be ACKed.
This management simplifies the enumeration process management. If a command is not supported or contains an error, the firmware requests a STALL and can return to the main task,
waiting for the next SETUP request.
30.7.2.10.2 STALL Handshake and Retry Mechanism
The retry mechanism has priority over the STALL handshake. A STALL handshake is sent if the
STALLRQ bit is set and if there is no retry required.
30.7.2.11
Management of Control Endpoints
30.7.2.11.1 Overview
A SETUP request is always ACKed. When a new SETUP packet is received, the Received
SETUP interrupt (RXSTPI) is raised, but not the Received OUT Data interrupt (RXOUTI).
The FIFOCON and RWALL bits are irrelevant for control endpoints. The firmware shall therefore
never use them on these endpoints. When read, their value is always 0.
Control endpoints are managed using:
•the Received SETUP interrupt (RXSTPI) which is raised when a new SETUP packet is
received and which shall be cleared by firmware to acknowledge the packet and to free the
bank;
514
32058K
AVR32-01/12
AT32UC3A
•the Received OUT Data interrupt (RXOUTI) which is raised when a new OUT packet is
received and which shall be cleared by firmware to acknowledge the packet and to free the
bank;
•the Transmitted IN Data interrupt (TXINI) which is raised when the current bank is ready to
accept a new IN packet and which shall be cleared by firmware to send the packet.
30.7.2.11.2 Control Write
Figure 30-15 shows a control write transaction. During the status stage, the controller will not
necessarily send a NAK on the first IN token:
•if the firmware knows the exact number of descriptor bytes that must be read, it can then
anticipate the status stage and send a zero-length packet after the next IN token;
•or it can read the bytes and wait for the NAKed IN interrupt (NAKINI) which tells that all the
bytes have been sent by the host and that the transaction is now in the status stage.
Figure 30-15. Control Write
SETUP
USB Bus
DATA
SETUP
RXSTPI
OUT
HW
STATUS
OUT
IN
IN
NAK
SW
RXOUTI
HW
SW
HW
SW
TXINI
SW
30.7.2.11.3 Control Read
Figure 30-16 shows a control read transaction. The USB controller has to manage the simultaneous write requests from the CPU and the USB host.
Figure 30-16. Control Read
SETUP
USB Bus
RXSTPI
DATA
SETUP
HW
IN
STATUS
IN
OUT
SW
RXOUTI
TXINI
OUT
NAK
HW
SW
HW
SW
SW
Wr Enable
HOST
Wr Enable
CPU
A NAK handshake is always generated on the first status stage command.
When the controller detects the status stage, all the data written by the CPU is lost and clearing
TXINI has no effect.
The firmware checks if the transmission or the reception is complete.
515
32058K
AVR32-01/12
AT32UC3A
The OUT retry is always ACKed. This reception sets RXOUTI and TXINI. Handle this with the
following software algorithm:
set TXINI
wait for RXOUTI OR TXINI
if RXOUTI, then clear flag and return
if TXINI, then continue
Once the OUT status stage has been received, the USB controller waits for a SETUP request.
The SETUP request has priority over any other request and has to be ACKed. This means that
any other flag should be cleared and the FIFO reset when a SETUP is received.
The firmware has to take care of the fact that the byte counter is reset when a zero-length OUT
packet is received.
30.7.2.12
Management of IN Endpoints
30.7.2.12.1 Overview
IN packets are sent by the USB device controller upon IN requests from the host. All the data
can be written by the firmware which acknowledges or not the bank when it is full.
The endpoint must be configured first.
The TXINI bit is set by hardware at the same time as FIFOCON when the current bank is free.
This triggers an EPXINT interrupt if TXINE = 1.
TXINI shall be cleared by software (by setting the TXINIC bit) to acknowledge the interrupt, what
has no effect on the endpoint FIFO.
The firmware then writes into the FIFO and clears the FIFOCON bit to allow the USB controller
to send the data. If the IN endpoint is composed of multiple banks, this also switches to the next
bank. The TXINI and FIFOCON bits are updated by hardware in accordance with the status of
the next bank.
TXINI shall always be cleared before clearing FIFOCON.
The RWALL bit is set by hardware when the current bank is not full, i.e. the software can write
further data into the FIFO.
Figure 30-17. Example of an IN Endpoint with 1 Data Bank
NAK
IN
DATA
(bank 0)
ACK
IN
HW
TXINI
FIFOCON
SW
write data to CPU
BANK 0
SW
SW
write data to CPU
BANK 0
SW
516
32058K
AVR32-01/12
AT32UC3A
Figure 30-18. Example of an IN Endpoint with 2 Data Banks
DATA
(bank 0)
IN
ACK
IN
DATA
(bank 1)
ACK
HW
TXINI
FIFOCON
SW
SW
write data to CPU
BANK 0
SW
write data to CPU
BANK 1
SW
SW
write data to CPU
BANK0
30.7.2.12.2 Detailed Description
The data is written by the firmware, following the next flow:
•when the bank is empty, TXINI and FIFOCON are set, what triggers an EPXINT interrupt if
TXINE = 1;
•the firmware acknowledges the interrupt by clearing TXINI;
•the firmware writes the data into the current bank by using the USB Pipe/Endpoint X FIFO
Data register (USB_FIFOX_DATA), until all the data frame is written or the bank is full (in
which case RWALL is cleared by hardware and BYCT reaches the endpoint size);
•the firmware allows the controller to send the bank and switches to the next bank (if any) by
clearing FIFOCON.
If the endpoint uses several banks, the current one can be written by the firmware while the previous one is being read by the host. Then, when the firmware clears FIFOCON, the following
bank may already be free and TXINI is set immediately.
An “Abort” stage can be produced when a zero-length OUT packet is received during an IN
stage of a control or isochronous IN transaction. The KILLBK bit is used to kill the last written
bank. The best way to manage this abort is to apply the algorithm represented on Figure 30-19.
517
32058K
AVR32-01/12
AT32UC3A
Figure 30-19. Abort Algorithm
Endpoint
Abort
Disable the TXINI interrupt.
TXINEC = 1
NBUSYBK
== 0?
Yes
No
EPRSTX = 1
Abort is based on the fact
that no bank is busy, i.e. that
nothing has to be sent.
KILLBKS = 1
Yes
KILLBK
== 1?
Kill the last written bank.
Wait for the end of the
procedure.
No
Abort Done
30.7.2.13
Management of OUT Endpoints
30.7.2.13.1 Overview
OUT packets are sent by the host. All the data can be read by the firmware which acknowledges
or not the bank when it is empty.
The endpoint must be configured first.
The RXOUTI bit is set by hardware at the same time as FIFOCON when the current bank is full.
This triggers an EPXINT interrupt if RXOUTE = 1.
RXOUTI shall be cleared by software (by setting the RXOUTIC bit) to acknowledge the interrupt,
what has no effect on the endpoint FIFO.
The firmware then reads from the FIFO and clears the FIFOCON bit to free the bank. If the OUT
endpoint is composed of multiple banks, this also switches to the next bank. The RXOUTI and
FIFOCON bits are updated by hardware in accordance with the status of the next bank.
RXOUTI shall always be cleared before clearing FIFOCON.
The RWALL bit is set by hardware when the current bank is not empty, i.e. the software can read
further data from the FIFO.
518
32058K
AVR32-01/12
AT32UC3A
Figure 30-20. Example of an OUT Endpoint with 1 Data Bank
DATA
(bank 0)
OUT
NAK
ACK
DATA
(bank 0)
OUT
HW
ACK
HW
RXOUTI
SW
SW
read data from CPU
BANK 0
FIFOCON
read data from CPU
BANK 0
SW
Figure 30-21. Example of an OUT Endpoint with 2 Data Banks
OUT
DATA
(bank 0)
ACK
OUT
DATA
(bank 1)
ACK
HW
RXOUTI
HW
SW
read data from CPU
BANK 0
FIFOCON
SW
SW read data from CPU
BANK 1
30.7.2.13.2 Detailed Description
The data is read by the firmware, following the next flow:
•when the bank is full, RXOUTI and FIFOCON are set, what triggers an EPXINT interrupt if
RXOUTE = 1;
•the firmware acknowledges the interrupt by clearing RXOUTI;
•the firmware can read the byte count of the current bank from BYCT to know how many bytes
to read, rather than polling RWALL;
•the firmware reads the data from the current bank by using the USB Pipe/Endpoint X FIFO
Data register (USB_FIFOX_DATA), until all the expected data frame is read or the bank is
empty (in which case RWALL is cleared by hardware and BYCT reaches 0);
•the firmware frees the bank and switches to the next bank (if any) by clearing FIFOCON.
If the endpoint uses several banks, the current one can be read by the firmware while the following one is being written by the host. Then, when the firmware clears FIFOCON, the following
bank may already be ready and RXOUTI is set immediately.
30.7.2.14
Underflow
This error exists only for isochronous IN/OUT endpoints. It raises the Underflow interrupt
(UNDERFI), what triggers an EPXINT interrupt if UNDERFE = 1.
An underflow can occur during IN stage if the host attempts to read from an empty bank. A zerolength packet is then automatically sent by the USB controller.
519
32058K
AVR32-01/12
AT32UC3A
An underflow can not occur during OUT stage on a CPU action, since the firmware may read
only if the bank is not empty (RXOUTI = 1 or RWALL = 1).
An underflow can also occur during OUT stage if the host sends a packet while the bank is
already full. Typically, the CPU is not fast enough. The packet is lost.
An underflow can not occur during IN stage on a CPU action, since the firmware may write only
if the bank is not full (TXINI = 1 or RWALL = 1).
30.7.2.15
Overflow
This error exists for all endpoint types. It raises the Overflow interrupt (OVERFI), what triggers
an EPXINT interrupt if OVERFE = 1.
An overflow can occur during OUT stage if the host attempts to write into a bank that is too small
for the packet. The packet is acknowledged and the Received OUT Data interrupt (RXOUTI) is
raised as if no overflow had occurred. The bank is filled with all the first bytes of the packet that
fit in.
An overflow can not occur during IN stage on a CPU action, since the firmware may write only if
the bank is not full (TXINI = 1 or RWALL = 1).
30.7.2.16
CRC Error
This error exists only for isochronous OUT endpoints. It raises the CRC Error interrupt
(CRCERRI), what triggers an EPXINT interrupt if CRCERRE = 1.
A CRC error can occur during OUT stage if the USB controller detects a corrupted received
packet. The OUT packet is stored in the bank as if no CRC error had occurred (RXOUTI is
raised).
30.7.2.17
Interrupts
See the structure of the USB device interrupt system on Figure 30-6 on page 504.
There are two kinds of device interrupts: processing, i.e. their generation is part of the normal
processing, and exception, i.e. errors (not related to CPU exceptions).
30.7.2.17.1 Global Interrupts
The processing device global interrupts are:
•the Suspend interrupt (SUSP);
•the Start of Frame interrupt (SOF) with no frame number CRC error (FNCERR = 0);
•the End of Reset interrupt (EORST);
•the Wake-Up interrupt (WAKEUP);
•the End of Resume interrupt (EORSM);
•the Upstream Resume interrupt (UPRSM);
•the Endpoint X interrupt (EPXINT);
•the DMA Channel X interrupt (DMAXINT).
The exception device global interrupts are:
•the Start of Frame interrupt (SOF) with a frame number CRC error (FNCERR = 1).
520
32058K
AVR32-01/12
AT32UC3A
30.7.2.17.2 Endpoint Interrupts
The processing device endpoint interrupts are:
•the Transmitted IN Data interrupt (TXINI);
•the Received OUT Data interrupt (RXOUTI);
•the Received SETUP interrupt (RXSTPI);
•the Short Packet interrupt (SHORTPACKET);
•the Number of Busy Banks interrupt (NBUSYBK).
The exception device endpoint interrupts are:
•the Underflow interrupt (UNDERFI);
•the NAKed OUT interrupt (NAKOUTI);
•the NAKed IN interrupt (NAKINI);
•the Overflow interrupt (OVERFI);
•the STALLed interrupt (STALLEDI);
•the CRC Error interrupt (CRCERRI);
30.7.2.17.3 DMA Interrupts
The processing device DMA interrupts are:
•the End of USB Transfer Status interrupt (EOT_STA);
•the End of Channel Buffer Status interrupt (EOCH_BUFF_STA);
•the Descriptor Loaded Status interrupt (DESC_LD_STA).
There is no exception device DMA interrupt.
521
32058K
AVR32-01/12
AT32UC3A
30.7.3
30.7.3.1
USB Host Operation
Description of Pipes
For the USB controller in host mode, the term “pipe” is used instead of “endpoint” (used in
device mode). A host pipe corresponds to a device endpoint, as described by the Figure 30-22
from the USB specification.
Figure 30-22. USB Communication Flow
In host mode, the USB controller associates a pipe to a device endpoint, considering the device
configuration descriptors.
30.7.3.2
Power-On and Reset
Figure 30-23 describes the USB controller host mode main states.
Figure 30-23. Host Mode States
Device
Disconnection
Macrooff
Clock stopped
<any
other
state>
Idle
Device
Connection
Device
Disconnection
Ready
SOFE = 0
SOFE = 1
Suspend
After a hardware reset, the USB controller host mode is in the Reset state.
When the USB macro is enabled (USBE = 1) in host mode (ID = 0), its host mode state goes to
the Idle state. In this state, the controller waits for device connection with minimal power consumption. The USB pad should be in the Idle state. Once a device is connected, the macro
enters the Ready state, what does not require the USB clock to be activated.
522
32058K
AVR32-01/12
AT32UC3A
The controller enters the Suspend state when the USB bus is in a “Suspend” state, i.e. when the
host mode does not generate the “Start of Frame”. In this state, the USB consumption is minimal. The host mode exits the Suspend state when starting to generate the SOF over the USB
line.
30.7.3.3
Device Detection
A device is detected by the USB controller host mode when D+ or D- is no longer tied low, i.e.
when the device D+ or D- pull-up resistor is connected. To enable this detection, the host controller has to provide the VBus power supply to the device by setting the VBUSRQ bit (by setting
the VBUSRQS bit).
The device disconnection is detected by the host controller when both D+ and D- are pulled
down.
30.7.3.4
USB Reset
The USB controller sends a USB bus reset when the firmware sets the RESET bit. The USB
Reset Sent interrupt (RSTI) is raised when the USB reset has been sent. In this case, all the
pipes are disabled and de-allocated.
If the bus was previously in a “Suspend” state (SOFE = 0), the USB controller automatically
switches it to the “Resume” state, the Host Wake-Up interrupt (HWUPI) is raised and the SOFE
bit is set by hardware in order to generate SOFs immediately after the USB reset.
30.7.3.5
Pipe Reset
A pipe can be reset at any time by setting its PRSTX bit in the UPRST register. This is recommended before using a pipe upon hardware reset or when a USB bus reset has been sent. This
resets:
•the internal state machine of this pipe;
•the receive and transmit bank FIFO counters;
•all the registers of this pipe (UPCFGX, UPSTAX, UPCONX), except its configuration (ALLOC,
PBK, PSIZE, PTOKEN, PTYPE, PEPNUM, INTFRQ) and its Data Toggle Sequence bit-field
(DTSEQ).
The pipe configuration remains active and the pipe is still enabled.
The pipe reset may be associated with a clear of the data toggle sequence. This can be
achieved by setting the RSTDT bit (by setting the RSTDTS bit).
In the end, the firmware has to clear the PRSTX bit to complete the reset operation and to start
using the FIFO.
30.7.3.6
Pipe Activation
The pipe is maintained inactive and reset (see Section 30.7.3.5 on page 523 for more details) as
long as it is disabled (PENX = 0). The Data Toggle Sequence bit-field (DTSEQ) is also reset.
The algorithm represented on Figure 30-24 must be followed in order to activate a pipe.
523
32058K
AVR32-01/12
AT32UC3A
Figure 30-24. Pipe Activation Algorithm
Pipe
Activation
PENX = 1
Enable the pipe.
UPCFGX
Configure the pipe:
- interrupt request frequency;
- endpoint number;
- type;
- token;
- size;
- number of banks.
Allocate the configured DPRAM
banks.
INTFRQ
PEPNUM
PTYPE
PTOKEN
PSIZE
PBK
ALLOC
CFGOK ==
1?
Yes
Pipe Activated
Test if the pipe configuration is
correct.
No
ERROR
As long as the pipe is not correctly configured (CFGOK = 0), the controller can not send packets
to the device through this pipe.
The CFGOK bit is set by hardware only if the configured size and number of banks are correct
compared to their maximal allowed values for the pipe (see Table 30-1 on page 497) and to the
maximal FIFO size (i.e. the DPRAM size).
See Section 30.7.1.6 on page 506 for more details about DPRAM management.
Once the pipe is correctly configured (CFGOK = 1), only the PTOKEN and INTFRQ bit-fields can
be modified by software. INTFRQ is meaningless for non-interrupt pipes.
When starting an enumeration, the firmware gets the device descriptor by sending a
GET_DESCRIPTOR USB request. This descriptor contains the maximal packet size of the
device default control endpoint (bMaxPacketSize0) and the firmware re-configures the size of
the default control pipe with this size parameter.
30.7.3.7
Address Setup
Once the device has answered the first host requests with the default device address 0, the host
assigns a new address to the device. The host controller has to send a USB reset to the device
and to send a SET_ADDRESS(addr) SETUP request with the new address to be used by the
device. Once this SETUP transaction is over, the firmware writes the new address into the
HADDR bit-field. All following requests, on all pipes, will be performed using this new address.
When the host controller sends a USB reset, the HADDR bit-field is reset by hardware and the
following host requests will be performed using the default device address 0.
30.7.3.8
Remote Wake-Up
The controller host mode enters the Suspend state when the SOFE bit is cleared. No more
“Start of Frame” is sent on the USB bus and the USB device enters the Suspend state 3 ms
later.
The device awakes the host by sending an Upstream Resume (Remote Wake-Up feature).
When the host controller detects a non-idle state on the USB bus, it raises the Host Wake-Up
524
32058K
AVR32-01/12
AT32UC3A
interrupt (HWUPI). If the non-idle bus state corresponds to an Upstream Resume (K state), the
Upstream Resume Received interrupt (RXRSMI) is raised. The firmware has to generate a
Downstream Resume within 1 ms and for at least 20 ms by setting the RESUME bit. It is mandatory to set SOFE before setting RESUME to enter the Ready state, else RESUME will have no
effect.
30.7.3.9
Management of Control Pipes
A control transaction is composed of three stages:
•SETUP;
•Data (IN or OUT);
•Status (OUT or IN).
The firmware has to change the pipe token according to each stage.
For the control pipe, and only for it, each token is assigned a specific initial data toggle
sequence:
•SETUP: Data0;
•IN: Data1;
•OUT: Data1.
30.7.3.10
Management of IN Pipes
IN packets are sent by the USB device controller upon IN requests from the host. All the data
can be read by the firmware which acknowledges or not the bank when it is empty.
The pipe must be configured first.
When the host requires data from the device, the firmware has to select beforehand the IN
request mode with the INMODE bit:
•when INMODE is cleared, the USB controller will perform (INRQ + 1) IN requests before
freezing the pipe;
•when INMODE is set, the USB controller will perform IN requests endlessly when the pipe is
not frozen by the firmware.
The generation of IN requests starts when the pipe is unfrozen (PFREEZE = 0).
The RXINI bit is set by hardware at the same time as FIFOCON when the current bank is full.
This triggers a PXINT interrupt if RXINE = 1.
RXINI shall be cleared by software (by setting the RXINIC bit) to acknowledge the interrupt, what
has no effect on the pipe FIFO.
The firmware then reads from the FIFO and clears the FIFOCON bit to free the bank. If the IN
pipe is composed of multiple banks, this also switches to the next bank. The RXINI and FIFOCON bits are updated by hardware in accordance with the status of the next bank.
RXINI shall always be cleared before clearing FIFOCON.
The RWALL bit is set by hardware when the current bank is not empty, i.e. the software can read
further data from the FIFO.
525
32058K
AVR32-01/12
AT32UC3A
Figure 30-25. Example of an IN Pipe with 1 Data Bank
DATA
(bank 0)
IN
ACK
DATA
(bank 0)
IN
HW
RXINI
ACK
HW
SW
SW
read data from CPU
BANK 0
FIFOCON
read data from CPU
BANK 0
SW
Figure 30-26. Example of an IN Pipe with 2 Data Banks
IN
DATA
(bank 0)
ACK
IN
DATA
(bank 1)
HW
RXINI
HW
SW
read data from CPU
BANK 0
FIFOCON
30.7.3.11
ACK
SW
SW
read data from CPU
BANK 1
Management of OUT Pipes
OUT packets are sent by the host. All the data can be written by the firmware which acknowledges or not the bank when it is full.
The pipe must be configured and unfrozen first.
The TXOUTI bit is set by hardware at the same time as FIFOCON when the current bank is free.
This triggers a PXINT interrupt if TXOUTE = 1.
TXOUTI shall be cleared by software (by setting the TXOUTIC bit) to acknowledge the interrupt,
what has no effect on the pipe FIFO.
The firmware then writes into the FIFO and clears the FIFOCON bit to allow the USB controller
to send the data. If the OUT pipe is composed of multiple banks, this also switches to the next
bank. The TXOUTI and FIFOCON bits are updated by hardware in accordance with the status of
the next bank.
TXOUTI shall always be cleared before clearing FIFOCON.
The RWALL bit is set by hardware when the current bank is not full, i.e. the software can write
further data into the FIFO.
Note that if the firmware decides to switch to the Suspend state (by clearing the SOFE bit) while
a bank is ready to be sent, the USB controller automatically exits this state and the bank is sent.
526
32058K
AVR32-01/12
AT32UC3A
Figure 30-27. Example of an OUT Pipe with 1 Data Bank
DATA
(bank 0)
OUT
ACK
OUT
HW
TXOUTI
SW
SW
write data to CPU
BANK 0
FIFOCON
write data to CPU
BANK 0
SW
SW
Figure 30-28. Example of an OUT Pipe with 2 Data Banks and no Bank Switching Delay
OUT
DATA
(bank 0)
ACK
OUT
DATA
(bank 1)
ACK
HW
TXOUTI
SW
SW
write data to CPU
BANK 0
FIFOCON
SW
write data to CPU
BANK 1
SW
write data to CPU
BANK0
SW
Figure 30-29. Example of an OUT Pipe with 2 Data Banks and a Bank Switching Delay
OUT
DATA
(bank 0)
ACK
OUT
DATA
(bank 1)
ACK
HW
TXOUTI
SW
FIFOCON
30.7.3.12
SW
write data to CPU
BANK 0
SW
SW
write data to CPU
BANK 1
SW
write data to CPU
BANK0
CRC Error
This error exists only for isochronous IN pipes. It raises the CRC Error interrupt (CRCERRI),
what triggers a PXINT interrupt if CRCERRE = 1.
A CRC error can occur during IN stage if the USB controller detects a corrupted received packet.
The IN packet is stored in the bank as if no CRC error had occurred (RXINI is raised).
527
32058K
AVR32-01/12
AT32UC3A
30.7.3.13
Interrupts
See the structure of the USB host interrupt system on Figure 30-6 on page 504.
There are two kinds of host interrupts: processing, i.e. their generation is part of the normal processing, and exception, i.e. errors (not related to CPU exceptions).
30.7.3.13.1 Global Interrupts
The processing host global interrupts are:
•the Device Connection interrupt (DCONNI);
•the Device Disconnection interrupt (DDISCI);
•the USB Reset Sent interrupt (RSTI);
•the Downstream Resume Sent interrupt (RSMEDI);
•the Upstream Resume Received interrupt (RXRSMI);
•the Host Start of Frame interrupt (HSOFI);
•the Host Wake-Up interrupt (HWUPI);
•the Pipe X interrupt (PXINT);
•the DMA Channel X interrupt (DMAXINT).
There is no exception host global interrupt.
30.7.3.13.2 Pipe Interrupts
The processing host pipe interrupts are:
•the Received IN Data interrupt (RXINI);
•the Transmitted OUT Data interrupt (TXOUTI);
•the Transmitted SETUP interrupt (TXSTPI);
•the Short Packet interrupt (SHORTPACKETI);
•the Number of Busy Banks interrupt (NBUSYBK).
The exception host pipe interrupts are:
•the Underflow interrupt (UNDERFI);
•the Pipe Error interrupt (PERRI);
•the NAKed interrupt (NAKEDI);
•the Overflow interrupt (OVERFI);
•the Received STALLed interrupt (RXSTALLDI);
•the CRC Error interrupt (CRCERRI).
30.7.3.13.3 DMA Interrupts
The processing host DMA interrupts are:
•the End of USB Transfer Status interrupt (EOT_STA);
•the End of Channel Buffer Status interrupt (EOCH_BUFF_STA);
•the Descriptor Loaded Status interrupt (DESC_LD_STA).
There is no exception host DMA interrupt.
528
32058K
AVR32-01/12
AT32UC3A
30.7.4
USB DMA Operation
USB packets of any length may be transferred when required by the USB controller. These
transfers always feature sequential addressing. These two characteristics mean that in case of
high USB controller throughput, both HSB ports will benefit from “incrementing burst of unspecified length” since the average access latency of HSB slaves can then be reduced.
The DMA uses word “incrementing burst of unspecified length” of up to 256 beats for both data
transfers and channel descriptor loading. A burst may last on the HSB busses for the duration of
a whole USB packet transfer, unless otherwise broken by the HSB arbitration or the HSB 1 kbyte
boundary crossing.
Packet data HSB bursts may be locked on a DMA buffer basis for drastic overall HSB bus bandwidth performance boost with paged memories. This is because these memories row (or bank)
changes, which are very clock-cycle consuming, will then likely not occur or occur once instead
of dozens of times during a single big USB packet DMA transfer in case other HSB masters
address the memory. This means up to 128 words single cycle unbroken HSB bursts for bulk
pipes/endpoints and 256 words single cycle unbroken bursts for isochronous pipes/endpoints.
This maximal burst length is then controlled by the lowest programmed USB pipe/endpoint size
(PSIZE/EPSIZE) and DMA channel byte length (CH_BYTE_LENGTH).
The USB controller average throughput may be up to nearly 1.5 Mbyte/s. Its average access
latency decreases as burst length increases due to the 0 wait-state side effect of unchanged
pipe/endpoint. Word access allows reducing the HSB bandwidth required for the USB by 4 compared to native byte access. If at least 0 wait-state word burst capability is also provided by the
other DMA HSB bus slaves, each of both DMA HSB busses need less than 1.1% bandwidth
allocation for full USB bandwidth usage at 33 MHz, and less than 0.6% at 66 MHz.
Figure 30-30. Example of DMA Chained List
Transfer Descriptor
USB DMA Channel X Registers
(Current Transfer Descriptor)
Next Descriptor Address
Next Descriptor Address
HSB Address
Transfer Descriptor
Control
Next Descriptor Address
HSB Address
Control
HSB Address
Transfer Descriptor
Control
Next Descriptor Address
HSB Address
Status
Control
NULL
Memory Area
Data Buffer 1
Data Buffer 2
Data Buffer 3
529
32058K
AVR32-01/12
AT32UC3A
30.8
USB User Interface
Table 30-5.
USB PB Memory Map
Offset
Register
Name
Access
Reset Value
0x0000
Device General Control Register
UDCON
Read/Write
0x00000100
0x0004
Device Global Interrupt Register
UDINT
Read-Only
0x00000000
0x0008
Device Global Interrupt Clear Register
UDINTCLR
Write-Only
0x00000000
0x000C
Device Global Interrupt Set Register
UDINTSET
Write-Only
0x00000000
0x0010
Device Global Interrupt Enable Register
UDINTE
Read-Only
0x00000000
0x0014
Device Global Interrupt Enable Clear Register
UDINTECLR
Write-Only
0x00000000
0x0018
Device Global Interrupt Enable Set Register
UDINTESET
Write-Only
0x00000000
0x001C
Endpoint Enable/Reset Register
UERST
Read/Write
0x00000000
0x0020
Device Frame Number Register
UDFNUM
Read-Only
0x00000000
–
–
–
0x0024 - 0x00FC
Reserved
0x0100
Endpoint 0 Configuration Register
UECFG0
Read/Write
0x00000000
0x0104
Endpoint 1 Configuration Register
UECFG1
Read/Write
0x00000000
0x0108
Endpoint 2 Configuration Register
UECFG2
Read/Write
0x00000000
0x010C
Endpoint 3 Configuration Register
UECFG3
Read/Write
0x00000000
0x0110
Endpoint 4 Configuration Register
UECFG4
Read/Write
0x00000000
0x0114
Endpoint 5 Configuration Register
UECFG5
Read/Write
0x00000000
0x0118
Endpoint 6 Configuration Register
UECFG6
Read/Write
0x00000000
–
–
–
+0x004 - 0x012C
Reserved
0x0130
Endpoint 0 Status Register
UESTA0
Read-Only
0x00000100
0x0134
Endpoint 1 Status Register
UESTA1
Read-Only
0x00000100
0x0138
Endpoint 2 Status Register
UESTA2
Read-Only
0x00000100
0x013C
Endpoint 3 Status Register
UESTA3
Read-Only
0x00000100
0x0140
Endpoint 4 Status Register
UESTA4
Read-Only
0x00000100
0x0144
Endpoint 5 Status Register
UESTA5
Read-Only
0x00000100
0x0148
Endpoint 6 Status Register
UESTA6
Read-Only
0x00000100
–
–
–
+0x004 - 0x015C
Reserved
0x0160
Endpoint 0 Status Clear Register
UESTA0CLR
Write-Only
0x00000000
0x0164
Endpoint 1 Status Clear Register
UESTA1CLR
Write-Only
0x00000000
0x0168
Endpoint 2 Status Clear Register
UESTA2CLR
Write-Only
0x00000000
0x016C
Endpoint 3 Status Clear Register
UESTA3CLR
Write-Only
0x00000000
0x0170
Endpoint 4 Status Clear Register
UESTA4CLR
Write-Only
0x00000000
0x0174
Endpoint 5 Status Clear Register
UESTA5CLR
Write-Only
0x00000000
0x0178
Endpoint 6 Status Clear Register
UESTA6CLR
Write-Only
0x00000000
–
–
–
UESTA0SET
Write-Only
0x00000000
+0x04 - 0x018C
0x0190
Reserved
Endpoint 0 Status Set Register
530
32058K
AVR32-01/12
AT32UC3A
Table 30-5.
USB PB Memory Map
Offset
Register
Name
Access
Reset Value
0x0194
Endpoint 1 Status Set Register
UESTA1SET
Write-Only
0x00000000
0x0198
Endpoint 2 Status Set Register
UESTA2SET
Write-Only
0x00000000
0x019C
Endpoint 3 Status Set Register
UESTA3SET
Write-Only
0x00000000
0x01A0
Endpoint 4 Status Set Register
UESTA4SET
Write-Only
0x00000000
0x01A4
Endpoint 5 Status Set Register
UESTA5SET
Write-Only
0x00000000
0x01A8
Endpoint 6 Status Set Register
UESTA6SET
Write-Only
0x00000000
–
–
–
+0x04 - 0x01BC
Reserved
0x01C0
Endpoint 0 Control Register
UECON0
Read-Only
0x00000000
0x01C4
Endpoint 1 Control Register
UECON1
Read-Only
0x00000000
0x01C8
Endpoint 2 Control Register
UECON2
Read-Only
0x00000000
0x01CC
Endpoint 3 Control Register
UECON3
Read-Only
0x00000000
0x01D0
Endpoint 4 Control Register
UECON4
Read-Only
0x00000000
0x01D4
Endpoint 5 Control Register
UECON5
Read-Only
0x00000000
0x01D8
Endpoint 6 Control Register
UECON6
Read-Only
0x00000000
–
–
–
+0x04 - 0x01EC
Reserved
0x01F0
Endpoint 0 Control Set Register
UECON0SET
Write-Only
0x00000000
0x01F4
Endpoint 1 Control Set Register
UECON1SET
Write-Only
0x00000000
0x01F8
Endpoint 2 Control Set Register
UECON2SET
Write-Only
0x00000000
0x01FC
Endpoint 3 Control Set Register
UECON3SET
Write-Only
0x00000000
0x0200
Endpoint 4 Control Set Register
UECON4SET
Write-Only
0x00000000
0x0204
Endpoint 5 Control Set Register
UECON5SET
Write-Only
0x00000000
0x0208
Endpoint 6 Control Set Register
UECON6SET
Write-Only
0x00000000
–
–
–
+0x04 - 0x021C
Reserved
0x0220
Endpoint 0 Control Clear Register
UECON0CLR
Write-Only
0x00000000
0x0224
Endpoint 1 Control Clear Register
UECON1CLR
Write-Only
0x00000000
0x0228
Endpoint 2 Control Clear Register
UECON2CLR
Write-Only
0x00000000
0x022C
Endpoint 3 Control Clear Register
UECON3CLR
Write-Only
0x00000000
0x0230
Endpoint 4 Control Clear Register
UECON4CLR
Write-Only
0x00000000
0x0234
Endpoint 5 Control Clear Register
UECON5CLR
Write-Only
0x00000000
0x0238
Endpoint 6 Control Clear Register
UECON6CLR
Write-Only
0x00000000
–
–
–
UDDMA1_
NEXTDESC
Read/Write
0x00000000
+0x04 - 0x030C
Reserved
0x0310
Device DMA Channel 1 Next Descriptor
Address Register
0x0314
Device DMA Channel 1 HSB Address Register
UDDMA1_
ADDR
Read/Write
0x00000000
0x0318
Device DMA Channel 1 Control Register
UDDMA1_
CONTROL
Read/Write
0x00000000
531
32058K
AVR32-01/12
AT32UC3A
Table 30-5.
USB PB Memory Map
Offset
Register
Name
Access
Reset Value
0x031C
Device DMA Channel 1 Status Register
UDDMA1_
STATUS
Read/Write
0x00000000
0x0320
Device DMA Channel 2 Next Descriptor
Address Register
UDDMA2_
NEXTDESC
Read/Write
0x00000000
0x0324
Device DMA Channel 2 HSB Address Register
UDDMA2_
ADDR
Read/Write
0x00000000
0x0328
Device DMA Channel 2 Control Register
UDDMA2_
CONTROL
Read/Write
0x00000000
0x032C
Device DMA Channel 2 Status Register
UDDMA2_
STATUS
Read/Write
0x00000000
0x0330
Device DMA Channel 3 Next Descriptor
Address Register
UDDMA3_
NEXTDESC
Read/Write
0x00000000
0x0334
Device DMA Channel 3 HSB Address Register
UDDMA3_
ADDR
Read/Write
0x00000000
0x0338
Device DMA Channel 3 Control Register
UDDMA3_
CONTROL
Read/Write
0x00000000
0x033C
Device DMA Channel 3 Status Register
UDDMA3_
STATUS
Read/Write
0x00000000
0x0340
Device DMA Channel 4 Next Descriptor
Address Register
UDDMA4_
NEXTDESC
Read/Write
0x00000000
0x0344
Device DMA Channel 4 HSB Address Register
UDDMA4_
ADDR
Read/Write
0x00000000
0x0348
Device DMA Channel 4 Control Register
UDDMA4_
CONTROL
Read/Write
0x00000000
0x034C
Device DMA Channel 4 Status Register
UDDMA4_
STATUS
Read/Write
0x00000000
0x0350
Device DMA Channel 5 Next Descriptor
Address Register
UDDMA5_
NEXTDESC
Read/Write
0x00000000
0x0354
Device DMA Channel 5 HSB Address Register
UDDMA5_
ADDR
Read/Write
0x00000000
0x0358
Device DMA Channel 5 Control Register
UDDMA5_
CONTROL
Read/Write
0x00000000
0x035C
Device DMA Channel 5 Status Register
UDDMA5_
STATUS
Read/Write
0x00000000
0x0360
Device DMA Channel 6 Next Descriptor
Address Register
UDDMA6_
NEXTDESC
Read/Write
0x00000000
0x0364
Device DMA Channel 6 HSB Address Register
UDDMA6_
ADDR
Read/Write
0x00000000
0x0368
Device DMA Channel 6 Control Register
UDDMA6_
CONTROL
Read/Write
0x00000000
0x036C
Device DMA Channel 6 Status Register
UDDMA6_
STATUS
Read/Write
0x00000000
–
–
–
0x0370 - 0x03FC
Reserved
532
32058K
AVR32-01/12
AT32UC3A
Table 30-5.
USB PB Memory Map
Offset
Register
Name
Access
Reset Value
0x0400
Host General Control Register
UHCON
Read/Write
0x00000000
0x0404
Host Global Interrupt Register
UHINT
Read-Only
0x00000000
0x0408
Host Global Interrupt Clear Register
UHINTCLR
Write-Only
0x00000000
0x040C
Host Global Interrupt Set Register
UHINTSET
Write-Only
0x00000000
0x0410
Host Global Interrupt Enable Register
UHINTE
Read-Only
0x00000000
0x0414
Host Global Interrupt Enable Clear Register
UHINTECLR
Write-Only
0x00000000
0x0418
Host Global Interrupt Enable Set Register
UHINTESET
Write-Only
0x00000000
UPRST
Read/Write
0x00000000
0x0041C
Pipe Enable/Reset Register
0x0420
Host Frame Number Register
UHFNUM
Read/Write
0x00000000
0x0424
Host Address 1 Register
UHADDR1
Read/Write
0x00000000
0x0428
Host Address 2 Register
UHADDR2
Read/Write
0x00000000
–
–
–
+0x04 - 0x04FC
Reserved
0x0500
Pipe 0 Configuration Register
UPCFG0
Read/Write
0x00000000
0x0504
Pipe 1 Configuration Register
UPCFG1
Read/Write
0x00000000
0x0508
Pipe 2 Configuration Register
UPCFG2
Read/Write
0x00000000
0x050C
Pipe 3 Configuration Register
UPCFG3
Read/Write
0x00000000
0x0510
Pipe 4 Configuration Register
UPCFG4
Read/Write
0x00000000
0x0514
Pipe 5 Configuration Register
UPCFG5
Read/Write
0x00000000
0x0518
Pipe 6 Configuration Register
UPCFG6
Read/Write
0x00000000
–
–
–
+0x04 - 0x052C
Reserved
0x0530
Pipe 0 Status Register
UPSTA0
Read-Only
0x00000000
0x0534
Pipe 1 Status Register
UPSTA1
Read-Only
0x00000000
0x0538
Pipe 2 Status Register
UPSTA2
Read-Only
0x00000000
0x053C
Pipe 3 Status Register
UPSTA3
Read-Only
0x00000000
0x0540
Pipe 4 Status Register
UPSTA4
Read-Only
0x00000000
0x0544
Pipe 5 Status Register
UPSTA5
Read-Only
0x00000000
0x0548
Pipe 6 Status Register
UPSTA6
Read-Only
0x00000000
–
–
–
+0x04 - 0x055C
Reserved
0x0560
Pipe 0 Status Clear Register
UPSTA0CLR
Write-Only
0x00000000
0x0564
Pipe 1 Status Clear Register
UPSTA1CLR
Write-Only
0x00000000
0x0568
Pipe 2 Status Clear Register
UPSTA2CLR
Write-Only
0x00000000
0x056C
Pipe 3 Status Clear Register
UPSTA3CLR
Write-Only
0x00000000
0x0570
Pipe 4 Status Clear Register
UPSTA4CLR
Write-Only
0x00000000
0x0574
Pipe 5 Status Clear Register
UPSTA5CLR
Write-Only
0x00000000
0x0578
Pipe 6 Status Clear Register
UPSTA6CLR
Write-Only
0x00000000
–
–
–
+0x04 - 0x058C
Reserved
533
32058K
AVR32-01/12
AT32UC3A
Table 30-5.
USB PB Memory Map
Offset
Register
Name
Access
Reset Value
0x0590
Pipe 0 Status Set Register
UPSTA0SET
Write-Only
0x00000000
0x0594
Pipe 1 Status Set Register
UPSTA1SET
Write-Only
0x00000000
0x0598
Pipe 2 Status Set Register
UPSTA2SET
Write-Only
0x00000000
0x059C
Pipe 3 Status Set Register
UPSTA3SET
Write-Only
0x00000000
0x05A0
Pipe 4 Status Set Register
UPSTA4SET
Write-Only
0x00000000
0x05A4
Pipe 5 Status Set Register
UPSTA5SET
Write-Only
0x00000000
0x05A8
Pipe 6 Status Set Register
UPSTA6SET
Write-Only
0x00000000
–
–
–
+0x04 - 0x05BC
Reserved
0x05C0
Pipe 0 Control Register
UPCON0
Read-Only
0x00000000
0x05C4
Pipe 1 Control Register
UPCON1
Read-Only
0x00000000
0x05C8
Pipe 2 Control Register
UPCON2
Read-Only
0x00000000
0x05CC
Pipe 3 Control Register
UPCON3
Read-Only
0x00000000
0x05D0
Pipe 4 Control Register
UPCON4
Read-Only
0x00000000
0x05D4
Pipe 5 Control Register
UPCON5
Read-Only
0x00000000
0x05D8
Pipe 6 Control Register
UPCON6
Read-Only
0x00000000
–
–
–
+0x04 - 0x05EC
Reserved
0x05F0
Pipe 0 Control Set Register
UPCON0SET
Write-Only
0x00000000
0x05F4
Pipe 1 Control Set Register
UPCON1SET
Write-Only
0x00000000
0x05F8
Pipe 2 Control Set Register
UPCON2SET
Write-Only
0x00000000
0x05FC
Pipe 3 Control Set Register
UPCON3SET
Write-Only
0x00000000
0x0600
Pipe 4 Control Set Register
UPCON4SET
Write-Only
0x00000000
0x0604
Pipe 5 Control Set Register
UPCON5SET
Write-Only
0x00000000
0x0608
Pipe 6 Control Set Register
UPCON6SET
Write-Only
0x00000000
–
–
–
+0x04 - 0x061C
Reserved
0x0620
Pipe 0 Control Clear Register
UPCON0CLR
Write-Only
0x00000000
0x0624
Pipe 1 Control Clear Register
UPCON1CLR
Write-Only
0x00000000
0x0628
Pipe 2 Control Clear Register
UPCON2CLR
Write-Only
0x00000000
0x062C
Pipe 3 Control Clear Register
UPCON3CLR
Write-Only
0x00000000
0x0630
Pipe 4 Control Clear Register
UPCON4CLR
Write-Only
0x00000000
0x0634
Pipe 5 Control Clear Register
UPCON5CLR
Write-Only
0x00000000
0x0638
Pipe 6 Control Clear Register
UPCON6CLR
Write-Only
0x00000000
–
–
–
+0x04 - 0x064C
Reserved
0x0650
Pipe 0 IN Request Register
UPINRQ0
Read/Write
0x00000000
0x0654
Pipe 1 IN Request Register
UPINRQ1
Read/Write
0x00000000
0x0658
Pipe 2 IN Request Register
UPINRQ2
Read/Write
0x00000000
0x065C
Pipe 3 IN Request Register
UPINRQ3
Read/Write
0x00000000
534
32058K
AVR32-01/12
AT32UC3A
Table 30-5.
USB PB Memory Map
Offset
Register
Name
Access
Reset Value
0x0660
Pipe 4 IN Request Register
UPINRQ4
Read/Write
0x00000000
0x0664
Pipe 5 IN Request Register
UPINRQ5
Read/Write
0x00000000
0x0668
Pipe 6 IN Request Register
UPINRQ6
Read/Write
0x00000000
–
–
–
0x066C - 0x067C
Reserved
0x0680
Pipe 0 Error Register
UPERR0
Read/Write
0x00000000
0x0684
Pipe 1 Error Register
UPERR1
Read/Write
0x00000000
0x0688
Pipe 2 Error Register
UPERR2
Read/Write
0x00000000
0x068C
Pipe 3 Error Register
UPERR3
Read/Write
0x00000000
0x0690
Pipe 4 Error Register
UPERR4
Read/Write
0x00000000
0x0694
Pipe 5 Error Register
UPERR5
Read/Write
0x00000000
0x0698
Pipe 6 Error Register
UPERR6
Read/Write
0x00000000
–
–
–
UHDMA1_
NEXTDESC
Read/Write
0x00000000
+0x04 - 0x070C
Reserved
0x0710
Host DMA Channel 1 Next Descriptor Address
Register
0x0714
Host DMA Channel 1 HSB Address Register
UHDMA1_
ADDR
Read/Write
0x00000000
0x0718
Host DMA Channel 1 Control Register
UHDMA1_
CONTROL
Read/Write
0x00000000
0x071C
Host DMA Channel 1 Status Register
UHDMA1_
STATUS
Read/Write
0x00000000
0x0720
Host DMA Channel 2 Next Descriptor Address
Register
UHDMA2_
NEXTDESC
Read/Write
0x00000000
0x0724
Host DMA Channel 2 HSB Address Register
UHDMA2_
ADDR
Read/Write
0x00000000
0x0728
Host DMA Channel 2 Control Register
UHDMA2_
CONTROL
Read/Write
0x00000000
0x072C
Host DMA Channel 2 Status Register
UHDMA2_
STATUS
Read/Write
0x00000000
0x0730
Host DMA Channel 3 Next Descriptor Address
Register
UHDMA3_
NEXTDESC
Read/Write
0x00000000
0x0734
Host DMA Channel 3 HSB Address Register
UHDMA3_
ADDR
Read/Write
0x00000000
0x0738
Host DMA Channel 3 Control Register
UHDMA3_
CONTROL
Read/Write
0x00000000
0x073C
Host DMA Channel 3Status Register
UHDMA3_
STATUS
Read/Write
0x00000000
0x0740
Host DMA Channel 4 Next Descriptor Address
Register
UHDMA4_
NEXTDESC
Read/Write
0x00000000
0x0744
Host DMA Channel 4 HSB Address Register
UHDMA4_
ADDR
Read/Write
0x00000000
535
32058K
AVR32-01/12
AT32UC3A
Table 30-5.
USB PB Memory Map
Offset
Register
Name
Access
Reset Value
0x0748
Host DMA Channel 4 Control Register
UHDMA4_
CONTROL
Read/Write
0x00000000
0x074C
Host DMA Channel 4 Status Register
UHDMA4_
STATUS
Read/Write
0x00000000
0x0750
Host DMA Channel 5 Next Descriptor Address
Register
UHDMA5_
NEXTDESC
Read/Write
0x00000000
0x0754
Host DMA Channel 5 HSB Address Register
UHDMA5_
ADDR
Read/Write
0x00000000
0x0758
Host DMA Channel 5 Control Register
UHDMA5_
CONTROL
Read/Write
0x00000000
0x075C
Host DMA Channel 5 Status Register
UHDMA5_
STATUS
Read/Write
0x00000000
0x0760
Host DMA Channel 6 Next Descriptor Address
Register
UHDMA6_
NEXTDESC
Read/Write
0x00000000
0x0764
Host DMA Channel 6 HSB Address Register
UHDMA6_
ADDR
Read/Write
0x00000000
0x0768
Host DMA Channel 6 Control Register
UHDMA6_
CONTROL
Read/Write
0x00000000
0x076C
Host DMA Channel 6 Status Register
UHDMA6_
STATUS
Read/Write
0x00000000
–
–
–
0x0770 - 0x07FC
Reserved
0x0800
General Control Register
USBCON
Read/Write
0x03004000
0x0804
General Status Register
USBSTA
Read-Only
0x00000400
0x0808
General Status Clear Register
USBSTACLR
Write-Only
0x00000000
0x080C
General Status Set Register
USBSTASET
Write-Only
0x00000000
–
–
–
0x0810-0x0814
Reserved
0x0818
IP Version Register
UVERS
Read-Only
0x00000311
0x081C
IP Features Register
UFEATURES
Read-Only
0x00012467
0x0820
IP PB Address Size Register
UADDRSIZE
Read-Only
0x00001000
0x0824
IP Name Register 1
UNAME1
Read-Only
0x48555342
(“HUSB”)
0x0828
IP Name Register 2
UNAME2
Read-Only
0x004F5447
(“\0OTG“)
0x082C
USB Finite State Machine Status Register
USBFSM
Read-Only
0x00000009
–
–
–
0x0830 - 0x0BFC
Reserved
536
32058K
AVR32-01/12
AT32UC3A
Table 30-6.
USB HSB Memory Map
Offset
Register
Name
Access
Reset Value
0x00000 0x0FFFC
Pipe/Endpoint 0 FIFO Data Register
USB_
FIFO0_DATA
Read/Write
Undefined
0x10000 0x1FFFC
Pipe/Endpoint 1 FIFO Data Register
USB_
FIFO1_DATA
Read/Write
Undefined
0x20000 0x2FFFC
Pipe/Endpoint 2 FIFO Data Register
USB_
FIFO2_DATA
Read/Write
Undefined
0x30000 0x3FFFC
Pipe/Endpoint 3 FIFO Data Register
USB_
FIFO3_DATA
Read/Write
Undefined
0x40000 0x4FFFC
Pipe/Endpoint 4 FIFO Data Register
USB_
FIFO4_DATA
Read/Write
Undefined
0x50000 0x5FFFC
Pipe/Endpoint 5 FIFO Data Register
USB_
FIFO5_DATA
Read/Write
Undefined
0x60000 0x6FFFC
Pipe/Endpoint 6 FIFO Data Register
USB_
FIFO6_DATA
Read/Write
Undefined
+0x00004 0xFFFFC
Reserved
–
–
–
In the following subsections, the bit and bit-field access types use the following flags:
•“r”: readable;
•“w”: writable;
•“u”: may be updated by hardware.
537
32058K
AVR32-01/12
AT32UC3A
30.8.1
USB General Registers
30.8.1.1
USB General Control Register (USBCON)
Offset:
0x0800
Register Name:
USBCON
Access Type:
Read/Write
Reset Value:
0x03004000
31
–
30
–
29
–
28
–
27
–
26
–
25
UIMOD
rw
1
23
–
22
UNLOCK
rw
0
21
20
19
–
18
–
17
0
0
15
USBE
14
FRZCLK
13
VBUSPO
12
OTGPADE
11
HNPREQ
rw
0
rw
1
rw
0
rw
0
7
STOE
rw
0
6
HNPERRE
rw
0
5
ROLEEXE
rw
0
4
BCERRE
rw
0
TIMPAGE
rw
24
UIDE
rw
1
16
TIMVALUE
rw
0
0
10
SRPREQ
9
SRPSEL
8
VBUSHWC
rwu
0
rwu
0
rw
0
rw
0
3
VBERRE
rw
0
2
SRPE
rw
0
1
VBUSTE
rw
0
0
IDTE
rw
0
• IDTE: ID Transition Interrupt Enable
Set to enable the ID Transition interrupt (IDTI).
Clear to disable the ID Transition interrupt (IDTI).
• VBUSTE: VBus Transition Interrupt Enable
Set to enable the VBus Transition interrupt (VBUSTI).
Clear to disable the VBus Transition interrupt (VBUSTI).
• SRPE: SRP Interrupt Enable
Set to enable the SRP interrupt (SRPI).
Clear to disable the SRP interrupt (SRPI).
• VBERRE: VBus Error Interrupt Enable
Set to enable the VBus Error interrupt (VBERRI).
Clear to disable the VBus Error interrupt (VBERRI).
• BCERRE: B-Connection Error Interrupt Enable
Set to enable the B-Connection Error interrupt (BCERRI).
Clear to disable the B-Connection Error interrupt (BCERRI).
• ROLEEXE: Role Exchange Interrupt Enable
Set to enable the Role Exchange interrupt (ROLEEXI).
538
32058K
AVR32-01/12
AT32UC3A
Clear to disable the Role Exchange interrupt (ROLEEXI).
• HNPERRE: HNP Error Interrupt Enable
Set to enable the HNP Error interrupt (HNPERRI).
Clear to disable the HNP Error interrupt (HNPERRI).
• STOE: Suspend Time-Out Interrupt Enable
Set to enable the Suspend Time-Out interrupt (STOI).
Clear to disable the Suspend Time-Out interrupt (STOI).
• VBUSHWC: VBus Hardware Control
Set to disable the hardware control over the USB_VBOF output pin.
Clear to enable the hardware control over the USB_VBOF output pin.
If cleared, then the USB macro considers VBus problems and resets the USB_VBOF output pin in that event.
• SRPSEL: SRP Selection
Set to choose VBus pulsing as SRP method.
Clear to choose data line pulsing as SRP method.
• SRPREQ: SRP Request
Set to initiate an SRP when the controller is in device mode.
Cleared by hardware when the controller is initiating an SRP.
• HNPREQ: HNP Request
When the controller is in device mode:
Set to initiate an HNP.
Cleared by hardware when the controller is initiating an HNP.
When the controller is in host mode:
Set to accept an HNP.
Clear otherwise.
• OTGPADE: OTG Pad Enable
Set to enable the OTG pad.
Clear to disable the OTG pad.
Note that this bit can be set/cleared even if USBE = 0 or FRZCLK = 1. Disabling the USB controller (by clearing the USBE
bit) does not reset this bit.
• VBUSPO: VBus Polarity
When 0, the USB_VBOF output signal is in its default mode (active high).
When 1, the USB_VBOF output signal is inverted (active low).
To be generic. May be useful to control an external VBus power module.
Note that this bit can be set/cleared even if USBE = 0 or FRZCLK = 1. Disabling the USB controller (by clearing the USBE
bit) does not reset this bit.
539
32058K
AVR32-01/12
AT32UC3A
• FRZCLK: Freeze USB Clock
Set to disable the clock inputs (the resume detection is still active). This reduces power consumption. Unless explicitly
stated, all registers then become read-only.
Clear to enable the clock inputs.
Note that this bit can be set/cleared even if USBE = 0 or FRZCLK = 1. Disabling the USB controller (by clearing the USBE
bit) does not reset this bit, but this freezes the clock inputs whatever its value.
• USBE: USB Macro Enable
Set to enable the USB controller.
Clear to disable and reset the USB controller, to disable the USB transceiver and to disable the USB controller clock inputs.
Unless explicitly stated, all registers then become read-only and are reset.
Note that this bit can be set/cleared even if USBE = 0 or FRZCLK = 1.
• TIMVALUE: Timer Value
Set to initialize the new value of the special timer register selected by TIMPAGE.
• TIMPAGE: Timer Page
Write the page value to access a special timer register.
• UNLOCK: Timer Access Unlock
Set to unlock the TIMPAGE and TIMVALUE fields before writing them.
Reset to lock the TIMPAGE and TIMVALUE fields.
Note that the TIMPAGE and TIMVALUE fields can always be read, whatever the value of UNLOCK.
• UIDE: USB_ID Pin Enable
Set to select the USB mode (device/host) from the USB_ID input pin.
Clear to select the USB mode (device/host) with the UIMOD bit.
Note that this bit can be set/cleared even if USBE = 0 or FRZCLK = 1. Disabling the USB controller (by clearing the USBE
bit) does not reset this bit.
• UIMOD: USB Macro Mode
This bit has no effect when UIDE = 1 (USB_ID input pin activated).
Set to select the USB device mode.
Clear to select the USB host mode.
Note that this bit can be set/cleared even if USBE = 0 or FRZCLK = 1. Disabling the USB controller (by clearing the USBE
bit) does not reset this bit.
540
32058K
AVR32-01/12
AT32UC3A
30.8.1.2
USB General Status Register (USBSTA)
Offset:
0x0804
Register Name:
USBSTA
Access Type:
Read-Only
Reset Value:
0x00000400
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
12
10
ID
ru
1
9
VBUSRQ
ru
0
8
–
2
SRPI
ru
0
1
VBUSTI
ru
0
0
IDTI
ru
0
7
STOI
ru
0
6
HNPERRI
ru
0
0
0
11
VBUS
ru
0
5
ROLEEXI
ru
0
4
BCERRI
ru
0
3
VBERRI
ru
0
SPEED
ru
• IDTI: ID Transition Interrupt Flag
Asynchronous interrupt.
Set by hardware when a transition (high to low, low to high) has been detected on the USB_ID input pin. This triggers a
USB interrupt if IDTE = 1.
Shall be cleared by software (by setting the IDTIC bit) to acknowledge the interrupt (USB clock inputs must be enabled
before).
Note that this interrupt is generated even if the clock is frozen by the FRZCLK bit.
• VBUSTI: VBus Transition Interrupt Flag
Asynchronous interrupt.
Set by hardware when a transition (high to low, low to high) has been detected on the VBUS pad. This triggers a USB interrupt if VBUSTE = 1.
Shall be cleared by software (by setting the VBUSTIC bit) to acknowledge the interrupt (USB clock inputs must be enabled
before).
Note that this interrupt is generated even if the clock is frozen by the FRZCLK bit.
• SRPI: SRP Interrupt Flag
Shall only be used in host mode.
Set by hardware when an SRP has been detected. This triggers a USB interrupt if SRPE = 1.
Shall be cleared by software (by setting the SRPIC bit) to acknowledge the interrupt.
541
32058K
AVR32-01/12
AT32UC3A
• VBERRI: VBus Error Interrupt Flag
In host mode, set by hardware when a VBus drop has been detected. This triggers a USB interrupt if VBERRE = 1.
Shall be cleared by software (by setting the VBERRIC bit) to acknowledge the interrupt.
Note that if a VBus problem occurs, then the VBERRI interrupt is generated even if the USB macro does not go to an error
state because of VBUSHWC = 1.
• BCERRI: B-Connection Error Interrupt Flag
In host mode, set by hardware when an error occurs during the B-connection. This triggers a USB interrupt if BCERRE = 1.
Shall be cleared by software (by setting the BCERRIC bit) to acknowledge the interrupt.
• ROLEEXI: Role Exchange Interrupt Flag
Set by hardware when the USB controller has successfully switched its mode because of an HNP negotiation (host to
device or device to host). This triggers a USB interrupt if ROLEEXE = 1.
Shall be cleared by software (by setting the ROLEEXIC bit) to acknowledge the interrupt.
• HNPERRI: HNP Error Interrupt Flag
In device mode, set by hardware when an error has been detected during an HNP negotiation. This triggers a USB interrupt
if HNPERRE = 1.
Shall be cleared by software (by setting the HNPERRIC bit) to acknowledge the interrupt.
• STOI: Suspend Time-Out Interrupt Flag
In host mode, set by hardware when a time-out error (more than 200ms) has been detected after a suspend. This triggers
a USB interrupt if STOE = 1.
Shall be cleared by software (by setting the STOIC bit) to acknowledge the interrupt.
• VBUSRQ: VBus Request
In host mode, set by software (by setting the VBUSRQS bit) to assert the USB_VBOF output pin in order to enable the
VBus power supply generation.
Cleared by software by setting the VBUSRQC bit.
Cleared by hardware when a VBus error occurs when VBUSHWC = 0.
• ID: USB_ID Pin State
Set/cleared by hardware and reflects the state of the USB_ID input pin, even if USBE = 0.
• VBUS: VBus Level
Set/cleared by hardware and reflects the level of the VBus line, even if USBE = 0.
This bit can be used in device mode to monitor the USB bus connection state of the application.
• SPEED: Speed Status
Set by hardware according to the controller speed mode:
SPEED
Speed Status
0
0
FULL-SPEED mode
1
0
LOW-SPEED mode
X
1
Reserved
Shall only be used in host mode.
542
32058K
AVR32-01/12
AT32UC3A
30.8.1.3
USB General Status Clear Register (USBSTACLR)
Offset:
0x0808
Register Name:
USBSTACLR
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
VBUSRQC
w
0
8
–
7
STOIC
w
0
6
HNPERRIC
w
0
5
ROLEEXIC
w
0
4
BCERRIC
w
0
3
VBERRIC
w
0
2
SRPIC
w
0
1
VBUSTIC
w
0
0
IDTIC
w
0
• IDTIC: ID Transition Interrupt Flag Clear
Set to clear IDTI.
Clearing has no effect.
Always read as 0.
• VBUSTIC: VBus Transition Interrupt Flag Clear
Set to clear VBUSTI.
Clearing has no effect.
Always read as 0.
• SRPIC: SRP Interrupt Flag Clear
Set to clear SRPI.
Clearing has no effect.
Always read as 0.
• VBERRIC: VBus Error Interrupt Flag Clear
Set to clear VBERRI.
Clearing has no effect.
Always read as 0.
• BCERRIC: B-Connection Error Interrupt Flag Clear
Set to clear BCERRI.
543
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• ROLEEXIC: Role Exchange Interrupt Flag Clear
Set to clear ROLEEXI.
Clearing has no effect.
Always read as 0.
• HNPERRIC: HNP Error Interrupt Flag Clear
Set to clear HNPERRI.
Clearing has no effect.
Always read as 0.
• STOIC: Suspend Time-Out Interrupt Flag Clear
Set to clear STOI.
Clearing has no effect.
Always read as 0.
• VBUSRQC: VBus Request Clear
Set to clear VBUSRQ.
Clearing has no effect.
Always read as 0.
544
32058K
AVR32-01/12
AT32UC3A
30.8.1.4
USB General Status Set Register (USBSTASET)
Offset:
0x080C
Register Name:
USBSTASET
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
VBUSRQS
w
0
8
–
7
STOIS
w
0
6
HNPERRIS
w
0
5
ROLEEXIS
w
0
4
BCERRIS
w
0
3
VBERRIS
w
0
2
SRPIS
w
0
1
VBUSTIS
w
0
0
IDTIS
w
0
• IDTIS: ID Transition Interrupt Flag Set
Set to set IDTI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• VBUSTIS: VBus Transition Interrupt Flag Set
Set to set VBUSTI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• SRPIS: SRP Interrupt Flag Set
Set to set SRPI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• VBERRIS: VBus Error Interrupt Flag Set
Set to set VBERRI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• BCERRIS: B-Connection Error Interrupt Flag Set
Set to set BCERRI, what may be useful for test or debug purposes.
545
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• ROLEEXIS: Role Exchange Interrupt Flag Set
Set to set ROLEEXI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• HNPERRIS: HNP Error Interrupt Flag Set
Set to set HNPERRI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• STOIS: Suspend Time-Out Interrupt Flag Set
Set to set STOI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• VBUSRQS: VBus Request Set
Set to set VBUSRQ.
Clearing has no effect.
Always read as 0.
546
32058K
AVR32-01/12
AT32UC3A
30.8.1.5
USB IP Version Register (UVERS)
Offset:
0x0818
Register Name:
UVERS
Access Type:
Read-Only
Read Value:
0x00000260
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
16
0
17
METAL_FIX_NUM
r
0
10
9
8
1
0
15
14
13
12
11
VERSION_NUM
r
0
7
6
0
2
5
4
3
2
VERSION_NUM
r
6
0
• VERSION_NUM: IP Version Number
This field indicates the version number of the USB macro IP, encoded with 1 version digit per nibble, e.g. 0x0260 for version 2.6.0.
• METAL_FIX_NUM: Number of Metal Fixes
This field indicates the number of metal fixes of the USB macro IP.
547
32058K
AVR32-01/12
AT32UC3A
30.8.1.6
USB IP Features Register (UFEATURES)
Offset:
0x081C
Register Name:
UFEATURES
Access Type:
Read-Only
Read Value:
0x00012467
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
BYTE_WRITE
_DPRAM
r
0
14
13
12
11
10
9
8
0
r
1
0
0
1
0
0
7
DMA_BUFFER
_SIZE
r
0
6
5
4
3
2
1
0
1
1
FIFO_MAX_SIZE
DMA_FIFO_WORD_DEPTH
r
DMA_CHANNEL_NBR
r
1
1
EPT_NBR_MAX
r
0
0
1
• EPT_NBR_MAX: Maximal Number of Pipes/Endpoints
This field indicates the number of hardware-implemented pipes/endpoints:
EPT_NBR_MAX
Maximal Number of Pipes/Endpoints
0
0
0
0
16
0
0
0
1
1
0
0
1
0
2
...
1
1
1
1
15
• DMA_CHANNEL_NBR: Number of DMA Channels
This field indicates the number of hardware-implemented DMA channels:
DMA_CHANNEL_NBR
Number of DMA Channels
0
0
0
Reserved
0
0
1
1
0
1
0
2
...
1
1
1
7
548
32058K
AVR32-01/12
AT32UC3A
• DMA_BUFFER_SIZE: DMA Buffer Size
This field indicates the size of the DMA buffer:
DMA_BUFFER_SIZE
DMA Buffer Size
0
16 bits
1
24 bits
• DMA_FIFO_WORD_DEPTH: DMA FIFO Depth in Words
This field indicates the DMA FIFO depth controller in words:
DMA_FIFO_WORD_DEPTH
DMA FIFO Depth in Words
0
0
0
0
16
0
0
0
1
1
0
0
1
0
2
...
1
1
1
1
15
• FIFO_MAX_SIZE: Maximal FIFO Size
This field indicates the maximal FIFO size, i.e. the DPRAM size:
FIFO_MAX_SIZE
Maximal FIFO Size
0
0
0
< 256 bytes
0
0
1
< 512 bytes
0
1
0
< 1024 bytes
0
1
1
< 2048 bytes
1
0
0
< 4096 bytes
1
0
1
< 8192 bytes
1
1
0
< 16384 bytes
1
1
1
>= 16384 bytes
• BYTE_WRITE_DPRAM: DPRAM Byte-Write Capability
This field indicates whether the DPRAM is byte-write capable:
BYTE_WRITE_DPRAM
DPRAM Byte-Write Capability
0
DPRAM byte write lanes have shadow logic implemented in the USB macro IP interface.
1
DPRAM is natively byte-write capable.
549
32058K
AVR32-01/12
AT32UC3A
30.8.1.7
USB IP PB Address Size Register (UADDRSIZE)
Offset:
0x0820
Register Name:
UADDRSIZE
Access Type:
Read-Only
Read Value:
0x00001000
31
30
29
28
27
26
25
24
0
0
0
0
19
18
17
16
0
0
0
0
11
10
9
8
0
0
0
0
3
2
1
0
0
0
0
0
UADDRSIZE
r
0
0
0
0
23
22
21
20
UADDRSIZE
r
0
0
0
0
15
14
13
12
UADDRSIZE
r
0
0
0
1
7
6
5
4
UADDRSIZE
r
0
0
0
0
• UADDRSIZE: IP PB Address Size
This field indicates the size of the PB address space reserved for the USB macro IP interface (2 at the power of the number
of bits reserved to encode the PB addresses of the USB macro IP interface relatively to its base address).
550
32058K
AVR32-01/12
AT32UC3A
30.8.1.8
USB IP Name Register 1 (UNAME1)
Offset:
0x0824
Register Name:
UNAME1
Access Type:
Read-Only
Read Value:
0x48555342 (“HUSB”)
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
UNAME1
r
“H”
23
22
21
20
UNAME1
r
“U”
15
14
13
12
UNAME1
r
“S”
7
6
5
4
UNAME1
r
“B”
• UNAME1: IP Name Part 1
This field indicates the 1st part of the ASCII-encoded name of the USB macro IP.
551
32058K
AVR32-01/12
AT32UC3A
30.8.1.9
USB IP Name Register 2 (UNAME2)
Offset:
0x0828
Register Name:
UNAME2
Access Type:
Read-Only
Read Value:
0x004F5447 (“\0OTG“)
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
UNAME2
r
“\0“
23
22
21
20
UNAME2
r
“O“
15
14
13
12
UNAME2
r
“T“
7
6
5
4
UNAME2
r
“G“
• UNAME2: IP Name Part 2
This field indicates the 2nd part of the ASCII-encoded name of the USB macro IP.
552
32058K
AVR32-01/12
AT32UC3A
30.8.1.10
USB Finite State Machine Status Register (USBFSM)
Offset:
0x082C
Register Name:
USBFSM
Access Type:
Read-Only
Read Value:
0x00000009
31
-
30
-
29
-
28
-
27
-
26
-
25
-
24
-
0
0
0
0
0
0
0
0
23
-
22
-
21
-
20
-
19
-
18
-
17
-
16
-
0
0
0
0
0
0
0
0
15
-
14
-
13
-
12
-
11
-
10
-
9
-
8
-
0
0
0
0
0
0
0
0
7
-
6
-
5
-
4
-
3
2
1
0
0
0
0
0
1
0
1
DRDSTATE
r
0
• DRDSTATE
This field indicates the state of the USB controller.
Refer to the OTG specification for more details.
USBFSM
Description
0
a_idle state : this is the start state for A-devices (when the ID pin is 0)
1
a_wait_vrise : In this state, the A-device waits for the voltage on VBus to rise above the Adevice VBus Valid threshold (4.4 V).
2
a_wait_bcon : In this state, the A-device waits for the B-device to signal a connection.
3
a_host : In this state, the A-device that operates in Host mode is operational.
4
a_suspend : The A-device operating as a host is in the suspend mode.
5
a_peripheral : The A-device operates as a peripheral.
6
a_wait_vfall : In this state, the A-device waits for the voltage on VBus to drop below the Adevice Session Valid threshold (1.4 V).
7
a_vbus_err : In this state, the A-device waits for recovery of the overcurrent condition that
caused it to enter this state.
8
a_wait_discharge : In this state, the A-device waits for the data usb line to discharge (100
us).
9
b_idle : this is the start state for B-device (when the ID pin is 1).
10
b_peripheral : In this state, the B-device acts as the peripheral.
553
32058K
AVR32-01/12
AT32UC3A
USBFSM
Description
11
b_wait_begin_hnp : In this state, the B-device is in suspend mode and waits until 3 ms
before initiating the HNP protocol if requested.
12
b_wait_discharge : In this state, the B-device waits for the data usb line to discharge (100
us) before becoming Host.
13
b_wait_acon : In this state, the B-device waits for the A-device to signal a connect before
becoming B-Host.
14
b_host : In this state, the B-device acts as the Host.
15
b_srp_init : In this state, the B-device attempts to start a session using the SRP protocol.
554
32058K
AVR32-01/12
AT32UC3A
30.8.2
USB Device Registers
30.8.2.1
USB Device General Control Register (UDCON)
Offset:
0x0000
Register Name:
UDCON
Access Type:
Read/Write
Reset Value:
0x00000100
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
LS
11
–
10
–
9
RMWKUP
8
DETACH
rwu
0
rw
1
2
1
0
0
0
0
rw
0
7
ADDEN
rwu
0
6
5
4
0
0
0
3
UADD
rwu
0
• UADD: USB Address
Set to configure the device address.
Cleared by hardware upon receiving a USB reset.
• ADDEN: Address Enable
Set to activate the UADD field (USB address).
Cleared by hardware upon receiving a USB reset.
Clearing by software has no effect.
• DETACH: Detach
Set to physically detach the device (disconnect internal pull-up resistor from D+ and D-).
Clear to reconnect the device.
• RMWKUP: Remote Wake-Up
Set to send an upstream resume to the host for a remote wake-up.
Cleared by hardware upon receiving a USB reset or once the upstream resume has been sent.
Clearing by software has no effect.
• LS: Low-Speed Mode Force
Set to force the low-speed mode.
555
32058K
AVR32-01/12
AT32UC3A
Clear to unforce the low-speed mode. Then, the full-speed mode is active.
Note that this bit can be set/cleared even if USBE = 0 or FRZCLK = 1. Disabling the USB controller (by clearing the USBE
bit) does not reset this bit.
556
32058K
AVR32-01/12
AT32UC3A
30.8.2.2
USB Device Global Interrupt Register (UDINT)
Offset:
0x0004
Register Name:
UDINT
Access Type:
Read-Only
Reset Value:
0x00000000
31
30
DMA6INT
ru
0
29
DMA5INT
ru
0
28
DMA4INT
ru
0
27
DMA3INT
ru
0
26
DMA2INT
ru
0
25
DMA1INT
ru
0
24
–
23
–
22
–
21
–
20
19
18
EP6INT
ru
0
17
EP5INT
ru
0
16
EP4INT
ru
0
15
EP3INT
ru
0
14
EP2INT
ru
0
13
EP1INT
ru
0
12
EP0INT
ru
0
11
–
10
–
9
–
8
–
7
–
6
UPRSM
ru
0
5
EORSM
ru
0
4
WAKEUP
ru
0
3
EORST
ru
0
2
SOF
ru
0
1
–
0
SUSP
ru
0
• SUSP: Suspend Interrupt Flag
Set by hardware when a USB “Suspend” idle bus state has been detected for 3 frame periods (J state for 3 ms). This triggers a USB interrupt if SUSPE = 1.
Shall be cleared by software (by setting the SUSPC bit) to acknowledge the interrupt.
Cleared by hardware when a Wake-Up interrupt (WAKEUP) is raised.
• SOF: Start of Frame Interrupt Flag
Set by hardware when a USB “Start of Frame” PID (SOF) has been detected (every 1 ms). This triggers a USB interrupt if
SOFE = 1. The FNUM field is updated.
Shall be cleared by software (by setting the SOFC bit) to acknowledge the interrupt.
• EORST: End of Reset Interrupt Flag
Set by hardware when a USB “End of Reset” has been detected. This triggers a USB interrupt if EORSTE = 1.
Shall be cleared by software (by setting the EORSTC bit) to acknowledge the interrupt.
• WAKEUP: Wake-Up Interrupt Flag
Asynchronous interrupt.
Set by hardware when the USB controller is reactivated by a filtered non-idle signal from the lines (not by an upstream
resume). This triggers an interrupt if WAKEUPE = 1.
Shall be cleared by software (by setting the WAKEUPC bit) to acknowledge the interrupt (USB clock inputs must be
enabled before).
Cleared by hardware when a Suspend interrupt (SUSP) is raised.
557
32058K
AVR32-01/12
AT32UC3A
Note that this interrupt is generated even if the clock is frozen by the FRZCLK bit.
• EORSM: End of Resume Interrupt Flag
Set by hardware when the USB controller detects a valid “End of Resume” signal initiated by the host. This triggers a USB
interrupt if EORSME = 1.
Shall be cleared by software (by setting the EORSMC bit) to acknowledge the interrupt.
• UPRSM: Upstream Resume Interrupt Flag
Set by hardware when the USB controller sends a resume signal called “Upstream Resume”. This triggers a USB interrupt
if UPRSME = 1.
Shall be cleared by software (by setting the UPRSMC bit) to acknowledge the interrupt (USB clock inputs must be enabled
before).
• EPXINT, X in [0..6]: Endpoint X Interrupt Flag
Set by hardware when an interrupt is triggered by the endpoint X (UESTAX, UECONX). This triggers a USB interrupt if
EPXINTE = 1.
Cleared by hardware when the interrupt source is serviced.
• DMAXINT, X in [1..6]: DMA Channel X Interrupt Flag
Set by hardware when an interrupt is triggered by the DMA channel X. This triggers a USB interrupt if DMAXINTE = 1.
Cleared by hardware when the UDDMAX_STATUS interrupt source is cleared.
558
32058K
AVR32-01/12
AT32UC3A
30.8.2.3
USB Device Global Interrupt Clear Register (UDINTCLR)
Offset:
0x0008
Register Name:
UDINTCLR
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
UPRSMC
w
0
5
EORSMC
w
0
4
WAKEUPC
w
0
3
EORSTC
w
0
2
SOFC
w
0
1
–
0
SUSPC
w
0
• SUSPC: Suspend Interrupt Flag Clear
Set to clear SUSP.
Clearing has no effect.
Always read as 0.
• SOFC: Start of Frame Interrupt Flag Clear
Set to clear SOF.
Clearing has no effect.
Always read as 0.
• EORSTC: End of Reset Interrupt Flag Clear
Set to clear EORST.
Clearing has no effect.
Always read as 0.
• WAKEUPC: Wake-Up Interrupt Flag Clear
Set to clear WAKEUP.
Clearing has no effect.
Always read as 0.
• EORSMC: End of Resume Interrupt Flag Clear
Set to clear EORSM.
559
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• UPRSMC: Upstream Resume Interrupt Flag Clear
Set to clear UPRSM.
Clearing has no effect.
Always read as 0.
560
32058K
AVR32-01/12
AT32UC3A
30.8.2.4
USB Device Global Interrupt Set Register (UDINTSET)
Offset:
0x000C
Register Name:
UDINTSET
Access Type:
Write-Only
Read Value:
0x00000000
31
30
DMA7INTS
w
0
29
DMA5INTS
w
0
28
DMA4INTS
w
0
27
DMA3INTS
w
0
26
DMA2INTS
w
0
25
DMA1INTS
w
0
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
UPRSMS
w
0
5
EORSMS
w
0
4
WAKEUPS
w
0
3
EORSTS
w
0
2
SOFS
w
0
1
–
0
SUSPS
w
0
• SUSPS: Suspend Interrupt Flag Set
Set to set SUSP, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• SOFS: Start of Frame Interrupt Flag Set
Set to set SOF, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• EORSTS: End of Reset Interrupt Flag Set
Set to set EORST, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• WAKEUPS: Wake-Up Interrupt Flag Set
Set to set WAKEUP, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• EORSMS: End of Resume Interrupt Flag Set
Set to set EORSM, what may be useful for test or debug purposes.
561
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• UPRSMS: Upstream Resume Interrupt Flag Set
Set to set UPRSM, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• DMAXINTS, X in [1..6]: DMA Channel X Interrupt Flag Set
Set to set DMAXINT, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
562
32058K
AVR32-01/12
AT32UC3A
30.8.2.5
USB Device Global Interrupt Enable Register (UDINTE)
Offset:
0x0010
Register Name:
UDINTE
Access Type:
Read-Only
Reset Value:
0x00000000
31
30
DMA6INTE
r
0
29
DMA5INTE
r
0
28
DMA4INTE
r
0
27
DMA3INTE
r
0
26
DMA2INTE
r
0
25
DMA1INTE
r
0
24
–
23
–
22
–
21
–
20
19
18
EP6INTE
r
0
17
EP5INTE
r
0
16
EP4INTE
r
0
15
EP3INTE
r
0
14
EP2INTE
r
0
13
EP1INTE
r
0
12
EP0INTE
r
0
11
–
10
–
9
–
8
–
7
–
6
UPRSME
r
0
5
EORSME
r
0
4
WAKEUPE
r
0
3
EORSTE
r
0
2
SOFE
r
0
1
–
0
SUSPE
r
0
• SUSPE: Suspend Interrupt Enable
Set by software (by setting the SUSPES bit) to enable the Suspend interrupt (SUSP).
Clear by software (by setting the SUSPEC bit) to disable the Suspend interrupt (SUSP).
• SOFE: Start of Frame Interrupt Enable
Set by software (by setting the SOFES bit) to enable the Start of Frame interrupt (SOF).
Clear by software (by setting the SOFEC bit) to disable the Start of Frame interrupt (SOF).
• EORSTE: End of Reset Interrupt Enable
Set by software (by setting the EORSTES bit) to enable the End of Reset interrupt (EORST).
Clear by software (by setting the EORSTEC bit) to disable the End of Reset interrupt (EORST).
• WAKEUPE: Wake-Up Interrupt Enable
Set by software (by setting the WAKEUPES bit) to enable the Wake-Up interrupt (WAKEUP).
Clear by software (by setting the WAKEUPEC bit) to disable the Wake-Up interrupt (WAKEUP).
• EORSME: End of Resume Interrupt Enable
Set by software (by setting the EORSMES bit) to enable the End of Resume interrupt (EORSM).
Clear by software (by setting the EORSMEC bit) to disable the End of Resume interrupt (EORSM).
• UPRSME: Upstream Resume Interrupt Enable
Set by software (by setting the UPRSMES bit) to enable the Upstream Resume interrupt (UPRSM).
Clear by software (by setting the UPRSMEC bit) to disable the Upstream Resume interrupt (UPRSM).
563
32058K
AVR32-01/12
AT32UC3A
• EPXINTE, X in [0..6]: Endpoint X Interrupt Enable
Set by software (by setting the EPXINTES bit) to enable the Endpoint X interrupt (EPXINT).
Clear by software (by setting the EPXINTEC bit) to disable the Endpoint X interrupt (EPXINT).
• DMAXINTE, X in [1..6]: DMA Channel X Interrupt Enable
Set by software (by setting the DMAXINTES bit) to enable the DMA Channel X interrupt (DMAXINT).
Clear by software (by setting the DMAXINTEC bit) to disable the DMA Channel X interrupt (DMAXINT).
564
32058K
AVR32-01/12
AT32UC3A
30.8.2.6
USB Device Global Interrupt Enable Clear Register (UDINTECLR)
Offset:
0x0014
Register Name:
UDINTECLR
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
DMA6INTEC
w
0
29
DMA5INTEC
w
0
28
DMA4INTEC
w
0
27
DMA3INTEC
w
0
26
DMA2INTEC
w
0
25
DMA1INTEC
w
0
24
–
23
–
22
–
21
–
20
–
19
–
18
EP6INTEC
w
0
17
EP5INTEC
w
0
16
EP4INTEC
w
0
15
EP3INTEC
w
0
14
EP2INTEC
w
0
13
EP1INTEC
w
0
12
EP0INTEC
w
0
11
–
10
–
9
–
8
–
7
–
6
UPRSMEC
w
0
5
EORSMEC
w
0
4
WAKEUPEC
w
0
3
EORSTEC
w
0
2
SOFEC
w
0
1
–
0
SUSPEC
w
0
• SUSPEC: Suspend Interrupt Enable Clear
Set to clear SUSPE.
Clearing has no effect.
Always read as 0.
• SOFEC: Start of Frame Interrupt Enable Clear
Set to clear SOFE.
Clearing has no effect.
Always read as 0.
• EORSTEC: End of Reset Interrupt Enable Clear
Set to clear EORSTE.
Clearing has no effect.
Always read as 0.
• WAKEUPEC: Wake-Up Interrupt Enable Clear
Set to clear WAKEUPE.
Clearing has no effect.
Always read as 0.
• EORSMEC: End of Resume Interrupt Enable Clear
Set to clear EORSME.
565
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• UPRSMEC: Upstream Resume Interrupt Enable Clear
Set to clear UPRSME.
Clearing has no effect.
Always read as 0.
• EPXINTEC, X in [0..6]: Endpoint X Interrupt Enable Clear
Set to clear EPXINTE.
Clearing has no effect.
Always read as 0.
• DMAXINTEC, X in [1..6]: DMA Channel X Interrupt Enable Clear
Set to clear DMAXINTE.
Clearing has no effect.
Always read as 0.
566
32058K
AVR32-01/12
AT32UC3A
30.8.2.7
USB Device Global Interrupt Enable Set Register (UDINTESET)
Offset:
0x0018
Register Name:
UDINTESET
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
DMA6INTES
w
0
29
DMA5INTES
w
0
28
DMA4INTES
w
0
27
DMA3INTES
w
0
26
DMA2INTES
w
0
25
DMA1INTES
w
0
24
–
23
–
22
–
21
–
20
–
19
–
18
EP6INTES
w
0
17
EP5INTES
w
0
16
EP4INTES
w
0
15
EP3INTES
w
0
14
EP2INTES
w
0
13
EP1INTES
w
0
12
EP0INTES
w
0
11
–
10
–
9
–
8
–
7
–
6
UPRSMES
w
0
5
EORSMES
w
0
4
WAKEUPES
w
0
3
EORSTES
w
0
2
SOFES
w
0
1
–
0
SUSPES
w
0
• SUSPES: Suspend Interrupt Enable Set
Set to set SUSPE.
Clearing has no effect.
Always read as 0.
• SOFES: Start of Frame Interrupt Enable Set
Set to set SOFE.
Clearing has no effect.
Always read as 0.
• EORSTES: End of Reset Interrupt Enable Set
Set to set EORSTE.
Clearing has no effect.
Always read as 0.
• WAKEUPES: Wake-Up Interrupt Enable Set
Set to set WAKEUPE.
Clearing has no effect.
Always read as 0.
• EORSMES: End of Resume Interrupt Enable Set
Set to set EORSME.
567
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• UPRSMES: Upstream Resume Interrupt Enable Set
Set to set UPRSME.
Clearing has no effect.
Always read as 0.
• EPXINTES, X in [0..6]: Endpoint X Interrupt Enable Set
Set to set EPXINTE.
Clearing has no effect.
Always read as 0.
• DMAXINTES, X in [1..6]: DMA Channel X Interrupt Enable Set
Set to set DMAXINTE.
Clearing has no effect.
Always read as 0.
568
32058K
AVR32-01/12
AT32UC3A
30.8.2.8
USB Device Frame Number Register (UDFNUM)
Offset:
0x0020
Register Name:
UDFNUM
Access Type:
Read-Only
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
FNCERR
ru
0
14
–
13
12
11
10
9
8
7
6
0
0
FNUM
ru
0
0
0
0
0
0
5
FNUM
ru
0
4
3
2
–
1
–
0
–
0
0
• FNUM: Frame Number
Set by hardware. These bits are the 11-bit frame number information. They are provided in the last received SOF packet.
Cleared by hardware upon receiving a USB reset.
Note that FNUM is updated even if a corrupted SOF is received.
• FNCERR: Frame Number CRC Error
Set by hardware when a corrupted frame number is received. This bit and the SOF interrupt flag are updated at the same
time.
Cleared by hardware upon receiving a USB reset.
569
32058K
AVR32-01/12
AT32UC3A
30.8.2.9
USB Endpoint Enable/Reset Register (UERST)
Offset:
0x001C
Register Name:
UERST
Access Type:
Read/Write
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
EPRST6
rwu
0
21
EPRST5
rwu
0
20
EPRST4
rwu
0
19
EPRST3
rwu
0
18
EPRST2
rwu
0
17
EPRST1
rwu
0
16
EPRST0
rwu
0
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
EPEN6
rw
0
5
EPEN5
rw
0
4
EPEN4
rw
0
3
EPEN3
rw
0
2
EPEN2
rw
0
1
EPEN1
rw
0
0
EPEN0
rw
0
• EPENX, X in [0..6]: Endpoint X Enable
Set to enable the endpoint X.
Clear to disable the endpoint X, what forces the endpoint X state to inactive (no answer to USB requests) and resets the
endpoint X registers (UECFGX, UESTAX, UECONX) but not the endpoint configuration (ALLOC, EPBK, EPSIZE, EPDIR,
EPTYPE).
• EPRSTX, X in [0..6]: Endpoint X Reset
Set by software to reset the endpoint X FIFO prior to any other operation, upon hardware reset or when a USB bus reset
has been received. This resets the endpoint X registers (UECFGX, UESTAX, UECONX) but not the endpoint configuration
(ALLOC, EPBK, EPSIZE, EPDIR, EPTYPE).
All the endpoint mechanism (FIFO counter, reception, transmission, etc.) is reset apart from the Data Toggle Sequence
field (DTSEQ) which can be cleared by setting the RSTDT bit (by setting the RSTDTS bit).
The endpoint configuration remains active and the endpoint is still enabled.
Then, clear by software to complete the reset operation and to start using the FIFO.
Cleared by hardware upon receiving a USB reset.
570
32058K
AVR32-01/12
AT32UC3A
30.8.2.10
USB Endpoint X Configuration Register (UECFGX)
Offset:
0x0100 + X . 0x04
Register Name:
UECFGX, X in [0..6]
Access Type:
Read/Write
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
11
10
–
9
AUTOSW
rwu
0
8
EPDIR
rwu
0
2
1
ALLOC
rwu
0
0
–
7
–
6
5
EPSIZE
rwu
0
0
EPTYPE
rwu
0
0
4
3
EPBK
rwu
0
0
0
• ALLOC: Endpoint Memory Allocate
Set to allocate the endpoint memory.
Clear to free the endpoint memory.
Cleared by hardware upon receiving a USB reset (except for the endpoint 0).
Note that after setting this bit, the user should check the CFGOK bit to know whether the allocation of this endpoint is
correct.
• EPBK: Endpoint Banks
Set to select the number of banks for the endpoint:
EPBK
Endpoint Banks
0
0
1 (single-bank endpoint)
0
1
2 (double-bank endpoint)
1
0
3 (triple-bank endpoint)
1
1
Reserved
For control endpoints, a single-bank endpoint (00b) should be selected.
Cleared by hardware upon receiving a USB reset (except for the endpoint 0).
571
32058K
AVR32-01/12
AT32UC3A
• EPSIZE: Endpoint Size
Set to select the size of each endpoint bank:
EPSIZE
Endpoint Size
0
0
0
8 bytes
0
0
1
16 bytes
0
1
0
32 bytes
0
1
1
64 bytes
1
0
0
128 bytes
1
0
1
256 bytes
1
1
0
512 bytes
1
1
1
1024 bytes
Cleared by hardware upon receiving a USB reset (except for the endpoint 0).
• EPDIR: Endpoint Direction
Set to select the endpoint direction:
EPDIR
Endpoint Direction
0
OUT
1
IN (not for control endpoints)
Cleared by hardware upon receiving a USB reset.
• AUTOSW: Automatic Switch
Set to automatically switch bank when it is ready.
Clear to disable the automatic bank switching.
Cleared by hardware upon receiving a USB reset.
• EPTYPE: Endpoint Type
Set to select the endpoint type:
EPTYPE
Endpoint Type
0
0
Control
0
1
Isochronous
1
0
Bulk
1
1
Interrupt
Cleared by hardware upon receiving a USB reset.
572
32058K
AVR32-01/12
AT32UC3A
30.8.2.11
USB Endpoint X Status Register (UESTAX)
Offset:
0x0130 + X . 0x04
Register Name:
UESTAX, X in [0..6]
Access Type:
Read-Only
Reset Value:
0x00000100
31
–
23
30
29
28
0
0
0
21
20
0
0
0
14
13
22
BYCT
ru
0
15
CURRBK
ru
12
NBUSYBK
ru
27
BYCT
ru
0
26
25
24
0
0
0
19
–
18
CFGOK
ru
0
17
CTRLDIR
ru
0
16
RWALL
ru
0
11
–
10
–
9
0
0
0
0
7
SHORT
PACKET
ru
0
6
STALLEDI/
CRCERRI
ru
0
5
4
3
OVERFI
NAKINI
NAKOUTI
ru
0
ru
0
ru
0
2
RXSTPI/
UNDERFI
ru
0
8
DTSEQ
ru
0
1
1
0
RXOUTI
TXINI
ru
0
ru
0
• TXINI: Transmitted IN Data Interrupt Flag
For control endpoints:
Set by hardware when the current bank is ready to accept a new IN packet. This triggers an EPXINT interrupt if
TXINE = 1.
Shall be cleared by software (by setting the TXINIC bit) to acknowledge the interrupt and to send the packet.
For isochronous, bulk and interrupt IN endpoints:
Set by hardware at the same time as FIFOCON when the current bank is free. This triggers an EPXINT interrupt if
TXINE = 1.
Shall be cleared by software (by setting the TXINIC bit) to acknowledge the interrupt, what has no effect on the endpoint FIFO.
The software then writes into the FIFO and clears the FIFOCON bit to allow the USB controller to send the data. If
the IN endpoint is composed of multiple banks, this also switches to the next bank. The TXINI and FIFOCON bits are
updated by hardware in accordance with the status of the next bank.
TXINI shall always be cleared before clearing FIFOCON.
This bit is inactive (cleared) for isochronous, bulk and interrupt OUT endpoints.
• RXOUTI: Received OUT Data Interrupt Flag
For control endpoints:
Set by hardware when the current bank contains a bulk OUT packet (data or status stage). This triggers an EPXINT
interrupt if RXOUTE = 1.
573
32058K
AVR32-01/12
AT32UC3A
Shall be cleared by software (by setting the RXOUTIC bit) to acknowledge the interrupt and to free the bank.
For isochronous, bulk and interrupt OUT endpoints:
Set by hardware at the same time as FIFOCON when the current bank is full. This triggers an EPXINT interrupt if
RXOUTE = 1.
Shall be cleared by software (by setting the RXOUTIC bit) to acknowledge the interrupt, what has no effect on the
endpoint FIFO.
The software then reads from the FIFO and clears the FIFOCON bit to free the bank. If the OUT endpoint is composed of multiple banks, this also switches to the next bank. The RXOUTI and FIFOCON bits are updated by
hardware in accordance with the status of the next bank.
RXOUTI shall always be cleared before clearing FIFOCON.
This bit is inactive (cleared) for isochronous, bulk and interrupt IN endpoints.
• RXSTPI: Received SETUP Interrupt Flag
For control endpoints, set by hardware to signal that the current bank contains a new valid SETUP packet. This triggers an
EPXINT interrupt if RXSTPE = 1.
Shall be cleared by software (by setting the RXSTPIC bit) to acknowledge the interrupt and to free the bank.
This bit is inactive (cleared) for bulk and interrupt IN/OUT endpoints and it means UNDERFI for isochronous IN/OUT
endpoints.
• UNDERFI: Underflow Interrupt Flag
For isochronous IN/OUT endpoints, set by hardware when an underflow error occurs. This triggers an EPXINT interrupt if
UNDERFE = 1.
An underflow can occur during IN stage if the host attempts to read from an empty bank. A zero-length packet is then automatically sent by the USB controller.
An underflow can also occur during OUT stage if the host sends a packet while the bank is already full. Typically, the CPU
is not fast enough. The packet is lost.
Shall be cleared by software (by setting the UNDERFIC bit) to acknowledge the interrupt.
This bit is inactive (cleared) for bulk and interrupt IN/OUT endpoints and it means RXSTPI for control endpoints.
• NAKOUTI: NAKed OUT Interrupt Flag
Set by hardware when a NAK handshake has been sent in response to an OUT request from the host. This triggers an
EPXINT interrupt if NAKOUTE = 1.
Shall be cleared by software (by setting the NAKOUTIC bit) to acknowledge the interrupt.
• NAKINI: NAKed IN Interrupt Flag
Set by hardware when a NAK handshake has been sent in response to an IN request from the host. This triggers an EPXINT interrupt if NAKINE = 1.
Shall be cleared by software (by setting the NAKINIC bit) to acknowledge the interrupt.
• OVERFI: Overflow Interrupt Flag
Set by hardware when an overflow error occurs. This triggers an EPXINT interrupt if OVERFE = 1.
For all endpoint types, an overflow can occur during OUT stage if the host attempts to write into a bank that is too small for
the packet. The packet is acknowledged and the Received OUT Data interrupt (RXOUTI) is raised as if no overflow had
occurred. The bank is filled with all the first bytes of the packet that fit in.
574
32058K
AVR32-01/12
AT32UC3A
Shall be cleared by software (by setting the OVERFIC bit) to acknowledge the interrupt.
• STALLEDI: STALLed Interrupt Flag
Set by hardware to signal that a STALL handshake has been sent. To do that, the software has to set the STALLRQ bit (by
setting the STALLRQS bit). This triggers an EPXINT interrupt if STALLEDE = 1.
Shall be cleared by software (by setting the STALLEDIC bit) to acknowledge the interrupt.
• CRCERRI: CRC Error Interrupt Flag
Set by hardware to signal that a CRC error has been detected in an isochronous OUT endpoint. The OUT packet is stored
in the bank as if no CRC error had occurred. This triggers an EPXINT interrupt if CRCERRE = 1.
Shall be cleared by software (by setting the CRCERRIC bit) to acknowledge the interrupt.
• SHORTPACKET: Short Packet Interrupt Flag
For non-control OUT endpoints, set by hardware when a short packet has been received.
For non-control IN endpoints, set by hardware when a short packet is transmitted upon ending a DMA transfer, thus signaling an end of isochronous frame or a bulk or interrupt end of transfer, this only if the End of DMA Buffer Output Enable bit
(DMAEND_EN) and the Automatic Switch bit (AUTOSW) are set.
This triggers an EPXINT interrupt if SHORTPACKETE = 1.
Shall be cleared by software (by setting the SHORTPACKETC bit) to acknowledge the interrupt.
• DTSEQ: Data Toggle Sequence
Set by hardware to indicate the PID of the current bank:
DTSEQ
Data Toggle Sequence
0
0
Data0
0
1
Data1
1
X
Reserved
For IN transfers, it indicates the data toggle sequence that will be used for the next packet to be sent. This is not relative to
the current bank.
For OUT transfers, this value indicates the last data toggle sequence received on the current bank.
Note that by default DTSEQ = 01b, as if the last data toggle sequence was Data1, so the next sent or expected data toggle
sequence should be Data0.
• NBUSYBK: Number of Busy Banks
Set by hardware to indicate the number of busy banks:
NBUSYBK
Number of Busy Banks
0
0
0 (all banks free)
0
1
1
1
0
2
1
1
3
For IN endpoints, it indicates the number of banks filled by the user and ready for IN transfer. When all banks are free, this
triggers an EPXINT interrupt if NBUSYBKE = 1.
575
32058K
AVR32-01/12
AT32UC3A
For OUT endpoints, it indicates the number of banks filled by OUT transactions from the host. When all banks are busy,
this triggers an EPXINT interrupt if NBUSYBKE = 1.
Note that when the FIFOCON bit is cleared (by setting the FIFOCONC bit) to validate a new bank, this field is updated 2 or
3 clock cycles later to calculate the address of the next bank.
An EPXINT interrupt is triggered if :
- for IN endpoint, NBUSYBKE=1 and all the banks are free.
- for OUT endpoint, NBUSYBKE=1 and all the banks are busy.
• CURRBK: Current Bank
For non-control endpoints, set by hardware to indicate the current bank:
CURRBK
Current Bank
0
0
Bank0
0
1
Bank1
1
0
Bank2
1
1
Reserved
Note that this field may be updated 1 clock cycle after the RWALL bit changes, so the user should not poll this field as an
interrupt flag.
• RWALL: Read/Write Allowed
For IN endpoints, set by hardware when the current bank is not full, i.e. the software can write further data into the FIFO.
For OUT endpoints, set by hardware when the current bank is not empty, i.e. the software can read further data from the
FIFO.
Never set if STALLRQ = 1 or in case of error.
Cleared by hardware otherwise.
This bit shall not be used for control endpoints.
• CTRLDIR: Control Direction
Set by hardware after a SETUP packet to indicate the direction of the following packet:
CTRLDIR
Control Direction
0
OUT
1
IN
Can not be set or cleared by software.
• CFGOK: Configuration OK Status
This bit is updated when the ALLOC bit is set.
Set by hardware if the endpoint X number of banks (EPBK) and size (EPSIZE) are correct compared to the maximal
allowed number of banks and size for this endpoint and to the maximal FIFO size (i.e. the DPRAM size).
If this bit is cleared by hardware, the user should reprogram the UECFGX register with correct EPBK and EPSIZE values.
• BYCT: Byte Count
Set by the hardware to indicate the byte count of the FIFO.
576
32058K
AVR32-01/12
AT32UC3A
For IN endpoints, incremented after each byte written by the software into the endpoint and decremented after each byte
sent to the host.
For OUT endpoints, incremented after each byte received from the host and decremented after each byte read by the software from the endpoint.
Note that this field may be updated 1 clock cycle after the RWALL bit changes, so the user should not poll this field as an
interrupt flag.
577
32058K
AVR32-01/12
AT32UC3A
30.8.2.12
USB Endpoint X Status Clear Register (UESTAXCLR)
Offset:
0x0160 + X . 0x04
Register Name:
UESTAXCLR, X in [0..6]
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
SHORT
PACKETC
w
0
6
STALLEDIC/
CRCERRIC
w
0
5
4
3
1
0
OVERFIC
NAKINIC
NAKOUTIC
RXOUTIC
TXINIC
w
0
w
0
w
0
2
RXSTPIC/
UNDERFIC
w
0
w
0
w
0
• TXINIC: Transmitted IN Data Interrupt Flag Clear
Set to clear TXINI.
Clearing has no effect.
Always read as 0.
• RXOUTIC: Received OUT Data Interrupt Flag Clear
Set to clear RXOUTI.
Clearing has no effect.
Always read as 0.
• RXSTPIC: Received SETUP Interrupt Flag Clear
Set to clear RXSTPI.
Clearing has no effect.
Always read as 0.
• UNDERFIC: Underflow Interrupt Flag Clear
Set to clear UNDERFI.
Clearing has no effect.
Always read as 0.
• NAKOUTIC: NAKed OUT Interrupt Flag Clear
Set to clear NAKOUTI.
578
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• NAKINIC: NAKed IN Interrupt Flag Clear
Set to clear NAKINI.
Clearing has no effect.
Always read as 0.
• OVERFIC: Overflow Interrupt Flag Clear
Set to clear OVERFI.
Clearing has no effect.
Always read as 0.
• STALLEDIC: STALLed Interrupt Flag Clear
Set to clear STALLEDI.
Clearing has no effect.
Always read as 0.
• CRCERRIC: CRC Error Interrupt Flag Clear
Set to clear CRCERRI.
Clearing has no effect.
Always read as 0.
• SHORTPACKETC: Short Packet Interrupt Flag Clear
Set to clear SHORTPACKET.
Clearing has no effect.
Always read as 0.
579
32058K
AVR32-01/12
AT32UC3A
30.8.2.13
USB Endpoint X Status Set Register (UESTAXSET)
Offset:
0x0190 + X . 0x04
Register Name:
UESTAXSET, X in [0..6]
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
NBUSYBKS
w
0
11
–
10
–
9
8
–
7
SHORT
PACKETS
w
0
6
STALLEDIS/
CRCERRIS
w
0
5
4
3
1
0
OVERFIS
NAKINIS
NAKOUTIS
RXOUTIS
TXINIS
w
0
w
0
w
0
2
RXSTPIS/
UNDERFIS
w
0
w
0
w
0
• TXINIS: Transmitted IN Data Interrupt Flag Set
Set to set TXINI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• RXOUTIS: Received OUT Data Interrupt Flag Set
Set to set RXOUTI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• RXSTPIS: Received SETUP Interrupt Flag Set
Set to set RXSTPI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• UNDERFIS: Underflow Interrupt Flag Set
Set to set UNDERFI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• NAKOUTIS: NAKed OUT Interrupt Flag Set
Set to set NAKOUTI, what may be useful for test or debug purposes.
580
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• NAKINIS: NAKed IN Interrupt Flag Set
Set to set NAKINI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• OVERFIS: Overflow Interrupt Flag Set
Set to set OVERFI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• STALLEDIS: STALLed Interrupt Flag Set
Set to set STALLEDI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• CRCERRIS: CRC Error Interrupt Flag Set
Set to set CRCERRI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• SHORTPACKETS: Short Packet Interrupt Flag Set
Set to set SHORTPACKET, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• NBUSYBKS: Number of Busy Banks Interrupt Flag Set
Set to force the Number of Busy Banks interrupt flag (NBUSYBK), what may be useful for test or debug purposes.
Set again to unforce the Number of Busy Banks interrupt flag (NBUSYBK).
Clearing has no effect.
Always read as 0.
581
32058K
AVR32-01/12
AT32UC3A
30.8.2.14
USB Endpoint X Control Register (UECONX)
Offset:
0x01C0 + X . 0x04
Register Name:
UECONX, X in [0..6]
Access Type:
Read-Only
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
STALLRQ
ru
0
18
RSTDT
ru
0
17
–
16
EPDISHDMA
r
0
15
–
14
FIFOCON
ru
0
13
KILLBK
ru
0
12
NBUSYBKE
r
0
11
–
10
–
9
–
8
–
7
SHORT
PACKETE
r
0
6
STALLEDE/
CRCERRE
r
0
5
4
3
1
0
OVERFE
NAKINE
NAKOUTE
RXOUTE
TXINE
r
0
r
0
r
0
2
RXSTPE/
UNDERFE
r
0
r
0
r
0
• TXINE: Transmitted IN Data Interrupt Enable
Set by software (by setting the TXINES bit) to enable the Transmitted IN Data interrupt (TXINI).
Clear by software (by setting the TXINEC bit) to disable the Transmitted IN Data interrupt (TXINI).
• RXOUTE: Received OUT Data Interrupt Enable
Set by software (by setting the RXOUTES bit) to enable the Received OUT Data interrupt (RXOUT).
Clear by software (by setting the RXOUTEC bit) to disable the Received OUT Data interrupt (RXOUT).
• RXSTPE: Received SETUP Interrupt Enable
Set by software (by setting the RXSTPES bit) to enable the Received SETUP interrupt (RXSTPI).
Clear by software (by setting the RXSTPEC bit) to disable the Received SETUP interrupt (RXSTPI).
• UNDERFE: Underflow Interrupt Enable
Set by software (by setting the UNDERFES bit) to enable the Underflow interrupt (UNDERFI).
Clear by software (by setting the UNDERFEC bit) to disable the Underflow interrupt (UNDERFI).
• NAKOUTE: NAKed OUT Interrupt Enable
Set by software (by setting the NAKOUTES bit) to enable the NAKed OUT interrupt (NAKOUTI).
Clear by software (by setting the NAKOUTEC bit) to disable the NAKed OUT interrupt (NAKOUTI).
• NAKINE: NAKed IN Interrupt Enable
Set by software (by setting the NAKINES bit) to enable the NAKed IN interrupt (NAKINI).
Clear by software (by setting the NAKINEC bit) to disable the NAKed IN interrupt (NAKINI).
582
32058K
AVR32-01/12
AT32UC3A
• OVERFE: Overflow Interrupt Enable
Set by software (by setting the OVERFES bit) to enable the Overflow interrupt (OVERFI).
Clear by software (by setting the OVERFEC bit) to disable the Overflow interrupt (OVERFI).
• STALLEDE: STALLed Interrupt Enable
Set by software (by setting the STALLEDES bit) to enable the STALLed interrupt (STALLEDI).
Clear by software (by setting the STALLEDEC bit) to disable the STALLed interrupt (STALLEDI).
• CRCERRE: CRC Error Interrupt Enable
Set by software (by setting the CRCERRES bit) to enable the CRC Error interrupt (CRCERRI).
Clear by software (by setting the CRCERREC bit) to disable the CRC Error interrupt (CRCERRI).
• SHORTPACKETE: Short Packet Interrupt Enable
Set by software (by setting the SHORTPACKETES bit) to enable the Short Packet interrupt (SHORTPACKET).
Clear by software (by setting the SHORTPACKETEC bit) to disable the Short Packet interrupt (SHORTPACKET).
• NBUSYBKE: Number of Busy Banks Interrupt Enable
Set by software (by setting the NBUSYBKES bit) to enable the Number of Busy Banks interrupt (NBUSYBK).
Clear by software (by setting the NBUSYBKEC bit) to disable the Number of Busy Banks interrupt (NBUSYBK).
• KILLBK: Kill IN Bank
Set by software (by setting the KILLBKS bit) to kill the last written bank.
Cleared by hardware when the bank is killed.
Caution: The bank is really cleared when the “kill packet” procedure is accepted by the USB macro core. This bit is automatically cleared after the end of the procedure:
– The bank is really cleared or the bank is sent (IN transfer): NBUSYBK is decremented.
– The bank is not cleared but sent (IN transfer): NBUSYBK is decremented.
– The bank is not cleared because it was empty.
The software shall wait for this bit to be cleared before trying to kill another packet.
Note that this kill request is refused if at the same time an IN token is coming and the last bank is the current one being sent
on the USB line. If at least 2 banks are ready to be sent, there is no problem to kill a packet even if an IN token is coming.
Indeed, in this case, the current bank is sent (IN transfer) while the last bank is killed.
• FIFOCON: FIFO Control
For control endpoints:
The FIFOCON and RWALL bits are irrelevant. The software shall therefore never use them on these endpoints.
When read, their value is always 0.
For IN endpoints:
Set by hardware when the current bank is free, at the same time as TXINI.
Clear by software (by setting the FIFOCONC bit) to send the FIFO data and to switch to the next bank.
For OUT endpoints:
Set by hardware when the current bank is full, at the same time as RXOUTI.
583
32058K
AVR32-01/12
AT32UC3A
Clear by software (by setting the FIFOCONC bit) to free the current bank and to switch to the next bank.
• EPDISHDMA: Endpoint Interrupts Disable HDMA Request Enable
Set by software (by setting the EPDISHDMAS bit) to pause the on-going DMA channel X transfer on any Endpoint X interrupt (EPXINT), whatever the state of the Endpoint X Interrupt Enable bit (EPXINTE).
The software then has to acknowledge or to disable the interrupt source (e.g. RXOUTI) or to clear the EPDISHDMA bit (by
setting the EPDISHDMAC bit) in order to complete the DMA transfer.
In ping-pong mode, if the interrupt is associated to a new system-bank packet (e.g. Bank1) and the current DMA transfer is
running on the previous packet (Bank0), then the previous-packet DMA transfer completes normally, but the new-packet
DMA transfer will not start (not requested).
If the interrupt is not associated to a new system-bank packet (NAKINI, NAKOUTI, etc.), then the request cancellation may
occur at any time and may immediately pause the current DMA transfer.
This may be used for example to identify erroneous packets, to prevent them from being transferred into a buffer, to complete a DMA transfer by software after reception of a short packet, etc.
• RSTDT: Reset Data Toggle
Set by software (by setting the RSTDTS bit) to clear the data toggle sequence, i.e. to set to Data0 the data toggle sequence
of the next sent (IN endpoints) or received (OUT endpoints) packet.
Cleared by hardware instantaneously.
The software does not have to wait for this bit to be cleared.
• STALLRQ: STALL Request
Set by software (by setting the STALLRQS bit) to request to send a STALL handshake to the host.
Cleared by hardware when a new SETUP packet is received.
Can also be cleared by software by setting the STALLRQC bit.
584
32058K
AVR32-01/12
AT32UC3A
30.8.2.15
USB Endpoint X Control Clear Register (UECONXCLR)
Offset:
0x0220 + X . 0x04
Register Name:
UECONXCLR, X in [0..6]
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
STALLRQC
w
0
18
–
17
–
16
EPDISHDMAC
w
0
15
–
14
FIFOCONC
w
0
13
–
12
NBUSYBKEC
w
0
11
–
10
–
9
–
8
–
7
SHORT
PACKETEC
w
0
6
STALLEDEC/
CRCERREC
w
0
5
4
3
1
0
OVERFEC
NAKINEC
NAKOUTEC
RXOUTEC
TXINEC
w
0
w
0
w
0
2
RXSTPEC/
UNDERFEC
w
0
w
0
w
0
• TXINEC: Transmitted IN Data Interrupt Enable Clear
Set to clear TXINE.
Clearing has no effect.
Always read as 0.
• RXOUTEC: Received OUT Data Interrupt Enable Clear
Set to clear RXOUTE.
Clearing has no effect.
Always read as 0.
• RXSTPEC: Received SETUP Interrupt Enable Clear
Set to clear RXSTPE.
Clearing has no effect.
Always read as 0.
• UNDERFEC: Underflow Interrupt Enable Clear
Set to clear UNDERFE.
Clearing has no effect.
Always read as 0.
• NAKOUTEC: NAKed OUT Interrupt Enable Clear
Set to clear NAKOUTE.
585
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• NAKINEC: NAKed IN Interrupt Enable Clear
Set to clear NAKINE.
Clearing has no effect.
Always read as 0.
• OVERFEC: Overflow Interrupt Enable Clear
Set to clear OVERFE.
Clearing has no effect.
Always read as 0.
• STALLEDEC: STALLed Interrupt Enable Clear
Set to clear STALLEDE.
Clearing has no effect.
Always read as 0.
• CRCERREC: CRC Error Interrupt Enable Clear
Set to clear CRCERRE.
Clearing has no effect.
Always read as 0.
• SHORTPACKETEC: Short Packet Interrupt Enable Clear
Set to clear SHORTPACKETE.
Clearing has no effect.
Always read as 0.
• NBUSYBKEC: Number of Busy Banks Interrupt Enable Clear
Set to clear NBUSYBKE.
Clearing has no effect.
Always read as 0.
• FIFOCONC: FIFO Control Clear
Set to clear FIFOCON.
Clearing has no effect.
Always read as 0.
• EPDISHDMAC: Endpoint Interrupts Disable HDMA Request Enable Clear
Set to clear EPDISHDMA.
Clearing has no effect.
Always read as 0.
586
32058K
AVR32-01/12
AT32UC3A
• STALLRQC: STALL Request Clear
Set to clear STALLRQ.
Clearing has no effect.
Always read as 0.
587
32058K
AVR32-01/12
AT32UC3A
30.8.2.16
USB Endpoint X Control Set Register (UECONXSET)
Offset:
0x01F0 + X . 0x04
Register Name:
UECONXSET, X in [0..6]
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
STALLRQS
w
0
18
RSTDTS
17
–
16
EPDISHDMAS
w
0
15
–
14
–
w
0
13
KILLBKS
12
NBUSYBKES
w
0
11
–
10
–
9
–
8
–
7
SHORT
PACKETES
w
0
6
STALLEDES/
CRCERRES
w
0
5
4
3
1
0
OVERFES
NAKINES
NAKOUTES
RXOUTES
TXINES
w
0
w
0
w
0
2
RXSTPES/
UNDERFES
w
0
w
0
w
0
• TXINES: Transmitted IN Data Interrupt Enable Set
Set to set TXINE.
Clearing has no effect.
Always read as 0.
• RXOUTES: Received OUT Data Interrupt Enable Set
Set to set RXOUTE.
Clearing has no effect.
Always read as 0.
• RXSTPES: Received SETUP Interrupt Enable Set
Set to set RXSTPE.
Clearing has no effect.
Always read as 0.
• UNDERFES: Underflow Interrupt Enable Set
Set to set UNDERFE.
Clearing has no effect.
Always read as 0.
• NAKOUTES: NAKed OUT Interrupt Enable Set
Set to set NAKOUTE.
588
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• NAKINES: NAKed IN Interrupt Enable Set
Set to set NAKINE.
Clearing has no effect.
Always read as 0.
• OVERFES: Overflow Interrupt Enable Set
Set to set OVERFE.
Clearing has no effect.
Always read as 0.
• STALLEDES: STALLed Interrupt Enable Set
Set to set STALLEDE.
Clearing has no effect.
Always read as 0.
• CRCERRES: CRC Error Interrupt Enable Set
Set to set CRCERRE.
Clearing has no effect.
Always read as 0.
• SHORTPACKETES: Short Packet Interrupt Enable Set
Set to set SHORTPACKETE.
Clearing has no effect.
Always read as 0.
• NBUSYBKES: Number of Busy Banks Interrupt Enable Set
Set to set NBUSYBKE.
Clearing has no effect.
Always read as 0.
• KILLBKS: Kill IN Bank Set
Set to set KILLBK.
Clearing has no effect.
Always read as 0.
• EPDISHDMAS: Endpoint Interrupts Disable HDMA Request Enable Set
Set to set EPDISHDMA.
Clearing has no effect.
589
32058K
AVR32-01/12
AT32UC3A
Always read as 0.
• RSTDTS: Reset Data Toggle Set
Set to set RSTDT.
Clearing has no effect.
Always read as 0.
• STALLRQS: STALL Request Set
Set to set STALLRQ.
Clearing has no effect.
Always read as 0.
590
32058K
AVR32-01/12
AT32UC3A
30.8.2.17
USB Device DMA Channel X Next Descriptor Address Register (UDDMAX_NEXTDESC)
Offset:
0x0310 + (X - 1) . 0x10
Register Name:
UDDMAX_NEXTDESC, X in [1..6]
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
0
0
0
23
22
21
0
0
0
15
14
13
0
0
0
7
6
5
NXT_DESC_ADDR
rwu
0
0
0
28
27
NXT_DESC_ADDR
rwu
0
0
26
25
24
0
0
0
20
19
NXT_DESC_ADDR
rwu
0
0
18
17
16
0
0
0
12
11
NXT_DESC_ADDR
rwu
0
0
10
9
8
0
0
0
4
2
–
1
–
0
–
3
–
0
• NXT_DESC_ADDR: Next Descriptor Address
This field contains the bits 31:4 of the 16-byte aligned address of the next channel descriptor to be processed.
Note that this field is written either by software or by descriptor loading.
591
32058K
AVR32-01/12
AT32UC3A
30.8.2.18
USB Device DMA Channel X HSB Address Register (UDDMAX_ADDR)
Offset:
0x0314 + (X - 1) . 0x10
Register Name:
UDDMAX_ADDR, X in [1..6]
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
0
0
0
0
19
18
17
16
0
0
0
0
11
10
9
8
0
0
0
0
3
2
1
0
0
0
0
0
HSB_ADDR
rwu
0
0
0
0
23
22
21
20
HSB_ADDR
rwu
0
0
0
0
15
14
13
12
HSB_ADDR
rwu
0
0
0
0
7
6
5
4
HSB_ADDR
rwu
0
0
0
0
• HSB_ADDR: HSB Address
This field determines the HSB bus current address of a channel transfer.
The address set on the HSB address bus is HSB_ADDR rounded down to the nearest word-aligned address, i.e.
HSB_ADDR[1:0] is considered as 00b since only word accesses are performed.
Channel HSB start and end addresses may be aligned on any byte boundary.
The software may write this field only when the Channel Enabled bit (CH_EN) of the UDDMAX_STATUS register is clear.
This field is updated at the end of the address phase of the current access to the HSB bus. It is incremented of the HSB
access byte-width.
The HSB access width is 4 bytes, or less at packet start or end if the start or end address is not aligned on a word
boundary.
The packet start address is either the channel start address or the next channel address to be accessed in the channel
buffer.
The packet end address is either the channel end address or the latest channel address accessed in the channel buffer.
The channel start address is written by software or loaded from the descriptor, whereas the channel end address is either
determined by the end of buffer or the end of USB transfer if the Buffer Close Input Enable bit (BUFF_CLOSE_IN_EN) is
set.
592
32058K
AVR32-01/12
AT32UC3A
30.8.2.19
USB Device DMA Channel X Control Register (UDDMAX_CONTROL)
Offset:
0x0318 + (X - 1) . 0x10
Register Name:
UDDMAX_CONTROL, X in [1..6]
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
CH_BYTE_LENGTH
rwu
0
0
26
25
24
0
0
0
0
0
0
23
22
21
18
17
16
0
20
19
CH_BYTE_LENGTH
rwu
0
0
0
0
0
0
0
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
BURST_LOCK
_EN
rwu
0
6
DESC_LD_
IRQ_EN
rwu
0
5
EOBUFF_
IRQ_EN
rwu
0
4
3
DMAEND_EN
rwu
0
rwu
0
1
LD_NXT_CH_
DESC_EN
rwu
0
0
EOT_IRQ_EN
2
BUFF_CLOSE
_IN_EN
rwu
0
CH_EN
rwu
0
• CH_EN: Channel Enable
Set this bit to enable this channel data transfer.
Clear this bit to disable the channel data transfer.
This may be used to start or resume any requested transfer.
This bit is cleared by hardware when the HSB source channel is disabled at end of dma buffer.
• LD_NXT_CH_DESC_EN: Load Next Channel Descriptor Enable
Set this bit to allow automatic next descriptor loading at the end of the channel transfer.
Clear this bit to disable this feature.
If set, the dma channel controller loads the next descriptor when the UDDMAX_STATUS.CH_EN bit is reset due to software of hardware event (for example at the end of the current transfer).
• BUFF_CLOSE_IN_EN: Buffer Close Input Enable
Set this bit to automatically closed the current dma transfer at the end of the usb OUT data transfer (received short packet).
Clear this bit to disable this feature.
• DMAEND_EN: End of DMA Buffer Output Enable
Set this bit to properly complete the usb transfer at the end of the dma transfer.
For IN endpoint, it means that a short packet (or a Zero Length Packet) will be sent to the usb line to properly closed the
usb transfer at the end of the dma transfer.
For OUT endpoint, it means that all the banks will be properly released. (NBUSYBK=0) at the end of the dma transfer.
593
32058K
AVR32-01/12
AT32UC3A
• EOT_IRQ_EN: End of USB Transfer Interrupt Enable
Set this bit to enable the end of usb OUT data transfer interrupt.
This interrupt is generated only if the BUFF_CLOSE_IN_EN bit is set.
Clear this bit to disable this interrupt.
• EOBUFF_IRQ_EN: End of Buffer Interrupt Enable
Set this bit to enable the end of buffer interrupt.
This interrupt is generated when the channel byte count reaches zero.
Clear this bit to disable this interrupt.
• DESC_LD_IRQ_EN: Descriptor Loaded Interrupt Enable
Set this bit to enable the Descripor Loaded interrupt.
This interrupt is generated when a Descriptor has been loaded from the system bus.
Clear this bit to disable this interrupt.
• BURST_LOCK_EN: Burst Lock Enable
Set this bit to lock the HSB data burst for maximum optimization of HSB busses bandwidth usage and maximization of flyby duration.
If clear, the DMA never locks HSB access.
• CH_BYTE_LENGTH: Channel Byte Length
This field determines the total number of bytes to be transferred for this buffer.
The maximum channel transfer size 64 kB is reached when this field is 0 (default value).
If the transfer size is unknown, the transfer end is controlled by the peripheral and this field should be set to 0.
This field can be written by software or descriptor loading only after the UDDMAX_STATUS.CH_EN bit has been cleared,
otherwise this field is ignored.
594
32058K
AVR32-01/12
AT32UC3A
30.8.2.20
USB Device DMA Channel X Status Register (UDDMAX_STATUS)
Offset:
0x031C + (X - 1) . 0x10
Register Name:
UDDMAX_STATUS, X in [1..6]
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
0
0
0
0
19
18
17
16
CH_BYTE_CNT
ru
0
0
0
0
23
22
21
20
CH_BYTE_CNT
ru
0
0
0
0
0
0
0
0
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
DESC_LD_
STA
ru
0
5
EOCH_BUFF_
STA
ru
0
4
3
2
1
0
EOT_STA
–
–
CH_ACTIVE
CH_EN
rwu
0
rwu
0
–
ru
0
• CH_EN: Channel Enabled
If set, the DMA channel is currently enabled.
If cleared, the DMA channel does no longer transfer data.
• CH_ACTIVE: Channel Active
If set, the DMA channel is currently trying to source USB data.
If cleared, the DMA channel is no longer trying to source USB data.
When a USB data transfer is completed, this bit is automatically reset.
• EOT_STA: End of USB Transfer Status
Set by hardware when the completion of the usb data transfer has closed the dma transfer. It is valid only if
BUFF_CLOSE_EN=1.
This bit is automatically cleared when read by software.
• EOCH_BUFF_STA: End of Channel Buffer Status
Set by hardware when the Channel Byte Count downcounts to zero.
This bit is automatically cleared when read by software.
• DESC_LD_STA: Descriptor Loaded Status
Set by hardware when a Descriptor has been loaded from the HSB bus.
This bit is automatically cleared when read by software.
595
32058K
AVR32-01/12
AT32UC3A
• CH_BYTE_CNT: Channel Byte Count
This field gives the current number of bytes still to be transferred for this buffer.
This field is decremented at each dma access.
This field is reliable (stable) only if the CH_EN flag is 0.
596
32058K
AVR32-01/12
AT32UC3A
30.8.3
USB Host Registers
30.8.3.1
USB Host General Control Register (UHCON)
Offset:
0x0400
Register Name:
UHCON
Access Type:
Read/Write
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
RESUME
9
RESET
8
SOFE
rwu
0
rwu
0
rwu
0
2
–
1
–
0
–
7
–
6
–
5
–
4
–
3
–
• SOFE: Start of Frame Generation Enable
Set this bit to generate SOF on the USB bus in full speed mode and keep alive in low speed mode.
Clear this bit to disable the SOF generation and to leave the USB bus in idle state.
This bit is set by hardware when a USB reset is requested or an upstream resume interrupt is detected (UHINT.TXRSMI).
• RESET: Send USB Reset
Set this bit to generate a USB Reset on the USB bus.
Cleared by hardware when the USB Reset has been sent.
It may be useful to clear this bit by software when a device disconnection is detected (UHINT.DDISCI is set) whereas a
USB Reset is being sent.
• RESUME: Send USB Resume
Set this bit to generate a USB Resume on the USB bus.
Cleared by hardware when the USB Resume has been sent or when a USB reset is requested. Clearing by software has
no effect. This bit should be set only when the start of frame generation is enable. (SOFE bit set).
597
32058K
AVR32-01/12
AT32UC3A
30.8.3.2
USB Host Global Interrupt Register (UHINT)
Offset:
0x0404
Register Name:
UHINT
Access Type:
Read-Only
Reset Value:
0x00000000
31
–
30
DMA6INT
r
0
29
DMA5INT
r
0
28
DMA4INT
r
0
27
DMA3INT
r
0
26
DMA2INT
r
0
25
DMA1INT
r
0
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
P6INT
r
0
13
P5INT
r
0
12
P4INT
r
0
11
P3INT
r
0
10
P2INT
r
0
9
P1INT
r
0
8
P0INT
r
0
7
–
6
HWUPI
r
0
5
HSOFI
r
0
4
RXRSMI
r
0
3
RSMEDI
r
0
2
RSTI
r
0
1
DDISCI
r
0
0
DCONNI
r
0
• DCONNI: Device Connection Interrupt Flag
Set by hardware when a new device has been connected to the USB bus.
Shall be cleared by software (by setting the DCONNIC bit).
• DDISCI: Device Disconnection Interrupt Flag
Set by hardware when the device has been removed from the USB bus.
Shall be cleared by software (by setting the DDISCIC bit).
• RSTI: USB Reset Sent Interrupt Flag
Set by hardware when a USB Reset has been sent to the device.
Shall be cleared by software (by setting the RSTIC bit).
• RSMEDI: Downstream Resume Sent Interrupt Flag
Set by hardware when a Downstream Resume has been sent to the Device.
Shall be cleared by software (by setting the RSMEDIC bit).
• RXRSMI: Upstream Resume Received Interrupt Flag
Set by hardware when an Upstream Resume has been received from the Device.
Shall be cleared by software (by setting the RXRSMIC bit).
• HSOFI: Host Start of Frame Interrupt Flag
Set by hardware when a SOF is issued by the Host controller. This triggers a USB interrupt when HSOFE is set. When
using the host controller in low speed mode, this bit is also set when a keep-alive is sent.
Shall be cleared by software (by setting the HSOFIC bit).
598
32058K
AVR32-01/12
AT32UC3A
• HWUPI: Host Wake-Up Interrupt Flag
Asynchronous interrupt.
Set by hardware in the following cases :
– The Host controller is in the suspend mode (SOFE=0) and an upstream resume from
the Peripheral is detected.
– The Host controller is in the suspend mode (SOFE=0) and a Peripheral disconnection
is detected.
– The Host controller is in the Idle state (VBUSRQ=0, no VBus is generated), and an
OTG SRP event initiated by the Peripheral is detected.
Note that this interrupt is generated even if the clock is frozen by the FRZCLK bit.
• PXINT, X in [0..6]: Pipe X Interrupt Flag
Set by hardware when an interrupt is triggered by the endpoint X (UPSTAX). This triggers a USB interrupt if the corresponding pipe interrupt enable bit is set (UHINTE register). Cleared by hardware when the interrupt source is served.
• DMAXINT, X in [1..6]: DMA Channel X Interrupt Flag
Set by hardware when an interrupt is triggered by the DMA channel X. This triggers a USB interrupt if the corresponding
DMAXINTE is set (UHINTE register).
Cleared by hardware when the UHDMAX_STATUS interrupt source is cleared.
599
32058K
AVR32-01/12
AT32UC3A
30.8.3.3
USB Host Global Interrupt Clear Register (UHINTCLR)
Offset:
0x0408
Register Name:
UHINTCLR
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
HWUPIC
w
0
5
HSOFIC
w
0
4
RXRSMIC
w
0
3
RSMEDIC
w
0
2
RSTIC
w
0
1
DDISCIC
w
0
0
DCONNIC
w
0
• DCONNIC: Device Connection Interrupt Flag Clear
Set to clear DCONNI.
Clearing has no effect.
Always read as 0.
• DDISCIC: Device Disconnection Interrupt Flag Clear
Set to clear DDISCI.
Clearing has no effect.
Always read as 0.
• RSTIC: USB Reset Sent Interrupt Flag Clear
Set to clear RSTI.
Clearing has no effect.
Always read as 0.
• RSMEDIC: Downstream Resume Sent Interrupt Flag Clear
Set to clear RSMEDI.
Clearing has no effect.
Always read as 0.
• RXRSMIC: Upstream Resume Received Interrupt Flag Clear
Set to clear RXRSMI.
600
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• HSOFIC: Host Start of Frame Interrupt Flag Clear
Set to clear HSOFI.
Clearing has no effect.
Always read as 0.
• HWUPIC: Host Wake-Up Interrupt Flag Clear
Set to clear HWUPI.
Clearing has no effect.
Always read as 0.
601
32058K
AVR32-01/12
AT32UC3A
30.8.3.4
USB Host Global Interrupt Set Register (UHINTSET)
Offset:
0x040C
Register Name:
UHINTSET
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
DMA6INTS
w
0
29
DMA5INTS
w
0
28
DMA4INTS
w
0
27
DMA3INTS
w
0
26
DMA2INTS
w
0
25
DMA1INTS
w
0
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
HWUPIS
w
0
5
HSOFIS
w
0
4
RXRSMIS
w
0
3
RSMEDIS
w
0
2
RSTIS
w
0
1
DDISCIS
w
0
0
DCONNIS
w
0
• DCONNIS: Device Connection Interrupt Flag Set
Set to set DCONNI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• DDISCIS: Device Disconnection Interrupt Flag Set
Set to set DDISCI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• RSTIS: USB Reset Sent Interrupt Flag Set
Set to set RSTI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• RSMEDIS: Downstream Resume Sent Interrupt Flag Set
Set to set RSMEDI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• RXRSMIS: Upstream Resume Received Interrupt Flag Set
Set to set RXRSMI, what may be useful for test or debug purposes.
602
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• HSOFIS: Host Start of Frame Interrupt Flag Set
Set to set HSOFI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• HWUPIS: Host Wake-Up Interrupt Flag Set
Set to set HWUPI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• DMAXINTS, X in [1..6]: DMA Channel X Interrupt Flag Set
Set to set DMAXINT, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
603
32058K
AVR32-01/12
AT32UC3A
30.8.3.5
USB Host Global Interrupt Enable Register (UHINTE)
Offset:
0x0410
Register Name:
UHINTE
Access Type:
Read-Only
Reset Value:
0x00000000
31
–
30
DMA6INTE
r
0
29
DMA5INTE
r
0
28
DMA4INTE
r
0
27
DMA3INTE
r
0
26
DMA2INTE
r
0
25
DMA1INTE
r
0
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
P6INTE
r
0
13
P5INTE
r
0
12
P4INTE
r
0
11
P3INTE
r
0
10
P2INTE
r
0
9
P1INTE
r
0
8
P0INTE
r
0
7
–
6
HWUPIE
r
0
5
HSOFIE
r
0
4
RXRSMIE
r
0
3
RSMEDIE
r
0
2
RSTIE
r
0
1
DDISCIE
r
0
0
DCONNIE
r
0
• DCONNIE: Device Connection Interrupt Enable
Set by software (by setting the DCONNIES bit) to enable the Device Connection interrupt (DCONNI).
Clear by software (by setting the DCONNIEC bit) to disable the Device Connection interrupt (DCONNI).
• DDISCIE: Device Disconnection Interrupt Enable
Set by software (by setting the DDISCIES bit) to enable the Device Disconnection interrupt (DDISCI).
Clear by software (by setting the DDISCIEC bit) to disable the Device Disconnection interrupt (DDISCI).
• RSTIE: USB Reset Sent Interrupt Enable
Set by software (by setting the RSTIES bit) to enable the USB Reset Sent interrupt (RSTI).
Clear by software (by setting the RSTIEC bit) to disable the USB Reset Sent interrupt (RSTI).
• RSMEDIE: Downstream Resume Sent Interrupt Enable
Set by software (by setting the RSMEDIES bit) to enable the Downstream Resume interrupt (RSMEDI).
Clear by software (by setting the RSMEDIEC bit) to disable the Downstream Resume interrupt (RSMEDI).
• RXRSMIE: Upstream Resume Received Interrupt Enable
Set by software (by setting the RXRSMIES bit) to enable the Upstream Resume Received interrupt (RXRSMI).
Clear by software (by setting the RXRSMIEC bit) to disable the Downstream Resume interrupt (RXRSMI).
• HSOFIE: Host Start of Frame Interrupt Enable
Set by software (by setting the HSOFIES bit) to enable the Host Start of Frame interrupt (HSOFI).
Clear by software (by setting the HSOFIEC bit) to disable the Host Start of Frame interrupt (HSOFI).
604
32058K
AVR32-01/12
AT32UC3A
• HWUPIE: Host Wake-Up Interrupt Enable
Set by software (by setting the HWUPIES bit) to enable the Host Wake-up Interrupt (HWUPI).
Clear by software (by setting the HWUPIEC bit) to disable the Host Wake-up Interrupt (HWUPI).
• PXINTE, X in [0..6]: Pipe X Interrupt Enable
Set by software (by setting the PXINTES bit) to enable the Pipe X Interrupt (PXINT).
Clear by software (by setting the PXINTEC bit) to disable the Pipe X Interrupt (PXINT).
• DMAXINTE, X in [1..6]: DMA Channel X Interrupt Enable
Set by software (by setting the DMAXINTES bit) to enable the DMA Channel X Interrupt (DMAXINT).
Clear by software (by setting the DMAXINTEC bit) to disable the DMA Channel X Interrupt (DMAXINT).
605
32058K
AVR32-01/12
AT32UC3A
30.8.3.6
USB Host Global Interrupt Enable Clear Register (UHINTECLR)
Offset:
0x0414
Register Name:
UHINTECLR
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
DMA6INTEC
w
0
29
DMA5INTEC
w
0
28
DMA4INTEC
w
0
27
DMA3INTEC
w
0
26
DMA2INTEC
w
0
25
DMA1INTEC
w
0
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
P6INTEC
w
0
13
P5INTEC
w
0
12
P4INTEC
w
0
11
P3INTEC
w
0
10
P2INTEC
w
0
9
P1INTEC
w
0
8
P0INTEC
w
0
7
–
6
HWUPIEC
w
0
5
HSOFIEC
w
0
4
RXRSMIEC
w
0
3
RSMEDIEC
w
0
2
RSTIEC
w
0
1
DDISCIEC
w
0
0
DCONNIEC
w
0
• DCONNIEC: Device Connection Interrupt Enable Clear
Set to clear DCONNIE.
Clearing has no effect.
Always read as 0.
• DDISCIEC: Device Disconnection Interrupt Enable Clear
Set to clear DDISCIEC.
Clearing has no effect.
Always read as 0.
• RSTIEC: USB Reset Sent Interrupt Enable Clear
Set to clear RSTIEC.
Clearing has no effect.
Always read as 0.
• RSMEDIEC: Downstream Resume Sent Interrupt Enable Clear
Set to clear RSMEDIEC.
Clearing has no effect.
Always read as 0.
• RXRSMIEC: Upstream Resume Received Interrupt Enable Clear
Set to clear RSTIEC.
606
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• HSOFIEC: Host Start of Frame Interrupt Enable Clear
Set to clear HSOFIEC.
Clearing has no effect.
Always read as 0.
• HWUPIEC: Host Wake-Up Interrupt Enable Clear
Set to clear HWUPIEC.
Clearing has no effect.
Always read as 0.
• PXINTEC, X in [0..6]: Pipe X Interrupt Enable Clear
Set to clear PXINTEC.
Clearing has no effect.
Always read as 0.
• DMAXINTEC, X in [1..6]: DMA Channel X Interrupt Enable Clear
Set to clear DMAXINTEC.
Clearing has no effect.
Always read as 0.
607
32058K
AVR32-01/12
AT32UC3A
30.8.3.7
USB Host Global Interrupt Enable Set Register (UHINTESET)
Offset:
0x0418
Register Name:
UHINTESET
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
DMA6INTES
w
0
29
DMA5INTES
w
0
28
DMA4INTES
w
0
27
DMA3INTES
w
0
26
DMA2INTES
w
0
25
DMA1INTES
w
0
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
P6INTES
w
0
13
P5INTES
w
0
12
P4INTES
w
0
11
P3INTES
w
0
10
P2INTES
w
0
9
P1INTES
w
0
8
P0INTES
w
0
7
–
6
HWUPIES
w
0
5
HSOFIES
w
0
4
RXRSMIES
w
0
3
RSMEDIES
w
0
2
RSTIES
w
0
1
DDISCIES
w
0
0
DCONNIES
w
0
• DCONNIES: Device Connection Interrupt Enable Set
Set to set DCONNIE.
Clearing has no effect.
Always read as 0.
• DDISCIES: Device Disconnection Interrupt Enable Set
Set to set DDISCIE.
Clearing has no effect.
Always read as 0.
• RSTIES: USB Reset Sent Interrupt Enable Set
Set to set RSTIE.
Clearing has no effect.
Always read as 0.
• RSMEDIES: Downstream Resume Sent Interrupt Enable Set
Set to set RSMEDIE.
Clearing has no effect.
Always read as 0.
• RXRSMIES: Upstream Resume Received Interrupt Enable Set
Set to set RXRSMIE.
608
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• HSOFIES: Host Start of Frame Interrupt Enable Set
Set to set HSOFIE.
Clearing has no effect.
Always read as 0.
• HWUPIES: Host Wake-Up Interrupt Enable Set
Set to set HWUPIE.
Clearing has no effect.
Always read as 0.
• PXINTES, X in [0..6]: Pipe X Interrupt Enable Set
Set to set PXINTE.
Clearing has no effect.
Always read as 0.
• DMAXINTES, X in [1..6]: DMA Channel X Interrupt Enable Set
Set to set DMAXINTE.
Clearing has no effect.
Always read as 0.
609
32058K
AVR32-01/12
AT32UC3A
30.8.3.8
USB Host Frame Number Register (UHFNUM)
Offset:
0x0420
Register Name:
UHFNUM
Access Type:
Read/Write
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
23
22
21
20
27
–
26
–
25
–
24
–
19
18
17
16
0
0
0
10
9
8
FLENHIGH
ru
0
0
0
0
0
15
–
14
–
13
12
11
7
6
0
0
FNUM
rwu
0
0
0
0
0
0
5
FNUM
rwu
0
4
3
2
–
1
–
0
–
0
0
• FNUM: Frame Number
The value contained in this register is the current SOF number.
This value can be modified by software.
• FLENHIGH: Frame Length
This register gives the 8 high-order bits of the 14-bits internal frame counter (frame counter at 12 Mhz, counter length is
12000 to ensure a SOF generation every 1 ms).
610
32058K
AVR32-01/12
AT32UC3A
30.8.3.9
USB Host Frame Number Register (UHADDR1)
Offset:
0x0424
Register Name:
UHADDR1
Access Type:
Read/Write
Reset Value:
0x00000000
31
–
30
29
28
–
0
0
0
23
–
22
21
20
–
0
0
0
15
–
14
13
12
–
0
0
0
7
–
6
5
4
–
0
0
0
27
UHADDR_P3
rwu
0
26
25
24
0
0
0
19
UHADDR_P2
rwu
0
18
17
16
0
0
0
11
UHADDR_P1
rwu
0
10
9
8
0
0
0
3
UHADDR_P0
rwu
0
2
1
0
• UHADDR_P0 : USB Host Address
These bits should contain the address of the Pipe0 of the USB Device.
Cleared by hardware when a USB reset is requested.
• UHADDR_P1 : USB Host Address
These bits should contain the address of the Pipe1 of the USB Device.
Cleared by hardware when a USB reset is requested.
• UHADDR_P2 : USB Host Address
These bits should contain the address of the Pipe2 of the USB Device.
Cleared by hardware when a USB reset is requested.
• UHADDR_P3 : USB Host Address
These bits should contain the address of the Pipe3 of the USB Device.
Cleared by hardware when a USB reset is requested.
611
32058K
AVR32-01/12
AT32UC3A
30.8.3.10
USB Host Frame Number Register (UHADDR2)
Offset:
0x0428
Register Name:
UHADDR2
Access Type:
Read/Write
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
–
0
0
0
23
–
22
21
20
–
0
0
0
15
–
14
13
12
–
0
0
0
7
–
6
5
4
–
0
0
0
27
–
rwu
0
26
–
25
–
24
–
0
0
0
19
UHADDR_P6
rwu
0
18
17
16
0
0
0
11
UHADDR_P5
rwu
0
10
9
8
0
0
0
3
UHADDR_P4
rwu
0
2
1
0
• UHADDR_P4 : USB Host Address
These bits should contain the address of the Pipe4 of the USB Device.
Cleared by hardware when a USB reset is requested.
• UHADDR_P5 : USB Host Address
These bits should contain the address of the Pipe5 of the USB Device.
Cleared by hardware when a USB reset is requested.
• UHADDR_P6 : USB Host Address
These bits should contain the address of the Pipe6 of the USB Device.
Cleared by hardware when a USB reset is requested.
• UHADDR_P7 : USB Host Address
These bits should contain the address of the Pipe7 of the USB Device.
Cleared by hardware when a USB reset is requested.
612
32058K
AVR32-01/12
AT32UC3A
30.8.3.11
USB Pipe Enable/Reset Register (UPRST)
Offset:
0x0041C
Register Name:
UPRST
Access Type:
Read/Write
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
PRST6
rwu
0
21
PRST5
rwu
0
20
PRST4
rwu
0
19
PRST3
rwu
0
18
PRST2
rwu
0
17
PRST1
rwu
0
16
PRST0
rwu
0
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
PEN6
rw
0
5
PEN5
rw
0
4
PEN4
rw
0
3
PEN3
rw
0
2
PEN2
rw
0
1
PEN1
rw
0
0
PEN0
rw
0
• PENX, X in [0..6]: Pipe X Enable
Set to enable the Pipe X.
Clear to disable the Pipe X, what forces the Pipe X state to inactive and resets the pipe X registers (UPCFGX, UPSTAX,
UPCONX) but not the pipe configuration (ALLOC, PBK, PSIZE).
• PRSTX, X in [0..6]: Pipe X Reset
Set by software to reset the Pipe X FIFO.
This resets the endpoint X registers (UPCFGX, UPSTAX, UPCONX) but not the endpoint configuration (ALLOC, PBK,
PSIZE, PTOKEN, PTYPE, PEPNUM, INTFRQ).
All the endpoint mechanism (FIFO counter, reception, transmission, etc.) is reset apart from the Data Toggle management
.
The endpoint configuration remains active and the endpoint is still enabled.
Then, clear by software to complete the reset operation and to start using the FIFO.
613
32058K
AVR32-01/12
AT32UC3A
30.8.3.12
USB Pipe X Configuration Register (UPCFGX)
Offset:
0x0500 + X . 0x04
Register Name:
UPCFGX, X in [0..6]
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
0
0
17
16
0
INTFRQ
rwu
0
0
0
0
0
0
23
–
22
–
21
–
20
–
19
18
15
–
14
–
7
–
13
12
PTYPE
rwu
6
0
0
0
5
PSIZE
rwu
0
4
PEPNUM
rwu
0
0
0
11
–
10
AUTOSW
rwu
0
9
0
0
2
1
ALLOC
rwu
0
0
–
3
PBK
rwu
0
0
0
8
PTOKEN
rwu
• ALLOC: Pipe Memory Allocate
Set to configure the pipe memory with the characteristics.
Clear to update the memory allocation.
This bit is cleared by hardware when a USB Reset is requested.
Refer to the DPRAM Management chapter for more details.
• PBK: Pipe Banks
Set to select the number of banks for the pipe:
PBK
Endpoint Banks
0
0
1 (single-bank pipe)
0
1
2 (double-bank pipe)
1
0
3 (triple-bank pipe)
1
1
Reserved
For control endpoints, a single-bank pipe (00b) should be selected.
Cleared by hardware upon sending a USB reset.
614
32058K
AVR32-01/12
AT32UC3A
• PSIZE: Pipe Size
Set to select the size of each pipe bank:
PSIZE
Endpoint Size
0
0
0
8 bytes
0
0
1
16 bytes
0
1
0
32 bytes
0
1
1
64 bytes
1
0
0
128 bytes
1
0
1
256 bytes
1
1
0
512 bytes
1
1
1
1024 bytes
Cleared by hardware upon sending a USB reset.
• PTOKEN: Pipe Token
Set to select the endpoint token:
PTOKEN
Endpoint Direction
00
SETUP
01
IN
10
OUT
11
reserved
• AUTOSW: Automatic Switch
Set to automatically switch bank when it is ready.
Clear to disable the automatic bank switching.
Cleared by hardware upon sending a USB reset.
• PTYPE: Pipe Type
Set to select the pipe type:
PTYPE
Pipe Type
0
0
Control
0
1
Isochronous
1
0
Bulk
1
1
Interrupt
Cleared by hardware upon sending a USB reset.
• PEPNUM: Pipe Endpoint Number
Set this field according to the Pipe configuration. Set the number of the endpoint targeted by the pipe. This value is from 0
to 15.
Cleared by hardware upon sending a USB reset.
615
32058K
AVR32-01/12
AT32UC3A
• INTFRQ: Pipe Interrupt Request Frequency
These bits are the maximum value in millisecond of the polling period for an Interrupt Pipe.
This value has no effect for a non-Interrupt Pipe.
Cleared by hardware upon sending a USB reset.
616
32058K
AVR32-01/12
AT32UC3A
30.8.3.13
USB Pipe X Status Register (UPSTAX)
Offset:
0x0530 + X . 0x04
Register Name:
UPSTAX, X in [0..6]
Access Type:
Read-Only
Reset Value:
0x00000000
31
–
23
30
29
28
0
0
0
21
20
0
0
0
14
13
22
PBYCT
r
0
15
CURRBK
r
12
NBUSYBK
r
27
PBYCT
r
0
26
25
24
0
0
0
19
–
18
CFGOK
r
0
17
–
16
RWALL
r
0
11
–
10
–
9
0
0
0
0
7
SHORT
PACKETI
r
0
6
RXSTALLDI/
CRCERRI
r
0
5
4
3
OVERFI
NAKEDI
PERRI
r
0
r
0
r
0
2
TXSTPI/
UNDERFI
r
0
8
DTSEQ
r
0
0
1
0
TXOUTI
RXINI
r
0
r
0
• RXINI: Received IN Data Interrupt Flag
Set by hardware when a new USB message is stored in the current bank of the Pipe. This triggers an interrupt if the RXINE
bit is set. Shall be cleared by software (by setting the RXINIC bit).
• TXOUTI: Transmitted OUT Data Interrupt Flag
Set by hardware when the current OUT bank is free and can be filled. This triggers an interrupt if the TXOUTE bit is set.
Shall be cleared by software (by setting the TXOUTIC bit).
• TXSTPI: Transmitted SETUP Interrupt Flag
For Control endpoints. Set by hardware when the current SETUP bank is free and can be filled. This triggers an interrupt if
the TXSTPE bit is set. Shall be cleared by software (by setting the TXSTPIC bit).
• UNDERFI: Underflow Interrupt Flag
Set by hardware when a transaction underflow occurs in the current isochronous or interrupt pipe. (the pipe can’t send the
OUT data packet in time because the current bank is not ready). A zero-length-packet (ZLP) will be send instead of. This
triggers an interrupt if the UNDERFLE bit is set. Shall be cleared by software (by setting the UNDERFIC bit).
• PERRI: Pipe Error Interrupt Flag
Set by hardware when an error occurs on the current bank of the Pipe. This triggers an interrupt if the PERRE bit is set.
Refers to the UPERRX register to determine the source of the error. Automatically cleared by hardware when the error
source bit is cleared.
• NAKEDI: NAKed Interrupt Flag
Set by hardware when a NAK has been received on the current bank of the Pipe. This triggers an interrupt if the NAKEDE
bit is set. Shall be cleared by software (by setting the NAKEDIC bit).
617
32058K
AVR32-01/12
AT32UC3A
• OVERFI: Overflow Interrupt Flag
Set by hardware when the current Pipe has received more data than the maximum length of the current Pipe. An interrupt
is triggered if the OVERFIE bit is set. Shall be cleared by software (by setting the OVERFIC bit).
• RXSTALLDI: Received STALLed Interrupt Flag
For all endpoints but isochronous. Set by hardware when a STALL handshake has been received on the current bank of
the Pipe. The Pipe is automatically frozen. This triggers an interrupt if the RXSTALLE bit is set. Shall be cleared to handshake the interrupt (by setting the RXSTALLDIC bit).
• CRCERRI: CRC Error Interrupt Flag
For isochronous endpoint, set by hardware when a CRC error occurs on the current bank of the Pipe. This triggers an interrupt if the TXSTPE bit is set. Shall be cleared to handshake the interrupt (by setting the CRCERRIC bit).
• SHORTPACKETI: Short Packet Interrupt Flag
Set by hardware when a short packet is received by the host controller (packet length inferior to the PSIZE programmed
field). Shall be cleared to handshake the interrupt (by setting the SHORTPACKETIC bit).
• DTSEQ: Data Toggle Sequence
Set by hardware to indicate the data PID of the current bank.
DTSEQ
Data toggle sequence
0
0
Data0
0
1
Data1
1
0
reserved
1
1
reserved
For OUT pipe, this field indicates the data toggle of the next packet that will be sent.
For IN pipe, this field indicates the data toggle of the received packet stored in the current bank.
• NBUSYBK: Number of Busy Banks
Set by hardware to indicate the number of busy bank.
For OUT Pipe, it indicates the number of busy bank(s), filled by the user, ready for OUT transfer. When all banks are busy,
this triggers an PXINT interrupt if UPCONX.NBUSYBKE = 1.
For IN Pipe, it indicates the number of busy bank(s) filled by IN transaction from the Device. When all banks are free, this
triggers an PXINT interrupt if UPCONX.NBUSYBKE = 1..
NBUSYBK
Number of busy bank
0
0
All banks are free.
0
1
1 busy bank
1
0
2 busy banks
1
1
reserved
618
32058K
AVR32-01/12
AT32UC3A
• CURRBK: Current Bank
For non-control pipe, set by hardware to indicate the number of the current bank.
CURRBK
Current Bank
0
0
Bank0
0
1
Bank1
1
0
Bank2
1
1
Reserved
Note that this field may be updated 1 clock cycle after the RWALL bit changes, so the user should not poll this field as an
interrupt flag.
• RWALL: Read/Write Allowed
For OUT pipe, set by hardware when the current bank is not full, i.e. the software can write further data into the FIFO.
For IN pipe, set by hardware when the current bank is not empty, i.e. the software can read further data from the FIFO.
Cleared by hardware otherwise.
This bit is also cleared by hardware when the RXSTALL or the PERR bit is set.
• CFGOK: Configuration OK Status
This bit is updated when the ALLOC bit is set.
Set by hardware if the pipe X number of banks (PBK) and size (PSIZE) are correct compared to the maximal allowed number of banks and size for this pipe and to the maximal FIFO size (i.e. the DPRAM size).
If this bit is cleared by hardware, the user should reprogram the UPCFGX register with correct PBK and PSIZE values.
• PBYCT: Pipe Byte Count
Set by the hardware to indicate the byte count of the FIFO.
For OUT pipe, incremented after each byte written by the software into the pipe and decremented after each byte sent to
the peripheral.
For In pipe, incremented after each byte received from the peripheral and decremented after each byte read by the software from the pipe.
Note that this field may be updated 1 clock cycle after the RWALL bit changes, so the user should not poll this field as an
interrupt flag.
619
32058K
AVR32-01/12
AT32UC3A
30.8.3.14
USB Pipe X Status Clear Register (UPSTAXCLR)
Offset:
0x0560 + X . 0x04
Register Name:
UPSTAXCLR, X in [0..6]
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
SHORT
PACKETIC
w
0
6
RXSTALLDIC/
CRCERRIC
w
0
5
4
3
1
0
OVERFIC
NAKEDIC
–
TXOUTIC
RXINIC
w
0
w
0
2
TXSTPIC/
UNDERFIC
w
0
w
0
w
0
• RXINIC: Received IN Data Interrupt Flag Clear
Set to clear RXINI.
Clearing has no effect.
Always read as 0.
• TXOUTIC: Transmitted OUT Data Interrupt Flag Clear
Set to clear TXOUTI.
Clearing has no effect.
Always read as 0.
• TXSTPIC: Transmitted SETUP Interrupt Flag Clear
Set to clear TXSTPI.
Clearing has no effect.
Always read as 0.
• UNDERFIC: Underflow Interrupt Flag Clear
Set to clear UNDERFI.
Clearing has no effect.
Always read as 0.
• NAKEDIC: NAKed Interrupt Flag Clear
Set to clear NAKEDI.
620
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• OVERFIC: Overflow Interrupt Flag Clear
Set to clear OVERFI.
Clearing has no effect.
Always read as 0.
• RXSTALLDIC: Received STALLed Interrupt Flag Clear
Set to clear RXSTALLDI.
Clearing has no effect.
Always read as 0.
• CRCERRIC: CRC Error Interrupt Flag Clear
Set to clear CRCERRI.
Clearing has no effect.
Always read as 0.
• SHORTPACKETIC: Short Packet Interrupt Flag Clear
Set to clear SHORTPACKETI.
Clearing has no effect.
Always read as 0.
621
32058K
AVR32-01/12
AT32UC3A
30.8.3.15
USB Pipe X Status Set Register (UPSTAXSET)
Offset:
0x0590 + X . 0x04
Register Name:
UPSTAXSET, X in [0..6]
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
NBUSYBKS
w
0
11
–
10
–
9
–
8
–
7
SHORT
PACKETIS
w
0
6
RXSTALLDIS/
CRCERRIS
w
0
5
4
3
1
0
OVERFIS
NAKEDIS
PERRIS
TXOUTIS
RXINIS
w
0
w
0
w
0
2
TXSTPIS/
UNDERFIS
w
0
w
0
w
0
• RXINIS: Received IN Data Interrupt Flag Set
Set to set RXINI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• TXOUTIS: Transmitted OUT Data Interrupt Flag Set
Set to set TXOUTI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• TXSTPIS: Transmitted SETUP Interrupt Flag Set
Set to set TXSTPI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• UNDERFIS: Underflow Interrupt Flag Set
Set to set UNDERFI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• PERRIS: Pipe Error Interrupt Flag Set
Set to set PERRI, what may be useful for test or debug purposes.
622
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• NAKEDIS: NAKed Interrupt Flag Set
Set to set NAKEDI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• OVERFIS: Overflow Interrupt Flag Set
Set to set OVERFI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• RXSTALLDIS: Received STALLed Interrupt Flag Set
Set to set RXSTALLDI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• CRCERRIS: CRC Error Interrupt Flag Set
Set to set CRCERRI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• SHORTPACKETIS: Short Packet Interrupt Flag Set
Set to set SHORTPACKETI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
• NBUSYBKS: Number of Busy Banks Interrupt Flag Set
Set to set NBUSYBKI, what may be useful for test or debug purposes.
Clearing has no effect.
Always read as 0.
623
32058K
AVR32-01/12
AT32UC3A
30.8.3.16
USB Pipe X Control Register (UPCONX)
Offset:
0x05C0 + X . 0x04
Register Name:
UPCONX, X in [0..6]
Access Type:
Read-Only
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
RSTDT
ru
0
17
PFREEZE
ru
0
16
PDISHDMA
ru
0
15
–
14
FIFOCON
ru
0
13
–
12
NBUSYBKE
ru
0
11
–
10
–
9
–
8
–
7
SHORT
PACKETIE
ru
0
6
RXSTALLDE/
CRCERRE
ru
0
5
4
3
1
0
OVERFIE
NAKEDE
PERRE
TXOUTE
RXINE
ru
0
ru
0
ru
0
2
TXSTPE/
UNDERFIE
ru
0
ru
0
ru
0
• RXINE: Received IN Data Interrupt Enable
Set by software (by setting the RXINES bit) to enable the Received IN Data interrupt (RXINE).
Clear by software (by setting the RXINEC bit) to disable the Received IN Data interrupt (RXINE).
• TXOUTE: Transmitted OUT Data Interrupt Enable
Set by software (by setting the TXOUTES bit) to enable the Transmitted OUT Data interrupt (TXOUTE).
Clear by software (by setting the TXOUTECbit) to disable the Transmitted OUT Data interrupt (TXOUTE).
• TXSTPE: Transmitted SETUP Interrupt Enable
Set by software (by setting the TXSTPES bit) to enable the Transmitted SETUP interrupt (TXSTPE).
Clear by software (by setting the TXSTPEC bit) to disable the Transmitted SETUP interrupt (TXSTPE).
• UNDERFIE: Underflow Interrupt Enable
Set by software (by setting the UNDERFIES bit) to enable the Underflow interrupt (UNDERFIE).
Clear by software (by setting the UNDERFIEC bit) to disable the Underflow interrupt (UNDERFIE).
• PERRE: Pipe Error Interrupt Enable
Set by software (by setting the PERRES bit) to enable the Pipe Error interrupt (PERRE).
Clear by software (by setting the PERREC bit) to disable the Pipe Error interrupt (PERRE).
• NAKEDE: NAKed Interrupt Enable
Set by software (by setting the NAKEDES bit) to enable the NAKed interrupt (NAKEDE).
Clear by software (by setting the NAKEDEC bit) to disable the NAKed interrupt (NAKEDE).
624
32058K
AVR32-01/12
AT32UC3A
• OVERFIE: Overflow Interrupt Enable
Set by software (by setting the OVERFIES bit) to enable the Overflow interrupt (OVERFIE).
Clear by software (by setting the OVERFIEC bit) to disable the Overflow interrupt (OVERFIE).
• RXSTALLDE: Received STALLed Interrupt Enable
Set by software (by setting the RXSTALLDES bit) to enable the Received STALLed interrupt (RXSTALLDE).
Clear by software (by setting the RXSTALLDEC bit) to disable the Received STALLed interrupt (RXSTALLDE).
• CRCERRE: CRC Error Interrupt Enable
Set by software (by setting the CRCERRES bit) to enable the CRC Error interrupt (CRCERRE).
Clear by software (by setting the CRCERREC bit) to disable the CRC Error interrupt (CRCERRE).
• SHORTPACKETIE: Short Packet Interrupt Enable
Set by software (by setting the SHORTPACKETES bit) to enable the Short Packet interrupt (SHORTPACKETIE).
Clear by software (by setting the SHORTPACKETEC bit) to disable the Short Packet interrupt (SHORTPACKETE).
• NBUSYBKE: Number of Busy Banks Interrupt Enable
Set by software (by setting the NBUSYBKES bit) to enable the Number of Busy Banks interrupt (NBUSYBKE).
Clear by software (by setting the NBUSYBKEC bit) to disable the Number of Busy Banks interrupt (NBUSYBKE).
• FIFOCON: FIFO Control
For OUT and SETUP Pipe :
Set by hardware when the current bank is free, at the same time than TXOUTI or TXSTPI.
Clear by software (by setting the FIFOCONC bit) to send the FIFO data and to switch the bank.
For IN Pipe:
Set by hardware when a new IN message is stored in the current bank, at the same time than RXINI.
Clear by software (by setting the FIFOCONC bit) to free the current bank and to switch to the next bank.
• PDISHDMA: Pipe Interrupts Disable HDMA Request Enable
See EPDISHDMA (UECONX register).
• PFREEZE: Pipe Freeze
Set by software (by setting the PFREEZES bit) to Freeze the Pipe requests generation.
Clear by software (by setting the PFREEZEC bit) to enable the Pipe request generation.
This bit is set by hardware when:
- the pipe is not configured
- a STALL handshake has been received on this Pipe
- An error occurs on the Pipe (PERR = 1)
- (INRQ+1) In requests have been processed
This bit is set at 1 by hardware after a Pipe reset or a Pipe enable.
• RSTDT: Reset Data Toggle
Set by software (by setting the RSTDTS bit) to reset the Data Toggle to its initial value for the current Pipe.
Cleared by hardware when proceed.
625
32058K
AVR32-01/12
AT32UC3A
30.8.3.17
USB Pipe X Control Clear Register (UPCONXCLR)
Offset:
0x0620 + X . 0x04
Register Name:
UPCONXCLR, X in [0..6]
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
PFREEZEC
w
0
16
PDISHDMAC
w
0
15
–
14
FIFOCONC
w
0
13
–
12
NBUSYBKEC
w
0
11
–
10
–
9
–
8
–
7
SHORT
PACKETIEC
w
0
6
RXSTALLDEC/
CRCERREC
w
0
5
4
3
1
0
OVERFIEC
NAKEDEC
PERREC
TXOUTEC
RXINEC
w
0
w
0
w
0
2
TXSTPEC/
UNDERFIEC
w
0
w
0
w
0
• RXINEC: Received IN Data Interrupt Enable Clear
Set to clear RXINE.
Clearing has no effect.
Always read as 0.
• TXOUTEC: Transmitted OUT Data Interrupt Enable Clear
Set to clear TXOUTE.
Clearing has no effect.
Always read as 0.
• TXSTPEC: Transmitted SETUP Interrupt Enable Clear
Set to clear TXSTPE.
Clearing has no effect.
Always read as 0.
• UNDERFIEC: Underflow Interrupt Enable Clear
Set to clear UNDERFIE.
Clearing has no effect.
Always read as 0.
• PERREC: Pipe Error Interrupt Enable Clear
Set to clear PERRE.
626
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• NAKEDEC: NAKed Interrupt Enable Clear
Set to clear NAKEDE.
Clearing has no effect.
Always read as 0.
• OVERFIEC: Overflow Interrupt Enable Clear
Set to clear OVERFIE.
Clearing has no effect.
Always read as 0.
• RXSTALLDEC: Received STALLed Interrupt Enable Clear
Set to clear RXSTALLDE.
Clearing has no effect.
Always read as 0.
• CRCERREC: CRC Error Interrupt Enable Clear
Set to clear CRCERRE.
Clearing has no effect.
Always read as 0.
• SHORTPACKETIEC: Short Packet Interrupt Enable Clear
Set to clear SHORTPACKETIE.
Clearing has no effect.
Always read as 0.
• NBUSYBKEC: Number of Busy Banks Interrupt Enable Clear
Set to clear NBUSYBKE.
Clearing has no effect.
Always read as 0.
• FIFOCONC: FIFO Control Clear
Set to clear FIFOCON.
Clearing has no effect.
Always read as 0.
• PDISHDMAC: Pipe Interrupts Disable HDMA Request Enable Clear
Set to clear PDISHDMA.
Clearing has no effect.
Always read as 0.
• PFREEZEC: Pipe Freeze Clear
Set to clear PFREEZE.
627
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
628
32058K
AVR32-01/12
AT32UC3A
30.8.3.18
USB Pipe X Control Set Register (UPCONXSET)
Offset:
0x05F0 + X . 0x04
Register Name:
UPCONXSET, X in [0..6]
Access Type:
Write-Only
Read Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
RSTDTS
w
0
17
PFREEZES
w
0
16
PDISHDMAS
w
0
15
–
14
–
13
–
12
NBUSYBKES
w
0
11
–
10
–
9
–
8
–
7
SHORT
PACKETIES
w
0
6
RXSTALLDES/
CRCERRES
w
0
5
4
3
1
0
OVERFIES
NAKEDES
PERRES
TXOUTES
RXINES
w
0
w
0
w
0
2
TXSTPES/
UNDERFIES
w
0
w
0
w
0
• RXINES: Received IN Data Interrupt Enable Set
Set to set RXINE.
Clearing has no effect.
Always read as 0.
• TXOUTE: Transmitted OUT Data Interrupt Enable Set
Set to set RXINE.
Clearing has no effect.
Always read as 0.
• TXSTPES: Transmitted SETUP Interrupt Enable Set
Set to set TXSTPE.
Clearing has no effect.
Always read as 0.
• UNDERFIES: Underflow Interrupt Enable Set
Set to set UNDERFIE.
Clearing has no effect.
Always read as 0.
• PERRES: Pipe Error Interrupt Enable Set
Set to set PERRE.
629
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
• NAKEDES: NAKed Interrupt Enable Set
Set to set NAKEDE.
Clearing has no effect.
Always read as 0.
• OVERFIES: Overflow Interrupt Enable Set
Set to set OVERFIE.
Clearing has no effect.
Always read as 0.
• RXSTALLDES: Received STALLed Interrupt Enable Set
Set to set RXSTALLDE.
Clearing has no effect.
Always read as 0.
• CRCERRES: CRC Error Interrupt Enable Set
Set to set CRCERRE.
Clearing has no effect.
Always read as 0.
• SHORTPACKETIES: Short Packet Interrupt Enable Set
Set to set SHORTPACKETIE.
Clearing has no effect.
Always read as 0.
• NBUSYBKES: Number of Busy Banks Interrupt Enable Set
Set to set NBUSYBKE.
Clearing has no effect.
Always read as 0.
• PDISHDMAS: Pipe Interrupts Disable HDMA Request Enable Set
Set to set PDISHDMA.
Clearing has no effect.
Always read as 0.
• PFREEZES: Pipe Freeze Set
Set to set PFREEZE.
Clearing has no effect.
Always read as 0.
• RSTDTS: Reset Data Toggle Set
Set to set RSTDT.
630
32058K
AVR32-01/12
AT32UC3A
Clearing has no effect.
Always read as 0.
631
32058K
AVR32-01/12
AT32UC3A
30.8.3.19
USB Pipe X IN Request Register (UPINRQX)
Offset:
0x0650 + X . 0x04
Register Name:
UPINRQX, X in [0..6]
Access Type:
Read/Write
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
INMODE
rw
0
7
6
5
4
3
2
1
0
0
0
0
0
INRQ
rwu
0
0
0
0
• INRQ: IN Request Number before Freeze
Enter the number of IN transactions before the USB controller freezes the pipe. The USB controller will perform (INRQ+1)
IN requests before to freeze the Pipe. This counter is automatically decreased by 1 each time a IN request has been successfully performed.
This register has no effect when the INMODE bit is set (infinite IN requests generation till the pipe is not frozen).
• INMODE: IN Request Mode
Set this bit to allow the USB controller to perform infinite IN requests when the Pipe is not frozen.
Clear this bit to perform a pre-defined number of IN requests. This number is the INRQ field.
632
32058K
AVR32-01/12
AT32UC3A
30.8.3.20
USB Pipe X Error Register (UPERRX)
Offset:
0x0680 + X . 0x04
Register Name:
UPERRX, X in [0..6]
Access Type:
Read/Write
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
5
4
CRC16
rwu
0
3
TIMEOUT
rwu
0
2
PID
rwu
0
1
DATAPID
rwu
0
0
DATATGL
rwu
0
COUNTER
rwu
0
0
• DATATGL: Data Toggle Error
Set by hardware when a data toggle error has been detected.
Shall be cleared by software. Setting by software has no effect.
• DATAPID: Data PID Error
Set by hardware when a data PID error has been detected.
Shall be cleared by software. Setting by software has no effect.
• PID: PID Error
Set by hardware when a PID error has been detected.
Shall be cleared by software. Setting by software has no effect.
• TIMEOUT: Time-Out Error
Set by hardware when a Timeout error has been detected.
Shall be cleared by software. Setting by software has no effect.
• CRC16: CRC16 Error
Set by hardware when a CRC16 error has been detected.
Shall be cleared by software. Setting by software has no effect.
• COUNTER: Error Counter
Set by hardware when a CRC16 error has been detected.
Shall be cleared by software. Setting by software has no effect.
633
32058K
AVR32-01/12
AT32UC3A
30.8.3.21
USB Host DMA Channel X Next Descriptor Address Register (UHDMAX_NEXTDESC)
Offset:
0x0710 + (X - 1) . 0x10
Register Name:
UHDMAX_NEXTDESC, X in [1..6]
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
0
0
0
23
22
21
0
0
0
15
14
13
0
0
0
7
6
5
NXT_DESC_ADDR
rwu
0
0
0
28
27
NXT_DESC_ADDR
rwu
0
0
26
25
24
0
0
0
20
19
NXT_DESC_ADDR
rwu
0
0
18
17
16
0
0
0
12
11
NXT_DESC_ADDR
rwu
0
0
10
9
8
0
0
0
4
2
–
1
–
0
–
3
–
0
Same as ”USB Device DMA Channel X Next Descriptor Address Register (UDDMAX_NEXTDESC)” on page 591.
634
32058K
AVR32-01/12
AT32UC3A
30.8.3.22
USB Host DMA Channel X HSB Address Register (UHDMAX_ADDR)
Offset:
0x0714 + (X - 1) . 0x10
Register Name:
UHDMAX_ADDR, X in [1..6]
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
0
0
0
0
19
18
17
16
0
0
0
0
11
10
9
8
0
0
0
0
3
2
1
0
0
0
0
0
HSB_ADDR
rwu
0
0
0
0
23
22
21
20
HSB_ADDR
rwu
0
0
0
0
15
14
13
12
HSB_ADDR
rwu
0
0
0
0
7
6
5
4
HSB_ADDR
rwu
0
0
0
0
Same as ”USB Device DMA Channel X HSB Address Register (UDDMAX_ADDR)” on page 592.
635
32058K
AVR32-01/12
AT32UC3A
30.8.3.23
USB Host DMA Channel X Control Register (UHDMAX_CONTROL)
Offset:
0x0718 + (X - 1) . 0x10
Register Name:
UHDMAX_CONTROL, X in [1..6]
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
CH_BYTE_LENGTH
rwu
0
0
26
25
24
0
0
0
0
0
0
23
22
21
18
17
16
0
20
19
CH_BYTE_LENGTH
rwu
0
0
0
0
0
0
0
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
BURST_LOCK
_EN
rwu
0
6
DESC_LD_
IRQ_EN
rwu
0
5
EOBUFF_
IRQ_EN
rwu
0
4
3
DMAEND_EN
rwu
0
rwu
0
1
LD_NXT_CH_
DESC_EN
rwu
0
0
EOT_IRQ_EN
2
BUFF_CLOSE
_IN_EN
rwu
0
CH_EN
rwu
0
Same as ”USB Device DMA Channel X Control Register (UDDMAX_CONTROL)” on page 593.
(just replace the IN endpoint term by OUT endpoint, and vice-versa)
636
32058K
AVR32-01/12
AT32UC3A
30.8.3.24
USB Host DMA Channel X Status Register (UHDMAX_STATUS)
Offset:
0x071C + (X - 1) . 0x10
Register Name:
UHDMAX_STATUS, X in [1..6]
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
0
0
0
0
19
18
17
16
CH_BYTE_CNT
ru
0
0
0
0
23
22
21
20
CH_BYTE_CNT
ru
0
0
0
0
0
0
0
0
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
6
DESC_LD_
STA
ru
0
5
EOCH_BUFF_
STA
ru
0
4
3
2
1
0
EOT_STA
–
–
CH_ACTIVE
CH_EN
rwu
0
rwu
0
–
ru
0
Same as ”USB Device DMA Channel X Status Register (UDDMAX_STATUS)” on page 595.
637
32058K
AVR32-01/12
AT32UC3A
30.8.4
USB Pipe/Endpoint X FIFO Data Register (USB_FIFOX_DATA)
Note that this register can be accessed even if USBE = 0 or FRZCLK = 1. Disabling the USB controller (by clearing the
USBE bit) does not reset the DPRAM.
638
32058K
AVR32-01/12
AT32UC3A
31. Timer/Counter (TC)
Rev: 2.2.2.1
31.1
Features
• Three 16-bit Timer Counter Channels
• A Wide Range of Functions Including:
– Frequency Measurement
– Event Counting
– Interval Measurement
– Pulse Generation
– Delay Timing
– Pulse Width Modulation
– Up/down Capabilities
• Each Channel is User-configurable and Contains:
– Three External Clock Inputs
– Five Internal Clock Inputs
– Two Multi-purpose Input/Output Signals
• Internal Interrupt Signal
• Two Global Registers that Act on All Three TC Channels
31.2
Description
The Timer Counter (TC) includes three identical 16-bit Timer Counter channels.
Each channel can be independently programmed to perform a wide range of functions including
frequency measurement, event counting, interval measurement, pulse generation, delay timing
and pulse width modulation.
Each channel has three external clock inputs, five internal clock inputs and two multi-purpose
input/output signals which can be configured by the user. Each channel drives an internal interrupt signal which can be programmed to generate processor interrupts.
The Timer Counter block has two global registers which act upon all three TC channels.
The Block Control Register allows the three channels to be started simultaneously with the same
instruction.
The Block Mode Register defines the external clock inputs for each channel, allowing them to be
chained.
639
32058K
AVR32-01/12
AT32UC3A
31.3
Block Diagram
Figure 31-1. Timer Counter Block Diagram
Parallel I/O
Controller
TIMER_CLOCK1
TCLK0
TIMER_CLOCK2
TIOA1
TIOA2
TIMER_CLOCK3
XC0
TCLK1
TIMER_CLOCK4
XC1
TCLK2
Timer/Counter
Channel 0
TIOB
XC2
TIMER_CLOCK5
TC0XC0S
TIOA
SYNC
TCLK0
TCLK1
TCLK2
TIOA0
TIOB0
TIOA0
TIOB0
INT0
TCLK0
TCLK1
XC0
TIOA0
XC1
TIOA2
XC2
TCLK2
TC1XC1S
TCLK0
XC0
TCLK1
XC1
TCLK2
XC2
TIOA0
TC2XC2S
TIOA1
Timer/Counter
Channel 1
TIOA
TIOB
SYNC
Timer/Counter
Channel 2
TIOB1
INT1
TIOA
TIOB
SYNC
TIOA1
TIOB1
TIOA1
TIOA2
TIOB2
TIOA2
TIOB2
INT2
Timer Counter
Interrupt
Controller
Table 31-1.
Signal Name Description
Block/Channel
Signal Name
XC0, XC1, XC2
Channel Signal
Description
External Clock Inputs
TIOA
Capture Mode: Timer Counter Input
Waveform Mode: Timer Counter Output
TIOB
Capture Mode: Timer Counter Input
Waveform Mode: Timer Counter Input/Output
INT
SYNC
Interrupt Signal Output
Synchronization Input Signal
640
32058K
AVR32-01/12
AT32UC3A
31.4
Pin Name List
Table 31-2.
31.5
TC pin list
Pin Name
Description
Type
TCLK0-TCLK2
External Clock Input
Input
TIOA0-TIOA2
I/O Line A
I/O
TIOB0-TIOB2
I/O Line B
I/O
Product Dependencies
31.5.1
I/O Lines
The pins used for interfacing the compliant external devices may be multiplexed with PIO lines.
The programmer must first program the PIO controllers to assign the TC pins to their peripheral
functions.
31.5.2
Debug operation
The Timer Counter clocks are frozen during debug operation, unless the OCD system keeps
peripherals running in debug operation.
31.5.3
Power Management
The Timer Counter clock is generated by the power manager. Before using the TC, the programmer must ensure that the TC clock is enabled in the power manager.
31.5.4
Interrupt
The TC has an interrupt line connected to the interrupt controller. Handling the TC interrupt
requires programming the interrupt controller before configuring the TC.
31.6
Functional Description
31.6.1
TC Description
The three channels of the Timer Counter are independent and identical in operation. The registers for channel programming are listed in Table 31-4 on page 654.
31.6.1.1
16-bit Counter
Each channel is organized around a 16-bit counter. The value of the counter is incremented at
each positive edge of the selected clock. When the counter has reached the value 0xFFFF and
passes to 0x0000, an overflow occurs and the COVFS bit in SR (Status Register) is set.
The current value of the counter is accessible in real time by reading the Counter Value Register, CV. The counter can be reset by a trigger. In this case, the counter value passes to 0x0000
on the next valid edge of the selected clock.
31.6.1.2
Clock Selection
At block level, input clock signals of each channel can either be connected to the external inputs
TCLK0, TCLK1 or TCLK2, or be connected to the configurable I/O signals TIOA0, TIOA1 or
TIOA2 for chaining by programming the BMR (Block Mode). See Figure 31-2.
641
32058K
AVR32-01/12
AT32UC3A
Each channel can independently select an internal or external clock source for its counter:
• Internal clock signals: TIMER_CLOCK1, TIMER_CLOCK2, TIMER_CLOCK3,
TIMER_CLOCK4, TIMER_CLOCK5. The Peripherals Chapter details the connection of these
clock sources.
• External clock signals: XC0, XC1 or XC2. The Peripherals Chapter details the connection of
these clock sources.
This selection is made by the TCCLKS bits in the TC Channel Mode Register .
The selected clock can be inverted with the CLKI bit in CMR. This allows counting on the opposite edges of the clock.
The burst function allows the clock to be validated when an external signal is high. The BURST
parameter in the Mode Register defines this signal (none, XC0, XC1, XC2).
Note:
In all cases, if an external clock is used, the duration of each of its levels must be longer than the
master clock period. The external clock frequency must be at least 2.5 times lower than the master clock
Figure 31-2. Clock Selection
TCCLKS
TIMER_CLOCK1
TIMER_CLOCK2
CLKI
TIMER_CLOCK3
TIMER_CLOCK4
TIMER_CLOCK5
Selected
Clock
XC0
XC1
XC2
BURST
1
31.6.1.3
Clock Control
The clock of each counter can be controlled in two different ways: it can be enabled/disabled
and started/stopped. See Figure 31-3.
• The clock can be enabled or disabled by the user with the CLKEN and the CLKDIS commands
in the Control Register. In Capture Mode it can be disabled by an RB load event if LDBDIS is
set to 1 in CMR. In Waveform Mode, it can be disabled by an RC Compare event if CPCDIS is
set to 1 in CMR. When disabled, the start or the stop actions have no effect: only a CLKEN
command in the Control Register can re-enable the clock. When the clock is enabled, the
CLKSTA bit is set in the Status Register.
• The clock can also be started or stopped: a trigger (software, synchro, external or compare)
always starts the clock. The clock can be stopped by an RB load event in Capture Mode
642
32058K
AVR32-01/12
AT32UC3A
(LDBSTOP = 1 in CMR) or a RC compare event in Waveform Mode (CPCSTOP = 1 in CMR).
The start and the stop commands have effect only if the clock is enabled.
Figure 31-3. Clock Control
Selected
Clock
Trigger
CLKSTA
CLKEN
Q
Q
S
CLKDIS
S
R
R
Counter
Clock
31.6.1.4
Stop
Event
Disable
Event
TC Operating Modes
Each channel can independently operate in two different modes:
• Capture Mode provides measurement on signals.
• Waveform Mode provides wave generation.
The TC Operating Mode is programmed with the WAVE bit in the TC Channel Mode Register.
In Capture Mode, TIOA and TIOB are configured as inputs.
In Waveform Mode, TIOA is always configured to be an output and TIOB is an output if it is not
selected to be the external trigger.
31.6.1.5
Trigger
A trigger resets the counter and starts the counter clock. Three types of triggers are common to
both modes, and a fourth external trigger is available to each mode.
The following triggers are common to both modes:
• Software Trigger: Each channel has a software trigger, available by setting SWTRG in CCR.
• SYNC: Each channel has a synchronization signal SYNC. When asserted, this signal has the
same effect as a software trigger. The SYNC signals of all channels are asserted
simultaneously by writing BCR (Block Control) with SYNC set.
• Compare RC Trigger: RC is implemented in each channel and can provide a trigger when the
counter value matches the RC value if CPCTRG is set in CMR.
The channel can also be configured to have an external trigger. In Capture Mode, the external
trigger signal can be selected between TIOA and TIOB. In Waveform Mode, an external event
can be programmed on one of the following signals: TIOB, XC0, XC1 or XC2. This external
event can then be programmed to perform a trigger by setting ENETRG in CMR.
643
32058K
AVR32-01/12
AT32UC3A
If an external trigger is used, the duration of the pulses must be longer than the master clock
period in order to be detected.
Regardless of the trigger used, it will be taken into account at the following active edge of the
selected clock. This means that the counter value can be read differently from zero just after a
trigger, especially when a low frequency signal is selected as the clock.
31.6.2
Capture Operating Mode
This mode is entered by clearing the WAVE parameter in CMR (Channel Mode Register).
Capture Mode allows the TC channel to perform measurements such as pulse timing, frequency, period, duty cycle and phase on TIOA and TIOB signals which are considered as
inputs.
Figure 31-4 shows the configuration of the TC channel when programmed in Capture Mode.
31.6.2.1
Capture Registers A and B
Registers A and B (RA and RB) are used as capture registers. This means that they can be
loaded with the counter value when a programmable event occurs on the signal TIOA.
The LDRA parameter in CMR defines the TIOA edge for the loading of register A, and the LDRB
parameter defines the TIOA edge for the loading of Register B.
RA is loaded only if it has not been loaded since the last trigger or if RB has been loaded since
the last loading of RA.
RB is loaded only if RA has been loaded since the last trigger or the last loading of RB.
Loading RA or RB before the read of the last value loaded sets the Overrun Error Flag (LOVRS)
in SR (Status Register). In this case, the old value is overwritten.
31.6.2.2
Trigger Conditions
In addition to the SYNC signal, the software trigger and the RC compare trigger, an external trigger can be defined.
The ABETRG bit in CMR selects TIOA or TIOB input signal as an external trigger. The
ETRGEDG parameter defines the edge (rising, falling or both) detected to generate an external
trigger. If ETRGEDG = 0 (none), the external trigger is disabled.
644
32058K
AVR32-01/12
32058K
AVR32-01/12
MTIOA
MTIOB
1
If RA is not loaded
or RB is Loaded
Edge
Detector
ETRGEDG
SWTRG
Timer/Counter Channel
ABETRG
BURST
CLKI
S
R
OVF
LDRB
Edge
Detector
Edge
Detector
Capture
Register A
LDBSTOP
R
S
CLKEN
LDRA
If RA is Loaded
CPCTRG
16-bit Counter
RESET
Trig
CLK
Q
Q
CLKSTA
LDBDIS
Capture
Register B
CLKDIS
TC1_SR
TIOA
TIOB
SYNC
XC2
XC1
XC0
TIMER_CLOCK5
TIMER_CLOCK4
TIMER_CLOCK3
TIMER_CLOCK2
TIMER_CLOCK1
TCCLKS
Compare RC =
Register C
COVFS
LDRBS
INT
AT32UC3A
Figure 31-4. Capture Mode
CPCS
LOVRS
ETRGS
LDRAS
TC1_IMR
645
AT32UC3A
31.6.3
Waveform Operating Mode
Waveform operating mode is entered by setting the WAVE parameter in CMR (Channel Mode
Register).
In Waveform Operating Mode the TC channel generates 1 or 2 PWM signals with the same frequency and independently programmable duty cycles, or generates different types of one-shot
or repetitive pulses.
In this mode, TIOA is configured as an output and TIOB is defined as an output if it is not used
as an external event (EEVT parameter in CMR).
Figure 31-5 shows the configuration of the TC channel when programmed in Waveform Operating Mode.
31.6.3.1
Waveform Selection
Depending on the WAVSEL parameter in CMR (Channel Mode Register), the behavior of CV
varies.
With any selection, RA, RB and RC can all be used as compare registers.
RA Compare is used to control the TIOA output, RB Compare is used to control the TIOB output
(if correctly configured) and RC Compare is used to control TIOA and/or TIOB outputs.
646
32058K
AVR32-01/12
32058K
AVR32-01/12
TIOB
SYNC
XC2
XC1
XC0
TIMER_CLOCK5
TIMER_CLOCK4
TIMER_CLOCK3
TIMER_CLOCK2
TIMER_CLOCK1
1
EEVT
BURST
ENETRG
Trig
CLK
R
S
OVF
WAVSEL
RESET
16-bit Counter
WAVSEL
Q
Q
CLKSTA
Compare RA =
Register A
TC1_SR
Timer/Counter Channel
Edge
Detector
EEVTEDG
SWTRG
CLKI
Compare RC =
Compare RB =
CPCSTOP
CPCDIS
Register C
CLKDIS
Register B
R
S
CLKEN
CPAS
INT
BSWTRG
BEEVT
BCPB
BCPC
ASWTRG
AEEVT
ACPA
ACPC
Output Controller
Output Controller
TCCLKS
TIOB
MTIOB
TIOA
MTIOA
AT32UC3A
Figure 31-5. Waveform Mode
CPCS
CPBS
COVFS
ETRGS
TC1_IMR
647
AT32UC3A
31.6.3.2
WAVSEL = 00
When WAVSEL = 00, the value of CV is incremented from 0 to 0xFFFF. Once 0xFFFF has been
reached, the value of CV is reset. Incrementation of CV starts again and the cycle continues.
See Figure 31-6.
An external event trigger or a software trigger can reset the value of CV. It is important to note
that the trigger may occur at any time. See Figure 31-7.
RC Compare cannot be programmed to generate a trigger in this configuration. At the same
time, RC Compare can stop the counter clock (CPCSTOP = 1 in CMR) and/or disable the counter clock (CPCDIS = 1 in CMR).
Figure 31-6. WAVSEL= 00 without trigger
Counter Value
Counter cleared by compare match with 0xFFFF
0xFFFF
RC
RB
RA
Waveform Examples
Time
TIOB
TIOA
648
32058K
AVR32-01/12
AT32UC3A
Figure 31-7. WAVSEL= 00 with trigger
Counter cleared by compare match with 0xFFFF
Counter Value
0xFFFF
Counter cleared by trigger
RC
RB
RA
Time
Waveform Examples
TIOB
TIOA
31.6.3.3
WAVSEL = 10
When WAVSEL = 10, the value of CV is incremented from 0 to the value of RC, then automatically reset on a RC Compare. Once the value of CV has been reset, it is then incremented and
so on. See Figure 31-8.
It is important to note that CV can be reset at any time by an external event or a software trigger
if both are programmed correctly. See Figure 31-9.
In addition, RC Compare can stop the counter clock (CPCSTOP = 1 in CMR) and/or disable the
counter clock (CPCDIS = 1 in CMR).
Figure 31-8. WAVSEL = 10 Without Trigger
Counter Value
0xFFFF
Counter cleared by compare match with RC
RC
RB
RA
Waveform Examples
Time
TIOB
TIOA
649
32058K
AVR32-01/12
AT32UC3A
Figure 31-9. WAVSEL = 10 With Trigger
Counter Value
0xFFFF
Counter cleared by compare match with RC
Counter cleared by trigger
RC
RB
RA
Waveform Examples
Time
TIOB
TIOA
31.6.3.4
WAVSEL = 01
When WAVSEL = 01, the value of CV is incremented from 0 to 0xFFFF. Once 0xFFFF is
reached, the value of CV is decremented to 0, then re-incremented to 0xFFFF and so on. See
Figure 31-10.
A trigger such as an external event or a software trigger can modify CV at any time. If a trigger
occurs while CV is incrementing, CV then decrements. If a trigger is received while CV is decrementing, CV then increments. See Figure 31-11.
RC Compare cannot be programmed to generate a trigger in this configuration.
At the same time, RC Compare can stop the counter clock (CPCSTOP = 1) and/or disable the
counter clock (CPCDIS = 1).
650
32058K
AVR32-01/12
AT32UC3A
Figure 31-10. WAVSEL = 01 Without Trigger
Counter decremented by compare match with 0xFFFF
Counter Value
0xFFFF
RC
RB
RA
Time
Waveform Examples
TIOB
TIOA
Figure 31-11. WAVSEL = 01 With Trigger
Counter Value
Counter decremented by compare match with 0xFFFF
0xFFFF
Counter decremented
by trigger
RC
RB
Counter incremented
by trigger
RA
Time
Waveform Examples
TIOB
TIOA
31.6.3.5
WAVSEL = 11
When WAVSEL = 11, the value of CV is incremented from 0 to RC. Once RC is reached, the
value of CV is decremented to 0, then re-incremented to RC and so on. See Figure 31-12.
A trigger such as an external event or a software trigger can modify CV at any time. If a trigger
occurs while CV is incrementing, CV then decrements. If a trigger is received while CV is decrementing, CV then increments. See Figure 31-13.
RC Compare can stop the counter clock (CPCSTOP = 1) and/or disable the counter clock
(CPCDIS = 1).
651
32058K
AVR32-01/12
AT32UC3A
Figure 31-12. WAVSEL = 11 Without Trigger
Counter Value
0xFFFF
Counter decremented by compare match with RC
RC
RB
RA
Time
Waveform Examples
TIOB
TIOA
Figure 31-13. WAVSEL = 11 With Trigger
Counter Value
0xFFFF
Counter decremented by compare match with RC
RC
Counter decremented
by trigger
RB
Counter incremented
by trigger
RA
Waveform Examples
Time
TIOB
TIOA
31.6.3.6
External Event/Trigger Conditions
An external event can be programmed to be detected on one of the clock sources (XC0, XC1,
XC2) or TIOB. The external event selected can then be used as a trigger.
The EEVT parameter in CMR selects the external trigger. The EEVTEDG parameter defines the
trigger edge for each of the possible external triggers (rising, falling or both). If EEVTEDG is
cleared (none), no external event is defined.
652
32058K
AVR32-01/12
AT32UC3A
If TIOB is defined as an external event signal (EEVT = 0), TIOB is no longer used as an output
and the compare register B is not used to generate waveforms and subsequently no IRQs. In
this case the TC channel can only generate a waveform on TIOA.
When an external event is defined, it can be used as a trigger by setting bit ENETRG in CMR.
As in Capture Mode, the SYNC signal and the software trigger are also available as triggers. RC
Compare can also be used as a trigger depending on the parameter WAVSEL.
31.6.3.7
Output Controller
The output controller defines the output level changes on TIOA and TIOB following an event.
TIOB control is used only if TIOB is defined as output (not as an external event).
The following events control TIOA and TIOB: software trigger, external event and RC compare.
RA compare controls TIOA and RB compare controls TIOB. Each of these events can be programmed to set, clear or toggle the output as defined in the corresponding parameter in CMR.
653
32058K
AVR32-01/12
AT32UC3A
31.7
Timer Counter (TC) User Interface
Table 31-3.
Offset
TC Global Memory Map
Channel/Register
Name
Access
Reset Value
0x00
TC Channel 0
See Table 31-4
0x40
TC Channel 1
See Table 31-4
0x80
TC Channel 2
See Table 31-4
0xC0
TC Block Control Register
BCR
Write-only
–
0xC4
TC Block Mode Register
BMR
Read/Write
0
BCR (Block Control Register) and BMR (Block Mode Register) control the whole TC block. TC
channels are controlled by the registers listed in Table 31-4. The offset of each of the channel
registers in Table 31-4 is in relation to the offset of the corresponding channel as mentioned in
Table 31-4.
Table 31-4.
Offset
TC Channel Memory Map
Register
Name
Access
Reset Value
0x00
Channel Control Register
CCR
Write-only
–
0x04
Channel Mode Register
CMR
Read/Write
0
0x08
Reserved
–
0x0C
Reserved
–
0x10
Counter Value
CV
Read-only
0
0x14
Register A
RA
Read/Write(1)
0
(1)
0
0x18
Register B
RB
0x1C
Register C
RC
Read/Write
0
0x20
Status Register
SR
Read-only
0
0x24
Interrupt Enable Register
IER
Write-only
–
0x28
Interrupt Disable Register
IDR
Write-only
–
0x2C
Interrupt Mask Register
IMR
Read-only
0
Notes:
Read/Write
1. Read-only if WAVE = 0
654
32058K
AVR32-01/12
AT32UC3A
31.7.1
TC Block Control Register
Register Name:
BCR
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
–
5
–
4
–
3
–
2
–
1
–
0
SYNC
• SYNC: Synchro Command
0 = No effect.
1 = Asserts the SYNC signal which generates a software trigger simultaneously for each of the channels.
655
32058K
AVR32-01/12
AT32UC3A
31.7.2
TC Block Mode Register
Register Name:
BMR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
–
5
4
3
2
1
TC2XC2S
TC1XC1S
0
TC0XC0S
• TC0XC0S: External Clock Signal 0 Selection
TC0XC0S
Signal Connected to XC0
0
0
TCLK0
0
1
none
1
0
TIOA1
1
1
TIOA2
• TC1XC1S: External Clock Signal 1 Selection
TC1XC1S
Signal Connected to XC1
0
0
TCLK1
0
1
none
1
0
TIOA0
1
1
TIOA2
• TC2XC2S: External Clock Signal 2 Selection
TC2XC2S
Signal Connected to XC2
0
0
TCLK2
0
1
none
1
0
TIOA0
1
1
TIOA1
656
32058K
AVR32-01/12
AT32UC3A
31.7.3
TC Channel Control Register
Register Name:
CCR
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
–
5
–
4
–
3
–
2
SWTRG
1
CLKDIS
0
CLKEN
• CLKEN: Counter Clock Enable Command
0 = No effect.
1 = Enables the clock if CLKDIS is not 1.
• CLKDIS: Counter Clock Disable Command
0 = No effect.
1 = Disables the clock.
• SWTRG: Software Trigger Command
0 = No effect.
1 = A software trigger is performed: the counter is reset and the clock is started.
657
32058K
AVR32-01/12
AT32UC3A
31.7.4
TC Channel Mode Register: Capture Mode
Register Name:
CMR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
23
–
22
–
21
–
20
–
19
15
WAVE = 0
14
CPCTRG
13
–
12
–
11
–
10
ABETRG
9
7
LDBDIS
6
LDBSTOP
5
4
3
CLKI
2
1
TCCLKS
BURST
26
–
25
–
18
17
24
–
16
LDRB
LDRA
8
ETRGEDG
0
• TCCLKS: Clock Selection
TCCLKS
Clock Selected
0
0
0
TIMER_CLOCK1
0
0
1
TIMER_CLOCK2
0
1
0
TIMER_CLOCK3
0
1
1
TIMER_CLOCK4
1
0
0
TIMER_CLOCK5
1
0
1
XC0
1
1
0
XC1
1
1
1
XC2
• CLKI: Clock Invert
0 = Counter is incremented on rising edge of the clock.
1 = Counter is incremented on falling edge of the clock.
• BURST: Burst Signal Selection
BURST
0
0
The clock is not gated by an external signal.
0
1
XC0 is ANDed with the selected clock.
1
0
XC1 is ANDed with the selected clock.
1
1
XC2 is ANDed with the selected clock.
• LDBSTOP: Counter Clock Stopped with RB Loading
0 = Counter clock is not stopped when RB loading occurs.
1 = Counter clock is stopped when RB loading occurs.
• LDBDIS: Counter Clock Disable with RB Loading
658
32058K
AVR32-01/12
AT32UC3A
0 = Counter clock is not disabled when RB loading occurs.
1 = Counter clock is disabled when RB loading occurs.
• ETRGEDG: External Trigger Edge Selection
ETRGEDG
Edge
0
0
none
0
1
rising edge
1
0
falling edge
1
1
each edge
• ABETRG: TIOA or TIOB External Trigger Selection
0 = TIOB is used as an external trigger.
1 = TIOA is used as an external trigger.
• CPCTRG: RC Compare Trigger Enable
0 = RC Compare has no effect on the counter and its clock.
1 = RC Compare resets the counter and starts the counter clock.
• WAVE
0 = Capture Mode is enabled.
1 = Capture Mode is disabled (Waveform Mode is enabled).
• LDRA: RA Loading Selection
LDRA
Edge
0
0
none
0
1
rising edge of TIOA
1
0
falling edge of TIOA
1
1
each edge of TIOA
• LDRB: RB Loading Selection
LDRB
Edge
0
0
none
0
1
rising edge of TIOA
1
0
falling edge of TIOA
1
1
each edge of TIOA
659
32058K
AVR32-01/12
AT32UC3A
31.7.5
TC Channel Mode Register: Waveform Mode
Register Name:
CMR
Access Type:
Read/Write
31
30
29
28
BSWTRG
23
27
BEEVT
22
21
20
ASWTRG
19
AEEVT
15
WAVE = 1
14
13
7
CPCDIS
6
CPCSTOP
WAVSEL
5
26
25
24
BCPC
BCPB
18
17
16
ACPC
12
ENETRG
11
4
3
CLKI
BURST
ACPA
10
EEVT
9
8
EEVTEDG
2
1
TCCLKS
0
• TCCLKS: Clock Selection
TCCLKS
Clock Selected
0
0
0
TIMER_CLOCK1
0
0
1
TIMER_CLOCK2
0
1
0
TIMER_CLOCK3
0
1
1
TIMER_CLOCK4
1
0
0
TIMER_CLOCK5
1
0
1
XC0
1
1
0
XC1
1
1
1
XC2
• CLKI: Clock Invert
0 = Counter is incremented on rising edge of the clock.
1 = Counter is incremented on falling edge of the clock.
• BURST: Burst Signal Selection
BURST
0
0
The clock is not gated by an external signal.
0
1
XC0 is ANDed with the selected clock.
1
0
XC1 is ANDed with the selected clock.
1
1
XC2 is ANDed with the selected clock.
• CPCSTOP: Counter Clock Stopped with RC Compare
0 = Counter clock is not stopped when counter reaches RC.
1 = Counter clock is stopped when counter reaches RC.
• CPCDIS: Counter Clock Disable with RC Compare
660
32058K
AVR32-01/12
AT32UC3A
0 = Counter clock is not disabled when counter reaches RC.
1 = Counter clock is disabled when counter reaches RC.
• EEVTEDG: External Event Edge Selection
EEVTEDG
Edge
0
0
none
0
1
rising edge
1
0
falling edge
1
1
each edge
• EEVT: External Event Selection
EEVT
Note:
Signal selected as external event
TIOB Direction
0
0
TIOB
input(1)
0
1
XC0
output
1
0
XC1
output
1
1
XC2
output
1. If TIOB is chosen as the external event signal, it is configured as an input and no longer generates waveforms and subsequently no IRQs.
• ENETRG: External Event Trigger Enable
0 = The external event has no effect on the counter and its clock. In this case, the selected external event only controls the
TIOA output.
1 = The external event resets the counter and starts the counter clock.
• WAVSEL: Waveform Selection
WAVSEL
Effect
0
0
UP mode without automatic trigger on RC Compare
1
0
UP mode with automatic trigger on RC Compare
0
1
UPDOWN mode without automatic trigger on RC Compare
1
1
UPDOWN mode with automatic trigger on RC Compare
• WAVE = 1
0 = Waveform Mode is disabled (Capture Mode is enabled).
1 = Waveform Mode is enabled.
661
32058K
AVR32-01/12
AT32UC3A
• ACPA: RA Compare Effect on TIOA
ACPA
Effect
0
0
none
0
1
set
1
0
clear
1
1
toggle
• ACPC: RC Compare Effect on TIOA
ACPC
Effect
0
0
none
0
1
set
1
0
clear
1
1
toggle
• AEEVT: External Event Effect on TIOA
AEEVT
Effect
0
0
none
0
1
set
1
0
clear
1
1
toggle
• ASWTRG: Software Trigger Effect on TIOA
ASWTRG
Effect
0
0
none
0
1
set
1
0
clear
1
1
toggle
• BCPB: RB Compare Effect on TIOB
BCPB
0
Effect
0
none
662
32058K
AVR32-01/12
AT32UC3A
BCPB
Effect
0
1
set
1
0
clear
1
1
toggle
• BCPC: RC Compare Effect on TIOB
BCPC
Effect
0
0
none
0
1
set
1
0
clear
1
1
toggle
• BEEVT: External Event Effect on TIOB
BEEVT
Effect
0
0
none
0
1
set
1
0
clear
1
1
toggle
• BSWTRG: Software Trigger Effect on TIOB
BSWTRG
Effect
0
0
none
0
1
set
1
0
clear
1
1
toggle
663
32058K
AVR32-01/12
AT32UC3A
31.7.6
TC Counter Value Register
Register Name:
CV
Access Type:
Read-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
CV
7
6
5
4
CV
• CV: Counter Value
CV contains the counter value in real time.
664
32058K
AVR32-01/12
AT32UC3A
31.7.7
TC Register A
Register Name:
RA
Access Type:
Read-only if WAVE = 0, Read/Write if WAVE = 1
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
RA
7
6
5
4
RA
• RA: Register A
RA contains the Register A value in real time.
665
32058K
AVR32-01/12
AT32UC3A
31.7.8
TC Register B
Register Name:
RB
Access Type:
Read-only if WAVE = 0, Read/Write if WAVE = 1
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
RB
7
6
5
4
RB
• RB: Register B
RB contains the Register B value in real time.
666
32058K
AVR32-01/12
AT32UC3A
31.7.9
TC Register C
Register Name:
RC
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
14
13
12
11
10
9
8
3
2
1
0
RC
7
6
5
4
RC
• RC: Register C
RC contains the Register C value in real time.
667
32058K
AVR32-01/12
AT32UC3A
31.7.10
TC Status Register
Register Name:
SR
Access Type:
Read-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
MTIOB
17
MTIOA
16
CLKSTA
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
ETRGS
6
LDRBS
5
LDRAS
4
CPCS
3
CPBS
2
CPAS
1
LOVRS
0
COVFS
Note: Reading the Status Register will also clear the interrupt flag for the corresponding interrupts.
• COVFS: Counter Overflow Status
0 = No counter overflow has occurred since the last read of the Status Register.
1 = A counter overflow has occurred since the last read of the Status Register.
• LOVRS: Load Overrun Status
0 = Load overrun has not occurred since the last read of the Status Register or WAVE = 1.
1 = RA or RB have been loaded at least twice without any read of the corresponding register since the last read of the Status Register, if WAVE = 0.
• CPAS: RA Compare Status
0 = RA Compare has not occurred since the last read of the Status Register or WAVE = 0.
1 = RA Compare has occurred since the last read of the Status Register, if WAVE = 1.
• CPBS: RB Compare Status
0 = RB Compare has not occurred since the last read of the Status Register or WAVE = 0.
1 = RB Compare has occurred since the last read of the Status Register, if WAVE = 1.
• CPCS: RC Compare Status
0 = RC Compare has not occurred since the last read of the Status Register.
1 = RC Compare has occurred since the last read of the Status Register.
• LDRAS: RA Loading Status
0 = RA Load has not occurred since the last read of the Status Register or WAVE = 1.
1 = RA Load has occurred since the last read of the Status Register, if WAVE = 0.
• LDRBS: RB Loading Status
0 = RB Load has not occurred since the last read of the Status Register or WAVE = 1.
1 = RB Load has occurred since the last read of the Status Register, if WAVE = 0.
• ETRGS: External Trigger Status
0 = External trigger has not occurred since the last read of the Status Register.
1 = External trigger has occurred since the last read of the Status Register.
668
32058K
AVR32-01/12
AT32UC3A
• CLKSTA: Clock Enabling Status
0 = Clock is disabled.
1 = Clock is enabled.
• MTIOA: TIOA Mirror
0 = TIOA is low. If WAVE = 0, this means that TIOA pin is low. If WAVE = 1, this means that TIOA is driven low.
1 = TIOA is high. If WAVE = 0, this means that TIOA pin is high. If WAVE = 1, this means that TIOA is driven high.
• MTIOB: TIOB Mirror
0 = TIOB is low. If WAVE = 0, this means that TIOB pin is low. If WAVE = 1, this means that TIOB is driven low.
1 = TIOB is high. If WAVE = 0, this means that TIOB pin is high. If WAVE = 1, this means that TIOB is driven high.
669
32058K
AVR32-01/12
AT32UC3A
31.7.11
TC Interrupt Enable Register
Register Name:
IER
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
ETRGS
6
LDRBS
5
LDRAS
4
CPCS
3
CPBS
2
CPAS
1
LOVRS
0
COVFS
Note: Reading the Status Register will also clear the interrupt flag for the corresponding interrupts.
• COVFS: Counter Overflow
0 = No effect.
1 = Enables the Counter Overflow Interrupt.
• LOVRS: Load Overrun
0 = No effect.
1 = Enables the Load Overrun Interrupt.
• CPAS: RA Compare
0 = No effect.
1 = Enables the RA Compare Interrupt.
• CPBS: RB Compare
0 = No effect.
1 = Enables the RB Compare Interrupt.
• CPCS: RC Compare
0 = No effect.
1 = Enables the RC Compare Interrupt.
• LDRAS: RA Loading
0 = No effect.
1 = Enables the RA Load Interrupt.
• LDRBS: RB Loading
0 = No effect.
1 = Enables the RB Load Interrupt.
• ETRGS: External Trigger
0 = No effect.
1 = Enables the External Trigger Interrupt.
670
32058K
AVR32-01/12
AT32UC3A
31.7.12
TC Interrupt Disable Register
Register Name:
IDR
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
ETRGS
6
LDRBS
5
LDRAS
4
CPCS
3
CPBS
2
CPAS
1
LOVRS
0
COVFS
Note: Reading the Status Register will also clear the interrupt flag for the corresponding interrupts.
• COVFS: Counter Overflow
0 = No effect.
1 = Disables the Counter Overflow Interrupt.
• LOVRS: Load Overrun
0 = No effect.
1 = Disables the Load Overrun Interrupt (if WAVE = 0).
• CPAS: RA Compare
0 = No effect.
1 = Disables the RA Compare Interrupt (if WAVE = 1).
• CPBS: RB Compare
0 = No effect.
1 = Disables the RB Compare Interrupt (if WAVE = 1).
• CPCS: RC Compare
0 = No effect.
1 = Disables the RC Compare Interrupt.
• LDRAS: RA Loading
0 = No effect.
1 = Disables the RA Load Interrupt (if WAVE = 0).
• LDRBS: RB Loading
0 = No effect.
1 = Disables the RB Load Interrupt (if WAVE = 0).
• ETRGS: External Trigger
0 = No effect.
1 = Disables the External Trigger Interrupt.
671
32058K
AVR32-01/12
AT32UC3A
31.7.13
TC Interrupt Mask Register
Register Name:
IMR
Access Type:
Read-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
ETRGS
6
LDRBS
5
LDRAS
4
CPCS
3
CPBS
2
CPAS
1
LOVRS
0
COVFS
Note: Reading the Status Register will also clear the interrupt flag for the corresponding interrupts.
• COVFS: Counter Overflow
0 = The Counter Overflow Interrupt is disabled.
1 = The Counter Overflow Interrupt is enabled.
• LOVRS: Load Overrun
0 = The Load Overrun Interrupt is disabled.
1 = The Load Overrun Interrupt is enabled.
• CPAS: RA Compare
0 = The RA Compare Interrupt is disabled.
1 = The RA Compare Interrupt is enabled.
• CPBS: RB Compare
0 = The RB Compare Interrupt is disabled.
1 = The RB Compare Interrupt is enabled.
• CPCS: RC Compare
0 = The RC Compare Interrupt is disabled.
1 = The RC Compare Interrupt is enabled.
• LDRAS: RA Loading
0 = The Load RA Interrupt is disabled.
1 = The Load RA Interrupt is enabled.
• LDRBS: RB Loading
0 = The Load RB Interrupt is disabled.
1 = The Load RB Interrupt is enabled.
• ETRGS: External Trigger
0 = The External Trigger Interrupt is disabled.
1 = The External Trigger Interrupt is enabled.
672
32058K
AVR32-01/12
AT32UC3A
32. Pulse Width Modulation Controller (PWM)
Rev: 1.3.0.1
32.1
Features
• 7 Channels
• One 20-bit Counter Per Channel
• Common Clock Generator Providing Thirteen Different Clocks
– A Modulo n Counter Providing Eleven Clocks
– Two Independent Linear Dividers Working on Modulo n Counter Outputs
• Independent Channels
– Independent Enable Disable Command for Each Channel
– Independent Clock Selection for Each Channel
– Independent Period and Duty Cycle for Each Channel
– Double Buffering of Period or Duty Cycle for Each Channel
– Programmable Selection of The Output Waveform Polarity for Each Channel
– Programmable Center or Left Aligned Output Waveform for Each Channel
32.2
Description
The PWM macrocell controls several channels independently. Each channel controls one
square output waveform. Characteristics of the output waveform such as period, duty-cycle and
polarity are configurable through the user interface. Each channel selects and uses one of the
clocks provided by the clock generator. The clock generator provides several clocks resulting
from the division of the PWM macrocell master clock.
All PWM macrocell accesses are made through registers mapped on the peripheral bus.
Channels can be synchronized, to generate non overlapped waveforms. All channels integrate a
double buffering system in order to prevent an unexpected output waveform while modifying the
period or the duty-cycle.
673
32058K
AVR32-01/12
AT32UC3A
32.3
Block Diagram
Figure 32-1. Pulse Width Modulation Controller Block Diagram
PWM
Controller
PWMx
Channel
Period
PWMx
Update
Clock
Selector
Duty Cycle
Comparator
PWMx
Counter
PIO
PWM0
Channel
Period
PWM0
Update
Clock
Selector
Power
Manager
MCK
Clock Generator
Duty Cycle
Comparator
PWM0
Counter
PB Interface
Interrupt Generator
Interrupt
Controller
Peripheral
Bus
32.4
I/O Lines Description
Each channel outputs one waveform on one external I/O line.
Table 32-1.
I/O Line Description
Name
Description
Type
PWMx
PWM Waveform Output for channel x
Output
674
32058K
AVR32-01/12
AT32UC3A
32.5
Product Dependencies
32.5.1
I/O Lines
The pins used for interfacing the PWM may be multiplexed with PIO lines. The programmer must
first program the PIO controller to assign the desired PWM pins to their peripheral function. If I/O
lines of the PWM are not used by the application, they can be used for other purposes by the
PIO controller.
Not all PWM outputs may be enabled. If an application requires only four channels, then only
four PIO lines will be assigned to PWM outputs.
32.5.2
Debug operation
The PWM clock is running during debug operation.
32.5.3
Power Management
The PWM clock is generated by the Power Manager. Before using the PWM, the programmer
must ensure that the PWM clock is enabled in the Power Manager. However, if the application
does not require PWM operations, the PWM clock can be stopped when not needed and be
restarted later. In this case, the PWM will resume its operations where it left off.
In the PWM description, Master Clock (MCK) is the clock of the peripheral bus to which the
PWM is connected.
32.5.4
Interrupt Sources
The PWM interrupt line is connected to the interrupt controller. Using the PWM interrupt requires
the interrupt controller to be programmed first.
675
32058K
AVR32-01/12
AT32UC3A
32.6
Functional Description
The PWM macrocell is primarily composed of a clock generator module and 7 channels.
– Clocked by the system clock, MCK, the clock generator module provides 13 clocks.
– Each channel can independently choose one of the clock generator outputs.
– Each channel generates an output waveform with attributes that can be defined
independently for each channel through the user interface registers.
32.6.1
PWM Clock Generator
Figure 32-2. Functional View of the Clock Generator Block Diagram
MCK
modulo n counter
MCK
MCK/2
MCK/4
MCK/8
MCK/16
MCK/32
MCK/64
MCK/128
MCK/256
MCK/512
MCK/1024
Divider A
PREA
clkA
DIVA
PWM_MR
Divider B
PREB
clkB
DIVB
PWM_MR
Caution: Before using the PWM macrocell, the programmer must ensure that the PWM clock in
the Power Manager is enabled.
The PWM macrocell master clock, MCK, is divided in the clock generator module to provide different clocks available for all channels. Each channel can independently select one of the
divided clocks.
676
32058K
AVR32-01/12
AT32UC3A
The clock generator is divided in three blocks:
– a modulo n counter which provides 11 clocks: FMCK, FMCK/2, FMCK/4, FMCK/8, FMCK/16,
FMCK/32, FMCK/64, FMCK/128, FMCK/256, FMCK/512, FMCK/1024
– two linear dividers (1, 1/2, 1/3, ... 1/255) that provide two separate clocks: clkA and
clkB
Each linear divider can independently divide one of the clocks of the modulo n counter. The
selection of the clock to be divided is made according to the PREA (PREB) field of the PWM
Mode register (MR). The resulting clock clkA (clkB) is the clock selected divided by DIVA (DIVB)
field value in the PWM Mode register (MR).
After a reset of the PWM controller, DIVA (DIVB) and PREA (PREB) in the PWM Mode register
are set to 0. This implies that after reset clkA (clkB) are turned off.
At reset, all clocks provided by the modulo n counter are turned off except clock “clk”. This situation is also true when the PWM master clock is turned off through the Power Management
Controller.
32.6.2
32.6.2.1
PWM Channel
Block Diagram
Figure 32-3. Functional View of the Channel Block Diagram
inputs
from clock
generator
Channel
Clock
Selector
Internal
Counter
Comparator
PWMx output waveform
inputs from
Peripheral
Bus
Each of the 7 channels is composed of three blocks:
• A clock selector which selects one of the clocks provided by the clock generator described in
Section 32.6.1 ”PWM Clock Generator” on page 676.
• An internal counter clocked by the output of the clock selector. This internal counter is
incremented or decremented according to the channel configuration and comparators events.
The size of the internal counter is 20 bits.
• A comparator used to generate events according to the internal counter value. It also computes
the PWMx output waveform according to the configuration.
32.6.2.2
Waveform Properties
The different properties of output waveforms are:
• the internal clock selection. The internal channel counter is clocked by one of the clocks
provided by the clock generator described in the previous section. This channel parameter is
defined in the CPRE field of the CMRx register. This field is reset at 0.
• the waveform period. This channel parameter is defined in the CPRD field of the CPRDx
register.
677
32058K
AVR32-01/12
AT32UC3A
- If the waveform is left aligned, then the output waveform period depends on the counter
source clock and can be calculated:
By using the Master Clock (MCK) divided by an X given prescaler value
(with X being 1, 2, 4, 8, 16, 32, 64, 128, 256, 512, or 1024), the resulting period formula will be:
(-----------------------------X × CPRD )MCK
By using a Master Clock divided by one of both DIVA or DIVB divider, the formula becomes,
respectively:
(----------------------------------------CRPD × DIVA )( CRPD × DIVAB )
or ---------------------------------------------MCK
MCK
If the waveform is center aligned then the output waveform period depends on the counter
source clock and can be calculated:
By using the Master Clock (MCK) divided by an X given prescaler value
(with X being 1, 2, 4, 8, 16, 32, 64, 128, 256, 512, or 1024). The resulting period formula will
be:
(---------------------------------------2 × X × CPRD-)
MCK
By using a Master Clock divided by one of both DIVA or DIVB divider, the formula becomes,
respectively:
(--------------------------------------------------2 × CPRD × DIVA-)
( 2 × CPRD × DIVB )
or ---------------------------------------------------MCK
MCK
• the waveform duty cycle. This channel parameter is defined in the CDTY field of the CDTYx
register.
If the waveform is left aligned then:
period – 1 ⁄ fchannel_x_clock × CDTY -)
duty cycle = (------------------------------------------------------------------------------------------------------period
If the waveform is center aligned, then:
( ( period ⁄ 2 ) – 1 ⁄ fchannel_x_clock × CDTY ) -)
duty cycle = ---------------------------------------------------------------------------------------------------------------------( period ⁄ 2 )
• the waveform polarity. At the beginning of the period, the signal can be at high or low level.
This property is defined in the CPOL field of the CMRx register. By default the signal starts by
a low level.
• the waveform alignment. The output waveform can be left or center aligned. Center aligned
waveforms can be used to generate non overlapped waveforms. This property is defined in the
CALG field of the CMRx register. The default mode is left aligned.
678
32058K
AVR32-01/12
AT32UC3A
Figure 32-4. Non Overlapped Center Aligned Waveforms
No overlap
PWM0
PWM1
Period
Note:
1. See Figure 32-5 on page 680 for a detailed description of center aligned waveforms.
When center aligned, the internal channel counter increases up to CPRD and.decreases down
to 0. This ends the period.
When left aligned, the internal channel counter increases up to CPRD and is reset. This ends
the period.
Thus, for the same CPRD value, the period for a center aligned channel is twice the period for a
left aligned channel.
Waveforms are fixed at 0 when:
• CDTY = CPRD and CPOL = 0
• CDTY = 0 and CPOL = 1
Waveforms are fixed at 1 (once the channel is enabled) when:
• CDTY = 0 and CPOL = 0
• CDTY = CPRD and CPOL = 1
The waveform polarity must be set before enabling the channel. This immediately affects the
channel output level. Changes on channel polarity are not taken into account while the channel
is enabled.
679
32058K
AVR32-01/12
AT32UC3A
Figure 32-5. Waveform Properties
PWM_MCKx
CHIDx(PWM_SR)
CHIDx(PWM_ENA)
CHIDx(PWM_DIS)
Center Aligned
CALG(PWM_CMRx) = 1
PWM_CCNTx
CPRD(PWM_CPRDx)
CDTY(PWM_CDTYx)
Period
Output Waveform PWMx
CPOL(PWM_CMRx) = 0
Output Waveform PWMx
CPOL(PWM_CMRx) = 1
CHIDx(PWM_ISR)
Left Aligned
CALG(PWM_CMRx) = 0
PWM_CCNTx
CPRD(PWM_CPRDx)
CDTY(PWM_CDTYx)
Period
Output Waveform PWMx
CPOL(PWM_CMRx) = 0
Output Waveform PWMx
CPOL(PWM_CMRx) = 1
CHIDx(PWM_ISR)
680
32058K
AVR32-01/12
AT32UC3A
32.6.3
32.6.3.1
PWM Controller Operations
Initialization
Before enabling the output channel, this channel must have been configured by the software
application:
• Configuration of the clock generator if DIVA and DIVB are required
• Selection of the clock for each channel (CPRE field in the CMRx register)
• Configuration of the waveform alignment for each channel (CALG field in the CMRx register)
• Configuration of the period for each channel (CPRD in the CPRDx register). Writing in CPRDx
Register is possible while the channel is disabled. After validation of the channel, the user must
use CUPDx Register to update CPRDx as explained below.
• Configuration of the duty cycle for each channel (CDTY in the CDTYx register). Writing in
CDTYx Register is possible while the channel is disabled. After validation of the channel, the
user must use CUPDx Register to update CDTYx as explained below.
• Configuration of the output waveform polarity for each channel (CPOL in the CMRx register)
• Enable Interrupts (Writing CHIDx in the IER register)
• Enable the PWM channel (Writing CHIDx in the ENA register)
It is possible to synchronize different channels by enabling them at the same time by means of
writing simultaneously several CHIDx bits in the ENA register.
In such a situation, all channels may have the same clock selector configuration and the same
period specified.
32.6.3.2
Source Clock Selection Criteria
The large number of source clocks can make selection difficult. The relationship between the
value in the Period Register (CPRDx) and the Duty Cycle Register (CDTYx) can help the user in
choosing. The event number written in the Period Register gives the PWM accuracy. The Duty
Cycle quantum cannot be lower than 1/CPRDx value. The higher the value of CPRDx, the
greater the PWM accuracy.
For example, if the user sets 15 (in decimal) in CPRDx, the user is able to set a value between 1
up to 14 in CDTYx Register. The resulting duty cycle quantum cannot be lower than 1/15 of the
PWM period.
32.6.3.3
Changing the Duty Cycle or the Period
It is possible to modulate the output waveform duty cycle or period.
To prevent unexpected output waveform, the user must use the update register (PWM_CUPDx)
to change waveform parameters while the channel is still enabled. The user can write a new
period value or duty cycle value in the update register (CUPDx). This register holds the new
value until the end of the current cycle and updates the value for the next cycle. Depending on
the CPD field in the CMRx register, CUPDx either updates CPRDx or CDTYx. Note that even if
the update register is used, the period must not be smaller than the duty cycle.
681
32058K
AVR32-01/12
AT32UC3A
Figure 32-6. Synchronized Period or Duty Cycle Update
User's Writing
PWM_CUPDx Value
0
1
PWM_CPRDx
PWM_CMRx. CPD
PWM_CDTYx
End of Cycle
To prevent overwriting the CUPDx by software, the user can use status events in order to synchronize his software. Two methods are possible. In both, the user must enable the dedicated
interrupt in IER at PWM Controller level.
The first method (polling method) consists of reading the relevant status bit in ISR Register
according to the enabled channel(s). See Figure 32-7.
The second method uses an Interrupt Service Routine associated with the PWM channel.
Note:
Reading the ISR register automatically clears CHIDx flags.
Figure 32-7. Polling Method
PWM_ISR Read
Acknowledgement and clear previous register state
Writing in CPD field
Update of the Period or Duty Cycle
CHIDx = 1
YES
Writing in PWM_CUPDx
The last write has been taken into account
Note:
Polarity and alignment can be modified only when the channel is disabled.
682
32058K
AVR32-01/12
AT32UC3A
32.6.3.4
Interrupts
Depending on the interrupt mask in the IMR register, an interrupt is generated at the end of the
corresponding channel period. The interrupt remains active until a read operation in the ISR register occurs.
A channel interrupt is enabled by setting the corresponding bit in the IER register. A channel
interrupt is disabled by setting the corresponding bit in the IDR register.
683
32058K
AVR32-01/12
AT32UC3A
32.7
Pulse Width Modulation (PWM) Controller User Interface
32.7.1
Register Mapping
Table 32-2.
PWM Controller Registers
Access
Peripheral
Reset Value
MR
Read/Write
0
PWM Enable Register
ENA
Write-only
-
0x08
PWM Disable Register
DIS
Write-only
-
0x0C
PWM Status Register
SR
Read-only
0
0x10
PWM Interrupt Enable Register
IER
Write-only
-
0x14
PWM Interrupt Disable Register
IDR
Write-only
-
0x18
PWM Interrupt Mask Register
IMR
Read-only
0
0x1C
PWM Interrupt Status Register
ISR
Read-only
0
0x4C - 0xF8
Reserved
–
–
–
0x4C - 0xFC
Reserved
–
–
–
0x100 - 0x1FC
Reserved
0x200
Channel 0 Mode Register
CMR0
Read/Write
0x0
0x204
Channel 0 Duty Cycle Register
CDTY0
Read/Write
0x0
0x208
Channel 0 Period Register
CPRD0
Read/Write
0x0
0x20C
Channel 0 Counter Register
CCNT0
Read-only
0x0
0x210
Channel 0 Update Register
CUPD0
Write-only
-
...
Reserved
0x220
Channel 1 Mode Register
CMR1
Read/Write
0x0
0x224
Channel 1 Duty Cycle Register
CDTY1
Read/Write
0x0
0x228
Channel 1 Period Register
CPRD1
Read/Write
0x0
0x22C
Channel 1 Counter Register
CCNT1
Read-only
0x0
0x230
Channel 1 Update Register
CUPD1
Write-only
-
...
...
...
...
...
Offset
Register
Name
0x00
PWM Mode Register
0x04
684
32058K
AVR32-01/12
AT32UC3A
32.7.2
PWM Mode Register
Register Name:
MR
Access Type:
Read/Write
31
–
30
–
29
–
28
–
23
22
21
20
27
26
25
24
PREB
19
18
17
16
11
10
9
8
1
0
DIVB
15
–
14
–
13
–
12
–
7
6
5
4
PREA
3
2
DIVA
• DIVA, DIVB: CLKA, CLKB Divide Factor
DIVA, DIVB
CLKA, CLKB
0
CLKA, CLKB clock is turned off
1
CLKA, CLKB clock is clock selected by PREA, PREB
2-255
CLKA, CLKB clock is clock selected by PREA, PREB divided by DIVA, DIVB factor.
• PREA, PREB
PREA, PREB
Divider Input Clock
0
0
0
0
MCK.
0
0
0
1
MCK/2
0
0
1
0
MCK/4
0
0
1
1
MCK/8
0
1
0
0
MCK/16
0
1
0
1
MCK/32
0
1
1
0
MCK/64
0
1
1
1
MCK/128
1
0
0
0
MCK/256
1
0
0
1
MCK/512
1
0
1
0
MCK/1024
Other
Reserved
685
32058K
AVR32-01/12
AT32UC3A
32.7.3
PWM Enable Register
Register Name:
ENA
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
CHID6
5
CHID5
4
CHID4
3
CHID3
2
CHID2
1
CHID1
0
CHID0
• CHIDx: Channel ID
0 = No effect.
1 = Enable PWM output for channel x.
686
32058K
AVR32-01/12
AT32UC3A
32.7.4
PWM Disable Register
Register Name:
DIS
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
CHID6
5
CHID5
4
CHID4
3
CHID3
2
CHID2
1
CHID1
0
CHID0
• CHIDx: Channel ID
0 = No effect.
1 = Disable PWM output for channel x.
687
32058K
AVR32-01/12
AT32UC3A
32.7.5
PWM Status Register
Register Name:
SR
Access Type:
Read-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
CHID6
5
CHID5
4
CHID4
3
CHID3
2
CHID2
1
CHID1
0
CHID0
• CHIDx: Channel ID
0 = PWM output for channel x is disabled.
1 = PWM output for channel x is enabled.
688
32058K
AVR32-01/12
AT32UC3A
32.7.6
PWM Interrupt Enable Register
Register Name:
IER
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
CHID6
5
CHID5
4
CHID4
3
CHID3
2
CHID2
1
CHID1
0
CHID0
• CHIDx: Channel ID.
0 = No effect.
1 = Enable interrupt for PWM channel x.
689
32058K
AVR32-01/12
AT32UC3A
32.7.7
PWM Interrupt Disable Register
Register Name:
IDR
Access Type:
Write-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
CHID6
5
CHID5
4
CHID4
3
CHID3
2
CHID2
1
CHID1
0
CHID0
• CHIDx: Channel ID.
0 = No effect.
1 = Disable interrupt for PWM channel x.
690
32058K
AVR32-01/12
AT32UC3A
32.7.8
PWM Interrupt Mask Register
Register Name:
IMR
Access Type:
Read-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
CHID6
5
CHID5
4
CHID4
3
CHID3
2
CHID2
1
CHID1
0
CHID0
• CHIDx: Channel ID.
0 = Interrupt for PWM channel x is disabled.
1 = Interrupt for PWM channel x is enabled.
691
32058K
AVR32-01/12
AT32UC3A
32.7.9
PWM Interrupt Status Register
Register Name:
ISR
Access Type:
Read-only
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
–
9
–
8
–
7
–
6
CHID6
5
CHID5
4
CHID4
3
CHID3
2
CHID2
1
CHID1
0
CHID0
• CHIDx: Channel ID
0 = No new channel period since the last read of the ISR register.
1 = At least one new channel period since the last read of the ISR register.
Note: Reading ISR automatically clears CHIDx flags.
692
32058K
AVR32-01/12
AT32UC3A
32.7.10
PWM Channel Mode Register
Register Name:
CMRx
Access Type:
Read/Write
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
–
15
–
14
–
13
–
12
–
11
–
10
CPD
9
CPOL
8
CALG
7
–
6
–
5
–
4
–
3
2
1
0
CPRE
• CPRE: Channel Pre-scaler
CPRE
Channel Pre-scaler
0
0
0
0
MCK
0
0
0
1
MCK/2
0
0
1
0
MCK/4
0
0
1
1
MCK/8
0
1
0
0
MCK/16
0
1
0
1
MCK/32
0
1
1
0
MCK/64
0
1
1
1
MCK/128
1
0
0
0
MCK/256
1
0
0
1
MCK/512
1
0
1
0
MCK/1024
1
0
1
1
CLKA
1
1
0
0
CLKB
Other
Reserved
• CALG: Channel Alignment
0 = The period is left aligned.
1 = The period is center aligned.
• CPOL: Channel Polarity
0 = The output waveform starts at a low level.
1 = The output waveform starts at a high level.
693
32058K
AVR32-01/12
AT32UC3A
• CPD: Channel Update Period
0 = Writing to the CUPDx will modify the duty cycle at the next period start event.
1 = Writing to the CUPDx will modify the period at the next period start event.
694
32058K
AVR32-01/12
AT32UC3A
32.7.11
PWM Channel Duty Cycle Register
Register Name:
CDTYx
Access Type:
Read/Write
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
CDTY
23
22
21
20
CDTY
15
14
13
12
CDTY
7
6
5
4
CDTY
Only the first 20 bits (internal channel counter size) are significant.
• CDTY: Channel Duty Cycle
Defines the waveform duty cycle. This value must be defined between 0 and CPRD (CPRx).
695
32058K
AVR32-01/12
AT32UC3A
32.7.12
PWM Channel Period Register
Register Name:
CPRDx
Access Type:
Read/Write
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
CPRD
23
22
21
20
CPRD
15
14
13
12
CPRD
7
6
5
4
CPRD
Only the first 20 bits (internal channel counter size) are significant.
• CPRD: Channel Period
If the waveform is left-aligned, then the output waveform period depends on the counter source clock and can be
calculated:
– By using the Master Clock (MCK) divided by an X given prescaler value (with X being
1, 2, 4, 8, 16, 32, 64, 128, 256, 512, or 1024). The resulting period formula will be:
(-----------------------------X × CPRD )MCK
– By using a Master Clock divided by one of both DIVA or DIVB divider, the formula
becomes, respectively:
(----------------------------------------CRPD × DIVA )( CRPD × DIVAB )
or ---------------------------------------------MCK
MCK
If the waveform is center-aligned, then the output waveform period depends on the counter source clock and can be
calculated:
– By using the Master Clock (MCK) divided by an X given prescaler value (with X being
1, 2, 4, 8, 16, 32, 64, 128, 256, 512, or 1024). The resulting period formula will be:
(---------------------------------------2 × X × CPRD-)
MCK
– By using a Master Clock divided by one of both DIVA or DIVB divider, the formula
becomes, respectively:
(--------------------------------------------------2 × CPRD × DIVA-)
( 2 × CPRD × DIVB )
or ---------------------------------------------------MCK
MCK
696
32058K
AVR32-01/12
AT32UC3A
32.7.13
PWM Channel Counter Register
Register Name:
CCNTx
Access Type:
Read-only
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
CNT
23
22
21
20
CNT
15
14
13
12
CNT
7
6
5
4
CNT
• CNT: Channel Counter Register
Internal counter value. This register is reset when:
• the channel is enabled (writing CHIDx in the ENA register).
• the counter reaches CPRD value defined in the CPRDx register if the waveform is left aligned.
697
32058K
AVR32-01/12
AT32UC3A
32.7.14
PWM Channel Update Register
Register Name:
CUPDx
Access Type:
Write-only
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
CUPD
23
22
21
20
CUPD
15
14
13
12
CUPD
7
6
5
4
CUPD
This register acts as a double buffer for the period or the duty cycle. This prevents an unexpected waveform when modifying the waveform period or duty-cycle.
Only the first 20 bits (internal channel counter size) are significant.
CPD (CMRx Register)
0
The duty-cycle (CDTY in the CDTYx register) is updated with the CUPD value at the beginning of
the next period.
1
The period (CPRD in the CPRDx register) is updated with the CUPD value at the beginning of the
next period.
698
32058K
AVR32-01/12
AT32UC3A
33. Analog-to-Digital Converter (ADC)
Rev: 1.0.0.3
33.1
Features
• Integrated Multiplexer Offering Up to Eight Independent Analog Inputs
• Individual Enable and Disable of Each Channel
• Hardware or Software Trigger
– External Trigger Pin
– Timer Counter Outputs (Corresponding TIOA Trigger)
• PDC Support
• Possibility of ADC Timings Configuration
• Sleep Mode and Conversion Sequencer
– Automatic Wakeup on Trigger and Back to Sleep Mode after Conversions of all Enabled
Channels
33.2
Overview
The ADC is based on a Successive Approximation Register (SAR) 10-bit Analog-to-Digital Converter (ADC). It also integrates an ADC_NB_CHANNELS-to-1 analog multiplexer, making
possible the analog-to-digital conversions of ADC_NB_CHANNELS analog lines. The conversions extend from 0V to ADVREF.
The ADC supports an 8-bit or 10-bit resolution mode, and conversion results are reported in a
common register for all channels, as well as in a channel-dedicated register. Software trigger,
external trigger on rising edge of the TRIGGER pin or internal triggers from Timer Counter output(s) are configurable.
The ADC also integrates a Sleep Mode and a conversion sequencer and connects with a PDC
channel. These features reduce both power consumption and processor intervention.
Finally, the user can configure ADC timings, such as Startup Time and Sample & Hold Time.
699
32058K
AVR32-01/12
AT32UC3A
33.3
Block Diagram
Figure 33-1. Analog-to-Digital Converter Block Diagram
Timer
Counter
Channels
ADC
Trigger
Selection
TRIGGER
Control
Logic
ADC Interrupt
INTC
VDDANA
ADVREF
HSB
ADDedicated
Analog
Inputs
PDC
ADAD-
Analog Inputs
Multiplexed
With I/O lines
AD-
PIO
AD-
Successive
Approximation
Register
Analog-to-Digital
Converter
User
Interface
Peripheral Bridge
PB
AD-
GND
33.4
I/O Lines Description
Table 33-1.
ADC Pins Description
Pin Name
Description
VDDANA
Analog power supply
ADVREF
Reference voltage
AD0 - AD[ADC_NB_CHANNELS-1]
Analog input channels
TRIGGER
External trigger
33.5
Product Dependencies
33.5.1
GPIO
The pin TRIGGER may be shared with other peripheral functions through the PIO Controller. In
this case, the PIO Controller should be set accordingly to assign the pin TRIGGER to the ADC
function.
700
32058K
AVR32-01/12
AT32UC3A
33.5.2
Analog Inputs
The analog input pins can be multiplexed with PIO lines. In this case, the assignment of the ADC
input is automatically done as soon as the corresponding channel is enabled by writing the
CHER register . By default, after reset, the PIO line is configured as input with its pull-up enabled
and the ADC input is connected to the GND.
33.5.3
Power Manager
The ADC is automatically clocked after the first conversion in Normal Mode. In Sleep Mode, the
ADC clock is automatically stopped after each conversion. As the logic is small and the ADC cell
can be put into Sleep Mode, the Power Manager(PM) has no effect on the ADC behavior.
33.5.4
Interrupt Controller
The ADC interrupt line is connected on one of the internal sources of the Interrupt Controller.
Using the ADC interrupt requires the INTC to be programmed first.
33.5.5
Timer Triggers
Timer Counters may or may not be used as hardware triggers depending on user requirements.
Thus, some or all of the timer counters may be non-connected.
33.5.6
Conversion Performances
For performance and electrical characteristics of the ADC, see the DC Characteristics section.
701
32058K
AVR32-01/12
AT32UC3A
33.6
Functional Description
33.6.1
Analog-to-digital Conversion
The ADC uses the ADC Clock to perform conversions. Converting a single analog value to a 10bit digital data requires Sample and Hold Clock cycles as defined in the field SHTIM of the MR
register and 10 ADC Clock cycles. The ADC Clock frequency is selected in the PRESCAL field
of the MR register.
The ADC clock range is between CLK_ADC/2, if PRESCAL is 0, and CLK_ADC/128, if PRESCAL is set to 63 (0x3F). PRESCAL must be programmed in order to provide an ADC clock
frequency according to the parameters given in the Product definition section.
33.6.2
Conversion Reference
The conversion is performed on a full range between 0V and the reference voltage pin ADVREF.
Analog inputs between these voltages convert to values based on a linear conversion.
33.6.3
Conversion Resolution
The ADC supports 8-bit or 10-bit resolutions. The 8-bit selection is performed by setting the bit
LOWRES in the ADC Mode Register (MR). By default, after a reset, the resolution is the highest
and the DATA field in the data registers is fully used. By setting the bit LOWRES, the ADC
switches in the lowest resolution and the conversion results can be read in the eight lowest significant bits of the data registers. The two highest bits of the DATA field in the corresponding
CDR register and of the LDATA field in the LCDR register read 0.
Moreover, when a PDC channel is connected to the ADC, 10-bit resolution sets the transfer
request sizes to 16-bit. Setting the bit LOWRES automatically switches to 8-bit data transfers. In
this case, the destination buffers are optimized.
702
32058K
AVR32-01/12
AT32UC3A
33.6.4
Conversion Results
When a conversion is completed, the resulting 10-bit digital value is stored in the Channel Data
Register (CDR) of the current channel and in the ADC Last Converted Data Register (LCDR).
The channel EOC bit in the Status Register (SR) is set and the DRDY is set. In the case of a
connected PDC channel, DRDY rising triggers a data transfer request. In any case, either EOC
and DRDY can trigger an interrupt.
Reading one of the CDR registers clears the corresponding EOC bit. Reading LCDR clears the
DRDY bit and the EOC bit corresponding to the last converted channel.
Figure 33-2. EOCx and DRDY Flag Behavior
Write CR
With START=1
Read CDRx
Write CR
With START=1
Read LCDR
CHx(CHSR)
EOCx(SR)
Conversion Time
Conversion Time
DRDY(SR)
703
32058K
AVR32-01/12
AT32UC3A
If the CDR is not read before further incoming data is converted, the corresponding Overrun
Error (OVRE) flag is set in the Status Register (SR).
In the same way, new data converted when DRDY is high sets the bit GOVRE (General Overrun
Error) in SR.
The OVRE and GOVRE flags are automatically cleared when SR is read.
Figure 33-3. GOVRE and OVREx Flag Behavior
Read SR
TRIGGER
CH0(CHSR)
CH1(CHSR)
LCDR
CRD0
CRD1
EOC0(SR)
EOC1(SR)
Undefined Data
Data B
Data A
Data C
Data A
Undefined Data
Data C
Undefined Data
Data B
Conversion
Conversion
Conversion
Read CDR0
Read CDR1
GOVRE(SR)
DRDY(ASR)
OVRE0(SR)
Warning: If the corresponding channel is disabled during a conversion or if it is disabled and
then reenabled during a conversion, its associated data and its corresponding EOC and OVRE
flags in SR are unpredictable.
704
32058K
AVR32-01/12
AT32UC3A
33.6.5
Conversion Triggers
Conversions of the active analog channels are started with a software or a hardware trigger. The
software trigger is provided by writing the Control Register (CR) with the bit START at 1.
The hardware trigger can be one of the TIOA outputs of the Timer Counter channels, or the
external trigger input of the ADC (TRIGGER). The hardware trigger is selected with the field
TRGSEL in the Mode Register (MR). The selected hardware trigger is enabled with the bit
TRGEN in the Mode Register (MR).
If a hardware trigger is selected, the start of a conversion is detected at each rising edge of the
selected signal. If one of the TIOA outputs is selected, the corresponding Timer Counter channel
must be programmed in Waveform Mode.
Only one start command is necessary to initiate a conversion sequence on all the channels. The
ADC hardware logic automatically performs the conversions on the active channels, then waits
for a new request. The Channel Enable (CHER) and Channel Disable (CHDR) Registers enable
the analog channels to be enabled or disabled independently.
If the ADC is used with a PDC, only the transfers of converted data from enabled channels are
performed and the resulting data buffers should be interpreted accordingly.
Warning: Enabling hardware triggers does not disable the software trigger functionality. Thus, if
a hardware trigger is selected, the start of a conversion can be initiated either by the hardware or
the software trigger.
33.6.6
Sleep Mode and Conversion Sequencer
The ADC Sleep Mode maximizes power saving by automatically deactivating the ADC when it is
not being used for conversions. Sleep Mode is selected by setting the bit SLEEP in the Mode
Register MR.
The SLEEP mode is automatically managed by a conversion sequencer, which can automatically process the conversions of all channels at lowest power consumption.
When a start conversion request occurs, the ADC is automatically activated. As the analog cell
requires a start-up time, the logic waits during this time and starts the conversion on the enabled
channels. When all conversions are complete, the ADC is deactivated until the next trigger. Triggers occurring during the sequence are not taken into account.
The conversion sequencer allows automatic processing with minimum processor intervention
and optimized power consumption. Conversion sequences can be performed periodically using
a Timer/Counter output. The periodic acquisition of several samples can be processed automatically without any intervention of the processor thanks to the PDC.
Note:
The reference voltage pins always remain connected in normal mode as in sleep mode.
705
32058K
AVR32-01/12
AT32UC3A
33.6.7
ADC Timings
Each ADC has its own minimal Startup Time that is programmed through the field STARTUP in
the Mode Register MR.
In the same way, a minimal Sample and Hold Time is necessary for the ADC to guarantee the
best converted final value between two channels selection. This time has to be programmed
through the bitfield SHTIM in the Mode Register MR.
Warning: No input buffer amplifier to isolate the source is included in the ADC. This must be
taken into consideration to program a precise value in the SHTIM field. See the section, ADC
Characteristics in the product datasheet.
706
32058K
AVR32-01/12
AT32UC3A
33.7
User Interface
Table 33-2.
Offset
ADC Register Mapping
Register
Name
Access
Reset State
0x00
Control Register
CR
Write-only
–
0x04
Mode Register
MR
Read/Write
0x00000000
0x10
Channel Enable Register
CHER
Write-only
–
0x14
Channel Disable Register
CHDR
Write-only
–
0x18
Channel Status Register
CHSR
Read-only
0x00000000
0x1C
Status Register
SR
Read-only
0x000C0000
0x20
Last Converted Data Register
LCDR
Read-only
0x00000000
0x24
Interrupt Enable Register
IER
Write-only
–
0x28
Interrupt Disable Register
IDR
Write-only
–
0x2C
Interrupt Mask Register
IMR
Read-only
0x00000000
0x30
Channel Data Register 0
CDR0
Read-only
0x00000000
...
...
...
CDR7
Read-only
0x00000000
VERSION
Read-only
–
...
...(if implemented)
0x4C
Channel Data Register 7(if implemented)
0xFC
Version Register
707
32058K
AVR32-01/12
AT32UC3A
33.7.1
Name:
Control Register
CR
Access Type:
Write-only
Offset:
0x00
Reset Value:
–
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
–
15
14
13
12
11
10
9
8
–
–
–
–
–
–
–
–
7
6
5
4
3
2
–
–
–
–
–
–
1
START
0
SWRST
• START: Start Conversion
0 = No effect.
1 = Begins analog-to-digital conversion.
• SWRST: Software Reset
0 = No effect.
1 = Resets the ADC simulating a hardware reset.
708
32058K
AVR32-01/12
AT32UC3A
33.7.2
Name:
Mode Register
MR
Access Type:
Read/Write
Offset:
0x04
Reset Value:
0x00000000
31
30
29
28
–
–
–
–
23
22
21
20
–
–
–
15
14
13
–
–
7
–
6
–
5
SLEEP
27
26
25
24
17
16
10
9
8
2
1
SHTIM
19
18
STARTUP
12
11
4
3
LOWRES
PRESCAL
TRGSEL
0
TRGEN
• SHTIM: Sample & Hold Time
Sample & Hold Time = (SHTIM+1) / ADCClock
• STARTUP: Start Up Time
Startup Time = (STARTUP+1) * 8 / ADCClock
• PRESCAL: Prescaler Rate Selection
ADCClock = CLK_ADC / ( (PRESCAL+1) * 2 )
• SLEEP: Sleep Mode
SLEEP
Selected Mode
0
Normal Mode
1
Sleep Mode
• LOWRES: Resolution
LOWRES
Selected Resolution
0
10-bit resolution
1
8-bit resolution
• TRGSEL: Trigger Selection
TRGSEL
Selected TRGSEL
0
0
0
Internal Trigger 0, depending of chip integration
0
0
1
Internal Trigger 1, depending of chip integration
0
1
0
Internal Trigger 2, depending of chip integration
0
1
1
Internal Trigger 3, depending of chip integration
1
0
0
Internal Trigger 4, depending of chip integration
709
32058K
AVR32-01/12
AT32UC3A
TRGSEL
Selected TRGSEL
1
0
1
Internal Trigger 5, depending of chip integration
1
1
0
External trigger
1
1
1
Reserved
• TRGEN: Trigger Enable
TRGEN
Selected TRGEN
0
Hardware triggers are disabled. Starting a conversion is only possible by software.
1
Hardware trigger selected by TRGSEL field is enabled.
710
32058K
AVR32-01/12
AT32UC3A
33.7.3
Name:
Channel Enable Register
CHER
Access Type:
Write-only
Offset:
0x10
Reset Value:
–
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
–
15
14
13
12
11
10
9
8
–
–
–
–
–
–
–
–
7
6
5
4
3
2
1
0
CH2
CH6
CH5
CH4
CH3
CH2
CH1
CH0
• CHx: Channel x Enable
0 = No effect.
1 = Enables the corresponding channel(if implemented).
711
32058K
AVR32-01/12
AT32UC3A
33.7.4
Name:
Channel Disable Register
CHDR
Access Type:
Write-only
Offset:
0x14
Reset Value:
–
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
–
15
14
13
12
11
10
9
8
–
–
–
–
–
–
–
–
7
6
5
4
3
2
1
0
CH7
CH6
CH5
CH4
CH3
CH2
CH1
CH0
• CHx: Channel x Disable
0 = No effect.
1 = Disables the corresponding channel(if implemented).
Warning: If the corresponding channel is disabled during a conversion or if it is disabled then reenabled during a conversion, its associated data and its corresponding EOC and OVRE flags in SR are unpredictable.
712
32058K
AVR32-01/12
AT32UC3A
33.7.5
Name:
Channel Status Register
CHSR
Access Type:
Read-only
Offset:
0x18
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
–
15
14
13
12
11
10
9
8
–
–
–
–
–
–
–
–
7
6
5
4
3
2
1
0
CH7
CH6
CH5
CH4
CH3
CH2
CH1
CH0
• CHx: Channel x Status
0 = Corresponding channel is disabled(if implemented).
1 = Corresponding channel is enabled(if implemented).
713
32058K
AVR32-01/12
AT32UC3A
33.7.6
Name:
Status Register
SR
Access Type:
Read-only
Offset:
0x1C
Reset Value:
0x000C0000
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
RXBUFF
ENDRX
GOVRE
DRDY
15
14
13
12
11
10
9
8
OVRE7
OVRE6
OVRE5
OVRE4
OVRE3
OVRE2
OVRE1
OVRE0
7
6
5
4
3
2
1
0
EOC7
EOC6
EOC5
EOC4
EOC3
EOC2
EOC1
EOC0
• RXBUFF: RX Buffer Full
0 = RCR or RNCR have a value other than 0.
1 = Both RCR and RNCR have a value of 0.
• ENDRX: End of RX Buffer
0 = The Receive Counter Register has not reached 0 since the last write in RCR or RNCR.
1 = The Receive Counter Register has reached 0 since the last write in RCR or RNCR.
• GOVRE: General Overrun Error
0 = No General Overrun Error occurred since the last read of SR.
1 = At least one General Overrun Error has occurred since the last read of SR.
• DRDY: Data Ready
0 = No data has been converted since the last read of LCDR.
1 = At least one data has been converted and is available in LCDR.
• OVREx: Overrun Error x
0 = No overrun error on the corresponding channel(if implemented) since the last read of SR.
1 = There has been an overrun error on the corresponding channel (if implemented) since the last read of SR.
• EOCx: End of Conversion x
0 = Corresponding analog channel (if implemented) is disabled, or the conversion is not finished.
1 = Corresponding analog channel (if implemented) is enabled and conversion is complete.
714
32058K
AVR32-01/12
AT32UC3A
33.7.7
Name:
Last Converted Data Register
LCDR
Access Type:
Read-only
Offset:
0x20
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
–
15
14
13
12
11
10
9
–
–
–
–
–
–
7
6
5
4
3
2
LDATA
1
LDATA
8
0
• LDATA: Last Data Converted
The analog-to-digital conversion data is placed into this register at the end of a conversion and remains until a new conversion is completed.
715
32058K
AVR32-01/12
AT32UC3A
33.7.8
Name:
Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x24
Reset Value:
–
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
RXBUFF
ENDRX
GOVRE
DRDY
15
14
13
12
11
10
9
8
OVRE7
OVRE6
OVRE5
OVRE4
OVRE3
OVRE2
OVRE1
OVRE0
7
6
5
4
3
2
1
0
EOC7
EOC6
EOC5
EOC4
EOC3
EOC2
EOC1
EOC0
• RXBUFF: Receive Buffer Full Interrupt Enable
0 = No effect.
1 = Enables the corresponding interrupt. ENDRX: End of Receive Buffer Interrupt Enable
• GOVRE: General Overrun Error Interrupt Enable
• DRDY: Data Ready Interrupt Enable
• OVREx: Overrun Error Interrupt Enable x
• EOCx: End of Conversion Interrupt Enable x
716
32058K
AVR32-01/12
AT32UC3A
33.7.9
Name:
Interrupt Disable Register
IDR
Access Type:
Write-only
Offset:
0x28
Reset Value:
–
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
RXBUFF
ENDRX
GOVRE
DRDY
15
14
13
12
11
10
9
8
OVRE7
OVRE6
OVRE5
OVRE4
OVRE3
OVRE2
OVRE1
OVRE0
7
6
5
4
3
2
1
0
EOC7
EOC6
EOC5
EOC4
EOC3
EOC2
EOC1
EOC0
• RXBUFF: Receive Buffer Full Interrupt Disable
0 = No effect.
1 = Disables the corresponding interrupt.
• ENDRX: End of Receive Buffer Interrupt Disable
• GOVRE: General Overrun Error Interrupt Disable
• DRDY: Data Ready Interrupt Disable
• OVREx: Overrun Error Interrupt Disable x
• EOCx: End of Conversion Interrupt Disable x
717
32058K
AVR32-01/12
AT32UC3A
33.7.10
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
Offset:
0x2C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
RXBUFF
ENDRX
GOVRE
DRDY
15
14
13
12
11
10
9
8
OVRE7
OVRE6
OVRE5
OVRE4
OVRE3
OVRE2
OVRE1
OVRE0
7
6
5
4
3
2
1
0
EOC7
EOC6
EOC5
EOC4
EOC3
EOC2
EOC1
EOC0
• RXBUFF: Receive Buffer Full Interrupt Mask
0 = The corresponding interrupt is disabled.
1 = The corresponding interrupt is enabled.
• ENDRX: End of Receive Buffer Interrupt Mask
• GOVRE: General Overrun Error Interrupt Mask
• DRDY: Data Ready Interrupt Mask
• OVREx: Overrun Error Interrupt Mask x
• EOCx: End of Conversion Interrupt Mask x
718
32058K
AVR32-01/12
AT32UC3A
33.7.11
Name:
Channel Data Register
CDRx
Access Type:
Read-only
Offset:
0x2C-0x4C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
–
–
–
–
15
14
13
12
11
10
9
–
–
–
–
–
–
7
6
5
4
3
2
DATA
1
DATA
8
0
• DATA: Converted Data
The analog-to-digital conversion data is placed into this register at the end of a conversion and remains until a new conversion is completed. The Convert Data Register (CDR) is only loaded if the corresponding analog channel is enabled.
719
32058K
AVR32-01/12
AT32UC3A
33.7.12
Name:
Version Register
VERSION
Access Type:
Read-only
Offset:
0xFC
Reset Value:
–
31
30
29
28
27
26
25
24
–
–
–
–
–
–
–
–
23
22
21
20
19
18
17
16
–
–
–
–
15
14
13
12
9
8
–
–
–
–
7
6
5
4
1
0
VARIANT
11
10
VERSION[11:8]
VERSION[7:0]
3
2
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the module. No functionality associated.
720
32058K
AVR32-01/12
AT32UC3A
34. Audio Bitstream DAC (ABDAC)
Rev: 1.0.1.1
34.1
Features
• Digital Stereo DAC
• Oversampled D/A conversion architecture
– Oversampling ratio fixed 128x
– FIR equalization filter
– Digital interpolation filter: Comb4
– 3rd Order Sigma-Delta D/A converters
• Digital bitstream outputs
• Parallel interface
• Connected to DMA Controller for background transfer without CPU intervention
34.2
Description
The Audio Bitstream DAC converts a 16-bit sample value to a digital bitstream with an average
value proportional to the sample value. Two channels are supported, making the Audio Bitstream DAC particularly suitable for stereo audio. Each channel has a pair of complementary
digital outputs, DACn and DACn_N, which can be connected to an external high input impedance amplifier.
The Audio Bitstream DAC is compromised of two 3rd order Sigma Delta D/A converter with an
oversampling ratio of 128. The samples are upsampled with a 4th order Sinc interpolation filter
(Comb4) before being input to the Sigmal Delta Modulator. In order to compensate for the pass
band frequency response of the interpolation filter and flatten the overall frequency response,
the input to the interpolation filter is first filtered with a simple 3-tap FIR filter.The total frequency
response of the Equalization FIR filter and the interpolation filter is given in Figure 34-2 on page
733. The digital output bitstreams from the Sigma Delta Modulators should be low-pass filtered
to remove high frequency noise inserted by the Modulation process.
The output DACn and DACn_N should be as ideal as possible before filtering, to achieve the
best SNR quality. The output can be connected to a class D amplifier output stage, or it can be
low pass filtered and connected to a high input impedance amplifier. A simple 1st order or higher
low pass filter that filters all the frequencies above 50 kHz should be adequate.
721
32058K
AVR32-01/12
AT32UC3A
34.3
Block Diagram
Figure 34-1. Functional Block Diagram
Audio Bitstream DAC
clk
Clock Generator
sample_clk
bit_clk
din1[15:0]
Equalization FIR
COMB
(INT=128)
Sigma-Delta
DA-MOD
bit_out1
din2[15:0]
Equalization FIR
COMB
(INT=128)
Sigma-Delta
DA-MOD
bit_out2
34.4
Pin Name List
Table 34-1.
I/O Lines Description
Pin Name
Pin Description
DATA0
Output from Audio Bitstream DAC Channel 0
Output
DATA1
Output from Audio Bitstream DAC Channel 1
Output
DATAN0
Inverted output from Audio Bitstream DAC Channel 0
Output
DATAN1
Inverted output from Audio Bitstream DAC Channel 1
Output
34.5
Type
Product Dependencies
34.5.1
I/O Lines
The output pins used for the output bitstream from the Audio Bitstream DAC may be multiplexed
with PIO lines.
Before using the Audio Bitstream DAC, the PIO controller must be configured in order for the
Audio Bitstream DAC I/O lines to be in Audio Bitstream DAC peripheral mode.
34.5.2
Power Management
The PB-bus clock to the Audio Bitstream DAC is generated by the power manager. Before using
the Audio Bitstream DAC, the programmer must ensure that the Audio Bitstream DAC clock is
enabled in the power manager.
722
32058K
AVR32-01/12
AT32UC3A
34.5.3
Clock Management
The Audio Bitstream DAC needs a separate clock for the D/A conversion operation. This clock
should be set up in the generic clock register in the power manager. The frequency of this clock
must be 256 times the frequency of the desired samplerate (fs). For fs=48kHz this means that the
clock must have a frequency of 12.288MHz.
34.5.4
Interrupts
The Audio Bitstream DAC interface has an interrupt line connected to the interrupt controller. In
order to handle interrupts, the interrupt controller must be programmed before configuring the
Audio Bitstream DAC.
All Audio Bitstream DAC interrupts can be enabled/disabled by writing to the Audio Bitstream
DAC Interrupt Enable/Disable Registers. Each pending and unmasked Audio Bitstream DAC
interrupt will assert the interrupt line. The Audio Bitstream DAC interrupt service routine can get
the interrupt source by reading the Interrupt Status Register.
34.5.5
DMA
The Audio Bitstream DAC is connected to the DMA controller. The DMA controller can be programmed to automatically transfer samples to the Audio Bitstream DAC Sample Data Register
(SDR) when the Audio Bitstream DAC is ready for new samples. This enables the Audio Bitstream DAC to operate without any CPU intervention such as polling the Interrupt Status
Register (ISR) or using interrupts. See the DMA controller documentation for details on how to
setup DMA transfers.
34.6
Functional Description
In order to use the Audio Bitstream DAC the product dependencies given in Section 34.5 on
page 722 must be resolved. Particular attention should be given to the configuration of clocks
and I/O lines in order to ensure correct operation of the Audio Bitstream DAC.
The Audio Bitstream DAC is enabled by writing the ENABLE bit in the Audio Bitstream DAC
Control Register (CR). The two 16-bit sample values for channel 0 and 1 can then be written to
the least and most significant halfword of the Sample Data Register (SDR), respectively. The
TX_READY bit in the Interrupt Status Register (ISR) will be set whenever the DAC is ready to
receive a new sample. A new sample value should be written to SDR before 256 DAC clock
cycles, or an underrun will occur, as indicated by the UNDERRUN status flags in ISR. ISR is
cleared when read, or when writing one to the corresponding bits in the Interrupt Clear Register
(ICR).
For interrupt-based operation, the relevant interrupts must be enabled by writing one to the corresponding bits in the Interrupt Enable Register (IER). Interrupts can be disabled by the Interrupt
Disable Register (IDR), and active interrupts are indicated in the read-only Interrupt Mask Register (IMR).
The Audio Bitstream DAC can also be configured for peripheral DMA access, in which case only
the enable bit in the control register needs to be set in the Audio Bitstream DAC module.
34.6.1
Equalization Filter
The equalization filter is a simple 3-tap FIR filter. The purpose of this filter is to compensate for
the pass band frequency response of the sinc interpolation filter. The equalization filter makes
the pass band response more flat and moves the -3dB corner a little higher.
723
32058K
AVR32-01/12
AT32UC3A
34.6.2
Interpolation filter
The interpolation filter interpolates from fs to 128fs. This filter is a 4th order Cascaded IntegratorComb filter, and the basic building blocks of this filter is a comb part and an integrator part.
34.6.3
Sigma Delta Modulator
This part is a 3rd order Sigma Delta Modulator consisting of three differentiators (delta blocks),
three integrators (sigma blocks) and a one bit quantizer. The purpose of the integrators is to
shape the noise, so that the noise is reduces in the band of interest and increased at the higher
frequencies, where it can be filtered.
34.6.4
Data Format
Input data is on two’s complement format.
724
32058K
AVR32-01/12
AT32UC3A
34.7
Audio Bitstream DAC User Interface
Register Mapping
Table 34-2.
Offset
Register
Register Name
Access
Reset
SDR
Read/Write
0x0
-
-
-
0x0
Sample Data Register
0x4
Reserved
0x8
Control Register
CR
Read/Write
0x0
0xc
Interrupt Mask Register
IMR
Read
0x0
0x10
Interrupt Enable Register
IER
Write
-
0x14
Interrupt Disable Register
IDR
Write
-
0x18
Interrupt Clear Register
ICR
Write
-
0x1C
Interrupt Status Register
ISR
Read
0x0
725
32058K
AVR32-01/12
AT32UC3A
34.7.1
Audio Bitstream DAC Sample Data Register
Name:
SDR
Access Type:
Read-Write
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
CHANNEL1
23
22
21
20
15
14
13
12
CHANNEL1
CHANNEL0
7
6
5
4
CHANNEL0
• CHANNEL0: Sample Data for Channel 0
Signed 16-bit Sample Data for channel 0. When the SWAP bit in the DAC Control Register (CR) is set writing to the Sample
Data Register (SDR) will cause the values written to CHANNEL0 and CHANNEL1 to be swapped.
• CHANNEL1: Sample Data for Channel 1
Signed 16-bit Sample Data for channel 1. When the SWAP bit in the DAC Control Register (CR) is set writing to the Sample
Data Register (SDR) will cause the values written to CHANNEL0 and CHANNEL1 to be swapped.
726
32058K
AVR32-01/12
AT32UC3A
34.7.2
Audio Bitstream DAC Control Register
Name:
CR
Access Type:
Read-Write
31
EN
23
15
7
-
30
SWAP
22
14
6
-
29
21
13
5
-
28
20
12
4
-
27
19
11
3
-
26
18
10
2
-
25
17
9
1
-
24
16
8
0
-
• SWAP: Swap Channels
0: The CHANNEL0 and CHANNEL1 samples will not be swapped when writing the Audio Bitstream DAC Sample Data
Register (SDR).
1: The CHANNEL0 and CHANNEL1 samples will be swapped when writing the Audio Bitstream DAC Sample Data Register (SDR).
• EN: Enable Audio Bitstream DAC
0: Audio Bitstream DAC is disabled.
1: Audio Bitstream DAC is enabled.
727
32058K
AVR32-01/12
AT32UC3A
34.7.3
Audio Bitstream DAC Interrupt Mask Register
Name:
IMR
Access Type:
Read-only
31
23
15
7
-
30
22
14
6
-
29
TX_READY
21
13
5
-
28
UNDERRUN
20
12
4
-
27
19
11
3
-
26
18
10
2
-
25
17
9
1
-
24
16
8
0
-
• UNDERRUN: Underrun Interrupt Mask
0: The Audio Bitstream DAC Underrun interrupt is disabled.
1: The Audio Bitstream DAC Underrun interrupt is enabled.
• TX_READY: TX Ready Interrupt Mask
0: The Audio Bitstream DAC TX Ready interrupt is disabled.
1: The Audio Bitstream DAC TX Ready interrupt is enabled.
728
32058K
AVR32-01/12
AT32UC3A
34.7.4
Audio Bitstream DAC Interrupt Enable Register
Name:
IER
Access Type:
Write-only
31
23
15
7
-
30
22
14
6
-
29
TX_READY
21
13
5
-
28
UNDERRUN
20
12
4
-
27
19
11
3
-
26
18
10
2
-
25
17
9
1
-
24
16
8
0
-
• UNDERRUN: Underrun Interrupt Enable
0: No effect.
1: Enables the Audio Bitstream DAC Underrun interrupt.
• TX_READY: TX Ready Interrupt Enable
0: No effect.
1: Enables the Audio Bitstream DAC TX Ready interrupt.
729
32058K
AVR32-01/12
AT32UC3A
34.7.5
Audio Bitstream DAC Interrupt Disable Register
Name:
IDR
Access Type:
Write-only
31
23
15
7
-
30
22
14
6
-
29
TX_READY
21
13
5
-
28
UNDERRUN
20
12
4
-
27
19
11
3
-
26
18
10
2
-
25
17
9
1
-
24
16
8
0
-
• UNDERRUN: Underrun Interrupt Disable
0: No effect.
1: Disable the Audio Bitstream DAC Underrun interrupt.
• TX_READY: TX Ready Interrupt Disable
0: No effect.
1: Disable the Audio Bitstream DAC TX Ready interrupt.
730
32058K
AVR32-01/12
AT32UC3A
34.7.6
Audio Bitstream DAC Interrupt Clear Register
Name:
ICR
Access Type:
Write-only
31
23
15
7
-
30
22
14
6
-
29
TX_READY
21
13
5
-
28
UNDERRUN
20
12
4
-
27
19
11
3
-
26
18
10
2
-
25
17
9
1
-
24
16
8
0
-
• UNDERRUN: Underrun Interrupt Clear
0: No effect.
1: Clear the Audio Bitstream DAC Underrun interrupt.
• TX_READY: TX Ready Interrupt Clear
0: No effect.
1: Clear the Audio Bitstream DAC TX Ready interrupt.
731
32058K
AVR32-01/12
AT32UC3A
34.7.7
Audio Bitstream DAC Interrupt Status Register
Name:
ISR
Access Type:
Read-only
31
23
15
7
-
30
22
14
6
-
29
TX_READY
21
13
5
-
28
UNDERRUN
20
12
4
-
27
19
11
3
-
26
18
10
2
-
25
17
9
1
-
24
16
8
0
-
• UNDERRUN: Underrun Interrupt Status
0: No Audio Bitstream DAC Underrun has occured since the last time ISR was read or since reset.
1: At least one Audio Bitstream DAC Underrun has occured since the last time ISR was read or since reset.
• TX_READY: TX Ready Interrupt Status
0: No Audio Bitstream DAC TX Ready has occuredt since the last time ISR was read.
1: At least one Audio Bitstream DAC TX Ready has occuredt since the last time ISR was read.
732
32058K
AVR32-01/12
AT32UC3A
34.8
Frequency Response
Figure 34-2. Frequecy response, EQ-FIR+COMB4
1 0
0
-1 0
-2 0
-3 0
-4 0
-5 0
-6 0
0
1
2
3
4
5
6
7
8
9
1 0
x
1 0
4
733
32058K
AVR32-01/12
AT32UC3A
35. On-Chip Debug
Rev: 1.3.0.0
35.1
35.2
Features
•
•
•
•
•
•
•
•
Debug interface in compliance with IEEE-ISTO 5001-2003 (Nexus 2.0) Class 2+
JTAG access to all on-chip debug functions
Advanced Program, Data, Ownership, and Watchpoint trace supported
NanoTrace JTAG-based trace access
Auxiliary port for high-speed trace information
Hardware support for 6 Program and 2 Data breakpoints
Unlimited number of software breakpoints supported
Automatic CRC check of memory regions
Overview
Debugging on the AT32UC3A is facilitated by a powerful On-Chip Debug (OCD) system. The
user accesses this through an external debug tool which connects to the JTAG port and the Auxiliary (AUX) port. The AUX port is primarily used for trace functions, and a JTAG-based
debugger is sufficient for basic debugging.
The debug system is based on the Nexus 2.0 standard, class 2+, which includes:
• Basic run-time control
• Program breakpoints
• Data breakpoints
• Program trace
• Ownership trace
• Data trace
In addition to the mandatory Nexus debug features, the AT32UC3A implements several useful
OCD features, such as:
• Debug Communication Channel between CPU and JTAG
• Run-time PC monitoring
• CRC checking
• NanoTrace
• Software Quality Assurance (SQA) support
The OCD features are controlled by OCD registers, which can be accessed by JTAG when the
NEXUS_ACCESS JTAG instruction is loaded. The CPU can also access OCD registers directly
using mtdr/mfdr instructions in any privileged mode. The OCD registers are implemented based
on the recommendations in the Nexus 2.0 standard, and are detailed in the AVR32UC Technical
Reference Manual.
734
32058K
AVR32-01/12
AT32UC3A
35.3
Block diagram
Figure 35-1. On-Chip Debug block diagram
JTAG
JTAG
AUX
On-Chip Debug
Memory
Service
Unit
Service Access Bus
Transmit Queue
Watchpoints
Debug PC
Debug
Instruction
Breakpoints
CPU
35.4
Program
Trace
Internal
SRAM
HSB Bus Matrix
Data Trace
Ownership
Trace
Memories and
peripherals
Functional description
35.4.1
JTAG-based debug features
A debugger can control all OCD features by writing OCD registers over the JTAG interface.
Many of these do not depend on output on the AUX port, allowing a JTAG-based debugger to be
used.
A JTAG-based debugger should connect to the device through a standard 10-pin IDC connector
as described in the AVR32UC Technical Reference Manual.
735
32058K
AVR32-01/12
AT32UC3A
Figure 35-2. JTAG-based debugger
PC
JTAG-based
debug tool
10-pin IDC
JTAG
AVR32
35.4.1.1
Debug Communication Channel
The Debug Communication Channel (DCC) consists of a pair OCD registers with associated
handshake logic, accessible to both CPU and JTAG. The registers can be used to exchange
data between the CPU and the JTAG master, both runtime as well as in debug mode.
35.4.1.2
Breakpoints
One of the most fundamental debug features is the ability to halt the CPU, to examine registers
and the state of the system. This is accomplished by breakpoints, of which many types are
available:
• Unconditional breakpoints are set by writing OCD registers by JTAG, halting the CPU
immediately.
• Program breakpoints halt the CPU when a specific address in the program is executed.
• Data breakpoints halt the CPU when a specific memory address is read or written, allowing
variables to be watched.
• Software breakpoints halt the CPU when the breakpoint instruction is executed.
When a breakpoint triggers, the CPU enters debug mode, and the D bit in the status register is
set. This is a privileged mode with dedicated return address and return status registers. All privileged instructions are permitted. Debug mode can be entered as either OCD Mode, running
instructions from JTAG, or Monitor Mode, running instructions from program memory.
736
32058K
AVR32-01/12
AT32UC3A
35.4.1.3
OCD Mode
When a breakpoint triggers, the CPU enters OCD mode, and instructions are fetched from the
Debug Instruction OCD register. Each time this register is written by JTAG, the instruction is
executed, allowing the JTAG to execute CPU instructions directly. The JTAG master can e.g.
read out the register file by issuing mtdr instructions to the CPU, writing each register to the
Debug Communication Channel OCD registers.
35.4.1.4
Monitor Mode
Since the OCD registers are directly accessible by the CPU, it is possible to build a softwarebased debugger that runs on the CPU itself. Setting the Monitor Mode bit in the Development
Control register causes the CPU to enter Monitor Mode instead of OCD mode when a breakpoint
triggers. Monitor Mode is similar to OCD mode, except that instructions are fetched from the
debug exception vector in regular program memory, instead of issued by JTAG.
35.4.1.5
Program Counter monitoring
Normally, the CPU would need to be halted for a JTAG-based debugger to examine the current
PC value. However, the AT32UC3A also proves a Debug Program Counter OCD register, where
the debugger can continuously read the current PC without affecting the CPU. This allows the
debugger to generate a simple statistic of the time spent in various areas of the code, easing
code optimization.
35.4.2
Memory Service Unit
The Memory Service Unit (MSU) is a block dedicated to test and debug functionality. It is controlled through a dedicated set of registers addressed through the MEMORY_SERVICE JTAG
command.
35.4.2.1
Cyclic Redundancy Check (CRC)
The MSU can be used to automatically calculate the CRC of a block of data in memory. The
OCD will then read out each word in the specified memory block and report the CRC32-value in
an OCD register.
35.4.2.2
NanoTrace
The MSU additionally supports NanoTrace. This is an AVR32-specific feature, in which trace
data is output to memory instead of the AUX port. This allows the trace data to be extracted by
JTAG MEMORY_ACCESS, enabling trace features for JTAG-based debuggers. The user must
write MSU registers to configure the address and size of the memory block to be used for NanoTrace. The NanoTrace buffer can be anywhere in the physical address range, including internal
and external RAM, through an EBI, if present. This area may not be used by the application running on the CPU.
35.4.3
AUX-based debug features
Utilizing the Auxiliary (AUX) port gives access to a wide range of advanced debug features. Of
prime importance are the trace features, which allow an external debugger to receive continuous
information on the program execution in the CPU. Additionally, Event In and Event Out pins
allow external events to be correlated with the program flow.
The AUX port contains a number of pins, as shown in Table 35-1 on page 738. These are multiplexed with PIO lines, and must explicitly be enabled by writing OCD registers before the debug
session starts. The AUX port is mapped to two different locations, selectable by OCD Registers,
minimizing the chance that the AUX port will need to be shared with an application.
737
32058K
AVR32-01/12
AT32UC3A
Debug tools utilizing the AUX port should connect to the device through a Nexus-compliant Mictor-38 connector, as described in the AVR32UC Technical Reference manual. This connector
includes the JTAG signals and the RESET_N pin, giving full access to the programming and
debug features in the device.
Table 35-1.
Auxiliary port signals
Signal
Direction
MCKO
Output
Trace data output clock
MDO[5:0]
Output
Trace data output
MSEO[1:0]
Output
Trace frame control
EVTI_N
Input
EVTO_N
Output
Description
Event In
Event Out
Figure 35-3. AUX+JTAG based debugger
PC
T ra c e b u ffe r
AU X+JTA G
d e b u g to o l
M ic to r 3 8
AUX
h ig h s p e e d
JTA G
AVR 32
35.4.3.1
Trace operation
Trace features are enabled by writing OCD registers by JTAG. The OCD extracts the trace information from the CPU, compresses this information and formats it into variable-length messages
according to the Nexus standard. The messages are buffered in a 16-frame transmit queue, and
are output on the AUX port one frame at a time.
738
32058K
AVR32-01/12
AT32UC3A
The trace features can be configured to be very selective, to reduce the bandwidth on the AUX
port. In case the transmit queue overflows, error messages are produced to indicate loss of
data. The transmit queue module can optionally be configured to halt the CPU when an overflow
occurs, to prevent the loss of messages, at the expense of longer run-time for the program.
35.4.3.2
Program Trace
Program trace allows the debugger to continuously monitor the program execution in the CPU.
Program trace messages are generated for every branch in the program, and contains compressed information, which allows the debugger to correlate the message with the source code
to identify the branch instruction and target address.
35.4.3.3
Data Trace
Data trace outputs a message every time a specific location is read or written. The message
contains information about the type (read/write) and size of the access, as well as the address
and data of the accessed location. The AT32UC3A contains two data trace channels, each of
which are controlled by a pair of OCD registers which determine the range of addresses (or single address) which should produce data trace messages.
35.4.3.4
Ownership Trace
Program and data trace operate on virtual addresses. In cases where an operating system runs
several processes in overlapping virtual memory segments, the Ownership Trace feature can be
used to identify the process switch. When the O/S activates a process, it will write the process ID
number to an OCD register, which produces an Ownership Trace Message, allowing the debugger to switch context for the subsequent program and data trace messages. As the use of this
feature depends on the software running on the CPU, it can also be used to extract other types
of information from the system.
35.4.3.5
Watchpoint messages
The breakpoint modules normally used to generate program and data breakpoints can also be
used to generate Watchpoint messages, allowing a debugger to monitor program and data
events without halting the CPU. Watchpoints can be enabled independently of breakpoints, so a
breakpoint module can optionally halt the CPU when the trigger condition occurs. Data trace
modules can also be configured to produce watchpoint messages instead of regular data trace
messages.
35.4.3.6
Event In and Event Out pins
The AUX port also contains an Event In pin (EVTI_N) and an Event Out pin (EVTO_N). EVTI_N
can be used to trigger a breakpoint when an external event occurs. It can also be used to trigger
specific program and data trace synchronization messages, allowing an external event to be
correlated to the program flow.
When the CPU enters debug mode, a Debug Status message is transmitted on the trace port.
All trace messages can be timestamped when they are received by the debug tool. However,
due to the latency of the transmit queue buffering, the timestamp will not be 100% accurate. To
improve this, EVTO_N can toggle every time a message is inserted into the transmit queue,
allowing trace messages to be timestamped precisely. EVTO_N can also toggle when a breakpoint module triggers, or when the CPU enters debug mode, for any reason. This can be used to
measure precisely when the respective internal event occurs.
739
32058K
AVR32-01/12
AT32UC3A
35.4.3.7
Software Quality Analysis (SQA)
Software Quality Analysis (SQA) deals with two important issues regarding embedded software
development. Code coverage involves identifying untested parts of the embedded code, to
improve test procedures and thus the quality of the released software. Performance analysis
allows the developer to precisely quantify the time spent in various parts of the code, allowing
bottlenecks to be identified and optimized.
Program trace must be used to accomplish these tasks without instrumenting (altering) the code
to be examined. However, traditional program trace cannot reconstruct the current PC value
without correlating the trace information with the source code, which cannot be done on-the-fly.
This limits program trace to a relatively short time segment, determined by the size of the trace
buffer in the debug tool.
The OCD system in AT32UC3A extends program trace with SQA capabilities, allowing the
debug tool to reconstruct the PC value on-the-fly. Code coverage and performance analysis can
thus be reported for an unlimited execution sequence.
740
32058K
AVR32-01/12
AT32UC3A
36. JTAG and Boundary Scan
Rev.: 2.0.0.2
36.1
36.2
Features
•
•
•
•
IEEE1149.1 compliant JTAG Interface
Boundary-Scan Chain for board-level testing
Direct memory access and programming capabilities through JTAG interface
On-Chip Debug access in compliance with IEEE-ISTO 5001-2003 (Nexus 2.0)
Overview
Figure 36-1 on page 742 shows how the JTAG is connected in an AVR32 device. The TAP Controller is a state machine controlled by the TCK and TMS signals. The TAP Controller selects
either the JTAG Instruction Register or one of several Data Registers as the scan chain (shift
register) between the TDI-input and TDO-output. The Instruction Register holds JTAG instructions controlling the behavior of a Data Register.
The ID Register, Bypass Register, and the Boundary-Scan Chain are the Data Registers used
for board-level testing. The Reset Register can be used to keep the device reset during test or
programming.
The Service Access Bus (SAB) interface contains address and data registers for the Service
Access Bus, which gives access to on-chip debug, programming, and other functions in the
device. The SAB offers several modes of access to the address and data registers, as discussed in Section 36.6.4.
Section 36.7 lists the supported JTAG instructions, with references to the description in this
document.
741
32058K
AVR32-01/12
AT32UC3A
36.3
Block diagram
Figure 36-1. JTAG and Boundary Scan access
AVR32 device
JTAG TAP
JTAG master
Boundary scan enable
TAP
Controller
TCK
TMS
TDI
TDO
Data register
scan enable
Instruction Register
Scan enable
Instruction Register
TMS TCK
TDO TDI
JTAG data registers
JTAG device
ID Register
Bypass
Reset
Register
...
Service Access Bus
interface
OCD Registers
Pins and analog blocks
Memories and
peripherals
Memory Service
Unit
Boundary Scan Chain
RAM
High Speed Bus
CPU
Service
Access
Bus
742
32058K
AVR32-01/12
AT32UC3A
36.4
I/O Lines Description
Table 36-1.
I/O Lines Description
Name
Description
Type
TCK
Test Clock Input. Fully asynchronous to system clock frequency.
Input
TMS
Test Mode Select, sampled on rising TCK
Input
TDI
Test Data In, sampled on rising TCK.
Input
TDO
Test Data Out, driven on falling TCK.
Output
36.5
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
36.5.1
I/O Lines
The JTAG interface pins are multiplexed with IO lines. When the JTAG is used the associated
pins must be enabled. To enable the JTAG pins TCK must be zero while RESET_N has a zero
to one transition. To disable the JTAG pins TCK must be one while RESET_N has a zero to one
transition.
While using the JTAG lines all normal peripheral activity on these lines are disabled. The user
must make sure that no external peripheral is blocking the JTAG lines while debugging.
36.6
Functional description
36.6.1
JTAG interface
The JTAG interface is accessed through the dedicated JTAG pins shown in Table 36-1 on page
743. The TMS control line navigates the TAP controller, as shown in Figure 36-2 on page 744.
The TAP controller manages the serial access to the JTAG Instruction and Data registers. Data
is scanned into the selected instruction or data register on TDI, and out of the register on TDO,
in the Shift-IR and Shift-DR states, respectively. The LSB is shifted in and out first. TDO is highZ in other states than Shift-IR and Shift-DR.
Independent of the initial state of the TAP Controller, the Test-Logic-Reset state can always be
entered by holding TMS high for 5 TCK clock periods. This sequence should always be applied
at the start of a JTAG session to bring the TAP Controller into a defined state before applying
JTAG commands. Applying a 0 on TMS for 1 TCK period brings the TAP Controller to the RunTest/Idle state, which is the starting point for JTAG operations.
The device implements a 5-bit Instruction Register (IR). A number of public JTAG instructions
defined by the JTAG standard are supported, as described in Section 36.8, as well as a number
of AVR32-specific private JTAG instructions described in Section 36.9. Each instruction selects
a specific data register for the Shift-DR path, as described for each instruction.
743
32058K
AVR32-01/12
AT32UC3A
Figure 36-2. TAP Controller State Diagram
1
Test-LogicReset
0
0
Run-Test/
Idle
1
Select-DR
Scan
1
Select-IR
Scan
1
0
0
Capture-DR
1
1
0
Shift-DR
0
0
Shift-IR
1
1
Exit1-DR
Exit1-IR
0
0
Pause-DR
1
Exit2-DR
0
0
Pause-IR
1
1
0
1
1
Update-DR
0
36.6.2
36.6.2.1
Capture-IR
0
0
1
Exit2-IR
1
1
Update-IR
0
Typical sequence
Assuming Run-Test/Idle is the present state, a typical scenario for using the JTAG interface is:
Scanning in JTAG instruction
At the TMS input, apply the sequence 1, 1, 0, 0 at the rising edges of TCK to enter the Shift
Instruction Register - Shift-IR state. While in this state, shift the 5 bits of the JTAG instructions
into the JTAG instruction register from the TDI input at the rising edge of TCK. The TMS input
must be held low during input of the 4 LSBs in order to remain in the Shift-IR state. The JTAG
Instruction selects a particular Data Register as path between TDI and TDO and controls the circuitry surrounding the selected Data Register.
Apply the TMS sequence 1, 1, 0 to re-enter the Run-Test/Idle state. The instruction is latched
onto the parallel output from the shift register path in the Update-IR state. The Exit-IR, Pause-IR,
and Exit2-IR states are only used for navigating the state machine.
744
32058K
AVR32-01/12
AT32UC3A
Figure 36-3. Scanning in JTAG instruction
TCK
TAP State
TLR
RTI
SelDR SelIR CapIR ShIR
Ex1IR UpdIR RTI
TMS
TDI
TDO
36.6.2.2
Instruction
ImplDefined
Scanning in/out data
At the TMS input, apply the sequence 1, 0, 0 at the rising edges of TCK to enter the Shift Data
Register - Shift-DR state. While in this state, upload the selected Data Register (selected by the
present JTAG instruction in the JTAG Instruction Register) from the TDI input at the rising edge
of TCK. In order to remain in the Shift-DR state, the TMS input must be held low. While the Data
Register is shifted in from the TDI pin, the parallel inputs to the Data Register captured in the
Capture-DR state is shifted out on the TDO pin.
Apply the TMS sequence 1, 1, 0 to re-enter the Run-Test/Idle state. If the selected Data Register
has a latched parallel-output, the latching takes place in the Update-DR state. The Exit-DR,
Pause-DR, and Exit2-DR states are only used for navigating the state machine.
As shown in the state diagram, the Run-Test/Idle state need not be entered between selecting
JTAG instruction and using Data Registers.
36.6.3
Boundary-Scan
The Boundary-Scan chain has the capability of driving and observing the logic levels on the digital I/O pins, as well as the boundary between digital and analog logic for analog circuitry having
off-chip connections. At system level, all ICs having JTAG capabilities are connected serially by
the TDI/TDO signals to form a long shift register. An external controller sets up the devices to
drive values at their output pins, and observe the input values received from other devices. The
controller compares the received data with the expected result. In this way, Boundary-Scan provides a mechanism for testing interconnections and integrity of components on Printed Circuits
Boards by using the 4 TAP signals only.
The four IEEE 1149.1 defined mandatory JTAG instructions IDCODE, BYPASS, SAMPLE/PRELOAD, and EXTEST can be used for testing the Printed Circuit Board. Initial scanning of the
data register path will show the ID-code of the device, since IDCODE is the default JTAG
instruction. It may be desirable to have the AVR32 device in reset during test mode. If not reset,
inputs to the device may be determined by the scan operations, and the internal software may
be in an undetermined state when exiting the test mode. Entering reset, the outputs of any Port
Pin will instantly enter the high impedance state, making the HIGHZ instruction redundant. If
needed, the BYPASS instruction can be issued to make the shortest possible scan chain
through the device. The device can be set in the reset state either by pulling the external
RESETn pin low, or issuing the AVR_RESET instruction with appropriate setting of the Reset
Data Register.
The EXTEST instruction is used for sampling external pins and loading output pins with data.
The data from the output latch will be driven out on the pins as soon as the EXTEST instruction
is loaded into the JTAG IR-register. Therefore, the SAMPLE/PRELOAD should also be used for
setting initial values to the scan ring, to avoid damaging the board when issuing the EXTEST
745
32058K
AVR32-01/12
AT32UC3A
instruction for the first time. SAMPLE/PRELOAD can also be used for taking a snapshot of the
external pins during normal operation of the part.
When using the JTAG interface for Boundary-Scan, the JTAG TCK clock is independent of the
internal chip clock, which is not required to run.
36.6.4
Service Access Bus
A number of private instructions are used to access Service Access Bus (SAB) resources. Each
of these are described in detail in SAB address map in the Service Access Bus chapter. The
MEMORY_SIZED_ACCESS instruction allows a sized read or write to any 36-bit address on the
bus. MEMORY_WORD_ACCESS is a shorthand instruction for 32-bit accesses to any 36-bit
address, while the NEXUS_ACCESS instruction is a Nexus-compliant shorthand instruction for
accessing the 32-bit OCD registers in the 7-bit address space reserved for these. These instructions require two passes through the Shift-DR TAP state: one for the address and control
information, and one for data.
To increase the transfer rate, consecutive memory accesses can be accomplished by the
MEMORY_BLOCK_ACCESS instruction, which only requires a single pass through Shift-DR for
data transfer only. The address is automatically incremented according to the size of the last
SAB transfer.
The access time to SAB resources depends on the type of resource being accessed. It is possible to read external memory through the EBI, in which case the latency may be very long. It is
possible to abort an ongoing SAB access by the CANCEL_ACCESS instruction, to avoid hanging the bus due to an extremely slow slave.
"The access time to SAB resources depends on the type of resource being
accessed. It is possible to abort an ongoing SAB access by the
CANCEL_ACCESS instruction, to avoid hanging the bus due to an extremely
slow slave."
36.6.4.1
Busy reporting
As the time taken to perform an access may vary depending on system activity and current chip
frequency, all the SAB access JTAG instructions can return a busy indicator. This indicates
whether a delay needs to be inserted, or an operation needs to be repeated in order to be successful. If a new access is requested while the SAB is busy, the request is ignored.
The SAB becomes busy when:
• Entering Update-DR in the address phase of any read operation, e.g. after scanning in a
NEXUS_ACCESS address with the read bit set.
• Entering Update-DR in the data phase of any write operation, e.g. after scanning in data for a
NEXUS_ACCESS write.
• Entering Update-DR during a MEMORY_BLOCK_ACCESS.
• Entering Update-DR after scanning in a counter value for SYNC.
• Entering Update-IR after scanning in a MEMORY_BLOCK_ACCESS if the previous access
was a read and data was scanned after scanning the address.
The SAB becomes ready again when:
• A read or write operation completes.
746
32058K
AVR32-01/12
AT32UC3A
• A SYNC countdown completed.
• A operation is cancelled by the CANCEL_ACCESS instruction.
What to do if the busy bit is set:
• During Shift-IR: The new instruction is selected, but the previous operation has not yet
completed and will continue (unless the new instruction is CANCEL_ACCESS). You may
continue shifting the same instruction until the busy bit clears, or start shifting data. If shifting
data, you must be prepared that the data shift may also report busy.
• During Shift-DR of an address: The new address is ignored. The SAB stays in address mode,
so no data must be shifted. Repeat the address until the busy bit clears.
• During Shift-DR of read data: The read data are invalid. The SAB stays in data mode. Repeat
scanning until the busy bit clears.
• During Shift-DR of write data: The write data are ignored. The SAB stays in data mode.
Repeat scanning until the busy bit clears.
36.6.4.2
Error reporting
The Service access port may not be able to complete all accesses as requested. This may be
because the address is invalid, the addressed area is read-only or cannot handle byte/halfword
accesses, or because the chip is set in a protected mode where only limited accesses are
allowed.
The error bit is updated when an access completes, and is cleared when a new access starts.
What to do if the error bit is set:
• During Shift-IR: The new instruction is selected. The last operation performed using the old
instruction did not complete successfully.
• During Shift-DR of an address: The previous operation failed. The new address is accepted.
If the read bit is set, a read operation is started.
• During Shift-DR of read data: The read operation failed, and the read data are invalid.
• During Shift-DR of write data: The previous write operation failed. The new data are accepted
and a write operation started. This should only occur during block writes or stream writes. No
error can occur between scanning a write address and the following write data.
• While polling with CANCEL_ACCESS: The previous access was cancelled. It may or may not
have actually completed.
36.6.5
Memory programming
The High-Speed Bus (HSB) in the device is mapped as a slave on the SAB. This enables all
HSB-mapped memories to be read or written through the SAB using JTAG instructions, as
described in Section 36.6.4.
Internal SRAM can always be directly accessed. External static memory or SDRAM can be
accessed if the EBI has been correctly configured to access this memory. It is also possible to
access the configuration registers for these modules to set up the correct configuration. Similarly, external parallel flash can be programmed by accessing the registers for the flash device
through the EBI.
The internal flash and fuses can likewise be programmed by accessing the registers in the Flash
Controller. When the security fuse is set, access to internal memory is blocked, and the
CHIP_ERASE instruction must be used to erase the fuse and flash contents. For detail see the
SAB address map section.
747
32058K
AVR32-01/12
AT32UC3A
Memory can be written while the CPU is executing, which can be utilized for debug purposes.
When downloading a new program, the JTAG HALT instruction should be used to freeze the
CPU, to prevent partially downloaded code from being executed.
36.7
JTAG Instruction Summary
The implemented JTAG instructions in the AVR32 are shown in the table below.
Table 36-2.
Instruction
OPCODE
JTAG Instruction Summary
Instruction
Description
0x01
IDCODE
Select the 32-bit ID register as data register.
749
0x02
SAMPLE_PRELOAD
Take a snapshot of external pin values without affecting system
operation.
749
0x03
EXTEST
Select boundary scan chain as data register for testing circuitry
external to the device.
749
0x04
INTEST
Select boundary scan chain for internal testing of the device.
749
0x06
CLAMP
Bypass device through Bypass register, while driving outputs from
boundary scan register.
750
0x0C
AVR_RESET
Apply or remove a static reset to the device
757
0x0F
CHIP_ERASE
Erase the device
757
0x10
NEXUS_ACCESS
Select the SAB Address and Data registers as data register for the
TAP. The registers are accessed in Nexus mode.
751
0x11
MEMORY_WORD_ACCESS
Select the SAB Address and Data registers as data register for the
TAP.
754
0x12
MEMORY_BLOCK_ACCESS
Select the SAB Data register as data register for the TAP. The
address is auto-incremented.
755
0x13
CANCEL_ACCESS
Cancel an ongoing Nexus or Memory access.
756
0x14
MEMORY_SERVICE
Select the SAB Address and Data registers as data register for the
TAP. The registers are accessed in Memory Service mode.
752
0x15
MEMORY_SIZED_ACCESS
Select the SAB Address and Data registers as data register for the
TAP.
753
0x17
SYNC
Synchronization counter
757
0x1C
HALT
Halt the CPU for safe programming.
758
0x1F
BYPASS
Bypass this device through the bypass register.
750
N/A
Acts as BYPASS
Others
36.7.1
Page
Security restrictions
When the security fuse in the Flash is programmed, the following JTAG instructions are
restricted:
• NEXUS_ACCESS
• MEMORY_WORD_ACCESS
• MEMORY_BLOCK_ACCESS
• MEMORY_SIZED_ACCESS
748
32058K
AVR32-01/12
AT32UC3A
For description of what memory locations remain accessible, please refer to the SAB address
map.
Full access to these instructions is re-enabled when the security fuse is erased by the
CHIP_ERASE JTAG instruction.
Note that the security bit will read as programmed and block these instructions also if the Flash
Controller is statically reset.
Other security mechanisms can also restrict these functions. If such mechanisms are present
they are listed in the SAB address map section.
36.8
Public JTAG instructions
36.8.1
IDCODE
This instruction selects the 32 bit ID register as Data Register. The ID register consists of a version number, a device number and the manufacturer code chosen by JEDEC. This is the default
instruction after power-up.
The active states are:
• Capture-DR: The static IDCODE value is latched into the shift register.
• Shift-DR: The IDCODE scan chain is shifted by the TCK input.
36.8.2
SAMPLE_PRELOAD
JTAG instruction for taking a snap-shot of the input/output pins without affecting the system
operation, and pre-loading the scan chain without updating the DR-latch. The Boundary-Scan
Chain is selected as Data Register.
The active states are:
• Capture-DR: Data on the external pins are sampled into the Boundary-Scan Chain.
• Shift-DR: The Boundary-Scan Chain is shifted by the TCK input.
36.8.3
EXTEST
JTAG instruction for selecting the Boundary-Scan Chain as Data Register for testing circuitry
external to the AVR32 package. The contents of the latched outputs of the Boundary-Scan chain
is driven out as soon as the JTAG IR-register is loaded with the EXTEST instruction.
The active states are:
• Capture-DR: Data on the external pins is sampled into the Boundary-Scan Chain.
• Shift-DR: The Internal Scan Chain is shifted by the TCK input.
• Update-DR: Data from the scan chain is applied to output pins.
36.8.4
INTEST
This instruction selects the Boundary-Scan Chain as Data Register for testing internal logic in
the device. The logic inputs are determined by the Boundary-Scan Chain, and the logic outputs
are captured by the Boundary-Scan chain. The device output pins are driven from the BoundaryScan Chain.
The active states are:
• Capture-DR: Data from the internal logic is sampled into the Boundary-Scan Chain.
749
32058K
AVR32-01/12
AT32UC3A
• Shift-DR: The Internal Scan Chain is shifted by the TCK input.
• Update-DR: Data from the scan chain is applied to internal logic inputs.
36.8.5
CLAMP
This instruction selects the Bypass register as Data Register. The device output pins are driven
from the Boundary-Scan Chain.
The active states are:
• Capture-DR: Loads a logic ‘0’ into the Bypass Register.
• Shift-DR: Data is scanned from TDI to TDO through the Bypass register.
36.8.6
BYPASS
JTAG instruction selecting the 1-bit Bypass Register for Data Register.
The active states are:
• Capture-DR: Loads a logic ‘0’ into the Bypass Register.
• Shift-DR: Data is scanned from TDI to TDO through the Bypass register.
36.9
Private JTAG Instructions
36.9.1
Notation
The AVR32 defines a number of private JTAG instructions. Each instruction is briefly described
in text, with details following in table form.
Table 36-4 on page 751 shows bit patterns to be shifted in a format like "peb01". Each character
corresponds to one bit, and eight bits are grouped together for readability. The rightmost bit is
always shifted first, and the leftmost bit shifted last. The symbols used are shown in Table 36-3.
Table 36-3.
Symbol
Symbol description
Description
0
Constant low value - always reads as zero.
1
Constant high value - always reads as one.
a
An address bit - always scanned with the least significant bit first
b
A busy bit. Reads as one if the SAB was busy, or zero if it was not. See Section 36.6.4.1 for
details on how the busy reporting works.
d
A data bit - always scanned with the least significant bit first.
e
An error bit. Reads as one if an error occurred, or zero if not. See Section 36.6.4.2 for details
on how the error reporting works.
p
The chip protected bit. Some devices may be set in a protected state where access to chip
internals are severely restricted. See the documentation for the specific device for details.
On devices without this possibility, this bit always reads as zero.
r
A direction bit. Set to one to request a read, set to zero to request a write.
s
A size bit. The size encoding is described where used.
x
A don’t care bit. Any value can be shifted in, and output data should be ignored.
In many cases, it is not required to shift all bits through the data register. Bit patterns are shown
using the full width of the shift register, but the suggested or required bits are emphasized using
750
32058K
AVR32-01/12
AT32UC3A
bold text. I.e. given the pattern "aaaaaaar xxxxxxxx xxxxxxxx xxxxxxxx xx", the shift register is
34 bits, but the test or debug unit may choose to shift only 8 bits "aaaaaaar".
The following describes how to interpret the fields in the instruction description tables:
Table 36-4.
36.9.2
Instruction description
Instruction
Description
IR input value
Shows the bit pattern to shift into IR in the Shift-IR state in order to select this
instruction. The pattern is show both in binary and in hexadecimal form for
convenience.
Example: 10000 (0x10)
IR output value
Shows the bit pattern shifted out of IR in the Shift-IR state when this instruction is
active.
Example: peb01
DR Size
Shows the number of bits in the data register chain when this instruction is active.
Example: 34 bits
DR input value
Shows which bit pattern to shift into the data register in the Shift-DR state when this
instruction is active. Multiple such lines may exist, e.g. to distinguish between reads
and writes.
Example: aaaaaaar xxxxxxxx xxxxxxxx xxxxxxxx xx
DR output value
Shows the bit pattern shifted out of the data register in the Shift-DR state when this
instruction is active. Multiple such lines may exist, e.g. to distinguish between reads
and writes.
Example: xx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb
NEXUS_ACCESS
This instruction allows Nexus-compliant access to on-chip debug registers through the SAB.
OCD registers are addressed by their register index, as listed in the AVR32 Technical Reference
Manual. The 7-bit register index and a read/write control bit, and the 32-bit data is accessed
through the JTAG port.
The data register is alternately interpreted by the SAB as an address register and a data register. The SAB starts in address mode after the NEXUS_ACCESS instruction is selected, and
toggles between address and data mode each time a data scan completes with the busy bit
cleared.
Starting in Run-Test/Idle, OCD registers are accessed in the following way:
1. Select the DR Scan path.
2. Scan in the 7-bit address for the OCD register and a direction bit (1=read, 0=write).
3. Go to Update-DR and re-enter Select-DR Scan.
4. For a read operation, scan out the contents of the addressed register. For a write operation, scan in the new contents of the register.
5. Return to Run-Test/Idle.
751
32058K
AVR32-01/12
AT32UC3A
For any operation, the full 7 bits of the address must be provided. For write operations, 32 data
bits must be provided, or the result will be undefined. For read operations, shifting may be terminated once the required number of bits have been acquired.
Table 36-5.
36.9.3
NEXUS_ACCESS details
Instructions
Details
IR input value
10000 (0x10)
IR output value
peb01
DR Size
34 bits
DR input value (Address phase)
aaaaaaar xxxxxxxx xxxxxxxx xxxxxxxx xx
DR input value (Data read phase)
xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xx
DR input value (Data write phase)
dddddddd dddddddd dddddddd dddddddd xx
DR output value (Address phase)
xx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb
DR output value (Data read phase)
eb dddddddd dddddddd dddddddd dddddddd
DR output value (Data write phase)
xx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb
MEMORY_SERVICE
This instruction allows access to registers in an optional Memory Service unit. Memory Service
registers are addressed by their register index, as listed in the Memory Service documentation.
The 7-bit register index and a read/write control bit, and the 32-bit data is accessed through the
JTAG port.
The Memory Service unit may offer features such as CRC calculation of memory, debug trace
support, and test features. Please refer to the Memory Service Unit documentation and the part
specific documentation for details.
The data register is alternately interpreted by the SAB as an address register and a data register. The SAB starts in address mode after the MEMORY_SERVICE instruction is selected, and
toggles between address and data mode each time a data scan completes with the busy bit
cleared.
Starting in Run-Test/Idle, Memory Service registers are accessed in the following way:
1. Select the DR Scan path.
2. Scan in the 7-bit address for the Memory Service register and a direction bit (1=read,
0=write).
3. Go to Update-DR and re-enter Select-DR Scan.
4. For a read operation, scan out the contents of the addressed register. For a write operation, scan in the new contents of the register.
5. Return to Run-Test/Idle.
752
32058K
AVR32-01/12
AT32UC3A
For any operation, the full 7 bits of the address must be provided. For write operations, 32 data
bits must be provided, or the result will be undefined. For read operations, shifting may be terminated once the required number of bits have been acquired.
Table 36-6.
36.9.4
MEMORY_SERVICE details
Instructions
Details
IR input value
10100 (0x14)
IR output value
peb01
DR Size
34 bits
DR input value (Address phase)
aaaaaaar xxxxxxxx xxxxxxxx xxxxxxxx xx
DR input value (Data read phase)
xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xx
DR input value (Data write phase)
dddddddd dddddddd dddddddd dddddddd xx
DR output value (Address phase)
xx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb
DR output value (Data read phase)
eb dddddddd dddddddd dddddddd dddddddd
DR output value (Data write phase)
xx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb
MEMORY_SIZED_ACCESS
This instruction allows access to the entire Service Access Bus data area. Data are accessed
through a 36-bit byte index, a 2-bit size, a direction bit, and 8, 16, or 32 bits of data. Not all units
mapped on the SAB bus may support all sizes of accesses, e.g. some may only support word
accesses.
The data register is alternately interpreted by the SAB as an address register and a data register. The SAB starts in address mode after the MEMORY_SIZED_ACCESS instruction is
selected, and toggles between address and data mode each time a data scan completes with
the busy bit cleared.
The size field is encoded as i Table 36-7.
Table 36-7.
Size Field Semantics
Size field value
00
Access size
Data alignment
Byte (8 bits)
Address modulo 4 : data alignment
0: dddddddd xxxxxxxx xxxxxxxx xxxxxxxx
1: xxxxxxxx dddddddd xxxxxxxx xxxxxxxx
2: xxxxxxxx xxxxxxxx dddddddd xxxxxxxx
3: xxxxxxxx xxxxxxxx xxxxxxxx dddddddd
753
32058K
AVR32-01/12
AT32UC3A
Table 36-7.
Size Field Semantics
Size field value
Access size
Data alignment
Halfword (16 bits)
Address modulo 4 : data alignment
0: dddddddd dddddddd xxxxxxxx xxxxxxxx
1: Not allowed
2: xxxxxxxx xxxxxxxx dddddddd dddddddd
3: Not allowed
10
Word (32 bits)
Address modulo 4 : data alignment
0: dddddddd dddddddd dddddddd dddddddd
1: Not allowed
2: Not allowed
3: Not allowed
11
Reserved
N/A
01
Starting in Run-Test/Idle, SAB data are accessed in the following way:
1. Select the DR Scan path.
2. Scan in the 36-bit address of the data to access, a 2-bit access size, and a direction bit
(1=read, 0=write).
3. Go to Update-DR and re-enter Select-DR Scan.
4. For a read operation, scan out the contents of the addressed area. For a write operation, scan in the new contents of the area.
5. Return to Run-Test/Idle.
For any operation, the full 36 bits of the address must be provided. For write operations, 32 data
bits must be provided, or the result will be undefined. For read operations, shifting may be terminated once the required number of bits have been acquired.
Table 36-8.
36.9.5
MEMORY_SIZED_ACCESS details
Instructions
Details
IR input value
10101 (0x15)
IR output value
peb01
DR Size
39 bits
DR input value (Address phase)
aaaaaaaa aaaaaaaa aaaaaaaa aaaaaaaa aaaassr
DR input value (Data read phase)
xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxx
DR input value (Data write phase)
dddddddd dddddddd dddddddd dddddddd xxxxxxx
DR output value (Address phase)
xxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb
DR output value (Data read phase)
xxxxxeb dddddddd dddddddd dddddddd dddddddd
DR output value (Data write phase)
xxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb
MEMORY_WORD_ACCESS
This instruction allows access to the entire Service Access Bus data area. Data are accessed
through a 34-bit word index, a direction bit, and 32 bits of data. This instruction is identical to
MEMORY_SIZED_ACCESS except that it always does word sized accesses. The size field is
implied, and the two lowest address bits are removed.
754
32058K
AVR32-01/12
AT32UC3A
Note: This instruction was previously known as MEMORY_ACCESS, and is provided for backwards compatibility.
The data register is alternately interpreted by the SAB as an address register and a data register. The SAB starts in address mode after the MEMORY_WORD_ACCESS instruction is
selected, and toggles between address and data mode each time a data scan completes with
the busy bit cleared.
Starting in Run-Test/Idle, SAB data are accessed in the following way:
1. Select the DR Scan path.
2. Scan in the 34-bit address of the data to access, and a direction bit (1=read, 0=write).
3. Go to Update-DR and re-enter Select-DR Scan.
4. For a read operation, scan out the contents of the addressed area. For a write operation, scan in the new contents of the area.
5. Return to Run-Test/Idle.
For any operation, the full 34 bits of the address must be provided. For write operations, 32 data
bits must be provided, or the result will be undefined. For read operations, shifting may be terminated once the required number of bits have been acquired.
Table 36-9.
36.9.6
MEMORY_WORD_ACCESS details
Instructions
Details
IR input value
10001 (0x11)
IR output value
peb01
DR Size
35 bits
DR input value (Address phase)
aaaaaaaa aaaaaaaa aaaaaaaa aaaaaaaa aar
DR input value (Data read phase)
xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xxx
DR input value (Data write phase)
dddddddd dddddddd dddddddd dddddddd xxx
DR output value (Address phase)
xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xeb
DR output value (Data read phase)
xeb dddddddd dddddddd dddddddd dddddddd
DR output value (Data write phase)
xxx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb
MEMORY_BLOCK_ACCESS
This instruction allows access to the entire SAB data area. Up to 32 bits of data are accessed at
a time, while the address is sequentially incremented from the previously used address.
In this mode, the SAB address, size, and access direction is not provided with each access.
Instead, the previous address is auto-incremented depending on the specified size and the previous operation repeated. The address must be set up in advance with
MEMORY_SIZE_ACCESS or MEMORY_WORD_ACCESS. It is allowed, but not required, to
shift data after shifting the address.
This instruction is primarily intended to speed up large quantities of sequential word accesses. It
is possible to use it also for byte and halfword accesses, but the overhead in this is case much
larger as 32 bits must still be shifted for each access.
The following sequence should be used:
755
32058K
AVR32-01/12
AT32UC3A
1. Use the MEMORY_SIZE_ACCESS or MEMORY_WORD_ACCESS to read or write the
first location.
2. Apply MEMORY_BLOCK_ACCESS in the IR Scan path.
3. Select the DR Scan path. The address will now have incremented by 1, 2, or 4 (corresponding to the next byte, halfword, or word location).
4. For a read operation, scan out the contents of the next addressed location. For a write
operation, scan in the new contents of the next addressed location.
5. Go to Update-DR.
6. If the block access is not complete, return to Select-DR Scan and repeat the access.
7. If the block access is complete, return to Run-Test/Idle.
For write operations, 32 data bits must be provided, or the result will be undefined. For read
operations, shifting may be terminated once the required number of bits have been acquired.
Table 36-10. MEMORY_BLOCK_ACCESS details
Instructions
Details
IR input value
10010 (0x12)
IR output value
peb01
DR Size
34 bits
DR input value (Data read phase)
xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx xx
DR input value (Data write phase)
dddddddd dddddddd dddddddd dddddddd xx
DR output value (Data read phase)
eb dddddddd dddddddd dddddddd dddddddd
DR output value (Data write phase)
xx xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxeb
The overhead using block word access is 4 cycles per 32 bits of data, resulting in an 88% transfer efficiency, or 2.1 MBytes per second with a 20 MHz TCK frequency.
36.9.7
CANCEL_ACCESS
If a very slow memory location is accessed during a SAB memory access, it could take a very
long time until the busy bit is cleared, and the SAB becomes ready for the next operation. The
CANCEL_ACCESS instruction provides a possibility to abort an ongoing transfer and report a
timeout to the user.
When the CANCEL_ACCESS instruction is selected, the current access will be terminated as
soon as possible. There are no guarantees about how long this will take, as the hardware may
not always be able to cancel the access immediately. The SAB is ready to respond to a new
command when the busy bit clears.
Table 36-11. CANCEL_ACCESS details
Instructions
Details
IR input value
10011 (0x13)
IR output value
peb01
DR Size
1
DR input value
x
DR output value
0
756
32058K
AVR32-01/12
AT32UC3A
36.9.8
SYNC
This instruction allows external debuggers and testers to measure the ratio between the external
JTAG clock and the internal system clock. The SYNC data register is a 16-bit counter that
counts down to zero using the internal system clock. The busy bit stays high until the counter
reaches zero.
Starting in Run-Test/Idle, SYNC instruction is used in the following way:
1. Select the DR Scan path.
2. Scan in an 16-bit counter value.
3. Go to Update-DR and re-enter Select-DR Scan.
4. Scan out the busy bit, and retry until the busy bit clears.
5. Calculate an approximation to the internal clock speed using the elapsed time and the
counter value.
6. Return to Run-Test/Idle.
The full 16-bit counter value must be provided when starting the synch operation, or the result
will be undefined. When reading status, shifting may be terminated once the required number of
bits have been acquired.
Table 36-12. SYNC_ACCESS details
36.9.9
Instructions
Details
IR input value
10111 (0x17)
IR output value
peb01
DR Size
16 bits
DR input value
dddddddd dddddddd
DR output value
xxxxxxxx xxxxxxeb
AVR_RESET
This instruction allows a debugger or tester to directly control separate reset domains inside the
chip. The shift register contains one bit for each controllable reset domain. Setting a bit to one
resets that domain and holds it in reset. Setting a bit to zero releases the reset for that domain.
See the device specific documentation for the number of reset domains, and what these
domains are.
For any operation, all bits must be provided or the result will be undefined.
Table 36-13. AVR_RESET details
36.9.10
Instructions
Details
IR input value
01100 (0x0C)
IR output value
p0001
DR Size
Device specific.
DR input value
Device specific.
DR output value
Device specific.
CHIP_ERASE
This instruction allows a programmer to completely erase all nonvolatile memories in a chip.
This will also clear any security bits that are set, so the device can be accessed normally. In
757
32058K
AVR32-01/12
AT32UC3A
devices without non-volatile memories this instruction does nothing, and appears to complete
immediately.
The erasing of non-volatile memories starts as soon as the CHIP_ERASE instruction is selected.
The CHIP_ERASE instruction selects a 1 bit bypass data register.
A chip erase operation should be performed as:
1. Scan in the HALT instruction
2. Scan in the value 1 to halt the CPU
3. Stay in Run-Test/Idle for 10 TCK cycles to let the halt command propagate properly
4. Scan in the CHIP_ERASE instruction
5. Keep scanning the CHIP_ERASE instruction until the busy bit is cleared and the protection bit is cleared.
6. Scan in the HALT instruction
7. Scan in the value 0 to release the CPU
8. Return to Run-Test/Idle
9. Stay in Run-Test/Idle for 10 TCK cycles to let the halt command propagate properly.
Table 36-14. CHIP_ERASE details
36.9.11
Instructions
Details
IR input value
01111 (0x0F)
IR output value
p0b01
Where b is the busy bit.
DR Size
1 bit
DR input value
x
DR output value
0
HALT
This instruction allows a programmer to easily stop the CPU to ensure that it does not execute
invalid code during programming.
This instruction selects a 1-bit halt register. Setting this bit to one halts the CPU. Setting this bit
to zero releases the CPU to run normally. The value shifted out from the data register is one if
the CPU is halted.
The HALT instruction can be used in the following way:
10. Scan in the value 1 to halt the CPU
11. Stay in Run-Test/Idle for 10 TCK cycles to let the command propagate properly
12. Use any MEMORY_* instructions to program the device
13. Scan in the HALT instruction
14. Scan in the value 0 to release the CPU
15. Return to Run-Test/Idle
758
32058K
AVR32-01/12
AT32UC3A
16. Stay in Run-Test/Idle for 10 TCK cycles to let the command propagate properly - the
device now runs with the new code.
Table 36-15. HALT details
Instructions
Details
IR input value
11100 (0x1C)
IR output value
p0001
DR Size
1 bit
DR input value
d
DR output value
d
759
32058K
AVR32-01/12
AT32UC3A
36.10 JTAG Data Registers
The following device specific registers can be selected as JTAG scan chain depending on the
instruction loaded in the JTAG Instruction Register. Additional registers exist, but are implicitly
described in the functional description of the relevant instructions.
36.10.1
Device Identification Register
The Device Identification Register contains a unique identifier for each product. The register is
selected by the IDCODE instruction, which is the default instruction after a JTAG reset.
MS
B
Bit
31
Device ID
36.10.1.1
LSB
28
27
12
11
1
0
Revision
Part Number
Manufacturer ID
1
4 bits
16 bits
11 bits
1 bit
Revision
This is a 4 bit number identifying the revision of the component.
Rev A = 0x0, B = 0x1, etc.
Part Number
The part number is a 16 bit code identifying the component.
Manufacturer ID
The Manufacturer ID is a 11 bit code identifying the manufacturer.
The JTAG manufacturer ID for ATMEL is 0x01F.
Device specific ID codes
The different device configurations have different JTAG ID codes, as shown in Table 36-16.
Note that if the flash controller is statically reset, the ID code will be undefined.
Table 36-16. Device and JTAG ID
36.10.2
Device name
JTAG ID code (r is the revision number)
AT32UC3A0512
0xr1EDC03F
AT32UC3A0256
0xr1EDF03F
AT32UC3A0128
0xr1EE203F
AT32UC3A1512
0xr1EDD03F
AT32UC3A1256
0xr1EE003F
AT32UC3A1128
0xr1EE303F
Reset register
The reset register is selected by the AVR_RESET instruction and contains one bit for each reset
domain in the device. Setting each bit to one will keep that domain reset until the bit is cleared.
LSB
Bit
Device ID
4
3
2
1
0
OCD
APP
RESERVED
RESERVED
CPU
760
32058K
AVR32-01/12
AT32UC3A
CPU
CPU
APP
HSB and PB buses
OCD
On-Chip Debug logic and registers
RSERVED
No effect
Note: This register is primarily intended for compatibility with other AVR32 devices. Certain operations may not function correctly when parts of the system are reset. It is generally
recommended to only write 0x11111 or 0x00000 to these bits to ensure no unintended side
effects occur.
36.10.3
Boundary-Scan Chain
The Boundary-Scan Chain has the capability of driving and observing the logic levels on the digital I/O pins, as well as driving and observing the logic levels between the digital I/O pins and the
internal logic. Typically, output value, output enable, and input data are all available in the
boundary scan chain.
The boundary scan chain is described in the BDSL (Boundary Scan Description Language) file
available at the Atmel web site.
36.11 SAB address map
The Service Access Bus (SAB) gives the user access to the internal address space and other
features through a 36 bits address space. The 4 MSBs identify the slave number, while the 32
LSBs are decoded within the slave’s address space. The SAB slaves are shown in Table 36-17.
Table 36-17. SAB Slaves, addresses and descriptions.
Slave
Address [35:32]
Description
Unallocated
0x0
Intentionally unallocated
OCD
0x1
OCD registers
HSB
0x4
HSB memory space, as seen by the CPU
HSB
0x5
Alternative mapping for HSB space, for compatibility with
other AVR32 devices.
Memory Service
Unit
0x6
Memory Service Unit registers
Reserved
Other
Unused
761
32058K
AVR32-01/12
AT32UC3A
37. Boot Sequence
This chapter summarizes the boot sequence of the AT32UC3A. The behaviour after power-up is
controlled by the Power Manager. For specific details, refer to Section 13. ”Power Manager
(PM)” on page 53.
37.1
Starting of clocks
After power-up, the device will be held in a reset state by the Power-On Reset circuitry, until the
power has stabilized throughout the device. Once the power has stabilized, the device will use
the internal RC Oscillator as clock source.
On system start-up, the PLLs are disabled. All clocks to all modules are running. No clocks have
a divided frequency, all parts of the system recieves a clock with the same frequency as the
internal RC Oscillator.
37.2
Fetching of initial instructions
After reset has been released, the AVR32 UC CPU starts fetching instructions from the reset
address, which is 0x8000_0000. This address points to the first address in the internal Flash.
The code read from the internal Flash is free to configure the system to use for example the
PLLs, to divide the frequency of the clock routed to some of the peripherals, and to gate the
clocks to unused peripherals.
762
32058K
AVR32-01/12
AT32UC3A
38. Electrical Characteristics
38.1
Absolute Maximum Ratings*
Operating Temperature.
Storage Temperature .
-60°C to +150°C
Voltage on Input Pin
with respect to Ground except for PC00, PC01, PC02, PC03,
PC04, PC05..........................................................-0.3V to 5.5V
Voltage on Input Pin
with respect to Ground for PC00, PC01, PC02, PC03, PC04,
PC05.....................................................................-0.3V to 3.6V
Maximum Operating Voltage (VDDCORE, VDDPLL) .
*NOTICE:
-40⋅C to +85⋅C
Stresses beyond those listed under “Absolute
Maximum Ratings” may cause permanent damage to the device. This is a stress rating only and
functional operation of the device at these or
other conditions beyond those indicated in the
operational sections of this specification is not
implied. Exposure to absolute maximum rating
conditions for extended periods may affect
device reliability.
1.95V
Maximum Operating Voltage (VDDIO, VDDIN, VDDANA).3.6V
Total DC Output Current on all I/O Pin
for TQFP100 package .
for LQGP144 package .
370 mA
470 mA
763
32058K
AVR32-01/12
AT32UC3A
38.2
DC Characteristics
The following characteristics are applicable to the operating temperature range: TA = -40°C to 85°C, unless otherwise specified and are certified for a junction temperature up to TJ = 100°C.
DC Characteristics
Table 38-1.
Symbol
Parameter
VVDDCOR
Max
Units
DC Supply Core
1.65
1.95
V
VVDDPLL
DC Supply PLL
1.65
1.95
V
VVDDIO
DC Supply Peripheral I/Os
3.0
3.6
V
VREF
Analog reference voltage
2.6
3.6
V
VIL
Input Low-level Voltage
-0.3
+0.8
V
All GPIOS except for PC00, PC01, PC02,
PC03, PC04, PC05.
2.0
5.5V
V
PC00, PC01, PC02, PC03, PC04, PC05.
2.0
3.6V
V
IOL=-4mA for PA0-PA20, PB0, PB4-PB9,
PB11-PB18, PB24-PB26, PB29-PB31,
PX0-PX39
0.4
V
IOL=-8mA for PA21-PA30, PB1-PB3,
PB10, PB19-PB23, PB27-PB28, PC0PC5
0.4
V
E
Input High-level Voltage
VIH
Output Low-level Voltage
VOL
Condition
IOH=4mA for PA0-PA20, PB0, PB4-PB9,
PB11-PB18, PB24-PB26, PB29-PB31,
PX0-PX39
Min.
Typ.
VVDDIO0.4
V
VVDDIO0.4
V
Output High-level Voltage
VOH
IOH=8mA for PA21-PA30, PB1-PB3,
PB10, PB19-PB23, PB27-PB28, PC0PC5
IOL
Output Low-level Current
IOH
Output HIgh-level Current
ILEAK
Input Leakage Current
PA0-PA20, PB0, PB4-PB9, PB11-PB18,
PB24-PB26, PB29-PB31, PX0-PX39
-4
mA
PA21-PA30, PB1-PB3, PB10, PB19PB23, PB27-PB28, PC0-PC5
-8
mA
PA0-PA20, PB0, PB4-PB9, PB11-PB18,
PB24-PB26, PB29-PB31, PX0-PX39
4
mA
PA21-PA30, PB1-PB3, PB10, PB19PB23, PB27-PB28, PC0-PC5
8
mA
Pullup resistors disabled
1
µA
CIN
Input Capacitance
TQFP100 Package
7
pF
LQFP144 Package
7
pF
RPULLUP
All GPIO and RESET_N pin.
15K
Ohm
Pull-up Resistance
10K
764
32058K
AVR32-01/12
AT32UC3A
38.3
Regulator characteristics
Table 38-2.
Symbol
Parameter
VVDDIN
Supply voltage (input)
VVDDOUT
Supply voltage (output)
IOUT
ISCR
Electrical characteristics
Condition
Min.
Typ.
Max.
Units
3
3.3
3.6
V
1.81
1.85
1.89
V
Maximum DC output current with VVDDIN = 3.3V
100
mA
Maximum DC output current with VVDDIN = 2.7V
90
mA
Low Power mode (stop, deep stop
or static) at TA =25°C
Static Current of internal regulator
Table 38-3.
Symbol
Parameter
CIN1
10
µA
Decoupling requirements
Typ.
Techno.
Units
Input Regulator Capacitor 1
1
NPO
nF
CIN2
Input Regulator Capacitor 2
4.7
X7R
uF
COUT1
Output Regulator Capacitor 1
470
NPO
pF
COUT2
Output Regulator Capacitor 2
2.2
X7R
uF
38.4
Condition
Analog characteristics
Table 38-4.
Symbol
Parameter
VADVREF
Analog voltage reference (input)
Condition
Table 38-5.
Min.
Typ.
Max.
Units
3.6
V
2.6
Decoupling requirements
Typ.
Techno
.
Units
Voltage reference Capacitor 1
10
-
nF
Voltage reference Capacitor 2
1
-
uF
Symbol
Parameter
CVREF1
CVREF2
38.4.1
Electrical characteristics
Condition
BOD
Table 38-6.
BODLEVEL Values
BODLEVEL Value
Typ.
Typ.
Typ.
Units.
00 0000b
1.40
1.47
1.55
V
01 0111b
1.45
1.52
1.6
V
01 1111b
1.55
1.6
1.65
V
10 0111b
1.65
1.69
1.75
V
The values in Table 38-6 describes the values of the BODLEVEL in the flash FGPFR register.
765
32058K
AVR32-01/12
AT32UC3A
Table 38-7.
38.4.2
BOD Timing
Symbol
Parameter
Test Conditions
Typ.
Max.
Units.
TBOD
Minimum time with
VDDCORE < VBOD to
detect power failure
Falling VDDCORE
from 1.8V to 1.1V
300
800
ns
POR
Table 38-8.
Electrical Characteristic
Symbol
Parameter
Test Conditions
VDDRR
VDDCORE rise rate to ensure power-on-reset
0.01
VSSFR
VDDCORE fall rate to ensure power-on-reset
0.01
VPOR+
Rising threshold voltage: voltage up to which
device is kept under reset by POR on rising
VDDCORE
VPOR-
Falling threshold voltage: voltage when POR
resets device on falling VDDCORE
VRESTART
On falling VDDCORE, voltage must go down to
this value before supply can rise again to ensure
reset signal is released at VPOR+
Falling VDDCORE:
1.8V -> VRESTART
TPOR
Minimum time with VDDCORE < VPOR-
Falling VDDCORE:
1.8V -> 1.1V
TRST
Time for reset signal to be propagated to system
Rising VDDCORE:
VRESTART -> VPOR+
Falling VDDCORE:
1.8V -> VPOR+
Min.
Typ.
Max.
Units.
V/ms
400
V/ms
1.35
1.5
1.6
V
1.25
1.3
1.4
V
0.5
V
-0.1
15
200
us
400
us
766
32058K
AVR32-01/12
AT32UC3A
38.5
Power Consumption
The values in Table 38-9 and Table 38-10 on page 769 are measured values of power consumption with operating conditions as follows:
•VDDIO = 3.3V
•VDDCORE = VDDPLL = 1.8V
•TA = 25°C, TA = 85°C
•I/Os are configured in input, pull-up enabled.
Figure 38-1. Measurement setup
VDDANA
VDDIO
Amp0
VDDIN
Internal
Voltage
Regulator
VDDOUT
Amp1
VDDCORE
VDDPLL
767
32058K
AVR32-01/12
AT32UC3A
These figures represent the power consumption measured on the power supplies.
Table 38-9.
Power Consumption for Different Modes
Mode
Conditions
Active
Typ : Ta =25 °C
CPU running from flash (1).
VDDIN=3.3 V. VDDCORE =1.8V.
CPU clocked from PLL0 at f MHz
Voltage regulator is on.
XIN0 : external clock. (1)
XIN1 stopped. XIN32 stopped
PLL0 running
All peripheral clocks activated.
GPIOs on internal pull-up.
JTAG unconnected with ext pull-up.
Idle
Frozen
Standby
Typ.
Unit
f = 12 MHz
9
mA
f = 24 MHz
15
mA
f = 36MHz
20
mA
f = 50 MHz
28
mA
f = 66 MHz
36.3
mA
Typ : Ta = 25 °C
CPU running from flash (1).
VDDIN=3.3 V. VDDCORE =1.8V.
CPU clocked from PLL0 at f MHz
Voltage regulator is on.
XIN0 : external clock.
XIN1 stopped. XIN32 stopped
PLL0 running
All peripheral clocks activated.
GPIOs on internal pull-up.
JTAG unconnected with ext pull-up.
f = 12 MHz
5
mA
f = 24 MHz
10
mA
f = 36MHz
14
mA
f = 50 MHz
19
mA
f = 66 MHz
25.5
mA
Typ : Ta = 25 °C
CPU running from flash (1).
CPU clocked from PLL0 at f MHz
Voltage regulator is on.
XIN0 : external clock.
XIN1 stopped. XIN32 stopped
PLL0 running
All peripheral clocks activated.
GPIOs on internal pull-up.
JTAG unconnected with ext pull-up.
f = 12 MHz
3
mA
f = 24 MHz
6
mA
f = 36MHz
9
mA
f = 50 MHz
13
mA
f = 66 MHz
16.8
mA
Typ : Ta = 25 °C
CPU running from flash (1).
CPU clocked from PLL0 at f MHz
Voltage regulator is on.
XIN0 : external clock.
XIN1 stopped. XIN32 stopped
PLL0 running
All peripheral clocks activated.
GPIOs on internal pull-up.
JTAG unconnected with ext pull-up.
f = 12 MHz
1
mA
f = 24 MHz
2
mA
f = 36MHz
3
mA
f = 50 MHz
4
mA
f = 66 MHz
4.8
mA
768
32058K
AVR32-01/12
AT32UC3A
Table 38-9.
Power Consumption for Different Modes
Mode
Conditions
Typ.
Unit
Stop
Typ : Ta = 25 °C.
CPU is in stop mode
GPIOs on internal pull-up.
All peripheral clocks de-activated.
DM and DP pins connected to ground.
XIN0,Xin1 and XIN2 are stopped
on Amp0
47
uA
on Amp1
40
uA
on Amp0
36
uA
Deepstop
Typ : Ta = 25 °C.CPU is in deepstop mode
GPIOs on internal pull-up.
All peripheral clocks de-activated.
DM and DP pins connected to ground.
XIN0,Xin1 and XIN2 are stopped
on Amp1
28
uA
on Amp0
25
uA
Static
Typ : Ta = 25 °C. CPU is in static mode
GPIOs on internal pull-up.
All peripheral clocks de-activated.
DM and DP pins connected to ground.
XIN0,Xin1 and XIN2 are stopped
on Amp1
14
uA
1.
Core frequency is generated from XIN0 using the PLL so that 140 MHz < fpll0 < 160 MHz and 10 MHz < fxin0
< 12MHz
Table 38-10. Power Consumption by Peripheral in Active Mode
Peripheral
38.6
Typ.
GPIO
37
SMC
10
SDRAMC
4
ADC
18
EBI
31
INTC
25
TWI
14
MACB
45
PDCA
30
PWM
36
RTC
7
SPI
13
SSC
13
TC
10
USART
35
USB
45
Unit
µA/MHz
Clock Characteristics
These parameters are given in the following conditions:
769
32058K
AVR32-01/12
AT32UC3A
• VDDCORE = 1.8V
• Ambient Temperature = 25°C
38.6.1
CPU/HSB Clock Characteristics
Table 38-11. Core Clock Waveform Parameters
Symbol
Parameter
1/(tCPCPU)
CPU Clock Frequency
tCPCPU
CPU Clock Period
38.6.2
Conditions
Min
Max
Units
66
MHz
15,15
ns
PBA Clock Characteristics
Table 38-12. PBA Clock Waveform Parameters
Symbol
Parameter
1/(tCPPBA)
PBA Clock Frequency
tCPPBA
PBA Clock Period
38.6.3
Conditions
Min
Max
Units
66
MHz
15,15
ns
PBB Clock Characteristics
Table 38-13. PBB Clock Waveform Parameters
Symbol
Parameter
1/(tCPPBB)
PBB Clock Frequency
tCPPBB
PBB Clock Period
38.7
Conditions
Min
Max
Units
66
MHz
15,15
ns
Crystal Oscillator Characteristis
The following characteristics are applicable to the operating temperature range: TA = -40°C to 85°C and worst case of
power supply, unless otherwise specified.
38.7.1
32 KHz Oscillator Characteristics
Table 38-14. 32 KHz Oscillator Characteristics
Symbol
Parameter
1/(tCP32KHz)
Crystal Oscillator Frequency
CL
Equivalent Load Capacitance
tST
Startup Time
IOSC
Current Consumption
Note:
Conditions
Min
Max
Unit
32 768
Hz
12.5
pF
600
1200
ms
Active mode
1.8
µA
Standby mode
0.1
µA
6
(1)
CL = 6pF
CL = 12.5pF(1)
Typ
1. CL is the equivalent load capacitance.
770
32058K
AVR32-01/12
AT32UC3A
38.7.2
Main Oscillators Characteristics
Table 38-15. Main Oscillator Characteristics
Symbol
Parameter
1/(tCPMAIN)
Crystal Oscillator Frequency
CL1, CL2
Internal Load Capacitance
(CL1 = CL2)
Conditions
Min
0.45
Max
Unit
16
MHz
12
Duty Cycle
40
tST
Startup Time
1/(tCPXIN)
XIN Clock Frequency
tCHXIN
XIN Clock High Half-period
0.4 x
tCPXIN
tCLXIN
XIN Clock Low Half-period
0.4 x
tCPXIN
CIN
XIN Input Capacitance
38.7.3
Typ
50
8 MHz
External clock
Crystal
0.45
pF
60
%
4
ms
50
MHz
16
MHz
0.6 x
tCPXIN
0.6 x
tCPXIN
7
pF
PLL Characteristics
Table 38-16. Phase Lock Loop Characteristics
Symbol
Parameter
FOUT
Output Frequency
FIN
Input Frequency
IPLL
Current Consumption
Conditions
Min
Typ
80
4
Max
Unit
240
MHz
16
MHz
active mode (Fout=80Mhz)
250
µA
active mode (Fout=240Mhz)
600
µA
771
32058K
AVR32-01/12
AT32UC3A
38.8
ADC Characteristics
Table 38-17. Channel Conversion Time and ADC Clock
Parameter
Conditions
ADC Clock Frequency
ADC Clock Frequency
Startup Time
Min
Max
Units
10-bit resolution mode
5
MHz
8-bit resolution mode
8
MHz
Return from Idle Mode
20
µs
Track and Hold Acquisition Time
Typ
600
Conversion Time
ADC Clock = 5 MHz
ns
2
µs
Conversion Time
ADC Clock = 8 MHz
1.25
µs
Throughput Rate
ADC Clock = 5 MHz
384(1)
kSPS
Throughput Rate
ADC Clock = 8 MHz
533(2)
kSPS
Notes:
1. Corresponds to 13 clock cycles at 5 MHz: 3 clock cycles for track and hold acquisition time and 10 clock cycles for
conversion.
2. Corresponds to 15 clock cycles at 8 MHz: 5 clock cycles for track and hold acquisition time and 10 clock cycles for
conversion.
Table 38-18. External Voltage Reference Input
Parameter
Conditions
Min
ADVREF Input Voltage Range
ADVREF Average Current
Typ
2.6
On 13 samples with ADC Clock = 5 MHz
200
Current Consumption on VDDANA
Note:
Max
Units
VDDANA
V
250
µA
1.25
mA
ADVREF should be connected to GND to avoid extra consumption in case ADC is not used.
Table 38-19. Analog Inputs
Parameter
Min
Input Voltage Range
Typ
0
Input Leakage Current
Max
1
Input Capacitance
Units
VADVREF
µA
17
pF
Table 38-20. Transfer Characteristics in 8-bit mode
Parameter
Conditions
Min
Resolution
Absolute Accuracy
Integral Non-linearity
Differential Non-linearity
Typ
Max
8
Units
Bit
f=5MHz
0.8
LSB
f=8MHz
1.5
LSB
f=5MHz
0.35
0.5
LSB
f=8MHz
0.5
1.0
LSB
f=5MHz
0.3
0.5
LSB
f=8MHz
0.5
1.0
LSB
Offset Error
f=5MHz
-0.5
0.5
LSB
Gain Error
f=5MHz
-0.5
0.5
LSB
772
32058K
AVR32-01/12
AT32UC3A
Table 38-21. Transfer Characteristics in 10-bit mode
Parameter
Conditions
Min
Resolution
Typ
Max
10
Units
Bit
Absolute Accuracy
f=5MHz
Integral Non-linearity
f=5MHz
1.5
2
LSB
f=5MHz
1
2
LSB
0.6
1
LSB
Differential Non-linearity
3
f=2.5MHz
LSB
Offset Error
f=5MHz
-2
2
LSB
Gain Error
f=5MHz
-2
2
LSB
773
32058K
AVR32-01/12
AT32UC3A
38.9
EBI Timings
These timings are given for worst case process, T = 85⋅C, VDDCORE = 1.65V, VDDIO = 3V and 40 pF load capacitance.
Table 38-22. SMC Clock Signal.
Symbol
Parameter
1/(tCPSMC)
SMC Controller Clock Frequency
Note:
Max(1)
Units
1/(tcpcpu)
MHz
1. The maximum frequency of the SMC interface is the same as the max frequency for the HSB.
Table 38-23. SMC Read Signals with Hold Settings
Symbol
Parameter
Min
Units
NRD Controlled (READ_MODE = 1)
SMC1
Data Setup before NRD High
SMC2
Data Hold after NRD High
12
0
SMC3
NRD High to NBS0/A0 Change
SMC4
NRD High to NBS1 Change(1)
(1)
nrd hold length * tCPSMC - 1.3
nrd hold length * tCPSMC - 1.3
SMC5
NRD High to NBS2/A1 Change
SMC6
NRD High to NBS3 Change(1)
(1)
nrd hold length * tCPSMC - 1.3
ns
nrd hold length * tCPSMC - 1.3
SMC7
NRD High to A2 - A25 Change
SMC8
NRD High to NCS Inactive(1)
SMC9
NRD Pulse Width
(1)
nrd hold length * tCPSMC - 1.3
(nrd hold length - ncs rd hold length) * tCPSMC - 2.3
nrd pulse length * tCPSMC - 1.4
NRD Controlled (READ_MODE = 0)
SMC10
Data Setup before NCS High
SMC11
Data Hold after NCS High
11.5
0
SMC12
NCS High to NBS0/A0 Change
(1)
ncs rd hold length * tCPSMC - 2.3
SMC13
NCS High to NBS0/A0 Change(1)
ncs rd hold length * tCPSMC - 2.3
SMC14
NCS High to NBS2/A1 Change
(1)
ncs rd hold length * tCPSMC - 2.3
SMC15
NCS High to NBS3 Change(1)
SMC16
NCS High to A2 - A25 Change
SMC17
NCS High to NRD Inactive(1)
SMC18
NCS Pulse Width
Note:
ns
ncs rd hold length * tCPSMC - 2.3
(1)
ncs rd hold length * tCPSMC - 4
ncs rd hold length - nrd hold length)* tCPSMC - 1.3
ncs rd pulse length * tCPSMC - 3.6
1. hold length = total cycle duration - setup duration - pulse duration. “hold length” is for “ncs rd hold length” or “nrd hold length”.
774
32058K
AVR32-01/12
AT32UC3A
Table 38-24. SMC Read Signals with no Hold Settings
Symbol
Parameter
Min
Units
NRD Controlled (READ_MODE = 1)
SMC19
Data Setup before NRD High
SMC20
Data Hold after NRD High
13.7
1
ns
NRD Controlled (READ_MODE = 0)
SMC21
Data Setup before NCS High
SMC22
Data Hold after NCS High
13.3
0
ns
Table 38-25. SMC Write Signals with Hold Settings
Symbol
Parameter
Min
Units
NRD Controlled (READ_MODE = 1)
SMC23
Data Out Valid before NWE High
(nwe pulse length - 1) * tCPSMC - 0.9
SMC24
Data Out Valid after NWE High(1)
nwe hold length * tCPSMC - 6
SMC25
NWE High to NBS0/A0 Change
nwe hold length * tCPSMC - 1.9
SMC26
NWE High to NBS1 Change(1)
SMC29
NWE High to NBS2/A1 Change
SMC30
NWE High to NBS3 Change(1)
SMC31
NWE High to A2 - A25 Change
SMC32
NWE High to NCS Inactive(1)
SMC33
NWE Pulse Width
(1)
nwe hold length * tCPSMC - 1.9
(1)
nwe hold length * tCPSMC - 1.9
ns
nwe hold length * tCPSMC - 1.9
(1)
nwe hold length * tCPSMC - 1.7
(nwe hold length - ncs wr hold length)* tCPSMC - 2.9
nwe pulse length * tCPSMC - 0.9
NRD Controlled (READ_MODE = 0)
SMC34
Data Out Valid before NCS High
(ncs wr pulse length - 1)* tCPSMC - 4.6
SMC35
Data Out Valid after NCS High(1)
ncs wr hold length * tCPSMC - 5.8
SMC36
NCS High to NWE Inactive
Note:
(1)
ns
(ncs wr hold length - nwe hold length)* tCPSMC - 0.6
1. hold length = total cycle duration - setup duration - pulse duration. “hold length” is for “ncs wr hold length” or “nwe hold
length"
775
32058K
AVR32-01/12
AT32UC3A
Table 38-26. SMC Write Signals with No Hold Settings (NWE Controlled only).
Symbol
Parameter
Min
SMC37
NWE Rising to A2-A25 Valid
5.4
SMC38
NWE Rising to NBS0/A0 Valid
5
SMC39
NWE Rising to NBS1 Change
5
SMC40
NWE Rising to A1/NBS2 Change
5
SMC41
NWE Rising to NBS3 Change
5
SMC42
NWE Rising to NCS Rising
SMC43
Data Out Valid before NWE Rising
SMC44
Data Out Valid after NWE Rising
SMC45
NWE Pulse Width
Units
ns
5.1
(nwe pulse length - 1) * tCPSMC - 1.2
5
nwe pulse length * tCPSMC - 0.9
Figure 38-2. SMC Signals for NCS Controlled Accesses.
SMC16
SMC16
SMC16
SMC12
SMC13
SMC14
SMC15
SMC12
SMC13
SMC14
SMC15
A2-A25
SMC12
SMC13
SMC14
SMC15
A0/A1/NBS[3:0]
NRD
SMC17
SMC17
NCS
SMC21
SMC18
SMC18
SMC18
SMC22
SMC10
SMC11
SMC34
SMC35
D0 - D15
SMC36
NWE
776
32058K
AVR32-01/12
AT32UC3A
Figure 38-3. SMC Signals for NRD and NRW Controlled Accesses.
SMC37
SMC7
SMC7
SMC31
A2-A25
SMC25
SMC26
SMC29
SMC30
SMC3
SMC4
SMC5
SMC6
SMC38
SMC39
SMC40
SMC41
SMC3
SMC4
SMC5
SMC6
A0/A1/NBS[3:0]
SMC42
SMC32
SMC8
NCS
SMC8
SMC9
SMC9
NRD
SMC19
SMC20
SMC43
SMC44
SMC1
SMC23
SMC2
SMC24
D0 - D15
SMC45
SMC33
NWE
38.9.1
SDRAM Signals
These timings are given for 10 pF load on SDCK and 40 pF on other signals.
Table 38-27. SDRAM Clock Signal.
Symbol
Parameter
1/(tCPSDCK)
SDRAM Controller Clock Frequency
Note:
Max(1)
Units
1/(tcpcpu)
MHz
1. The maximum frequency of the SDRAMC interface is the same as the max frequency for the
HSB.
Table 38-28. SDRAM Clock Signal.
Symbol
Parameter
Min
Units
SDRAMC1
SDCKE High before SDCK Rising Edge
7.4
ns
SDRAMC2
SDCKE Low after SDCK Rising Edge
3.2
SDRAMC3
SDCKE Low before SDCK Rising Edge
SDRAMC4
SDCKE High after SDCK Rising Edge
2.9
SDRAMC5
SDCS Low before SDCK Rising Edge
7.5
SDRAMC6
SDCS High after SDCK Rising Edge
1.6
SDRAMC7
RAS Low before SDCK Rising Edge
7.2
SDRAMC8
RAS High after SDCK Rising Edge
2.3
SDRAMC9
SDA10 Change before SDCK Rising Edge
7.6
SDRAMC10
SDA10 Change after SDCK Rising Edge
1.9
7
777
32058K
AVR32-01/12
AT32UC3A
Table 38-28. SDRAM Clock Signal.
Symbol
Parameter
Min
SDRAMC11
Address Change before SDCK Rising Edge
6.2
SDRAMC12
Address Change after SDCK Rising Edge
2.2
SDRAMC13
Bank Change before SDCK Rising Edge
6.3
SDRAMC14
Bank Change after SDCK Rising Edge
2.4
SDRAMC15
CAS Low before SDCK Rising Edge
7.4
SDRAMC16
CAS High after SDCK Rising Edge
1.9
SDRAMC17
DQM Change before SDCK Rising Edge
6.4
SDRAMC18
DQM Change after SDCK Rising Edge
2.2
SDRAMC19
D0-D15 in Setup before SDCK Rising Edge
9
SDRAMC20
D0-D15 in Hold after SDCK Rising Edge
0
SDRAMC23
SDWE Low before SDCK Rising Edge
7.6
SDRAMC24
SDWE High after SDCK Rising Edge
1.8
SDRAMC25
D0-D15 Out Valid before SDCK Rising Edge
7.1
SDRAMC26
D0-D15 Out Valid after SDCK Rising Edge
1.5
Units
ns
778
32058K
AVR32-01/12
AT32UC3A
Figure 38-4. SDRAMC Signals relative to SDCK.
SDCK
SDRAMC1
SDRAMC2
SDRAMC3
SDRAMC4
SDCKE
SDRAMC5
SDRAMC6
SDRAMC7
SDRAMC8
SDRAMC5
SDRAMC6
SDRAMC5
SDRAMC6
SDCS
RAS
SDRAMC15 SDRAMC16
SDRAMC15 SDRAMC16
CAS
SDRAMC23 SDRAMC24
SDWE
SDRAMC9 SDRAMC10
SDRAMC9 SDRAMC10
SDRAMC9 SDRAMC10
SDRAMC11 SDRAMC12
SDRAMC11 SDRAMC12
SDRAMC11 SDRAMC12
SDRAMC13 SDRAMC14
SDRAMC13 SDRAMC14
SDRAMC13 SDRAMC14
SDRAMC17 SDRAMC18
SDRAMC17 SDRAMC18
SDA10
A0 - A9,
A11 - A13
BA0/BA1
DQM0 DQM3
SDRAMC19 SDRAMC20
D0 - D15
Read
SDRAMC25 SDRAMC26
D0 - D15
to Write
779
32058K
AVR32-01/12
AT32UC3A
38.10 JTAG Timings
38.10.1
JTAG Interface Signals
Table 38-29. JTAG Interface Timing specification
Symbol
JTAG0
JTAG1
JTAG2
JTAG3
JTAG4
JTAG5
JTAG6
JTAG7
JTAG8
JTAG9
JTAG10
Note:
Parameter
Conditions
Min
TCK Low Half-period
(1)
Max
6
ns
TCK High Half-period
(1)
3
ns
TCK Period
(1)
9
ns
TDI, TMS Setup before TCK High
(1)
1
ns
TDI, TMS Hold after TCK High
(1)
0
ns
TDO Hold Time
(1)
4
ns
TCK Low to TDO Valid
(1)
Device Inputs Setup Time
(1)
ns
Device Inputs Hold Time
(1)
ns
Device Outputs Hold Time
(1)
ns
TCK to Device Outputs Valid
(1)
ns
6
Units
ns
1. VVDDIO from 3.0V to 3.6V, maximum external capacitor = 40pF
780
32058K
AVR32-01/12
AT32UC3A
Figure 38-5. JTAG Interface Signals
JTAG2
TCK
JTAG
JTAG1
0
TMS/TDI
JTAG3
JTAG4
JTAG7
JTAG8
TDO
JTAG5
JTAG6
Device
Inputs
Device
Outputs
JTAG9
JTAG10
38.11 SPI Characteristics
Figure 38-6. SPI Master mode with (CPOL = NCPHA = 0) or (CPOL= NCPHA= 1)
SPCK
SPI0
SPI1
MISO
SPI2
MOSI
781
32058K
AVR32-01/12
AT32UC3A
Figure 38-7. SPI Master mode with (CPOL=0 and NCPHA=1) or (CPOL=1 and NCPHA=0)
SPCK
SPI3
SPI4
MISO
SPI5
MOSI
Figure 38-8. SPI Slave mode with (CPOL=0 and NCPHA=1) or (CPOL=1 and NCPHA=0)
SPCK
SPI6
MISO
SPI7
SPI8
MOSI
Figure 38-9. SPI Slave mode with (CPOL = NCPHA = 0) or (CPOL= NCPHA= 1)
SPCK
SPI9
MISO
SPI10
SPI11
MOSI
782
32058K
AVR32-01/12
AT32UC3A
Table 38-30. SPI Timings
Symbol
SPI0
Parameter
MISO Setup time before SPCK rises (master)
SPI1
MISO Hold time after SPCK rises (master)
SPI2
SPCK rising to MOSI Delay (master)
Conditions
(1)
3.3V domain
3.3V domain
(1)
3.3V domain
(1)
(1)
SPI3
MISO Setup time before SPCK falls (master)
3.3V domain
SPI4
MISO Hold time after SPCK falls (master)
3.3V domain (1)
SPI5
SPCK falling to MOSI Delay (master)
3.3V domain (1)
SPI6
SPCK falling to MISO Delay (slave)
SPI7
MOSI Setup time before SPCK rises (slave)
SPI8
MOSI Hold time after SPCK rises (slave)
SPI9
SPCK rising to MISO Delay (slave)
SPI10
MOSI Setup time before SPCK falls (slave)
SPI11
Notes:
MOSI Hold time after SPCK falls (slave)
Min
Max
Units
(2)
22 + (tCPMCK)/2
ns
0
ns
7
ns
(2)
22 + (tCPMCK)/2
ns
0
ns
7
ns
26.5
ns
3.3V domain
(1)
3.3V domain
(1)
0
ns
3.3V domain
(1)
1.5
ns
3.3V domain
(1)
3.3V domain
(1)
0
ns
3.3V domain
(1)
1
ns
27
ns
1. 3.3V domain: VVDDIO from 3.0V to 3.6V, maximum external capacitor = 40 pF.
2. tCPMCK: Master Clock period in ns.
38.12 MACB Characteristics
Table 38-31. Ethernet MAC Signals
Symbol
Parameter
Conditions
EMAC1
Setup for EMDIO from EMDC rising
Load: 20pF
EMAC2
Hold for EMDIO from EMDC rising
Load: 20pF(2)
EMAC3
EMDIO toggling from EMDC falling
Load: 20pF(2)
Notes:
Min (ns)
Max (ns)
Min (ns)
Max (ns)
(2)
1. f: MCK frequency (MHz)
2. VVDDIO from 3.0V to 3.6V, maximum external capacitor = 20 pF
Table 38-32. Ethernet MAC MII Specific Signals
Symbol
EMAC4
EMAC5
Parameter
Setup for ECOL from ETXCK rising
Hold for ECOL from ETXCK rising
Conditions
Load: 20pF
(1)
3
Load: 20pF
(1)
0
(1)
3
0
EMAC6
Setup for ECRS from ETXCK rising
Load: 20pF
EMAC7
Hold for ECRS from ETXCK rising
Load: 20pF (1)
EMAC8
EMAC9
ETXER toggling from ETXCK rising
ETXEN toggling from ETXCK rising
Load: 20pF
(1)
15
Load: 20pF
(1)
15
(1)
15
EMAC10
ETX toggling from ETXCK rising
Load: 20pF
EMAC11
Setup for ERX from ERXCK
Load: 20pF (1)
1
783
32058K
AVR32-01/12
AT32UC3A
Table 38-32. Ethernet MAC MII Specific Signals
Symbol
EMAC12
EMAC13
Parameter
Conditions
Hold for ERX from ERXCK
Setup for ERXER from ERXCK
Min (ns)
Load: 20pF
(1)
1.5
Load: 20pF
(1)
1
(1)
0.5
EMAC14
Hold for ERXER from ERXCK
Load: 20pF
EMAC15
Setup for ERXDV from ERXCK
Load: 20pF (1)
1.5
EMAC16
Hold for ERXDV from ERXCK
Load: 20pF (1)
1
Note:
Max (ns)
1. VVDDIO from 3.0V to 3.6V, maximum external capacitor = 20 pF
Figure 38-10. Ethernet MAC MII Mode
EMDC
EMAC1
EMAC3
EMAC2
EMDIO
EMAC4
EMAC5
EMAC6
EMAC7
ECOL
ECRS
ETXCK
EMAC8
ETXER
EMAC9
ETXEN
EMAC10
ETX[3:0]
ERXCK
EMAC11
EMAC12
ERX[3:0]
EMAC13
EMAC14
EMAC15
EMAC16
ERXER
ERXDV
784
32058K
AVR32-01/12
AT32UC3A
Table 38-33. Ethernet MAC RMII Specific Signals
Symbol
Parameter
Min (ns)
Max (ns)
EMAC21
ETXEN toggling from EREFCK rising
7
14.5
EMAC22
ETX toggling from EREFCK rising
7
14.7
EMAC23
Setup for ERX from EREFCK
1.5
EMAC24
Hold for ERX from EREFCK
0
EMAC25
Setup for ERXER from EREFCK
1.5
EMAC26
Hold for ERXER from EREFCK
0
EMAC27
Setup for ECRSDV from EREFCK
1.5
EMAC28
Hold for ECRSDV from EREFCK
0
Figure 38-11. Ethernet MAC RMII Mode
EREFCK
EMAC21
ETXEN
EMAC22
ETX[1:0]
EMAC23
EMAC24
ERX[1:0]
EMAC25
EMAC26
EMAC27
EMAC28
ERXER
ECRSDV
38.13 Flash Characteristics
The following table gives the device maximum operating frequency depending on the field FWS
of the Flash FSR register. This field defines the number of wait states required to access the
Flash Memory.
Table 38-34.
Flash Wait States
FWS
Read Operations
Maximum Operating Frequency (MHz)
0
1 cycle
33
1
2 cycles
66
785
32058K
AVR32-01/12
AT32UC3A
Table 38-35.
Programming Time
Temperature Operating Range
Part
Page Programming Time (ms)
Chip Erase Time (ms)
Industrial
4
4
Automotive
16
16
786
32058K
AVR32-01/12
AT32UC3A
39. Mechanical Characteristics
39.1
Thermal Considerations
39.1.1
Thermal Data
Table 39-1 summarizes the thermal resistance data depending on the package.
Table 39-1.
39.1.2
Thermal Resistance Data
Symbol
Parameter
Condition
Package
Typ
θJA
Junction-to-ambient thermal resistance
Still Air
TQFP100
43.4
θJC
Junction-to-case thermal resistance
TQFP100
5.5
θJA
Junction-to-ambient thermal resistance
LQFP144
39.8
θJC
Junction-to-case thermal resistance
LQFP144
8.9
Still Air
Unit
⋅C/W
⋅C/W
Junction Temperature
The average chip-junction temperature, TJ, in °C can be obtained from the following:
1.
2.
T J = T A + ( P D × θ JA )
T J = T A + ( P D × ( θ HEATSINK + θ JC ) )
where:
• θJA = package thermal resistance, Junction-to-ambient (°C/W), provided in Table 39-1 on page
787.
• θJC = package thermal resistance, Junction-to-case thermal resistance (°C/W), provided in
Table 39-1 on page 787.
• θHEAT SINK = cooling device thermal resistance (°C/W), provided in the device datasheet.
• PD = device power consumption (W) estimated from data provided in the section ”Power
Consumption” on page 767.
• TA = ambient temperature (°C).
From the first equation, the user can derive the estimated lifetime of the chip and decide if a
cooling device is necessary or not. If a cooling device is to be fitted on the chip, the second
equation should be used to compute the resulting average chip-junction temperature TJ in °C.
787
32058K
AVR32-01/12
AT32UC3A
39.2
Package Drawings
Figure 39-1. TQFP-100 package drawing
Table 39-2.
Device and Package Maximum Weight
500
mg
Table 39-3.
Package Characteristics
Moisture Sensitivity Level
Table 39-4.
Jdec J-STD0-20D - MSL 3
Package Reference
JEDEC Drawing Reference
MS-026
JESD97 Classification
E3
788
32058K
AVR32-01/12
AT32UC3A
Figure 39-2. LQFP-144 package drawing
Table 39-5.
Device and Package Maximum Weight
1300
Table 39-6.
mg
Package Characteristics
Moisture Sensitivity Level
Table 39-7.
Jdec J-STD0-20D - MSL 3
Package Reference
JEDEC Drawing Reference
MS-026
JESD97 Classification
E3
789
32058K
AVR32-01/12
AT32UC3A
Figure 39-3. FFBGA-144 package drawing
Table 39-8.
Device and Package Maximum Weight
1300
Table 39-9.
mg
Package Characteristics
Moisture Sensitivity Level
MSL3
Table 39-10. Package Reference
JEDEC Drawing Reference
MS-026
JESD97 Classification
E3
790
32058K
AVR32-01/12
AT32UC3A
39.3
Soldering Profile
Table 39-11 gives the recommended soldering profile from J-STD-20.
Table 39-11. Soldering Profile
Profile Feature
Green Package
Average Ramp-up Rate (217°C to Peak)
3°C/sec
Preheat Temperature 175°C ±25°C
Min. 150 °C, Max. 200 °C
Time Maintained Above 217°C
60-150 sec
Time within 5⋅C of Actual Peak Temperature
30 sec
Peak Temperature Range
260 °C
Ramp-down Rate
6 °C/sec
Time 25⋅C to Peak Temperature
Max. 8 minutes
Note:
It is recommended to apply a soldering temperature higher than 250°C.
A maximum of three reflow passes is allowed per component.
791
32058K
AVR32-01/12
AT32UC3A
40. Ordering Information
Table 40-1.
Ordering Information
Device
AT32UC3A0512
AT32UC3A0256
AT32UC3A0128
AT32UC3A1512
AT32UC3A1256
AT32UC3A1128
40.1
Ordering Code
Package
Conditioning
Temperature Operating Range
AT32UC3A0512-ALUT
144 LQFP
Tray
Industrial (-40⋅C to 85⋅C)
AT32UC3A0512-ALUR
144 LQFP
Reel
Industrial (-40⋅C to 85⋅C)
AT32UC3A0512-ALTRA
144 LQFP
Reel
Automotive (-40⋅C to 85⋅C)
AT32UC3A0512-ALTTA
144 LQFP
Tray
Automotive (-40⋅C to 85⋅C)
AT32UC3A0512-CTUT
144 FFBGA
Tray
Industrial (-40⋅C to 85⋅C)
AT32UC3A0512-CTUR
144 FFBGA
Reel
Industrial (-40⋅C to 85⋅C)
AT32UC3A0256-ALUT
144 LQFP
Tray
Industrial (-40⋅C to 85⋅C)
AT32UC3A0256-ALUR
144 LQFP
Reel
Industrial (-40⋅C to 85⋅C)
AT32UC3A0256-CTUT
144 FFBGA
Tray
Industrial (-40⋅C to 85⋅C)
AT32UC3A0256-CTUR
144 FFBGA
Reel
Industrial (-40⋅C to 85⋅C)
AT32UC3A0128-ALUT
144 LQFP
Tray
Industrial (-40⋅C to 85⋅C)
AT32UC3A0128-ALUR
144 LQFP
Reel
Industrial (-40⋅C to 85⋅C)
AT32UC3A0128-CTUT
144 FFBGA
Tray
Industrial (-40⋅C to 85⋅C)
AT32UC3A0128-CTUR
144 FFBGA
Reel
Industrial (-40⋅C to 85⋅C)
AT32UC3A1512-AUT
100 TQFP
Tray
Industrial (-40⋅C to 85⋅C)
AT32UC3A1512-AUR
100 TQFP
Reel
Industrial (-40⋅C to 85⋅C)
AT32UC3A1256-AUT
100 TQFP
Tray
Industrial (-40⋅C to 85⋅C)
AT32UC3A1256-AUR
100 TQFP
Reel
Industrial (-40⋅C to 85⋅C)
AT32UC3A1128-AUT
100 TQFP
Tray
Industrial (-40⋅C to 85⋅C)
AT32UC3A1128-AUR
100 TQFP
Reel
Industrial (-40⋅C to 85⋅C)
Automotive Quality Grade
The AT32UC3A have been developed and manufactured according to the most stringent
requirements of the international standard ISO-TS-16949. This data sheet will contain limit values extracted from the results of extensive characterization (Temperature and Voltage). The
quality and reliability of the AT32UC3A is verified during regular product qualification as per
AEC-Q100 grade 3.
As indicated in the ordering information paragraph, the product is available in only one temperature grade T: -40°C / + 85°C.
792
32058K
AVR32-01/12
AT32UC3A
41. Errata
All industrial parts labelled with -UES (engineering samples) are revision E parts.
All automotive parts labelled with AT32UC3A0512-ALTRA or AT32UC3A0512-ALTTA are revision K parts.
41.1
Rev. K, L, M
41.1.1
PWM
1. PWM channel interrupt enabling triggers an interrupt
When enabling a PWM channel that is configured with center aligned period (CALG=1), an
interrupt is signalled.
Fix/Workaround
When using center aligned mode, enable the channel and read the status before channel
interrupt is enabled.
2. PWM counter restarts at 0x0001
The PWM counter restarts at 0x0001 and not 0x0000 as specified. Because of this the first
PWM period has one more clock cycle.
Fix/Workaround
- The first period is 0x0000, 0x0001, ..., period
- Consecutive periods are 0x0001, 0x0002, ..., period
3. PWM update period to a 0 value does not work
It is impossible to update a period equal to 0 by the using the PWM update register
(PWM_CUPD).
Fix/Workaround
Do not update the PWM_CUPD register with a value equal to 0.
41.1.2
ADC
1. Sleep Mode activation needs additional A to D conversion
If the ADC sleep mode is activated when the ADC is idle the ADC will not enter sleep mode
before after the next AD conversion.
Fix/Workaround
Activate the sleep mode in the mode register and then perform an AD conversion.
41.1.3
SPI
1. SPI Slave / PDCA transfer: no TX UNDERRUN flag
There is no TX UNDERRUN flag available, therefore in SPI slave mode, there is no way to
be informed of a character lost in transmission.
Fix/Workaround
For PDCA transfer: none.
2. SPI FDIV option does not work
Selecting clock signal using FDIV = 1 does not work as specified.
Fix/Workaround
Do not set FDIV = 1.
793
32058K
AVR32-01/12
AT32UC3A
3. SPI Bad Serial Clock Generation on 2nd chip_select when SCBR = 1, CPOL=1 and
NCPHA=0
When multiple CS are in use, if one of the baudrate equals to 1 and one of the others doesn't
equal to 1, and CPOL=1 and CPHA=0, then an aditional pulse will be generated on SCK.
Fix/workaround
When multiple CS are in use, if one of the baudrate equals 1, the other must also equal 1 if
CPOL=1 and CPHA=0.
4. SPI Glitch on RXREADY flag in slave mode when enabling the SPI or during the first
transfer
In slave mode, the SPI can generate a false RXREADY signal during enabling of the SPI or
during the first transfer.
Fix/Workaround
1. Set slave mode, set required CPOL/CPHA.
2. Enable SPI.
3. Set the polarity CPOL of the line in the opposite value of the required one.
4. Set the polarity CPOL to the required one.
5. Read the RXHOLDING register.
Transfers can now befin and RXREADY will now behave as expected.
5. SPI Disable does not work in Slave mode
Fix/workaround
Read the last received data then perform a Software reset.
41.1.4
Power Manager
1. If the BOD level is higher than VDDCORE, the part is constantly under reset
If the BOD level is set to a value higher than VDDCORE and enabled by fuses, the part will
be in constant reset.
Fix/Workaround
Apply an external voltage on VDDCORE that is higher than the BOD level and is lower than
VDDCORE max and disable the BOD.
41.1.5
PDCA
1. Wrong PDCA behavior when using two PDCA channels with the same PID.
Fix/Workaround
The same PID should not be assigned to more than one channel.
41.1.6
TWI
1. The TWI RXRDY flag in SR register is not reset when a software reset is performed.
Fix/Workaround
After a Software Reset, the register TWI RHR must be read.
41.1.7
USART
41.1.8
1. ISO7816 info register US_NER cannot be read
The NER register always returns zero.
Fix/Workaround
None
Processor and Architecture
1. LDM instruction with PC in the register list and without ++ increments Rp
794
32058K
AVR32-01/12
AT32UC3A
For LDM with PC in the register list: the instruction behaves as if the ++ field is always set, ie
the pointer is always updated. This happens even if the ++ field is cleared. Specifically, the
increment of the pointer is done in parallel with the testing of R12.
Fix/Workaround
None.
41.1.9
FLASHC
1. Reading from on-chip flash may fail after a flash fuse write operation (FLASHC LP,
UP, WGPB, EGPB, SSB, PGPFB, EAGPF commands).
After a flash fuse write operation (FLASHC LP, UP, WGPB, EGPB, SSB, PGPFB, EAGPF
commands), the following flash read access may return corrupted data. This erratum does
not affect write operations to regular flash memory.
Fix/Workaround
The flash fuse write operation (FLASHC LP, UP, WGPB, EGPB, SSB, PGPFB, EAGPF
commands) must be issued from internal RAM. After the write operation, perform a dummy
flash page write operation (FLASHC WP). Content and location of this page is not important
and filling the write buffer with all one (FFh) will leave the current flash content unchanged. It
is then safe to read and fetch code from the flash.
795
32058K
AVR32-01/12
AT32UC3A
41.2
Rev. J
41.2.1
PWM
1. PWM channel interrupt enabling triggers an interrupt
When enabling a PWM channel that is configured with center aligned period (CALG=1), an
interrupt is signalled.
Fix/Workaround
When using center aligned mode, enable the channel and read the status before channel
interrupt is enabled.
2. PWM counter restarts at 0x0001
The PWM counter restarts at 0x0001 and not 0x0000 as specified. Because of this the first
PWM period has one more clock cycle.
Fix/Workaround
- The first period is 0x0000, 0x0001, ..., period
- Consecutive periods are 0x0001, 0x0002, ..., period
3. PWM update period to a 0 value does not work
It is impossible to update a period equal to 0 by the using the PWM update register
(PWM_CUPD).
Fix/Workaround
Do not update the PWM_CUPD register with a value equal to 0.
41.2.2
ADC
1. Sleep Mode activation needs additional A to D conversion
If the ADC sleep mode is activated when the ADC is idle the ADC will not enter sleep mode
before after the next AD conversion.
Fix/Workaround
Activate the sleep mode in the mode register and then perform an AD conversion.
41.2.3
SPI
1. SPI Slave / PDCA transfer: no TX UNDERRUN flag
There is no TX UNDERRUN flag available, therefore in SPI slave mode, there is no way to
be informed of a character lost in transmission.
Fix/Workaround
For PDCA transfer: none.
2. SPI FDIV option does not work
Selecting clock signal using FDIV = 1 does not work as specified.
Fix/Workaround
Do not set FDIV = 1.
3. SPI Bad Serial Clock Generation on 2nd chip_select when SCBR = 1, CPOL=1 and
NCPHA=0
When multiple CS are in use, if one of the baudrate equals to 1 and one of the others doesn't
equal to 1, and CPOL=1 and CPHA=0, then an aditional pulse will be generated on SCK.
Fix/workaround
796
32058K
AVR32-01/12
AT32UC3A
When multiple CS are in use, if one of the baudrate equals 1, the other must also equal 1 if
CPOL=1 and CPHA=0.
4. SPI Glitch on RXREADY flag in slave mode when enabling the SPI or during the first
transfer
In slave mode, the SPI can generate a false RXREADY signal during enabling of the SPI or
during the first transfer.
Fix/Workaround
1. Set slave mode, set required CPOL/CPHA.
2. Enable SPI.
3. Set the polarity CPOL of the line in the opposite value of the required one.
4. Set the polarity CPOL to the required one.
5. Read the RXHOLDING register.
Transfers can now befin and RXREADY will now behave as expected.
5. SPI Disable does not work in Slave mode
Fix/workaround
Read the last received data then perform a Software reset.
41.2.4
Power Manager
1. If the BOD level is higher than VDDCORE, the part is constantly under reset
If the BOD level is set to a value higher than VDDCORE and enabled by fuses, the part will
be in constant reset.
Fix/Workaround
Apply an external voltage on VDDCORE that is higher than the BOD level and is lower than
VDDCORE max and disable the BOD.
41.2.5
PDCA
1. Wrong PDCA behavior when using two PDCA channels with the same PID.
Fix/Workaround
The same PID should not be assigned to more than one channel.
41.2.6
TWI
1. The TWI RXRDY flag in SR register is not reset when a software reset is performed.
Fix/Workaround
After a Software Reset, the register TWI RHR must be read.
41.2.7
SDRAMC
1.
Code execution from external SDRAM does not work
Code execution from SDRAM does not work.
Fix/Workaround
Do not run code from SDRAM.
41.2.8
GPIO
1. PA29 (TWI SDA) and PA30 (TWI SCL) GPIO VIH (input high voltage) is 3.6V max
instead of 5V tolerant
The following GPIOs are not 5V tolerant : PA29 and PA30.
Fix/Workaround
797
32058K
AVR32-01/12
AT32UC3A
None.
41.2.9
USART
41.2.10
1. ISO7816 info register US_NER cannot be read
The NER register always returns zero.
Fix/Workaround
None
Processor and Architecture
1. LDM instruction with PC in the register list and without ++ increments Rp
For LDM with PC in the register list: the instruction behaves as if the ++ field is always set, ie
the pointer is always updated. This happens even if the ++ field is cleared. Specifically, the
increment of the pointer is done in parallel with the testing of R12.
Fix/Workaround
None.
2. RETE instruction does not clear SREG[L] from interrupts.
The RETE instruction clears SREG[L] as expected from exceptions.
Fix/Workaround
When using the STCOND instruction, clear SREG[L] in the stacked value of SR before
returning from interrupts with RETE.
3.
41.2.11
Exceptions when system stack is protected by MPU
RETS behaves incorrectly when MPU is enabled and MPU is configured so that
system stack is not readable in unprivileged mode.
Fix/Woraround
Workaround 1: Make system stack readable in unprivileged mode,
or
Workaround 2: Return from supervisor mode using rete instead of rets. This
requires :
1. Changing the mode bits from 001b to 110b before issuing the instruction.
Updating the mode bits to the desired value must be done using a single mtsr instruction so it is done atomically. Even if this step is described in general
as not safe in the UC technical reference guide, it is safe in this very
specific case.
2. Execute the RETE instruction.
FLASHC
1. Reading from on-chip flash may fail after a flash fuse write operation (FLASHC LP,
UP, WGPB, EGPB, SSB, PGPFB, EAGPF commands).
After a flash fuse write operation (FLASHC LP, UP, WGPB, EGPB, SSB, PGPFB, EAGPF
commands), the following flash read access may return corrupted data. This erratum does
not affect write operations to regular flash memory.
Fix/Workaround
The flash fuse write operation (FLASHC LP, UP, WGPB, EGPB, SSB, PGPFB, EAGPF
commands) must be issued from internal RAM. After the write operation, perform a dummy
flash page write operation (FLASHC WP). Content and location of this page is not important
and filling the write buffer with all one (FFh) will leave the current flash content unchanged. It
is then safe to read and fetch code from the flash.
798
32058K
AVR32-01/12
AT32UC3A
41.3
Rev. I
41.3.1
PWM
1. PWM channel interrupt enabling triggers an interrupt
When enabling a PWM channel that is configured with center aligned period (CALG=1), an
interrupt is signalled.
Fix/Workaround
When using center aligned mode, enable the channel and read the status before channel
interrupt is enabled.
2. PWM counter restarts at 0x0001
The PWM counter restarts at 0x0001 and not 0x0000 as specified. Because of this the first
PWM period has one more clock cycle.
Fix/Workaround
- The first period is 0x0000, 0x0001, ..., period
- Consecutive periods are 0x0001, 0x0002, ..., period
3. PWM update period to a 0 value does not work
It is impossible to update a period equal to 0 by the using the PWM update register
(PWM_CUPD).
Fix/Workaround
Do not update the PWM_CUPD register with a value equal to 0.
41.3.2
ADC
1. Sleep Mode activation needs additional A to D conversion
If the ADC sleep mode is activated when the ADC is idle the ADC will not enter sleep mode
before after the next AD conversion.
Fix/Workaround
Activate the sleep mode in the mode register and then perform an AD conversion.
41.3.3
SPI
1. SPI Slave / PDCA transfer: no TX UNDERRUN flag
There is no TX UNDERRUN flag available, therefore in SPI slave mode, there is no way to
be informed of a character lost in transmission.
Fix/Workaround
For PDCA transfer: none.
2. SPI FDIV option does not work
Selecting clock signal using FDIV = 1 does not work as specified.
Fix/Workaround
Do not set FDIV = 1.
3. SPI Bad Serial Clock Generation on 2nd chip_select when SCBR = 1, CPOL=1 and
NCPHA=0
When multiple CS are in use, if one of the baudrate equals to 1 and one of the others doesn't
equal to 1, and CPOL=1 and CPHA=0, then an aditional pulse will be generated on SCK.
Fix/workaround
799
32058K
AVR32-01/12
AT32UC3A
When multiple CS are in use, if one of the baudrate equals 1, the other must also equal 1 if
CPOL=1 and CPHA=0.
4. SPI Glitch on RXREADY flag in slave mode when enabling the SPI or during the first
transfer
In slave mode, the SPI can generate a false RXREADY signal during enabling of the SPI or
during the first transfer.
Fix/Workaround
1. Set slave mode, set required CPOL/CPHA.
2. Enable SPI.
3. Set the polarity CPOL of the line in the opposite value of the required one.
4. Set the polarity CPOL to the required one.
5. Read the RXHOLDING register.
Transfers can now befin and RXREADY will now behave as expected.
5. SPI Disable does not work in Slave mode
Fix/workaround
Read the last received data then perform a Software reset.
41.3.4
Power Manager
1. If the BOD level is higher than VDDCORE, the part is constantly under reset
If the BOD level is set to a value higher than VDDCORE and enabled by fuses, the part will
be in constant reset.
Fix/Workaround
Apply an external voltage on VDDCORE that is higher than the BOD level and is lower than
VDDCORE max and disable the BOD.
41.3.5
Flashc
1. On AT32UC3A0512 and AT32UC3A1512, corrupted read in flash after FLASHC WP,
EP, EA, WUP, EUP commands may happen
- After a FLASHC Write Page (WP) or Erase Page (EP) command applied to a page in a
given half of the flash (first or last 256 kB of flash), reading (data read or code fetch) the
other half of the flash may fail. This may lead to an exception or to other errors derived from
this corrupted read access.
- After a FLASHC Erase All (EA) command, reading (data read or code fetch) the flash may
fail. This may lead to an exception or to other errors derived from this corrupted read access.
- After a FLASHC Write User Page (WUP) or Erase User Page (EUP) command, reading
(data read or code fetch) the second half (last 256 kB) of the flash may fail. This may lead to
an exception or to other errors derived from this corrupted read access.
Fix/Workaround
Flashc WP, EP, EA, WUP, EUP commands: these commands must be issued from RAM or
through the EBI. After these commands, read twice one flash page initialized to 00h in each
half part of the flash.
41.3.6
PDCA
1. Wrong PDCA behavior when using two PDCA channels with the same PID.
800
32058K
AVR32-01/12
AT32UC3A
Workaround/fix
The same PID should not be assigned to more than one channel.
41.3.7
GPIO
1. Some GPIO VIH (input high voltage) are 3.6V max instead of 5V tolerant
Only 11 GPIOs remain 5V tolerant (VIHmax=5V):PB01, PB02, PB03, PB10, PB19, PB20,
PB21, PB22, PB23, PB27, PB28.
Workaround/fix
None.
41.3.8
USART
1. ISO7816 info register US_NER cannot be read
The NER register always returns zero.
Fix/Workaround
None.
41.3.9
TWI
1. The TWI RXRDY flag in SR register is not reset when a software reset is performed.
Fix/Workaround
After a Software Reset, the register TWI RHR must be read.
41.3.10
SDRAMC
1.
41.3.11
Code execution from external SDRAM does not work
Code execution from SDRAM does not work.
Fix/Workaround
Do not run code from SDRAM.
Processor and Architecture
1. LDM instruction with PC in the register list and without ++ increments Rp
For LDM with PC in the register list: the instruction behaves as if the ++ field is always set, ie
the pointer is always updated. This happens even if the ++ field is cleared. Specifically, the
increment of the pointer is done in parallel with the testing of R12.
Fix/Workaround
None.
2. RETE instruction does not clear SREG[L] from interrupts.
The RETE instruction clears SREG[L] as expected from exceptions.
Fix/Workaround
When using the STCOND instruction, clear SREG[L] in the stacked value of SR before
returning from interrupts with RETE.
3.
Exceptions when system stack is protected by MPU
RETS behaves incorrectly when MPU is enabled and MPU is configured so that
system stack is not readable in unprivileged mode.
Fix/Woraround
Workaround 1: Make system stack readable in unprivileged mode,
or
Workaround 2: Return from supervisor mode using rete instead of rets. This
requires :
1. Changing the mode bits from 001b to 110b before issuing the instruction.
Updating the mode bits to the desired value must be done using a single mtsr
instruction so it is done atomically. Even if this step is described in general
as not safe in the UC technical reference guide, it is safe in this very
801
32058K
AVR32-01/12
AT32UC3A
specific case.
2. Execute the RETE instruction.
41.3.12
FLASHC
1. Reading from on-chip flash may fail after a flash fuse write operation (FLASHC LP,
UP, WGPB, EGPB, SSB, PGPFB, EAGPF commands).
After a flash fuse write operation (FLASHC LP, UP, WGPB, EGPB, SSB, PGPFB, EAGPF
commands), the following flash read access may return corrupted data. This erratum does
not affect write operations to regular flash memory.
Fix/Workaround
The flash fuse write operation (FLASHC LP, UP, WGPB, EGPB, SSB, PGPFB, EAGPF
commands) must be issued from internal RAM. After the write operation, perform a dummy
flash page write operation (FLASHC WP). Content and location of this page is not important
and filling the write buffer with all one (FFh) will leave the current flash content unchanged. It
is then safe to read and fetch code from the flash.
802
32058K
AVR32-01/12
AT32UC3A
41.4
Rev. H
41.4.1
PWM
1. PWM channel interrupt enabling triggers an interrupt
When enabling a PWM channel that is configured with center aligned period (CALG=1), an
interrupt is signalled.
Fix/Workaround
When using center aligned mode, enable the channel and read the status before channel
interrupt is enabled.
2. PWM counter restarts at 0x0001
The PWM counter restarts at 0x0001 and not 0x0000 as specified. Because of this the first
PWM period has one more clock cycle.
Fix/Workaround
- The first period is 0x0000, 0x0001, ..., period
- Consecutive periods are 0x0001, 0x0002, ..., period
3. PWM update period to a 0 value does not work
It is impossible to update a period equal to 0 by the using the PWM update register
(PWM_CUPD).
Fix/Workaround
Do not update the PWM_CUPD register with a value equal to 0.
41.4.2
ADC
1. Sleep Mode activation needs additional A to D conversion
If the ADC sleep mode is activated when the ADC is idle the ADC will not enter sleep mode
before after the next AD conversion.
Fix/Workaround
Activate the sleep mode in the mode register and then perform an AD conversion.
41.4.3
SPI
1. SPI Slave / PDCA transfer: no TX UNDERRUN flag
There is no TX UNDERRUN flag available, therefore in SPI slave mode, there is no way to
be informed of a character lost in transmission.
Fix/Workaround
For PDCA transfer: none.
2. SPI FDIV option does not work
Selecting clock signal using FDIV = 1 does not work as specified.
Fix/Workaround
Do not set FDIV = 1
3. SPI disable does not work in SLAVE mode.
Fix/Workaround
Read the last received data, then perform a Software Reset.
803
32058K
AVR32-01/12
AT32UC3A
4. SPI Bad Serial Clock Generation on 2nd chip_select when SCBR = 1, CPOL=1 and
NCPHA=0
When multiple CS are in use, if one of the baudrate equals to 1 and one of the others doesn't
equal to 1, and CPOL=1 and CPHA=0, then an aditional pulse will be generated on SCK.
Fix/workaround
When multiple CS are in use, if one of the baudrate equals 1, the other must also equal 1 if
CPOL=1 and CPHA=0.
5. SPI Glitch on RXREADY flag in slave mode when enabling the SPI or during the first
transfer
In slave mode, the SPI can generate a false RXREADY signal during enabling of the SPI or
during the first transfer.
Fix/Workaround
1. Set slave mode, set required CPOL/CPHA.
2. Enable SPI.
3. Set the polarity CPOL of the line in the opposite value of the required one.
4. Set the polarity CPOL to the required one.
5. Read the RXHOLDING register.
Transfers can now befin and RXREADY will now behave as expected.
41.4.4
6. SPI Disable does not work in Slave mode
Fix/workaround
Read the last received data then perform a Software reset.
Power Manager
1. Wrong reset causes when BOD is activated
Setting the BOD enable fuse will cause the Reset Cause Register to list BOD reset as the
reset source even though the part was reset by another source.
Fix/Workaround
Do not set the BOD enable fuse, but activate the BOD as soon as your program starts.
2. If the BOD level is higher than VDDCORE, the part is constantly under reset
If the BOD level is set to a value higher than VDDCORE and enabled by fuses, the part will
be in constant reset.
Fix/Workaround
Apply an external voltage on VDDCORE that is higher than the BOD level and is lower than
VDDCORE max and disable the BOD.
41.4.5
FLASHC
1. On AT32UC3A0512 and AT32UC3A1512, corrupted read in flash after FLASHC WP,
EP, EA, WUP, EUP commands may happen
- After a FLASHC Write Page (WP) or Erase Page (EP) command applied to a page in a
given half of the flash (first or last 256 kB of flash), reading (data read or code fetch) the
other half of the flash may fail. This may lead to an exception or to other errors derived from
this corrupted read access.
- After a FLASHC Erase All (EA) command, reading (data read or code fetch) the flash may
fail. This may lead to an exception or to other errors derived from this corrupted read access.
- After a FLASHC Write User Page (WUP) or Erase User Page (EUP) command, reading
804
32058K
AVR32-01/12
AT32UC3A
(data read or code fetch) the second half (last 256 kB) of the flash may fail. This may lead to
an exception or to other errors derived from this corrupted read access.
Fix/Workaround
Flashc WP, EP, EA, WUP, EUP commands: these commands must be issued from RAM or
through the EBI. After these commands, read twice one flash page initialized to 00h in each
half part of the flash.
41.4.6
PDCA
1. Wrong PDCA behavior when using two PDCA channels with the same PID.
Workaround/fix
The same PID should not be assigned to more than one channel.
41.4.7
TWI
1. The TWI RXRDY flag in SR register is not reset when a software reset is performed.
Fix/Workaround
After a Software Reset, the register TWI RHR must be read.
41.4.8
SDRAMC
1.
Code execution from external SDRAM does not work
Code execution from SDRAM does not work.
Fix/Workaround
Do not run code from SDRAM.
41.4.9
GPIO
1. Some GPIO VIH (input high voltage) are 3.6V max instead of 5V tolerant
Only 11 GPIOs remain 5V tolerant (VIHmax=5V):PB01, PB02, PB03, PB10, PB19, PB20,
PB21, PB22, PB23, PB27, PB28.
Workaround/fix
None.
41.4.10
USART
41.4.11
1. ISO7816 info register US_NER cannot be read
The NER register always returns zero.
Fix/Workaround
None.
Processor and Architecture
1. LDM instruction with PC in the register list and without ++ increments Rp
For LDM with PC in the register list: the instruction behaves as if the ++ field is always set, ie
the pointer is always updated. This happens even if the ++ field is cleared. Specifically, the
increment of the pointer is done in parallel with the testing of R12.
Fix/Workaround
None.
2. RETE instruction does not clear SREG[L] from interrupts.
The RETE instruction clears SREG[L] as expected from exceptions.
Fix/Workaround
When using the STCOND instruction, clear SREG[L] in the stacked value of SR before
returning from interrupts with RETE.
3.
Exceptions when system stack is protected by MPU
805
32058K
AVR32-01/12
AT32UC3A
RETS behaves incorrectly when MPU is enabled and MPU is configured so that
system stack is not readable in unprivileged mode.
Fix/Woraround
Workaround 1: Make system stack readable in unprivileged mode,
or
Workaround 2: Return from supervisor mode using rete instead of rets. This
requires :
1. Changing the mode bits from 001b to 110b before issuing the instruction.
Updating the mode bits to the desired value must be done using a single mtsr
instruction so it is done atomically. Even if this step is described in general
as not safe in the UC technical reference guide, it is safe in this very
specific case.
2. Execute the RETE instruction.
41.4.12
FLASHC
1. Reading from on-chip flash may fail after a flash fuse write operation (FLASHC LP,
UP, WGPB, EGPB, SSB, PGPFB, EAGPF commands).
After a flash fuse write operation (FLASHC LP, UP, WGPB, EGPB, SSB, PGPFB, EAGPF
commands), the following flash read access may return corrupted data. This erratum does
not affect write operations to regular flash memory.
Fix/Workaround
The flash fuse write operation (FLASHC LP, UP, WGPB, EGPB, SSB, PGPFB, EAGPF
commands) must be issued from internal RAM. After the write operation, perform a dummy
flash page write operation (FLASHC WP). Content and location of this page is not important
and filling the write buffer with all one (FFh) will leave the current flash content unchanged. It
is then safe to read and fetch code from the flash.
806
32058K
AVR32-01/12
AT32UC3A
41.5
Rev. E
41.5.1
SPI
1. SPI FDIV option does not work
Selecting clock signal using FDIV = 1 does not work as specified.
Fix/Workaround
Do not set FDIV = 1.
2. SPI Slave / PDCA transfer: no TX UNDERRUN flag
There is no TX UNDERRUN flag available, therefore in SPI slave mode, there is no way to
be informed of a character lost in transmission.
Fix/Workaround
For PDCA transfer: none.
3. SPI Bad serial clock generation on 2nd chip select when SCBR=1, CPOL=1 and
CNCPHA=0
When multiple CS are in use, if one of the baudrate equals to 1 and one of the others
doesn’t equal to 1, and CPOL=1 and CPHA=0, then an additional pulse will be generated on
SCK.
Fix/Workaround
When multiple CS are in use, if one of the baudrate equals to 1, the other must also equal 1
if CPOL=1 and CPHA=0.
4. SPI Glitch on RXREADY flag in slave mode when enabling the SPI or during the first
transfer
In slave mode, the SPI can generate a false RXREADY signal during enabling of the SPI or
during the first transfer.
Fix/Workaround
1. Set slave mode, set required CPOL/CPHA.
2. Enable SPI.
3. Set the polarity CPOL of the line in the opposite value of the required one.
4. Set the polarity CPOL to the required one.
5. Read the RXHOLDING register.
Transfers can now befin and RXREADY will now behave as expected.
5. SPI CSNAAT bit 2 in register CSR0...CSR3 is not available.
Fix/Workaround
Do not use this bit.
6. SPI disable does not work in SLAVE mode.
Fix/Workaround
Read the last received data, then perform a Software Reset.
7. SPI Bad Serial Clock Generation on 2nd chip_select when SCBR = 1, CPOL=1 and
NCPHA=0
When multiple CS are in use, if one of the baudrate equals to 1 and one of the others doesn't
equal to 1, and CPOL=1 and CPHA=0, then an aditional pulse will be generated on SCK.
807
32058K
AVR32-01/12
AT32UC3A
Fix/workaround
When multiple CS are in use, if one of the baudrate equals 1, the other must also equal 1 if
CPOL=1 and CPHA=0.
41.5.2
PWM
1. PWM counter restarts at 0x0001
The PWM counter restarts at 0x0001 and not 0x0000 as specified. Because of this the first
PWM period has one more clock cycle.
Fix/Workaround
- The first period is 0x0000, 0x0001, ..., period
- Consecutive periods are 0x0001, 0x0002, ..., period
2. PWM channel interrupt enabling triggers an interrupt
When enabling a PWM channel that is configured with center aligned period (CALG=1), an
interrupt is signalled.
Fix/Workaround
When using center aligned mode, enable the channel and read the status before channel
interrupt is enabled.
3. PWM update period to a 0 value does not work
It is impossible to update a period equal to 0 by the using the PWM update register
(PWM_CUPD).
Fix/Workaround
Do not update the PWM_CUPD register with a value equal to 0.
4.
PWM channel status may be wrong if disabled before a period has elapsed
Before a PWM period has elapsed, the read channel status may be wrong. The CHIDx-bit
for a PWM channel in the PWM Enable Register will read '1' for one full PWM period even if
the channel was disabled before the period elapsed. It will then read '0' as expected.
Fix/Workaround
Reading the PWM channel status of a disabled channel is only correct after a PWM period
has elapsed.
41.5.3
SSC
1. SSC does not trigger RF when data is low
The SSC cannot transmit or receive data when CKS = CKDIV and CKO = none, in TCMR or
RCMR respectively.
Fix/Workaround
Set CKO to a value that is not "none" and bypass the output of the TK/RK pin with the PIO.
2. SSC Data is not sent unless clock is set as output
The SSC cannot transmit or receive data when CKS = CKDIV and CKO = none, in TCMR or
RCMR respectively.
Fix/Workaround
Set CKO to a value that is not "none" and bypass the output of the TK/RK pin with the PIO.
808
32058K
AVR32-01/12
AT32UC3A
41.5.4
USB
1.
USB No end of host reset signaled upon disconnection
In host mode, in case of an unexpected device disconnection whereas a usb reset is being
sent by the usb controller, the UHCON.RESET bit may not been cleared by the hardware at
the end of the reset.
Fix/Workaround
A software workaround consists in testing (by polling or interrupt) the disconnection
(UHINT.DDISCI == 1) while waiting for the end of reset (UHCON.RESET == 0) to avoid
being stuck.
2. USBFSM and UHADDR1/2/3 registers are not available.
Do not use USBFSM register.
Fix/Workaround
Do not use USBFSM register and use HCON[6:0] field instead for all the pipes.
41.5.5
Processor and Architecture
1. Incorrect Processor ID
The processor ID reads 0x01 and not 0x02 as it should.
Fix/Workaround
None.
2. Bus error should be masked in Debug mode
If a bus error occurs during debug mode, the processor will not respond to debug commands through the DINST register.
Fix/Workaround
A reset of the device will make the CPU respond to debug commands again.
3. Read Modify Write (RMW) instructions on data outside the internal RAM does not
work.
Read Modify Write (RMW) instructions on data outside the internal RAM does not work.
Fix/Workaround
Do not perform RMW instructions on data outside the internal RAM.
4.
CRC calculation of a locked device will calculate CRC for 512 kB of flash memory,
even though the part has less flash.
Fix/Workaround
The flash address space is wrapping, so it is possible to use the CRC value by calculating
CRC of the flash content concatenated with itself N times. Where N is 512 kB/flash size.
5.
Need two NOPs instruction after instructions masking interrupts
The instructions following in the pipeline the instruction masking the interrupt through SR
may behave abnormally.
Fix/Workaround
Place two NOPs instructions after each SSRF or MTSR instruction setting IxM or GM in SR.
809
32058K
AVR32-01/12
AT32UC3A
6. CPU Cycle Counter does not reset the COUNT system register on COMPARE match.
The device revision E does not reset the COUNT system register on COMPARE match. In
this revision, the COUNT register is clocked by the CPU clock, so when the CPU clock
stops, so does incrementing of COUNT.
Fix/Workaround
None.
7. Memory Protection Unit (MPU) is non functional.
Fix/Workaround
Do not use the MPU.
8. The following alternate GPIO function C are not available in revE
MACB-WOL on GPIO9 (PA09), MACB-WOL on GPIO18 (PA18), USB-USB_ID on GPIO21
(PA21), USB-USB_VBOF on GPIO22 (PA22), and all function B and C on GPIO70 to
GPIO101 (PX00 to PX39).
Fix/Workaround
Do not use these alternate B and C functions on the listed GPIO pins.
9.
Clock connection table on Rev E
Here is the table of Rev E
Figure 41-1. Timer/Counter clock connections on RevE
Source
Name
Connection
Internal
TIMER_CLOCK1
32 KHz Oscillator
TIMER_CLOCK2
PBA Clock / 4
TIMER_CLOCK3
PBA Clock / 8
TIMER_CLOCK4
PBA Clock / 16
TIMER_CLOCK5
PBA Clock / 32
External
XC0
XC1
XC2
10. Local Bus fast GPIO not available in RevE.
Fix/Workaround
Do not use on this silicon revision.
11. Spurious interrupt may corrupt core SR mode to exception
If the rules listed in the chapter `Masking interrupt requests in peripheral modules' of the
AVR32UC Technical Reference Manual are not followed, a spurious interrupt may occur. An
interrupt context will be pushed onto the stack while the core SR mode will indicate an
exception. A RETE instruction would then corrupt the stack..
Fix/Workaround
Follow the rules of the AVR32UC Technical Reference Manual. To increase software
robustness, if an exception mode is detected at the beginning of an interrupt handler,
change the stack interrupt context to an exception context and issue a RETE instruction.
810
32058K
AVR32-01/12
AT32UC3A
12. CPU cannot operate on a divided slow clock (internal RC oscillator)
Fix/Workaround
Do not run the CPU on a divided slow clock.
13. LDM instruction with PC in the register list and without ++ increments Rp
For LDM with PC in the register list: the instruction behaves as if the ++ field is always set, ie
the pointer is always updated. This happens even if the ++ field is cleared. Specifically, the
increment of the pointer is done in parallel with the testing of R12.
Fix/Workaround
None.
14. RETE instruction does not clear SREG[L] from interrupts.
The RETE instruction clears SREG[L] as expected from exceptions.
Fix/Workaround
When using the STCOND instruction, clear SREG[L] in the stacked value of SR before
returning from interrupts with RETE.
15. Exceptions when system stack is protected by MPU
RETS behaves incorrectly when MPU is enabled and MPU is configured so that
system stack is not readable in unprivileged mode.
Fix/Woraround
Workaround 1: Make system stack readable in unprivileged mode,
or
Workaround 2: Return from supervisor mode using rete instead of rets. This
requires :
1. Changing the mode bits from 001b to 110b before issuing the instruction.
Updating the mode bits to the desired value must be done using a single mtsr
instruction so it is done atomically. Even if this step is described in general
as not safe in the UC technical reference guide, it is safe in this very
specific case.
2. Execute the RETE instruction.
41.5.6
SDRAMC
1.
Code execution from external SDRAM does not work
Code execution from SDRAM does not work.
Fix/Workaround
Do not run code from SDRAM.
2.
SDRAM SDCKE rise at the same time as SDCK while exiting self-refresh mode
SDCKE rise at the same time as SDCK while exiting self-refresh mode.
Fix/Workaround
None.
41.5.7
USART
1.
USART Manchester Encoder Not Working
Manchester encoding/decoding is not working.
Fix/Workaround
Do not use manchester encoding.
811
32058K
AVR32-01/12
AT32UC3A
2.
USART RXBREAK problem when no timeguard
In asynchronous mode the RXBREAK flag is not correctly handled when the timeguard is 0
and the break character is located just after the stop bit.
Fix/Workaround
If the NBSTOP is 1, timeguard should be different from 0.
3.
USART Handshaking: 2 characters sent / CTS rises when TX
If CTS switches from 0 to 1 during the TX of a character, if the Holding register is not empty,
the TXHOLDING is also transmitted.
Fix/Workaround
None.
4.
USART PDC and TIMEGUARD not supported in MANCHESTER
Manchester encoding/decoding is not working.
Fix/Workaround
Do not use manchester encoding.
5. USART SPI mode is non functional on this revision.
Fix/Workaround
Do not use the USART SPI mode.
6. DCD is active High instead of Low.
In modem mode the DCD signal is assumed to be active high by the USART, butshould
have been active low.
Fix/Workaround
Add an external inverter to the DCD line.
41.5.8
7. ISO7816 info register US_NER cannot be read
The NER register always returns zero.
Fix/Workaround
None.
Power Manager
1.
Voltage regulator input and output is connected to VDDIO and VDDCORE inside the
device
The voltage regulator input and output is connected to VDDIO and VDDCORE respectively
inside the device.
Fix/Workaround
Do not supply VDDCORE externally, as this supply will work in paralell with the regulator.
2. Wrong reset causes when BOD is activated
Setting the BOD enable fuse will cause the Reset Cause Register to list BOD reset as the
reset source even though the part was reset by another source.
Fix/Workaround
Do not set the BOD enable fuse, but activate the BOD as soon as your program starts.
3. PLL0/1 Lock control does not work
Lock Control does not work for PLL0 and PLL1.
812
32058K
AVR32-01/12
AT32UC3A
Fix/Workaround
In PLL0/1 Control register, the bit 7 should be set in order to prevent unexpected behaviour.
4. Peripheral Bus A maximum frequency is 33MHz instead of 66MHz.
Fix/Workaround
Do not set PBA frequency higher than 33 MHz.
5.
PCx pins go low in stop mode
In sleep mode stop all PCx pins will be controlled by GPIO module instead of oscillators.
This can cause drive contention on the XINx in worst case.
Fix/Workaround
Before entering stop mode set all PCx pins to input and GPIO controlled.
6.
On some rare parts, the maximum HSB and CPU speed is 50MHz instead of 66MHz.
Fix/Workaround
Do not set the HSB/CPU speed higher than 50MHz when the firmware generate exceptions.
7. If the BOD level is higher than VDDCORE, the part is constantly under reset
If the BOD level is set to a value higher than VDDCORE and enabled by fuses, the part will
be in constant reset.
Fix/Workaround
Apply an external voltage on VDDCORE that is higher than the BOD level and is lower than
VDDCORE max and disable the BOD.
8. System Timer mask (Bit 16) of the PM CPUMASK register is not available.
Fix/Workaround
Do not use this bit.
41.5.9
HMatrix
1. HMatrix fixed priority arbitration does not work
Fixed priority arbitration does not work.
Fix/Workaround
Use Round-Robin arbitration instead.
41.5.10
ADC
1.
ADC possible miss on DRDY when disabling a channel
The ADC does not work properly when more than one channel is enabled.
Fix/Workaround
Do not use the ADC with more than one channel enabled at a time.
2.
ADC OVRE flag sometimes not reset on Status Register read
The OVRE flag does not clear properly if read simultaneously to an end of conversion.
Fix/Workaround
None.
3. Sleep Mode activation needs additional A to D conversion
813
32058K
AVR32-01/12
AT32UC3A
If the ADC sleep mode is activated when the ADC is idle the ADC will not enter sleep mode
before after the next AD conversion.
Fix/Workaround
Activate the sleep mode in the mode register and then perform an AD conversion.
41.5.11
ABDAC
1. Audio Bitstream DAC is not functional.
Fix/Workaround
Do not use the ABDAC on revE.
41.5.12
FLASHC
1. The address of Flash General Purpose Fuse Register Low (FGPFRLO) is 0xFFFE140C
on revE instead of 0xFFFE1410.
Fix/Workaround
None.
2. The command Quick Page Read User Page(QPRUP) is not functional.
Fix/Workaround
None.
3. PAGEN Semantic Field for Program GP Fuse Byte is WriteData[7:0], ByteAddress[1:0]
on revision E instead of WriteData[7:0], ByteAddress[2:0].
Fix/Workaround
None.
4. On AT32UC3A0512 and AT32UC3A1512, corrupted read in flash after FLASHC WP,
EP, EA, WUP, EUP commands may happen
- After a FLASHC Write Page (WP) or Erase Page (EP) command applied to a page in a
given half of the flash (first or last 256 kB of flash), reading (data read or code fetch) the
other half of the flash may fail. This may lead to an exception or to other errors derived from
this corrupted read access.
- After a FLASHC Erase All (EA) command, reading (data read or code fetch) the flash may
fail. This may lead to an exception or to other errors derived from this corrupted read access.
- After a FLASHC Write User Page (WUP) or Erase User Page (EUP) command, reading
(data read or code fetch) the second half (last 256 kB) of the flash may fail. This may lead to
an exception or to other errors derived from this corrupted read access.
Fix/Workaround
Flashc WP, EP, EA, WUP, EUP commands: these commands must be issued from RAM or
through the EBI. After these commands, read twice one flash page initialized to 00h in each
half part of the flash.
41.5.13
RTC
1. Writes to control (CTRL), top (TOP) and value (VAL) in the RTC are discarded if the
RTC peripheral bus clock (PBA) is divided by a factor of four or more relative to the
HSB clock.
Fix/Workaround
Do not write to the RTC registers using the peripheral bus clock (PBA) divided by a factor of
four or more relative to the HSB clock.
814
32058K
AVR32-01/12
AT32UC3A
2. The RTC CLKEN bit (bit number 16) of CTRL register is not available.
Fix/Workaround
Do not use the CLKEN bit of the RTC on Rev E.
41.5.14
OCD
1. Stalled memory access instruction writeback fails if followed by a HW breakpoint.
Consider the following assembly code sequence:
A
B
If a hardware breakpoint is placed on instruction B, and instruction A is a memory access
instruction, register file updates from instruction A can be discarded.
Fix/Workaround
Do not place hardware breakpoints, use software breakpoints instead.
Alternatively, place a hardware breakpoint on the instruction before the memory
access instruction and then single step over the memory access instruction.
41.5.15
PDCA
1. Wrong PDCA behavior when using two PDCA channels with the same PID.
Workaround/fix
The same PID should not be assigned to more than one channel.
41.5.16
TWI
1. The TWI RXRDY flag in SR register is not reset when a software reset is performed.
Fix/Workaround
After a Software Reset, the register TWI RHR must be read.
41.5.17
FLASHC
1. Reading from on-chip flash may fail after a flash fuse write operation (FLASHC LP,
UP, WGPB, EGPB, SSB, PGPFB, EAGPF commands).
After a flash fuse write operation (FLASHC LP, UP, WGPB, EGPB, SSB, PGPFB, EAGPF
commands), the following flash read access may return corrupted data. This erratum does
not affect write operations to regular flash memory.
Fix/Workaround
The flash fuse write operation (FLASHC LP, UP, WGPB, EGPB, SSB, PGPFB, EAGPF
commands) must be issued from internal RAM. After the write operation, perform a dummy
flash page write operation (FLASHC WP). Content and location of this page is not important
and filling the write buffer with all one (FFh) will leave the current flash content unchanged. It
is then safe to read and fetch code from the flash.
815
32058K
AVR32-01/12
AT32UC3A
42. Datasheet Revision History
Please note that the referring page numbers in this section are referred to this document. The
referring revision in this section are referring to the document revision.
42.1
Rev. K – 01/12
42.2
42.3
42.4
42.5
1.
Update Errata Section
2.
Update Electrical characteristic Section
1.
Remove ordering code for automotive engineering samples
2.
Replace old automotive odering codes AT32UC3A0512-ALTR (revision I) by
AT32UC3A0512-ALTRA (revision K).
Replace old automotive odering codes AT32UC3A0512-ALTT (revision I) by
AT32UC3A0512-ALTTA (revision K).
1.
Update ”Errata” on page 779.
2.
Update eletrical characteristic in ”DC Characteristics” on page 2.
3.
Add BGA144 package information.
1.
Update ”Errata” on page 779.
2.
Update GPIO eletrical characteristic in ”DC Characteristics” on page 2.
1.
Add revision J to ”Errata” on page 779.
2.
Update DMIPS number in ”Features” on page 1.
Rev. I – 11/09
Rev. H – 03/09
Rev. G – 01/09
Rev. F – 08/08
816
32058K
AVR32-01/12
AT32UC3A
42.6
Rev. E – 04/08
42.7
1.
Open Drain Mode removed from ”General-Purpose Input/Output Controller (GPIO)”
on page 151.
1.
Updated ”Signal Description List” on page 8. Removed RXDN and TXDN from
USART section.
2.
Updated ”Errata” on page 779. Rev G replaced by rev H.
1.
Updated ”Signal Description List” on page 8. Removed RXDN and TXDN from
USART section.
2.
Updated ”Errata” on page 779. Rev G replaced by rev H.
1.
Updated ”Features” on page 1.
2.
Update ”Blockdiagram” on page 4 with local bus.
3.
Updated ”Peripherals” on page 34 with local bus.
4.
Add SPI feature in ”Universial Synchronous/Asynchronous Receiver/Transmitter
(USART)” on page 315.
5.
Updated ”USB On-The-Go Interface (USBB)” on page 517.
6.
Updated ”JTAG and Boundary Scan” on page 750 with programming procedure .
7.
Add description for silicon Rev G.
1.
Initial revision.
Rev. D – 04/08
42.8
Rev. C – 10/07
42.9
Rev. B – 10/07
42.10
Rev. A – 03/07
817
32058K
AVR32-01/12
AT32UC3A
Table of Contents
1
Description . ............................................................................................. 3
2
Configuration Summary . ........................................................................ 4
3
Abbreviations . ......................................................................................... 4
4
Blockdiagram . ......................................................................................... 5
4.1Processor and architecture . .......................................................................................6
5
Signals Description . ................................................................................ 8
6
Power Considerations . ......................................................................... 13
6.1Power Supplies . .......................................................................................................13
6.2Voltage Regulator . ...................................................................................................14
6.3Analog-to-Digital Converter (A.D.C) reference. . ......................................................15
7
Package and Pinout . ............................................................................. 16
8
I/O Line Considerations . ....................................................................... 20
8.1JTAG pins .................................................................................................................20
8.2RESET_N pin . .........................................................................................................20
8.3TWI pins . .................................................................................................................20
8.4GPIO pins . ...............................................................................................................20
9
Processor and Architecture . ................................................................ 21
9.1AVR32 Architecture . ................................................................................................21
9.2The AVR32UC CPU . ...............................................................................................21
9.3Programming Model . .................................