AT32UC3L016/32/64

Features
• High-performance, Low-power 32-bit Atmel® AVR® Microcontroller
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
•
– Compact Single-cycle RISC Instruction Set Including DSP Instructions
– Read-modify-write Instructions and Atomic Bit Manipulation
– Performance
• Up to 64DMIPS Running at 50MHz from Flash (1 Flash Wait State)
• Up to 36DMIPS Running at 25MHz from Flash (0 Flash Wait State)
– Memory Protection Unit (MPU)
• Secure Access Unit (SAU) providing User-defined Peripheral Protection
picoPower® Technology for Ultra-low Power Consumption
Multi-hierarchy Bus System
– High-performance Data Transfers on Separate Buses for Increased Performance
– 12 Peripheral DMA Channels Improve Speed for Peripheral Communication
Internal High-speed Flash
– 64Kbytes, 32Kbytes, and 16Kbytes Versions
– Single-cycle Access up to 25MHz
– FlashVault Technology Allows Pre-programmed Secure Library Support for End
User Applications
– Prefetch Buffer Optimizing Instruction Execution at Maximum Speed
– 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
– 16Kbytes (64Kbytes and 32Kbytes Flash), or 8Kbytes (16Kbytes Flash)
Interrupt Controller (INTC)
– Autovectored Low-latency Interrupt Service with Programmable Priority
External Interrupt Controller (EIC)
Peripheral Event System for Direct Peripheral to Peripheral Communication
System Functions
– Power and Clock Manager
– SleepWalking Power Saving Control
– Internal System RC Oscillator (RCSYS)
– 32KHz Oscillator
– Multipurpose Oscillator and Digital Frequency Locked Loop (DFLL)
Windowed Watchdog Timer (WDT)
Asynchronous Timer (AST) with Real-time Clock Capability
– Counter or Calendar Mode Supported
Frequency Meter (FREQM) for Accurate Measuring of Clock Frequency
Six 16-bit Timer/Counter (TC) Channels
– External Clock Inputs, PWM, Capture, and Various Counting Capabilities
36 PWM Channels (PWMA)
– 8-bit PWM with a Source Clock up to 150MHz
Four Universal Synchronous/Asynchronous Receiver/Transmitters (USART)
– Independent Baudrate Generator, Support for SPI
– Support for Hardware Handshaking
One Master/Slave Serial Peripheral Interfaces (SPI) with Chip Select Signals
– Up to 15 SPI Slaves can be Addressed
Two Master and Two Slave Two-wire Interface (TWI), 400kbit/s I2C-compatible
One 8-channel Analog-to-digital Converter (ADC) with up to 12 Bits Resolution
– Internal Temperature Sensor
32-bit Atmel
AVR
Microcontroller
AT32UC3L064
AT32UC3L032
AT32UC3L016
32099I–01/2012
AT32UC3L016/32/64
• Eight Analog Comparators (AC) with Optional Window Detection
• Capacitive Touch (CAT) Module
•
•
•
•
•
– Hardware-assisted Atmel® AVR® QTouch® and Atmel® AVR® QMatrix Touch Acquisition
– Supports QTouch and QMatrix Capture from Capacitive Touch Sensors
QTouch Library Support
– Capacitive Touch Buttons, Sliders, and Wheels
– QTouch and QMatrix Acquisition
On-chip Non-intrusive Debug System
– Nexus Class 2+, Runtime Control, Non-intrusive Data and Program Trace
– aWire Single-pin Programming Trace and Debug Interface Muxed with Reset Pin
– NanoTrace Provides Trace Capabilities through JTAG or aWire Interface
48-pin TQFP/QFN/TLLGA (36 GPIO Pins)
Five High-drive I/O Pins
Single 1.62-3.6 V Power Supply
2
32099I–01/2012
AT32UC3L016/32/64
1. Description
The Atmel® AVR® AT32UC3L016/32/64 is a complete system-on-chip microcontroller based on
the AVR32 UC RISC processor running at frequencies up to 50MHz. 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 and real-time operating systems. The Secure Access Unit (SAU) is
used together with the MPU to provide the required security and integrity.
Higher computation capability is achieved using a rich set of DSP instructions.
The AT32UC3L016/32/64 embeds state-of-the-art picoPower technology for ultra-low power
consumption. Combined power control techniques are used to bring active current consumption
down to 165 µA/MHz, and leakage down to 9nA while still retaining a bank of backup registers.
The device allows a wide range of trade-offs between functionality and power consumption, giving the user the ability to reach the lowest possible power consumption with the feature set
required for the application.
The Peripheral Direct Memory Access (DMA) controller enables data transfers between peripherals and memories without processor involvement. The Peripheral DMA controller drastically
reduces processing overhead when transferring continuous and large data streams.
The AT32UC3L016/32/64 incorporates on-chip Flash and SRAM memories for secure and fast
access. The FlashVault technology allows secure libraries to be programmed into the device.
The secure libraries can be executed while the CPU is in Secure State, but not read by nonsecure software in the device. The device can thus be shipped to end customers, who will be
able to program their own code into the device to access the secure libraries, but without risk of
compromising the proprietary secure code.
The External Interrupt Controller (EIC) allows pins to be configured as external interrupts. Each
external interrupt has its own interrupt request and can be individually masked.
The Peripheral Event System allows peripherals to receive, react to, and send peripheral events
without CPU intervention. Asynchronous interrupts allow advanced peripheral operation in low
power sleep modes.
The Power Manager (PM) improves design flexibility and security. The Power Manager supports
SleepWalking functionality, by which a module can be selectively activated based on peripheral
events, even in sleep modes where the module clock is stopped. Power monitoring is supported
by on-chip Power-on Reset (POR), Brown-out Detector (BOD), and Supply Monitor (SM). The
device features several oscillators, such as Digital Frequency Locked Loop (DFLL), Oscillator 0
(OSC0), and system RC oscillator (RCSYS). Either of these oscillators can be used as source
for the system clock. The DFLL is a programmable internal oscillator from 40 to 150MHz. It can
be tuned to a high accuracy if an accurate refernce clock is running, e.g. the 32 KHz crystal
oscillator.
The Watchdog Timer (WDT) will reset the device unless it is periodically serviced by the software. This allows the device to recover from a condition that has caused the system to be
unstable.
The Asynchronous Timer (AST) combined with the 32KHz crystal oscillator supports powerful
real-time clock capabilities, with a maximum timeout of up to 136 years. The AST can operate in
counter mode or calendar mode.
3
32099I–01/2012
AT32UC3L016/32/64
The Frequency Meter (FREQM) allows accurate measuring of a clock frequency by comparing it
to a known reference clock.
The device includes six identical 16-bit Timer/Counter (TC) channels. Each channel can be independently programmed to perform frequency measurement, event counting, interval
measurement, pulse generation, delay timing, and pulse width modulation.
The Pulse Width Modulation controller (PWMA) provides 8-bit PWM channels which can be synchronized and controlled from a common timer. One PWM channel is available for each I/O pin
on the device, enabling applications that require multiple PWM outputs, such as LCD backlight
control. The PWM channels can operate independently, with duty cycles set independently from
each other, or in interlinked mode, with multiple channels changed at the same time.
The AT32UC3L016/32/64 also features many communication interfaces, like USART, SPI, and
TWI, for communication intensive applications. The USART supports different communication
modes, like SPI Mode and LIN Mode.
A general purpose 8-channel ADC is provided, as well as eight analog comparators (AC). The
ADC can operate in 10-bit mode at full speed or in enhanced mode at reduced speed, offering
up to 12-bit resolution. The ADC also provides an internal temperature sensor input channel.
The analog comparators can be paired to detect when the sensing voltage is within or outside
the defined reference window.
The Capacitive Touch (CAT) module senses touch on external capacitive touch sensors, using
the QTouch technology. Capacitive touch sensors use no external mechanical components,
unlike normal push buttons, and therefore demand less maintenance in the user application.
The CAT module allows up to 17 touch sensors, or up to 16 by 8 matrix sensors to be interfaced.
One touch sensor can be configured to operate autonomously without software interaction,
allowing wakeup from sleep modes when activated.
Atmel offers the QTouch library for embedding capacitive touch buttons, sliders, and wheels
functionality into AVR microcontrollers. The patented charge-transfer signal acquisition offers
robust sensing and includes fully debounced reporting of touch keys as well as Adjacent Key
Suppression® (AKS®) technology for unambiguous detection of key events. The easy-to-use
QTouch Suite toolchain allows you to explore, develop, and debug your own touch applications.
The AT32UC3L016/32/64 integrates a class 2+ Nexus 2.0 On-chip Debug (OCD) System, with
non-intrusive real-time trace and full-speed read/write memory access, in addition to basic runtime control. The NanoTrace interface enables trace feature for aWire- or JTAG-based
debuggers. The single-pin aWire interface allows all features available through the JTAG interface to be accessed through the RESET pin, allowing the JTAG pins to be used for GPIO or
peripherals.
4
32099I–01/2012
AT32UC3L016/32/64
2. Overview
Block Diagram
DATAOUT
aWire
NEXUS
CLASS 2+
OCD
MEMORY PROTECTION UNIT
INSTR
INTERFACE
DATA
INTERFACE
M
M
M
SAU
S
HSB-PB
BRIDGE B
GENERALPURPOSE I/Os
S
POWER MANAGER
CLOCK
CONTROLLER
SLEEP
CONTROLLER
RESET
CONTROLLER
64/32/16 KB
FLASH
M
S
CONFIGURATION
PA
PB
S
HIGH SPEED
BUS MATRIX
S/M
LOCAL BUS
16/8 KB
SRAM
REGISTERS BUS
PERIPHERAL
DMA
CONTROLLER
HSB-PB
BRIDGE A
DMA
RESET_N
JTAG
INTERFACE
LOCAL BUS
INTERFACE
CAPACITIVE TOUCH
MODULE
DMA
EVTO_N
TCK
TDO
TDI
TMS
AVR32UC CPU
FLASH
CONTROLLER
MCKO
MDO[5..0]
MSEO[1..0]
EVTI_N
MEMORY INTERFACE
Block Diagram
USART0
USART1
USART2
USART3
DMA
Figure 2-1.
SPI
DMA
2.1
TWI MASTER 0
TWI MASTER 1
DIS
VDIVEN
CSA[16:0]
CSB[16:0]
SMP
SYNC
RXD
TXD
CLK
RTS, CTS
RCSYS
RC32K
XIN32
XOUT32
OSC32K
XIN0
XOUT0
OSC0
SYSTEM CONTROL
INTERFACE
DFLL
INTERRUPT
CONTROLLER
EXTINT[5..1]
NMI
PWMA[35..0]
NPCS[3..0]
TWCK
TWD
TWALM
TWCK
DMA
RC120M
TWI SLAVE 0
TWI SLAVE 1
8-CHANNEL ADC
INTERFACE
FREQUENCY METER
TRIGGER
AD[8..0]
A[2..0]
TIMER/COUNTER 0
TIMER/COUNTER 1
B[2..0]
CLK[2..0]
ASYNCHRONOUS
TIMER
WATCHDOG
TIMER
TWALM
ADVREFP
EXTERNAL INTERRUPT
CONTROLLER
PWM CONTROLLER
TWD
PA
PB
ADP[1..0]
DMA
RC32OUT
MISO, MOSI
GENERAL PURPOSE I/Os
SCK
GCLK[4..0]
AC INTERFACE
GLUE LOGIC
CONTROLLER
ACBP[3..0]
ACBN[3..0]
ACAP[3..0]
ACAN[3..0]
ACREFN
OUT[1:0]
IN[7..0]
5
32099I–01/2012
AT32UC3L016/32/64
2.2
Configuration Summary
Table 2-1.
Configuration Summary
Feature
AT32UC3L064
AT32UC3L032
AT32UC3L016
Flash
64KB
32KB
16KB
SRAM
16KB
16KB
8KB
GPIO
36
High-drive pins
5
External Interrupts
6
TWI
2
USART
4
Peripheral DMA Channels
12
Peripheral Event System
1
SPI
1
Asynchronous Timers
1
Timer/Counter Channels
6
PWM channels
36
Frequency Meter
1
Watchdog Timer
1
Power Manager
1
Secure Access Unit
1
Glue Logic Controller
1
Oscillators
ADC
Digital Frequency Locked Loop 40-150MHz (DFLL)
Crystal Oscillator 3-16MHz (OSC0)
Crystal Oscillator 32KHz (OSC32K)
RC Oscillator 120MHz (RC120M)
RC Oscillator 115kHz (RCSYS)
RC Oscillator 32kHz (RC32K)
8-channel 12-bit
Temperature Sensor
1
Analog Comparators
8
Capacitive Touch Module
1
JTAG
1
aWire
1
Max Frequency
Packages
50MHz
TQFP48/QFN48/TLLGA48
6
32099I–01/2012
AT32UC3L016/32/64
3. Package and Pinout
3.1
Package
The device pins are multiplexed with peripheral functions as described in Section 3.2.
TQFP48/QFN48 Pinout
36
35
34
33
32
31
30
29
28
27
26
25
PA14
VDDANA
ADVREFP
GNDANA
PB08
PB07
PB06
PB09
PA04
PA11
PA13
PA20
Figure 3-1.
PA15
PA16
PA17
PA19
PA18
VDDIO
GND
PB11
GND
PA10
PA12
VDDIO
37
38
39
40
41
42
43
44
45
46
47
48
24
23
22
21
20
19
18
17
16
15
14
13
PA21
PB10
RESET_N
PB04
PB05
GND
VDDCORE
VDDIN
PB01
PA07
PA01
PA02
12
11
10
9
8
7
6
5
4
3
2
1
PA05
PA00
PA06
PA22
PB03
PB02
PB00
PB12
PA03
PA08
PA09
GND
7
32099I–01/2012
AT32UC3L016/32/64
TLLGA48 Pinout
37
36
35
34
33
32
31
30
29
28
27
26
25
PA15
PA14
VDDANA
ADVREFP
GNDANA
PB08
PB07
PB06
PB09
PA04
PA11
PA13
PA20
Figure 3-2.
PA16
PA17
PA19
PA18
VDDIO
GND
PB11
GND
PA10
PA12
VDDIO
24
23
22
21
20
19
18
17
16
15
14
38
39
40
41
42
43
44
45
46
47
48
PA21
PB10
RESET_N
PB04
PB05
GND
VDDCORE
VDDIN
PB01
PA07
PA01
13
12
11
10
9
8
7
6
5
4
3
2
1
PA02
PA05
PA00
PA06
PA22
PB03
PB02
PB00
PB12
PA03
PA08
PA09
GND
8
32099I–01/2012
AT32UC3L016/32/64
3.2
Peripheral Multiplexing on I/O lines
3.2.1
Multiplexed signals
Each GPIO line can be assigned to one of the peripheral functions.The following table describes
the peripheral signals multiplexed to the GPIO lines.
Table 3-1.
GPIO Controller Function Multiplexing
48pin
PIN
G
P
I
O
11
PA00
0
14
PA01
13
GPIO Function
Pin
Type
A
B
C
VDDIO
Normal
I/O
USART0
TXD
USART1
RTS
SPI
NPCS[2]
1
VDDIO
Normal
I/O
USART0
RXD
USART1
CTS
SPI
NPCS[3]
USART1
CLK
PWMA
PWMA[1]
PA02
2
VDDIO
Highdrive I/O
USART0
RTS
ADCIFB
TRIGGER
USART2
TXD
TC0
A0
4
PA03
3
VDDIO
Normal
I/O
USART0
CTS
SPI
NPCS[1]
USART2
TXD
28
PA04
4
VDDIO
Normal
I/O
SPI
MISO
TWIMS0
TWCK
12
PA05
5
VDDIO
Normal
I/O (TWI)
SPI
MOSI
10
PA06
6
VDDIO
Highdrive I/O,
5V
tolerant
15
PA07
7
VDDIO
3
PA08
8
2
PA09
46
Supply
D
E
F
G
H
SCIF
GCLK[0]
CAT
CSA[2]
ACIFB
ACAP[0]
TWIMS0
TWALM
CAT
CSA[1]
PWMA
PWMA[2]
ACIFB
ACBP[0]
USART0
CLK
CAT
CSA[3]
TC0
B0
PWMA
PWMA[3]
ACIFB
ACBN[3]
USART0
CLK
CAT
CSB[3]
USART1
RXD
TC0
B1
PWMA
PWMA[4]
ACIFB
ACBP[1]
TWIMS1
TWCK
USART1
TXD
TC0
A1
PWMA
PWMA[5]
ACIFB
ACBN[0]
SPI
SCK
USART2
TXD
USART1
CLK
TC0
B0
PWMA
PWMA[6]
Normal
I/O (TWI)
SPI
NPCS[0]
USART2
RXD
TWIMS1
TWALM
TWIMS0
TWCK
PWMA
PWMA[7]
VDDIO
Highdrive I/O
USART1
TXD
SPI
NPCS[2]
TC0
A2
ADCIFB
ADP[0]
PWMA
PWMA[8]
9
VDDIO
Highdrive I/O
USART1
RXD
SPI
NPCS[3]
TC0
B2
ADCIFB
ADP[1]
PWMA
PWMA[9]
SCIF
GCLK[2]
EIC
EXTINT[1]
CAT
CSB[4]
PA10
10
VDDIO
Normal
I/O
TWIMS0
TWD
PWMA
PWMA[10]
ACIFB
ACAP[1]
SCIF
GCLK[2]
CAT
CSA[5]
27
PA11
11
VDDIN
Normal
I/O
47
PA12
12
VDDIO
Normal
I/O
26
PA13
13
VDDIN
Normal
I/O
36
PA14
14
VDDIO
37
PA15
15
38
PA16
39
41
PWMA
PWMA[0]
TC0
A0
ACIFB
ACAN[0]
CAT
CSA[7]
TWIMS0
TWD
CAT
CSB[7]
SCIF
GCLK[1]
CAT
CSB[1]
EIC
EXTINT[0]
CAT
CSB[2]
CAT
CSA[4]
PWMA
PWMA[11]
USART2
CLK
TC0
CLK1
CAT
SMP
PWMA
PWMA[12]
ACIFB
ACAN[1]
SCIF
GCLK[3]
CAT
CSB[5]
GLOC
OUT[0]
GLOC
IN[7]
TC0
A0
SCIF
GCLK[2]
PWMA
PWMA[13]
CAT
SMP
EIC
EXTINT[2]
CAT
CSA[0]
Normal
I/O
ADCIFB
AD[0]
TC0
CLK2
USART2
RTS
CAT
SMP
PWMA
PWMA[14]
SCIF
GCLK[4]
CAT
CSA[6]
VDDIO
Normal
I/O
ADCIFB
AD[1]
TC0
CLK1
GLOC
IN[6]
PWMA
PWMA[15]
CAT
SYNC
EIC
EXTINT[3]
CAT
CSB[6]
16
VDDIO
Normal
I/O
ADCIFB
AD[2]
TC0
CLK0
GLOC
IN[5]
PWMA
PWMA[16]
ACIFB
ACREFN
EIC
EXTINT[4]
CAT
CSA[8]
PA17
17
VDDIO
Normal
I/O (TWI)
TWIMS1
TWD
PWMA
PWMA[17]
CAT
SMP
CAT
DIS
CAT
CSB[8]
PA18
18
VDDIO
Normal
I/O
GLOC
IN[4]
PWMA
PWMA[18]
CAT
SYNC
EIC
EXTINT[5]
CAT
CSB[0]
TC0
A1
ADCIFB
AD[4]
TC0
B1
USART2
CTS
9
32099I–01/2012
AT32UC3L016/32/64
Table 3-1.
GPIO Controller Function Multiplexing
40
PA19
19
VDDIO
Normal
I/O
ADCIFB
AD[5]
TC0
A2
TWIMS1
TWALM
PWMA
PWMA[19]
CAT
SYNC
25
PA20
20
VDDIN
Normal
I/O
USART2
TXD
TC0
A1
GLOC
IN[3]
PWMA
PWMA[20]
SCIF
RC32OUT
USART2
RXD
TWIMS0
TWD
TC0
B1
ADCIFB
TRIGGER
PWMA
PWMA[21]
PWMA
PWMAOD[21]
24
PA21
21
VDDIN
Normal
I/O (TWI,
5V
tolerant
SMBus)
9
PA22
22
VDDIO
Normal
I/O
USART0
CTS
USART2
CLK
TC0
B2
CAT
SMP
PWMA
PWMA[22]
ACIFB
ACBN[2]
6
PB00
32
VDDIO
Normal
I/O
USART3
TXD
ADCIFB
ADP[0]
SPI
NPCS[0]
TC0
A1
PWMA
PWMA[23]
ACIFB
ACAP[2]
16
PB01
33
VDDIO
Highdrive I/O
USART3
RXD
ADCIFB
ADP[1]
SPI
SCK
TC0
B1
PWMA
PWMA[24]
7
PB02
34
VDDIO
Normal
I/O
USART3
RTS
USART3
CLK
SPI
MISO
TC0
A2
PWMA
PWMA[25]
8
PB03
35
VDDIO
Normal
I/O
USART3
CTS
USART3
CLK
SPI
MOSI
TC0
B2
VDDIN
Normal
I/O (TWI,
5V
tolerant
SMBus)
TC1
A0
USART1
RTS
USART1
CLK
TC1
B0
USART1
CTS
21
PB04
36
CAT
CSA[10]
CAT
CSA[12]
SCIF
GCLK[0]
CAT
SMP
CAT
CSB[10]
TC1
A0
CAT
CSA[9]
TC1
A1
CAT
CSB[9]
ACIFB
ACAN[2]
SCIF
GCLK[1]
CAT
CSB[11]
PWMA
PWMA[26]
ACIFB
ACBP[2]
TC1
A2
CAT
CSA[11]
TWIMS0
TWALM
PWMA
PWMA[27]
PWMA
PWMAOD[27]
TWIMS1
TWCK
CAT
CSA[14]
USART1
CLK
TWIMS0
TWCK
PWMA
PWMA[28]
PWMA
PWMAOD[28]
SCIF
GCLK[3]
CAT
CSB[14]
20
PB05
37
VDDIN
Normal
I/O (TWI,
5V
tolerant
SMBus)
30
PB06
38
VDDIO
Normal
I/O
TC1
A1
USART3
TXD
ADCIFB
AD[6]
GLOC
IN[2]
PWMA
PWMA[29]
ACIFB
ACAN[3]
EIC
EXTINT[0]
CAT
CSB[13]
31
PB07
39
VDDIO
Normal
I/O
TC1
B1
USART3
RXD
ADCIFB
AD[7]
GLOC
IN[1]
PWMA
PWMA[30]
ACIFB
ACAP[3]
EIC
EXTINT[1]
CAT
CSA[13]
32
PB08
40
VDDIO
Normal
I/O
TC1
A2
USART3
RTS
ADCIFB
AD[8]
GLOC
IN[0]
PWMA
PWMA[31]
CAT
SYNC
EIC
EXTINT[2]
CAT
CSB[12]
29
PB09
41
VDDIO
Normal
I/O
TC1
B2
USART3
CTS
USART3
CLK
PWMA
PWMA[32]
ACIFB
ACBN[1]
EIC
EXTINT[3]
CAT
CSB[15]
23
PB10
42
VDDIN
Normal
I/O
TC1
CLK0
USART1
TXD
USART3
CLK
EIC
EXTINT[4]
CAT
CSB[16]
44
PB11
43
VDDIO
Normal
I/O
TC1
CLK1
USART1
RXD
5
PB12
44
VDDIO
Normal
I/O
TC1
CLK2
TWIMS1
TWALM
GLOC
OUT[1]
PWMA
PWMA[33]
ADCIFB
TRIGGER
PWMA
PWMA[34]
CAT
VDIVEN
EIC
EXTINT[5]
CAT
CSA[16]
CAT
SYNC
PWMA
PWMA[35]
ACIFB
ACBP[3]
SCIF
GCLK[4]
CAT
CSA[15]
See Section 3.3 for a description of the various peripheral signals.
Refer to ”Electrical Characteristics” on page 770 for a description of the electrical properties of
the pin types used.
3.2.1.1
TWI, 5V Tolerant, and SMBUS Pins
Some Normal I/O pins offer TWI, 5V Tolerant, and SMBUS features. These features are only
available when either of the TWI functions or the PWMAOD function in the PWMA are selected
for these pins.
10
32099I–01/2012
AT32UC3L016/32/64
Refer to the ”TWI Pin Characteristics(1)” on page 778 for a description of the electrical properties of the TWI, 5V Tolerant, and SMBUS pins.
3.2.2
Peripheral Functions
Each GPIO line can be assigned to one of several peripheral functions. The following table
describes how the various peripheral functions are selected. The last listed function has priority
in case multiple functions are enabled on the same pin.
Table 3-2.
3.2.3
Function
Description
GPIO Controller Function multiplexing
GPIO and GPIO peripheral selection A to H
Nexus OCD AUX port connections
OCD trace system
aWire DATAOUT
aWire output in two-pin mode
JTAG port connections
JTAG debug port
Oscillators
OSC0, OSC32
JTAG Port Connections
If the JTAG is enabled, the JTAG will take control over a number of pins, irrespectively of the I/O
Controller configuration.
Table 3-3.
3.2.4
Peripheral Functions
JTAG Pinout
48-pin
Pin Name
JTAG Pin
11
PA00
TCK
14
PA01
TMS
13
PA02
TDO
4
PA03
TDI
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 I/O Controller 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 3-4.
Nexus OCD AUX Port Connections
Pin
AXS=1
AXS=0
EVTI_N
PA05
PB08
MDO[5]
PA10
PB00
MDO[4]
PA18
PB04
MDO[3]
PA17
PB05
MDO[2]
PA16
PB03
11
32099I–01/2012
AT32UC3L016/32/64
Table 3-4.
3.2.5
Pin
AXS=1
AXS=0
MDO[1]
PA15
PB02
MDO[0]
PA14
PB09
EVTO_N
PA04
PA04
MCKO
PA06
PB01
MSEO[1]
PA07
PB11
MSEO[0]
PA11
PB12
Oscillator Pinout
The oscillators are not mapped to the normal GPIO functions and their muxings are controlled
by registers in the System Control Interface (SCIF). Please refer to the SCIF chapter for more
information about this.
Table 3-5.
3.2.6
Nexus OCD AUX Port Connections
Oscillator Pinout
48-pin
Pin
Oscillator Function
3
PA08
XIN0
46
PA10
XIN32
26
PA13
XIN32_2
2
PA09
XOUT0
47
PA12
XOUT32
25
PA20
XOUT32_2
Other Functions
The functions listed in Table 3-6 are not mapped to the normal GPIO functions. The aWire DATA
pin will only be active after the aWire is enabled. The aWire DATAOUT pin will only be active
after the aWire is enabled and the 2_PIN_MODE command has been sent. The WAKE_N pin is
always enabled. Please refer to Section 6.1.4 on page 40 for constraints on the WAKE_N pin.
Table 3-6.
Other Functions
48-pin
Pin
Function
27
PA11
WAKE_N
22
RESET_N
aWire DATA
11
PA00
aWire DATAOUT
12
32099I–01/2012
AT32UC3L016/32/64
3.3
Signal Descriptions
The following table gives details on signal name classified by peripheral.
Table 3-7.
Signal Descriptions List
Signal Name
Function
Type
Active
Level
Comments
Analog Comparator Interface - ACIFB
ACAN3 - ACAN0
Negative inputs for comparators "A"
Analog
ACAP3 - ACAP0
Positive inputs for comparators "A"
Analog
ACBN3 - ACBN0
Negative inputs for comparators "B"
Analog
ACBP3 - ACBP0
Positive inputs for comparators "B"
Analog
ACREFN
Common negative reference
Analog
ADC Interface - ADCIFB
AD8 - AD0
Analog Signal
Analog
ADP1 - ADP0
Drive Pin for resistive touch screen
Output
TRIGGER
External trigger
Input
aWire - AW
DATA
aWire data
I/O
DATAOUT
aWire data output for 2-pin mode
I/O
Capacitive Touch Module - CAT
CSA16 - CSA0
Capacitive Sense A
I/O
CSB16 - CSB0
Capacitive Sense B
I/O
DIS
Discharge current control
Analog
SMP
SMP signal
Output
SYNC
Synchronize signal
VDIVEN
Voltage divider enable
Input
Output
External Interrupt Controller - EIC
NMI
Non-Maskable Interrupt
Input
EXTINT5 - EXTINT1
External interrupt
Input
Glue Logic Controller - GLOC
IN7 - IN0
Inputs to lookup tables
OUT1 - OUT0
Outputs from lookup tables
Input
Output
JTAG module - JTAG
TCK
Test Clock
Input
TDI
Test Data In
Input
TDO
Test Data Out
TMS
Test Mode Select
Output
Input
13
32099I–01/2012
AT32UC3L016/32/64
Table 3-7.
Signal Descriptions List
Power Manager - PM
RESET_N
Reset
Input
Low
Pulse Width Modulation Controller - PWMA
PWMA35 - PWMA0
PWMA channel waveforms
Output
PWMAOD35 PWMAOD0
PWMA channel waveforms, open drain
mode
Output
Not all channels support open
drain mode
System Control Interface - SCIF
GCLK4 - GCLK0
Generic Clock Output
Output
RC32OUT
RC32K output at startup
Output
XIN0
Crystal 0 Input
Analog/
Digital
XIN32
Crystal 32 Input (primary location)
Analog/
Digital
XIN32_2
Crystal 32 Input (secondary location)
Analog/
Digital
XOUT0
Crystal 0 Output
Analog
XOUT32
Crystal 32 Output (primary location)
Analog
XOUT32_2
Crystal 32 Output (secondary location)
Analog
Serial Peripheral Interface - SPI
MISO
Master In Slave Out
I/O
MOSI
Master Out Slave In
I/O
NPCS3 - NPCS0
SPI Peripheral Chip Select
I/O
SCK
Clock
I/O
Low
Timer/Counter - TC0, TC1
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 - TWIMS0, TWIMS1
TWALM
SMBus SMBALERT
I/O
TWCK
Two-wire Serial Clock
I/O
TWD
Two-wire Serial Data
I/O
Low
Universal Synchronous/Asynchronous Receiver/Transmitter - USART0, USART1, USART2, USART3
14
32099I–01/2012
AT32UC3L016/32/64
Table 3-7.
Signal Descriptions List
CLK
Clock
CTS
Clear To Send
RTS
Request To Send
RXD
Receive Data
Input
TXD
Transmit Data
Output
Note:
I/O
Input
Low
Output
Low
1. ADCIFB: AD3 does not exist.
Table 3-8.
Signal Description List, Continued
Signal Name
Function
Type
Active
Level
Comments
Power
VDDCORE
Core Power Supply / Voltage Regulator Output
Power
Input/Output
1.62V to 1.98V
VDDIO
I/O Power Supply
Power Input
1.62V to 3.6V. VDDIO should
always be equal to or lower than
VDDIN.
VDDANA
Analog Power Supply
Power Input
1.62V to 1.98V
ADVREFP
Analog Reference Voltage
Power Input
1.62V to 1.98V
VDDIN
Voltage Regulator Input
Power Input
1.62V to 3.6V (1)
GNDANA
Analog Ground
Ground
GND
Ground
Ground
Auxiliary Port - AUX
MCKO
Trace Data Output Clock
Output
MDO5 - MDO0
Trace Data Output
Output
MSEO1 - MSEO0
Trace Frame Control
Output
EVTI_N
Event In
EVTO_N
Event Out
Input
Low
Output
Low
General Purpose I/O pin
PA22 - PA00
Parallel I/O Controller I/O Port 0
I/O
PB12 - PB00
Parallel I/O Controller I/O Port 1
I/O
1.
See Section 6.1 on page 36
15
32099I–01/2012
AT32UC3L016/32/64
3.4
3.4.1
I/O Line Considerations
JTAG Pins
The JTAG is enabled if TCK is low while the RESET_N pin is released. The TCK, TMS, and TDI
pins have pull-up resistors when JTAG is enabled. The TCK pin always has pull-up enabled during reset. The TDO pin is an output, driven at VDDIO, and has no pull-up resistor. The JTAG
pins can be used as GPIO pins and multiplexed with peripherals when the JTAG is disabled.
Please refer to Section 3.2.3 on page 11 for the JTAG port connections.
3.4.2
PA00
Note that PA00 is multiplexed with TCK. PA00 GPIO function must only be used as output in the
application.
3.4.3
RESET_N Pin
The RESET_N pin is a schmitt input and integrates a permanent pull-up resistor to VDDIN. As
the product integrates a power-on reset detector, the RESET_N pin can be left unconnected in
case no reset from the system needs to be applied to the product.
The RESET_N pin is also used for the aWire debug protocol. When the pin is used for debugging, it must not be driven by external circuitry.
3.4.4
TWI Pins PA21/PB04/PB05
When these pins are used for TWI, the pins are open-drain outputs with slew-rate limitation and
inputs with spike filtering. When used as GPIO pins or used for other peripherals, the pins have
the same characteristics as other GPIO pins. Selected pins are also SMBus compliant (refer to
Section 3.2 on page 9). As required by the SMBus specification, these pins provide no leakage
path to ground when the AT32UC3L016/32/64 is powered down. This allows other devices on
the SMBus to continue communicating even though the AT32UC3L016/32/64 is not powered.
After reset a TWI function is selected on these pins instead of the GPIO. Please refer to the
GPIO Module Configuration chapter for details.
3.4.5
TWI Pins PA05/PA07/PA17
When these pins are used for TWI, the pins are open-drain outputs with slew-rate limitation and
inputs with spike filtering. When used as GPIO pins or used for other peripherals, the pins have
the same characteristics as other GPIO pins.
After reset a TWI function is selected on these pins instead of the GPIO. Please refer to the
GPIO Module Configuration chapter for details.
3.4.6
GPIO Pins
All the I/O lines integrate a pull-up resistor. Programming of this pull-up resistor is performed
independently for each I/O line through the GPIO Controller. After reset, I/O lines default as
inputs with pull-up resistors disabled, except PA00. PA20 selects SCIF-RC32OUT (GPIO Function F) as default enabled after reset.
3.4.7
High-Drive Pins
The five pins PA02, PA06, PA08, PA09, and PB01 have high-drive output capabilities. Refer to
Section 32. on page 770 for electrical characteristics.
16
32099I–01/2012
AT32UC3L016/32/64
3.4.8
RC32OUT Pin
3.4.8.1
Clock output at startup
After power-up, the clock generated by the 32kHz RC oscillator (RC32K) will be output on PA20,
even when the device is still reset by the Power-On Reset Circuitry. This clock can be used by
the system to start other devices or to clock a switching regulator to rise the power supply voltage up to an acceptable value.
The clock will be available on PA20, but will be disabled if one of the following conditions are
true:
• PA20 is configured to use a GPIO function other than F (SCIF-RC32OUT)
• PA20 is configured as a General Purpose Input/Output (GPIO)
• The bit FRC32 in the Power Manager PPCR register is written to zero (refer to the Power
Manager chapter)
The maximum amplitude of the clock signal will be defined by VDDIN.
Once the RC32K output on PA20 is disabled it can never be enabled again.
3.4.8.2
3.4.9
XOUT32_2 function
PA20 selects RC32OUT as default enabled after reset. This function is not automatically disabled when the user enables the XOUT32_2 function on PA20. This disturbs the oscillator and
may result in the wrong frequency. To avoid this, RC32OUT must be disabled when XOUT32_2
is enabled.
ADC Input Pins
These pins are regular I/O pins powered from the VDDIO. However, when these pins are used
for ADC inputs, the voltage applied to the pin must not exceed 1.98V. Internal circuitry ensures
that the pin cannot be used as an analog input pin when the I/O drives to VDD. When the pins
are not used for ADC inputs, the pins may be driven to the full I/O voltage range.
17
32099I–01/2012
AT32UC3L016/32/64
4. Processor and Architecture
Rev: 2.1.0.0
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.
4.1
Features
• 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 extension with saturating arithmetic, and a wide variety of multiply instructions
• 3-stage pipeline allowing one instruction per clock cycle for most instructions
– Byte, halfword, word, and double word memory access
– Multiple interrupt priority levels
• MPU allows for operating systems with memory protection
• Secure State for supporting FlashVault technology
4.2
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 halfword data types without penalty in code size and
performance.
Memory load and store operations are provided for byte, halfword, word, and double word data
with automatic sign- or zero extension of halfword 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
18
32099I–01/2012
AT32UC3L016/32/64
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.
4.3
The AVR32UC CPU
The AVR32UC CPU targets low- and medium-performance applications, and provides an
advanced On-Chip Debug (OCD) system, no caches, and a Memory Protection Unit (MPU).
Java acceleration hardware is not implemented.
AVR32UC 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.
A local bus interface is provided for connecting the CPU to device-specific high-speed systems,
such as floating-point units and I/O controller ports. This local bus has to be enabled by writing a
one to 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
CPU Local Bus section in the Memories chapter.
Figure 4-1 on page 20 displays the contents of AVR32UC.
19
32099I–01/2012
AT32UC3L016/32/64
OCD interface
Reset interface
Overview of the AVR32UC CPU
Interrupt controller interface
Figure 4-1.
OCD
system
Power/
Reset
control
AVR32UC CPU pipeline
MPU
4.3.1
High
Speed
Bus slave
CPU Local
Bus
master
CPU Local Bus
High Speed
Bus master
High Speed Bus
High Speed Bus
High Speed Bus master
High Speed Bus
Data memory controller
Instruction memory controller
CPU RAM
Pipeline Overview
AVR32UC has 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 4-2 on page 21 shows an overview of the AVR32UC pipeline stages.
20
32099I–01/2012
AT32UC3L016/32/64
Figure 4-2.
The AVR32UC Pipeline
MUL
IF
ID
Prefetch unit
Decode unit
Regfile
Read
ALU
LS
4.3.2
Multiply unit
Regfile
write
ALU 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.
4.3.2.1
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.
4.3.2.2
Java Support
AVR32UC does not provide Java hardware acceleration.
4.3.2.3
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.
4.3.2.4
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
21
32099I–01/2012
AT32UC3L016/32/64
address exception. Doubleword-sized accesses with word-aligned pointers will automatically be
performed as two word-sized accesses.
The following table shows the instructions with support for unaligned addresses. All other
instructions require aligned addresses.
Table 4-1.
4.3.2.5
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 if no coprocessors are present
• retj, incjosp, popjc, pushjc
• tlbr, tlbs, tlbw
• cache
4.3.2.6
CPU and Architecture Revision
Three major revisions of the AVR32UC CPU currently exist. The device described in this
datasheet uses CPU revision 3.
The Architecture Revision field in the CONFIG0 system register identifies which architecture
revision is implemented in a specific device.
AVR32UC CPU revision 3 is fully backward-compatible with revisions 1 and 2, ie. code compiled
for revision 1 or 2 is binary-compatible with revision 3 CPUs.
22
32099I–01/2012
AT32UC3L016/32/64
4.4
4.4.1
Programming Model
Register File Configuration
The AVR32UC register file is shown below.
Figure 4-3.
The AVR32UC Register File
Application
Supervisor
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
Exception
NMI
Bit 31
Bit 31
Bit 0
Secure
Bit 0
Bit 31
Bit 0
PC
LR
SP_APP
R12
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SMPC
R4
R3
R2
R1
R0
PC
LR
SP_SYS
R12
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SMPC
R4
R3
R2
R1
R0
PC
LR
SP_SYS
R12
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SMPC
R4
R3
R2
R1
R0
PC
LR
SP_SYS
R12
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SMPC
R4
R3
R2
R1
R0
PC
LR
SP_SYS
R12
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SMPC
R4
R3
R2
R1
R0
PC
LR
SP_SYS
R12
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SMPC
R4
R3
R2
R1
R0
PC
LR
SP_SYS
R12
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SMPC
R4
R3
R2
R1
R0
PC
LR
SP_SYS
R12
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SMPC
R4
R3
R2
R1
R0
PC
LR
SP_SEC
R12
R11
R10
R9
R8
INT0PC
R7
INT1PC
R6
FINTPC
R5
SMPC
R4
R3
R2
R1
R0
SR
SR
SR
SR
SR
SR
SR
SR
SR
SS_STATUS
SS_ADRF
SS_ADRR
SS_ADR0
SS_ADR1
SS_SP_SYS
SS_SP_APP
SS_RAR
SS_RSR
4.4.2
Status Register Configuration
The Status Register (SR) is split into two halfwords, one upper and one lower, see Figure 4-4
and Figure 4-5. 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.
Figure 4-4.
The Status Register High Halfword
Bit 31
Bit 16
SS
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 nam e
Initial value
G lobal Interrupt M ask
Interrupt Level 0 M ask
Interrupt Level 1 M ask
Interrupt Level 2 M ask
Interrupt Level 3 M ask
Exception M ask
M ode Bit 0
M ode Bit 1
M ode Bit 2
Reserved
Debug State
Debug State M ask
Reserved
Secure State
23
32099I–01/2012
AT32UC3L016/32/64
Figure 4-5.
The Status Register Low Halfword
Bit 15
Bit 0
-
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
Reserved
4.4.3
Processor States
4.4.3.1
Normal RISC State
The AVR32 processor supports several different execution contexts as shown in Table 4-2.
Table 4-2.
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.
4.4.3.2
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.
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.
24
32099I–01/2012
AT32UC3L016/32/64
Debug state can be entered as described in the AVR32UC Technical Reference Manual.
Debug state is exited by the retd instruction.
4.4.3.3
4.4.4
Secure State
The AVR32 can be set in a secure state, that allows a part of the code to execute in a state with
higher security levels. The rest of the code can not access resources reserved for this secure
code. Secure State is used to implement FlashVault technology. Refer to the AVR32UC Technical Reference Manual for details.
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 4-3.
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
25
32099I–01/2012
AT32UC3L016/32/64
Table 4-3.
System Registers (Continued)
Reg #
Address
Name
Function
24
96
JAVA_LV1
Unused in AVR32UC
25
100
JAVA_LV2
Unused in AVR32UC
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
26
32099I–01/2012
AT32UC3L016/32/64
Table 4-3.
4.5
System Registers (Continued)
Reg #
Address
Name
Function
90
360
MPUPSR2
MPU Privilege Select Register region 2
91
364
MPUPSR3
MPU Privilege Select Register region 3
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
412
SS_STATUS
Secure State Status Register
104
416
SS_ADRF
Secure State Address Flash Register
105
420
SS_ADRR
Secure State Address RAM Register
106
424
SS_ADR0
Secure State Address 0 Register
107
428
SS_ADR1
Secure State Address 1 Register
108
432
SS_SP_SYS
Secure State Stack Pointer System Register
109
436
SS_SP_APP
Secure State Stack Pointer Application Register
110
440
SS_RAR
Secure State Return Address Register
111
444
SS_RSR
Secure State Return Status Register
112-191
448-764
Reserved
Reserved for future use
192-255
768-1020
IMPL
IMPLEMENTATION DEFINED
Exceptions and Interrupts
In the AVR32 architecture, events are used as a common term for exceptions and interrupts.
AVR32UC incorporates a powerful event handling scheme. The different event sources, like Illegal Op-code and interrupt requests, have different priority levels, ensuring a well-defined
behavior when multiple events are received simultaneously. Additionally, pending events of a
higher priority class may preempt handling of ongoing events of a lower priority class.
When an event occurs, the execution of the instruction stream is halted, and execution is passed
to an event handler at an address specified in Table 4-4 on page 31. 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 interrupt sources have autovectored interrupt service routine (ISR)
addresses. This allows the interrupt controller to directly specify the ISR address as an address
27
32099I–01/2012
AT32UC3L016/32/64
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
interrupt requests, yielding a uniform event handling scheme.
An interrupt controller does the priority handling of the interrupts and provides the autovector offset to the CPU.
4.5.1
System Stack Issues
Event handling in AVR32UC 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.
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.
4.5.2
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 4-4 on
page 31, 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.
28
32099I–01/2012
AT32UC3L016/32/64
4.5.3
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.
4.5.4
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
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.
4.5.5
Entry Points for Events
Several different event handler entry points exist. In AVR32UC, the reset address is
0x80000000. 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 interrupt requests have entry points located at an offset relative to EVBA. This autovector offset is specified by an 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 4-4 on page 31. 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
29
32099I–01/2012
AT32UC3L016/32/64
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 4-4 on page 31. Some of
the exceptions are unused in AVR32UC since it has no MMU, coprocessor interface, or floatingpoint unit.
30
32099I–01/2012
AT32UC3L016/32/64
Table 4-4.
Priority and Handler Addresses for Events
Priority
Handler Address
Name
Event source
Stored Return Address
1
0x80000000
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
PC of offending instruction
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
PC of offending instruction
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
Instruction
PC of offending instruction
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
PC of offending instruction
25
EVBA+0x70
DTLB Miss (Write)
MPU
PC of offending instruction
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
31
32099I–01/2012
AT32UC3L016/32/64
5. Memories
5.1
Embedded Memories
• Internal high-speed flash
– 64Kbytes (AT32UC3L064)
– 32Kbytes (AT32UC3L032)
– 16Kbytes (AT32UC3L016)
• 0 wait state access at up to 25mhz in worst case conditions
• 1 wait state access at up to 50mhz 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 8% compared to 0 wait state operation
• 100 000 write cycles, 15-year data retention capability
• 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
– 16Kbytes (AT32UC3L064, AT32UC3L032)
– 8Kbytes (AT32UC3L016)
5.2
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 5-1.
AT32UC3L016/32/64 Physical Memory Map
Device
Table 5-2.
Start Address
Size
AT32UC3L064
AT32UC3L032
AT32UC3L016
Embedded SRAM
0x00000000
16Kbytes
16Kbytes
8Kbytes
Embedded Flash
0x80000000
64Kbytes
32Kbytes
16Kbytes
SAU Channels
0x90000000
256 bytes
256 bytes
256 bytes
HSB-PB Bridge B
0xFFFE0000
64Kbytes
64Kbytes
64Kbytes
HSB-PB Bridge A
0xFFFF0000
64Kbytes
64Kbytes
64Kbytes
Flash Memory Parameters
Part Number
Flash Size (FLASH_PW)
Number of pages
(FLASH_P)
Page size
(FLASH_W)
AT32UC3L064
64Kbytes
256
256 bytes
AT32UC3L032
32Kbytes
128
256 bytes
AT32UC3L016
16Kbytes
64
256 bytes
32
32099I–01/2012
AT32UC3L016/32/64
5.3
Peripheral Address Map
Table 5-3.
Peripheral Address Mapping
Address
Peripheral Name
0xFFFE0000
FLASHCDW
Flash Controller - FLASHCDW
0xFFFE0400
HMATRIX
HSB Matrix - HMATRIX
0xFFFE0800
SAU
Secure Access Unit - SAU
0xFFFF0000
PDCA
Peripheral DMA Controller - PDCA
INTC
Interrupt controller - INTC
0xFFFF1000
0xFFFF1400
PM
Power Manager - PM
0xFFFF1800
SCIF
System Control Interface - SCIF
AST
Asynchronous Timer - AST
WDT
Watchdog Timer - WDT
EIC
External Interrupt Controller - EIC
0xFFFF1C00
0xFFFF2000
0xFFFF2400
0xFFFF2800
FREQM
Frequency Meter - FREQM
0xFFFF2C00
GPIO
0xFFFF3000
General Purpose Input/Output Controller - GPIO
USART0
Universal Synchronous/Asynchronous
Receiver/Transmitter - USART0
USART1
Universal Synchronous/Asynchronous
Receiver/Transmitter - USART1
USART2
Universal Synchronous/Asynchronous
Receiver/Transmitter - USART2
USART3
Universal Synchronous/Asynchronous
Receiver/Transmitter - USART3
0xFFFF3400
0xFFFF3800
0xFFFF3C00
0xFFFF4000
SPI
Serial Peripheral Interface - SPI
0xFFFF4400
TWIM0
Two-wire Master Interface - TWIM0
33
32099I–01/2012
AT32UC3L016/32/64
Table 5-3.
Peripheral Address Mapping
0xFFFF4800
TWIM1
Two-wire Master Interface - TWIM1
TWIS0
Two-wire Slave Interface - TWIS0
TWIS1
Two-wire Slave Interface - TWIS1
PWMA
Pulse Width Modulation Controller - PWMA
0xFFFF4C00
0xFFFF5000
0xFFFF5400
0xFFFF5800
TC0
Timer/Counter - TC0
TC1
Timer/Counter - TC1
0xFFFF5C00
0xFFFF6000
ADCIFB
ADC Interface - ADCIFB
0xFFFF6400
ACIFB
Analog Comparator Interface - ACIFB
0xFFFF6800
CAT
Capacitive Touch Module - CAT
0xFFFF6C00
GLOC
Glue Logic Controller - GLOC
0xFFFF7000
AW
5.4
aWire - AW
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.
34
32099I–01/2012
AT32UC3L016/32/64
The following GPIO registers are mapped on the local bus:
Table 5-4.
Local Bus Mapped GPIO Registers
Port
Register
Mode
Local Bus
Address
Access
0
Output Driver Enable Register (ODER)
WRITE
0x40000040
Write-only
SET
0x40000044
Write-only
CLEAR
0x40000048
Write-only
TOGGLE
0x4000004C
Write-only
WRITE
0x40000050
Write-only
SET
0x40000054
Write-only
CLEAR
0x40000058
Write-only
TOGGLE
0x4000005C
Write-only
Pin Value Register (PVR)
-
0x40000060
Read-only
Output Driver Enable Register (ODER)
WRITE
0x40000140
Write-only
SET
0x40000144
Write-only
CLEAR
0x40000148
Write-only
TOGGLE
0x4000014C
Write-only
WRITE
0x40000150
Write-only
SET
0x40000154
Write-only
CLEAR
0x40000158
Write-only
TOGGLE
0x4000015C
Write-only
-
0x40000160
Read-only
Output Value Register (OVR)
1
Output Value Register (OVR)
Pin Value Register (PVR)
35
32099I–01/2012
AT32UC3L016/32/64
6. Supply and Startup Considerations
6.1
6.1.1
Supply Considerations
Power Supplies
The AT32UC3L016/32/64 has several types of power supply pins:
•VDDIO: Powers I/O lines. Voltage is 1.8 to 3.3V nominal.
•VDDIN: Powers I/O lines and the internal regulator. Voltage is 1.8 to 3.3V nominal.
•VDDANA: Powers the ADC. Voltage is 1.8V nominal.
•VDDCORE: Powers the core, memories, and peripherals. Voltage is 1.8V nominal.
The ground pins GND are common to VDDCORE, VDDIO, and VDDIN. The ground pin for
VDDANA is GNDANA.
When VDDCORE is not connected to VDDIN, the VDDIN voltage must be higher than 1.98V.
Refer to Section 32. on page 770 for power consumption on the various supply pins.
For decoupling recommendations for the different power supplies, please refer to the schematic
checklist.
6.1.2
Voltage Regulator
The AT32UC3L016/32/64 embeds a voltage regulator that converts from 3.3V nominal to 1.8V
with a load of up to 60mA. The regulator supplies the output voltage on VDDCORE. The regulator may only be used to drive internal circuitry in the device. VDDCORE should be externally
connected to the 1.8V domains. See Section 6.1.3 for regulator connection figures.
Adequate output supply decoupling is mandatory for VDDCORE to reduce ripple and avoid
oscillations. The best way to achieve this is to use two capacitors in parallell between
VDDCORE and GND as close to the device as possible. Please refer to Section 32.8.1 on page
784 for decoupling capacitors values and regulator characteristics.
Figure 6-1.
Supply Decoupling
3.3V
VDDIN
C IN3
CIN2
CIN1
1.8V
VDDCORE
COUT2
6.1.3
1.8V
Regulator
COUT1
Regulator Connection
The AT32UC3L016/32/64 supports three power supply configurations:
• 3.3V single supply mode
• 1.8V single supply mode
• 3.3V supply mode, with 1.8V regulated I/O lines
36
32099I–01/2012
AT32UC3L016/32/64
6.1.3.1
3.3V Single Supply Mode
In 3.3V single supply mode the internal regulator is connected to the 3.3V source (VDDIN pin)
and its output feeds VDDCORE. Figure 6-2 shows the power schematics to be used for 3.3V
single supply mode. All I/O lines will be powered by the same power (VDDIN=VDDIO).
Figure 6-2.
3.3V Single Supply Mode
+
1.98-3.6V
-
VDDIN
VDDIO
I/O Pins
I/O Pins
VDDCORE
OSC32K
RC32K
AST
Wake
POR33
SM33
Linear
regulator
VDDANA
GND
ADC
CPU,
Peripherals,
Memories,
SCIF, BOD,
RCSYS,
DFLL
GNDANA
37
32099I–01/2012
AT32UC3L016/32/64
6.1.3.2
1.8V Single Supply Mode
In 1.8V single supply mode the internal regulator is not used, and VDDIO and VDDCORE are
powered by a single 1.8V supply as shown in Figure 6-3. All I/O lines will be powered by the
same power (VDDIN = VDDIO = VDDCORE).
Figure 6-3.
1.8V Single Supply Mode.
+
1.62-1.98V
-
VDDIN
VDDIO
I/O Pins
I/O Pins
VDDCORE
Linear
Regulator
VDDANA
OSC32K
RC32K
AST
Wake
POR33
SM33
ADC
GNDANA
GND
CPU,
Peripherals,
Memories,
SCIF, BOD,
RCSYS,
DFLL
38
32099I–01/2012
AT32UC3L016/32/64
6.1.3.3
3.3V Supply Mode with 1.8V Regulated I/O Lines
In this mode, the internal regulator is connected to the 3.3V source and its output is connected
to both VDDCORE and VDDIO as shown in Figure 6-4. This configuration is required in order to
use Shutdown mode.
Figure 6-4.
3.3V Supply Mode with 1.8V Regulated I/O Lines
1.98-3.6V
+
-
VDDIN
VDDIO
I/O Pins
VDDCORE
I/O Pins
Linear
Regulator
VDDANA
OSC32K
RC32K
AST
Wake
POR33
SM33
ADC
GNDANA
GND
CPU,
Peripherals,
Memories,
SCIF, BOD,
RCSYS,
DFLL
In this mode, some I/O lines are powered by VDDIN while other I/O lines are powered by VDDIO.
Refer to Section 3.2 on page 9 for description of power supply for each I/O line.
Refer to the Power Manager chapter for a description of what parts of the system are powered in
Shutdown mode.
Important note: As the regulator has a maximum output current of 60mA, this mode can only be
used in applications where the maximum I/O current is known and compatible with the core and
peripheral power consumption. Typically, great care must be used to ensure that only a few I/O
lines are toggling at the same time and drive very small loads.
39
32099I–01/2012
AT32UC3L016/32/64
6.1.4
Power-up Sequence
6.1.4.1
Maximum Rise Rate
To avoid risk of latch-up, the rise rate of the power supplies must not exceed the values
described in Table 32-3 on page 771.
Recommended order for power supplies is also described in this chapter.
6.1.4.2
Minimum Rise Rate
The integrated Power-on Reset (POR33) circuitry monitoring the VDDIN powering supply
requires a minimum rise rate for the VDDIN power supply.
See Table 32-3 on page 771 for the minimum rise rate value.
If the application can not ensure that the minimum rise rate condition for the VDDIN power supply is met, one of the following configurations can be used:
• A logic “0” value is applied during power-up on pin PA11 until VDDIN rises above 1.2V.
• A logic “0” value is applied during power-up on pin RESET_N until VDDIN rises above 1.2V.
6.2
Startup Considerations
This chapter summarizes the boot sequence of the AT32UC3L016/32/64. The behavior after
power-up is controlled by the Power Manager. For specific details, refer to the Power Manager
chapter.
6.2.1
Starting of Clocks
After power-up, the device will be held in a reset state by the Power-on Reset (POR18 and
POR33) circuitry for a short time to allow the power to stabilize throughout the device. After
reset, the device will use the System RC Oscillator (RCSYS) as clock source. Please refer to
Table 32-17 on page 783 for the frequency for this oscillator.
On system start-up, the DFLL is disabled. All clocks to all modules are running. No clocks have
a divided frequency; all parts of the system receive a clock with the same frequency as the System RC Oscillator.
When powering up the device, there may be a delay before the voltage has stabilized, depending on the rise time of the supply used. The CPU can start executing code as soon as the supply
is above the POR18 and POR33 thresholds, and before the supply is stable. Before switching to
a high-speed clock source, the user should use the BOD to make sure the VDDCORE is above
the minimum level (1.62V).
6.2.2
Fetching of Initial Instructions
After reset has been released, the AVR32 UC CPU starts fetching instructions from the reset
address, which is 0x80000000. This address points to the first address in the internal flash.
The code read from the internal flash is free to configure the clock system and clock sources .
Please refer to the Power Manager and SCIF chapters for details.
40
32099I–01/2012
AT32UC3L016/32/64
7. Peripheral DMA Controller (PDCA)
Rev: 1.2.2.1
7.1
Features
•
•
•
•
•
•
7.2
Multiple channels
Generates transfers between memories and peripherals such as USART and SPI
Two address pointers/counters per channel allowing double buffering
Performance monitors to measure average and maximum transfer latency
Optional synchronizing of data transfers with extenal peripheral events
Ring buffer functionality
Overview
The Peripheral DMA Controller (PDCA) transfers data between on-chip peripheral modules such
as USART, SPI and memories (those memories may be 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 multiple DMA channels. Each channel has:
• A Peripheral Select Register
• 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 set of handshake interfaces. The
peripheral signals the PDCA when it is ready to receive or transmit data. The PDCA acknowledges the request when the transmission has started.
When a transmit buffer is empty or a receive buffer is full, an optional interrupt request can be
generated.
41
32099I–01/2012
AT32UC3L016/32/64
7.3
Block Diagram
Figure 7-1.
PDCA Block Diagram
Peripheral
0
Memory
HSB to PB
Bridge
HSB
Peripheral Bus
HSB
High Speed
Bus Matrix
HSB
Interrupt
Controller
IRQ
Peripheral
2
...
Peripheral DMA
Controller
(PDCA)
Peripheral
1
Peripheral
(n-1)
Handshake Interfaces
7.4
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
7.4.1
Power Management
If the CPU enters a sleep mode that disables the PDCA clocks, the PDCA will stop functioning
and resume operation after the system wakes up from sleep mode.
7.4.2
Clocks
The PDCA has two bus clocks connected: One High Speed Bus clock (CLK_PDCA_HSB) and
one Peripheral Bus clock (CLK_PDCA_PB). These clocks are generated by the Power Manager. Both clocks are enabled at reset, and can be disabled in the Power Manager. It is
recommended to disable the PDCA before disabling the clocks, to avoid freezing the PDCA in
an undefined state.
7.4.3
Interrupts
The PDCA interrupt request lines are connected to the interrupt controller. Using the PDCA
interrupts requires the interrupt controller to be programmed first.
42
32099I–01/2012
AT32UC3L016/32/64
7.4.4
7.5
7.5.1
Peripheral Events
The PDCA peripheral events are connected via the Peripheral Event System. Refer to the
Peripheral Event System chapter for details.
Functional Description
Basic Operation
The PDCA consists of multiple independent PDCA channels, each capable of handling DMA
requests in parallel. Each PDCA channels contains a set of configuration registers which must
be configured to start a DMA transfer.
In this section the steps necessary to configure one PDCA channel is outlined.
The peripheral to transfer data to or from must be configured correctly in the Peripheral Select
Register (PSR). This is performed by writing the Peripheral Identity (PID) value for the corresponding peripheral to the PID field in the PSR register. The PID also encodes the transfer
direction, i.e. memory to peripheral or peripheral to memory. See Section 7.5.6.
The transfer size must be written to the Transfer Size field in the Mode Register (MR.SIZE). The
size must match the data size produced or consumed by the selected peripheral. See Section
7.5.7.
The memory address to transfer to or from, depending on the PSR, must be written to the Memory Address Register (MAR). For each transfer the memory address is increased by either a
one, two or four, depending on the size set in MR. See Section 7.5.2.
The number of data items to transfer is written to the TCR register. If the PDCA channel is
enabled, a transfer will start immediately after writing a non-zero value to TCR or the reload version of TCR, TCRR. After each transfer the TCR value is decreased by one. Both MAR and TCR
can be read while the PDCA channel is active to monitor the DMA progress. See Section 7.5.3.
The channel must be enabled for a transfer to start. A channel is enable by writing a one to the
EN bit in the Control Register (CR).
7.5.2
Memory Pointer
Each channel has a 32-bit Memory Address Register (MAR). This register holds the memory
address for the next transfer to be performed. The register is automatically updated after each
transfer. The address will be increased by either one, two or four depending on the size of the
DMA transfer (byte, halfword or word). The MAR can be read at any time during transfer.
7.5.3
Transfer Counter
Each channel has a 16-bit Transfer Counter Register (TCR). This register must be written with
the number of transfers to be performed. The TCR register should contain the number of data
items to be transferred independently of the transfer size. The TCR can be read at any time during transfer to see the number of remaining transfers.
7.5.4
Reload Registers
Both the MAR and the TCR 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.
After reload, the MARR and TCRR registers are cleared.
43
32099I–01/2012
AT32UC3L016/32/64
If TCR is zero when writing to TCRR, the TCR and MAR are automatically updated with the
value written in TCRR and MARR.
7.5.5
Ring Buffer
When Ring Buffer mode is enabled the TCRR and MARR registers will not be cleared when
TCR and MAR registers reload. This allows the PDCA to read or write to the same memory
region over and over again until the transfer is actively stopped by the user. Ring Buffer mode is
enabled by writing a one to the Ring Buffer bit in the Mode Register (MR.RING).
7.5.6
Peripheral Selection
The Peripheral Select Register (PSR) decides which peripheral should be connected to the
PDCA channel. A peripheral is selected by writing the corresponding Peripheral Identity (PID) to
the PID field in the PSR register. Writing the PID 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. Refer to the Peripheral Identity (PID) table in the Module Configuration section for the peripheral PID values.
7.5.7
Transfer Size
The transfer size can be set individually for each channel to be either byte, halfword or word (8bit, 16-bit or 32-bit respectively). Transfer size is set by writing the desired value to the Transfer
Size field in the Mode Register (MR.SIZE).
When the PDCA moves data between peripherals and memory, data is automatically sized and
aligned. When memory is accessed, the size specified in MR.SIZE and system alignment is
used. When a peripheral register is accessed the data to be transferred is converted to a word
where bit n in the data corresponds to bit n in the peripheral register. If the transfer size is byte or
halfword, bits greater than 8 and16 respectively are set to zero.
Refer to the Module Configuration section for information regarding what peripheral registers are
used for the different peripherals and then to the peripheral specific chapter for information
about the size option available for the different registers.
7.5.8
Enabling and Disabling
Each DMA channel is enabled by writing a one to the Transfer Enable bit in the Control Register
(CR.TEN) and disabled by writing a one to the Transfer Disable bit (CR.TDIS). The current status can be read from the Status Register (SR).
While the PDCA channel is enabled all DMA request will be handled as long the TCR and TCRR
is not zero.
7.5.9
Interrupts
Interrupts can be enabled by writing a one to the corresponding bit in the Interrupt Enable Register (IER) and disabled by writing a one to the corresponding bit in the 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 TCRR register is zero.
• Transfer Finished - Both the TCR and TCRR registers are zero.
• Transfer Error - An error has occurred in accessing memory.
44
32099I–01/2012
AT32UC3L016/32/64
7.5.10
Priority
If more than 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 zero the highest priority.
7.5.11
Error Handling
If the Memory Address Register (MAR) 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
bit in the Interrupt Status Register (ISR.TERR) 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 a one to the Error Clear bit in the Control
Register (CR.ECLR). If the Transfer Error interrupt is enabled, an interrupt request will be generated when a transfer error occurs.
7.5.12
7.6
Peripheral Event Trigger
Peripheral events can be used to trigger PDCA channel transfers. Peripheral Event synchronizations are enabled by writing a one to the Event Trigger bit in the Mode Register (MR.ETRIG).
When set, all DMA requests will be blocked until a peripheral event is received. For each peripheral event received, only one data item is transferred. If no DMA requests are pending when a
peripheral event is received, the PDCA will start a transfer as soon as a peripheral event is
detected. If multiple events are received while the PDCA channel is busy transferring data, an
overflow condition will be signaled in the Peripheral Event System. Refer to the Peripheral Event
System chapter for more information.
Performance Monitors
Up to two performance monitors allow the user to measure the activity and stall cycles for PDCA
transfers. To monitor a PDCA channel, the corresponding channel number must be written to
one of the MON0/1CH fields in the Performance Control Register (PCONTROL) and a one must
be written to the corresponding CH0/1EN bit in the same register.
Due to performance monitor hardware resource sharing, the two monitor channels should NOT
be programmed to monitor the same PDCA channel. This may result in UNDEFINED performance monitor behavior.
7.6.1
Measuring mechanisms
Three different parameters can be measured by each channel:
• The number of data transfer cycles since last channel reset, both for read and write
• The number of stall cycles since last channel reset, both for read and write
• The maximum latency since last channel reset, both for read and write
These measurements can be extracted by software and used to generate indicators for bus
latency, bus load, and maximum bus latency.
Each of the counters has a fixed width, and may therefore overflow. When an overflow is
encountered in either the Performance Channel Data Read/Write Cycle registers (PRDATA0/1
and PWDATA0/1) or the Performance Channel Read/Write Stall Cycles registers (PRSTALL0/1
and PWSTALL0/1) of a channel, all registers in the channel are reset. This behavior is altered if
the Channel Overflow Freeze bit is one in the Performance Control register (PCONTROL.CH0/1OVF). If this bit is one, the channel registers are frozen when either DATA or
STALL reaches its maximum value. This simplifies one-shot readout of the counter values.
45
32099I–01/2012
AT32UC3L016/32/64
The registers can also be manually reset by writing a one to the Channel Reset bit in the PCONTROL register (PCONTROL.CH0/1RES). The Performance Channel Read/Write Latency
registers (PRLAT0/1 and PWLAT0/1) are saturating when their maximum count value is
reached. The PRLAT0/1 and PWLAT0/1 registers can only be reset by writing a one to the corresponding reset bit in PCONTROL (PCONTROL.CH0/1RES).
A counter is enabled by writing a one to the Channel Enable bit in the Performance Control Register (PCONTROL.CH0/1EN).
46
32099I–01/2012
AT32UC3L016/32/64
7.7
User Interface
7.7.1
Memory Map Overview
Table 7-1.
PDCA Register Memory Map
Address Range
Contents
0x000 - 0x03F
DMA channel 0 configuration registers
0x040 - 0x07F
DMA channel 1 configuration registers
...
...
(0x000 - 0x03F)+m*0x040
DMA channel m configuration registers
0x800-0x830
Performance Monitor registers
0x834
Version register
The channels are mapped as shown in Table 7-1. Each channel has a set of configuration registers, shown in Table 7-2, where n is the channel number.
7.7.2
Channel Memory Map
Table 7-2.
PDCA Channel Configuration Registers
Offset
Register
Register Name
Access
Reset
0x000 + n*0x040
Memory Address Register
MAR
Read/Write
0x00000000
0x004 + n*0x040
Peripheral Select Register
PSR
Read/Write
- (1)
0x008 + n*0x040
Transfer Counter Register
TCR
Read/Write
0x00000000
0x00C + n*0x040
Memory Address Reload Register
MARR
Read/Write
0x00000000
0x010 + n*0x040
Transfer Counter Reload Register
TCRR
Read/Write
0x00000000
0x014 + n*0x040
Control Register
CR
Write-only
0x00000000
0x018 + n*0x040
Mode Register
MR
Read/Write
0x00000000
0x01C + n*0x040
Status Register
SR
Read-only
0x00000000
0x020 + n*0x040
Interrupt Enable Register
IER
Write-only
0x00000000
0x024 + n*0x040
Interrupt Disable Register
IDR
Write-only
0x00000000
0x028 + n*0x040
Interrupt Mask Register
IMR
Read-only
0x00000000
0x02C + n*0x040
Interrupt Status Register
ISR
Read-only
0x00000000
Note:
1. The reset values are device specific. Please refer to the Module Configuration section at the
end of this chapter.
47
32099I–01/2012
AT32UC3L016/32/64
7.7.3
Performance Monitor Memory Map
PDCA Performance Monitor Registers(1)
Table 7-3.
Offset
Register
Register Name
Access
Reset
0x800
Performance Control Register
PCONTROL
Read/Write
0x00000000
0x804
Channel0 Read Data Cycles
PRDATA0
Read-only
0x00000000
0x808
Channel0 Read Stall Cycles
PRSTALL0
Read-only
0x00000000
0x80C
Channel0 Read Max Latency
PRLAT0
Read-only
0x00000000
0x810
Channel0 Write Data Cycles
PWDATA0
Read-only
0x00000000
0x814
Channel0 Write Stall Cycles
PWSTALL0
Read-only
0x00000000
0x818
Channel0 Write Max Latency
PWLAT0
Read-only
0x00000000
0x81C
Channel1 Read Data Cycles
PRDATA1
Read-only
0x00000000
0x820
Channel1 Read Stall Cycles
PRSTALL1
Read-only
0x00000000
0x824
Channel1 Read Max Latency
PRLAT1
Read-only
0x00000000
0x828
Channel1 Write Data Cycles
PWDATA1
Read-only
0x00000000
0x82C
Channel1 Write Stall Cycles
PWSTALL1
Read-only
0x00000000
0x830
Channel1 Write Max Latency
PWLAT1
Read-only
0x00000000
Note:
7.7.4
Version Register Memory Map
Table 7-4.
Note:
1. The number of performance monitors is device specific. If the device has only one performance monitor, the Channel1 registers are not available. Please refer to the Module
Configuration section at the end of this chapter for the number of performance monitors on this
device.
PDCA Version Register Memory Map
Offset
Register
Register Name
Access
Reset
0x834
Version Register
VERSION
Read-only
- (1)
1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter.
48
32099I–01/2012
AT32UC3L016/32/64
7.7.5
Name:
Memory Address Register
MAR
Access Type:
Read/Write
Offset:
0x000 + n*0x040
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
18
17
16
11
10
9
8
3
2
1
0
MADDR[31:24]
23
22
21
20
19
MADDR[23:16]
15
14
13
12
MADDR[15:8]
7
6
5
4
MADDR[7:0]
• 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.
49
32099I–01/2012
AT32UC3L016/32/64
7.7.6
Name:
Peripheral Select Register
PSR
Access Type:
Read/Write
Offset:
0x004 + n*0x040
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
PID
• PID: Peripheral Identifier
The Peripheral Identifier selects which peripheral should be connected to the DMA channel. Writing a 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. See the Module Configuration section of PDCA for details. The width of the PID field is device specific and
dependent on the number of peripheral modules in the device.
50
32099I–01/2012
AT32UC3L016/32/64
7.7.7
Name:
Transfer Counter Register
TCR
Access Type:
Read/Write
Offset:
0x008 + n*0x040
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
TCV[15:8]
7
6
5
4
TCV[7:0]
• TCV: Transfer Counter Value
Number of data items to be transferred by the 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.
51
32099I–01/2012
AT32UC3L016/32/64
7.7.8
Name:
Memory Address Reload Register
MARR
Access Type:
Read/Write
Offset:
0x00C + n*0x040
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
MARV[31:24]
23
22
21
20
MARV[23:16]
15
14
13
12
MARV[15:8]
7
6
5
4
MARV[7:0]
• MARV: Memory Address Reload Value
Reload Value for the MAR register. This value will be loaded into MAR when TCR reaches zero if the TCRR register has a nonzero value.
52
32099I–01/2012
AT32UC3L016/32/64
7.7.9
Name:
Transfer Counter Reload Register
TCRR
Access Type:
Read/Write
Offset:
0x010 + n*0x040
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
TCRV[15:8]
7
6
5
4
TCRV[7:0]
• TCRV: Transfer Counter Reload Value
Reload value for the TCR register. 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 reloaded, the TCRR register is cleared.
53
32099I–01/2012
AT32UC3L016/32/64
7.7.10
Name:
Control Register
CR
Access Type:
Write-only
Offset:
0x014 + n*0x040
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
-
-
-
-
-
-
-
ECLR
7
6
5
4
3
2
1
0
-
-
-
-
-
-
TDIS
TEN
• ECLR: Transfer Error Clear
Writing a zero to this bit has no effect.
Writing a one to this bit will clear the Transfer Error bit in the Status Register (SR.TERR). Clearing the SR.TERR bit will allow the
channel to transmit data. The memory address must first be set to point to a valid location.
• TDIS: Transfer Disable
Writing a zero to this bit has no effect.
Writing a one to this bit will disable transfer for the DMA channel.
• TEN: Transfer Enable
Writing a zero to this bit has no effect.
Writing a one to this bit will enable transfer for the DMA channel.
54
32099I–01/2012
AT32UC3L016/32/64
7.7.11
Name:
Mode Register
MR
Access Type:
Read/Write
Offset:
0x018 + n*0x040
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
-
-
-
-
RING
ETRIG
SIZE
• RING: Ring Buffer
0:The Ring buffer functionality is disabled.
1:The Ring buffer functionality is enabled. When enabled, the reload registers, MARR and TCRR will not be cleared after reload.
• ETRIG: Event Trigger
0:Start transfer when the peripheral selected in Peripheral Select Register (PSR) requests a transfer.
1:Start transfer only when or after a peripheral event is received.
• SIZE: Size of Transfer
Table 7-5.
Size of Transfer
SIZE
Size of Transfer
0
Byte
1
Halfword
2
Word
3
Reserved
55
32099I–01/2012
AT32UC3L016/32/64
7.7.12
Name:
Status Register
SR
Access Type:
Read-only
Offset:
0x01C + n*0x040
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
-
-
-
-
-
-
-
TEN
• TEN: Transfer Enabled
This bit is cleared when the TDIS bit in CR is written to one.
This bit is set when the TEN bit in CR is written to one.
0: Transfer is disabled for the DMA channel.
1: Transfer is enabled for the DMA channel.
56
32099I–01/2012
AT32UC3L016/32/64
7.7.13
Name:
Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x020 + n*0x040
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
-
-
-
-
-
TERR
TRC
RCZ
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
57
32099I–01/2012
AT32UC3L016/32/64
7.7.14
Name:
Interrupt Disable Register
IDR
Access Type:
Write-only
Offset:
0x024 + n*0x040
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
-
-
-
-
-
TERR
TRC
RCZ
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
58
32099I–01/2012
AT32UC3L016/32/64
7.7.15
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
Offset:
0x028 + n*0x040
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
-
-
-
-
-
TERR
TRC
RCZ
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
A bit in this register is cleared when the corresponding bit in IDR is written to one.
A bit in this register is set when the corresponding bit in IER is written to one.
59
32099I–01/2012
AT32UC3L016/32/64
7.7.16
Name:
Interrupt Status Register
ISR
Access Type:
Read-only
Offset:
0x02C + n*0x040
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
-
-
-
-
-
TERR
TRC
RCZ
• TERR: Transfer Error
This bit is cleared when no transfer errors have occurred since the last write to CR.ECLR.
This bit is set when one or more transfer errors has occurred since reset or the last write to CR.ECLR.
• TRC: Transfer Complete
This bit is cleared when the TCR and/or the TCRR holds a non-zero value.
This bit is set when both the TCR and the TCRR are zero.
• RCZ: Reload Counter Zero
This bit is cleared when the TCRR holds a non-zero value.
This bit is set when TCRR is zero.
60
32099I–01/2012
AT32UC3L016/32/64
7.7.17
Name:
Performance Control Register
PCONTROL
Access Type:
Read/Write
Offset:
0x800
Reset Value:
0x00000000
31
30
29
28
-
-
23
22
-
-
15
14
13
12
-
-
-
7
6
-
-
27
26
25
24
18
17
16
11
10
9
8
-
-
-
CH1RES
CH0RES
5
4
3
2
1
0
CH1OF
CH0OF
-
-
CH1EN
CH0EN
MON1CH
21
20
19
MON0CH
• MON1CH: Performance Monitor Channel 1
• MON0CH: Performance Monitor Channel 0
The PDCA channel number to monitor with counter n
Due to performance monitor hardware resource sharing, the two performance monitor channels should NOT be programmed to
monitor the same PDCA channel. This may result in UNDEFINED monitor behavior.
• CH1RES: Performance Channel 1 Counter Reset
Writing a zero to this bit has no effect.
Writing a one to this bit will reset the counter in the channel specified in MON1CH.
This bit always reads as zero.
• CH0RES: Performance Channel 0 Counter Reset
Writing a zero to this bit has no effect.
Writing a one to this bit will reset the counter in the channel specified in MON0CH.
This bit always reads as zero.
• CH1OF: Channel 1 Overflow Freeze
0: The performance channel registers are reset if DATA or STALL overflows.
1: All performance channel registers are frozen just before DATA or STALL overflows.
• CH1OF: Channel 0 Overflow Freeze
0: The performance channel registers are reset if DATA or STALL overflows.
1: All performance channel registers are frozen just before DATA or STALL overflows.
• CH1EN: Performance Channel 1 Enable
0: Performance channel 1 is disabled.
1: Performance channel 1 is enabled.
• CH0EN: Performance Channel 0 Enable
0: Performance channel 0 is disabled.
1: Performance channel 0 is enabled.
61
32099I–01/2012
AT32UC3L016/32/64
7.7.18
Name:
Performance Channel 0 Read Data Cycles
PRDATA0
Access Type:
Read-only
Offset:
0x804
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
DATA[31:24]
23
22
21
20
DATA[23:16]
15
14
13
12
DATA[15:8]
7
6
5
4
DATA[7:0]
• DATA: Data Cycles Counted Since Last Reset
Clock cycles are counted using the CLK_PDCA_HSB clock
62
32099I–01/2012
AT32UC3L016/32/64
7.7.19
Name:
Performance Channel 0 Read Stall Cycles
PRSTALL0
Access Type:
Read-only
Offset:
0x808
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
STALL[31:24]
23
22
21
20
STALL[23:16]
15
14
13
12
STALL[15:8]
7
6
5
4
STALL[7:0]
• STALL: Stall Cycles Counted Since Last Reset
Clock cycles are counted using the CLK_PDCA_HSB clock
63
32099I–01/2012
AT32UC3L016/32/64
7.7.20
Name:
Performance Channel 0 Read Max Latency
PRLAT0
Access Type:
Read/Write
Offset:
0x80C
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
LAT[15:8]
7
6
5
4
LAT[7:0]
• LAT: Maximum Transfer Initiation Cycles Counted Since Last Reset
Clock cycles are counted using the CLK_PDCA_HSB clock
This counter is saturating. The register is reset only when PCONTROL.CH0RES is written to one.
64
32099I–01/2012
AT32UC3L016/32/64
7.7.21
Name:
Performance Channel 0 Write Data Cycles
PWDATA0
Access Type:
Read-only
Offset:
0x810
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
DATA[31:24]
23
22
21
20
DATA[23:16]
15
14
13
12
DATA[15:8]
7
6
5
4
DATA[7:0]
• DATA: Data Cycles Counted Since Last Reset
Clock cycles are counted using the CLK_PDCA_HSB clock
65
32099I–01/2012
AT32UC3L016/32/64
7.7.22
Name:
Performance Channel 0 Write Stall Cycles
PWSTALL0
Access Type:
Read-only
Offset:
0x814
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
STALL[31:24]
23
22
21
20
STALL[23:16]
15
14
13
12
STALL[15:8]
7
6
5
4
STALL[7:0]
• STALL: Stall Cycles Counted Since Last Reset
Clock cycles are counted using the CLK_PDCA_HSB clock
66
32099I–01/2012
AT32UC3L016/32/64
7.7.23
Name:
Performance Channel 0 Write Max Latency
PWLAT0
Access Type:
Read/Write
Offset:
0x818
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
LAT[15:8]
7
6
5
4
LAT[7:0]
• LAT: Maximum Transfer Initiation Cycles Counted Since Last Reset
Clock cycles are counted using the CLK_PDCA_HSB clock
This counter is saturating. The register is reset only when PCONTROL.CH0RES is written to one.
67
32099I–01/2012
AT32UC3L016/32/64
7.7.24
Name:
Performance Channel 1 Read Data Cycles
PRDATA1
Access Type:
Read-only
Offset:
0x81C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
DATA[31:24]
23
22
21
20
DATA[23:16]
15
14
13
12
DATA[15:8]
7
6
5
4
DATA[7:0]
• DATA: Data Cycles Counted Since Last Reset
Clock cycles are counted using the CLK_PDCA_HSB clock
68
32099I–01/2012
AT32UC3L016/32/64
7.7.25
Name:
Performance Channel 1 Read Stall Cycles
PRSTALL1
Access Type:
Read-only
Offset:
0x820
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
STALL[31:24]
23
22
21
20
STALL[23:16]
15
14
13
12
STALL[15:8]
7
6
5
4
STALL[7:0]
• STALL: Stall Cycles Counted Since Last Reset
Clock cycles are counted using the CLK_PDCA_HSB clock
69
32099I–01/2012
AT32UC3L016/32/64
7.7.26
Name:
Performance Channel 1 Read Max Latency
PRLAT1
Access Type:
Read/Write
Offset:
0x824
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
LAT[15:8]
7
6
5
4
LAT[7:0]
• LAT: Maximum Transfer Initiation Cycles Counted Since Last Reset
Clock cycles are counted using the CLK_PDCA_HSB clock
This counter is saturating. The register is reset only when PCONTROL.CH1RES is written to one.
70
32099I–01/2012
AT32UC3L016/32/64
7.7.27
Name:
Performance Channel 1 Write Data Cycles
PWDATA1
Access Type:
Read-only
Offset:
0x828
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
DATA[31:24]
23
22
21
20
DATA[23:16]
15
14
13
12
DATA[15:8]
7
6
5
4
DATA[7:0]
• DATA: Data Cycles Counted Since Last Reset
Clock cycles are counted using the CLK_PDCA_HSB clock
71
32099I–01/2012
AT32UC3L016/32/64
7.7.28
Name:
Performance Channel 1 Write Stall Cycles
PWSTALL1
Access Type:
Read-only
Offset:
0x82C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
STALL[31:24]
23
22
21
20
STALL[23:16]
15
14
13
12
STALL[15:8]
7
6
5
4
STALL[7:0]
• STALL: Stall Cycles Counted Since Last Reset
Clock cycles are counted using the CLK_PDCA_HSB clock
72
32099I–01/2012
AT32UC3L016/32/64
7.7.29
Name:
Performance Channel 1 Write Max Latency
PWLAT1
Access Type:
Read/Write
Offset:
0x830
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
LAT[15:8]
7
6
5
4
LAT[7:0]
• LAT: Maximum Transfer Initiation Cycles Counted Since Last Reset
Clock cycles are counted using the CLK_PDCA_HSB clock
This counter is saturating. The register is reset only when PCONTROL.CH1RES is written to one.
73
32099I–01/2012
AT32UC3L016/32/64
7.7.30
Name:
PDCA Version Register
VERSION
Access Type:
Read-only
Offset:
0x834
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the module. No functionality associated.
74
32099I–01/2012
AT32UC3L016/32/64
7.8
Module Configuration
The specific configuration for each PDCA instance is listed in the following tables.The module
bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details.
Table 7-6.
PDCA Configuration
Feature
PDCA
Number of channels
12
Number of performance monitors
1
Table 7-7.
Module Clock Name
Module name
Clock Name
Description
CLK_PDCA_HSB
Clock for the PDCA HSB interface
CLK_PDCA_PB
Clock for the PDCA PB interface
PDCA
Table 7-8.
Register Reset Values
Register
Reset Value
PSR CH 0
0
PSR CH 1
1
PSR CH 2
2
PSR CH 3
3
PSR CH 4
4
PSR CH 5
5
PSR CH 6
6
PSR CH 7
7
PSR CH 8
8
PSR CH 9
9
PSR CH 10
10
PSR CH 11
11
VERSION
122
The table below defines the valid Peripheral Identifiers (PIDs). The direction is specified as
observed from the memory, so RX means transfers from peripheral to memory and TX means
from memory to peripheral.
Table 7-9.
Peripheral Identity Values
PID
Direction
Peripheral Instance
Peripheral Register
0
RX
USART0
RHR
1
RX
USART1
RHR
2
RX
USART2
RHR
3
RX
USART3
RHR
75
32099I–01/2012
AT32UC3L016/32/64
Table 7-9.
Peripheral Identity Values
PID
Direction
Peripheral Instance
Peripheral Register
4
RX
SPI
RDR
5
RX
TWIM0
RHR
6
RX
TWIM1
RHR
7
RX
TWIS0
RHR
8
RX
TWIS1
RHR
9
RX
ADCIFB
LCDR
10
RX
AW
RHR
11
RX
CAT
ACOUNT
12
TX
USART0
THR
13
TX
USART1
THR
14
TX
USART2
THR
15
TX
USART3
THR
16
TX
SPI
TDR
17
TX
TWIM0
THR
18
TX
TWIM1
THR
19
TX
TWIS0
THR
20
TX
TWIS1
THR
21
TX
AW
THR
22
TX
CAT
MBLEN
76
32099I–01/2012
AT32UC3L016/32/64
8. Flash Controller (FLASHCDW)
Rev: 1.0.2.0
8.1
Features
•
•
•
•
•
•
•
•
•
•
•
•
•
8.2
Controls on-chip flash memory
Supports 0 and 1 wait state bus access
Buffers reducing penalty of wait state in sequential code or loops
Allows interleaved burst reads for systems with one wait state, outputting one 32-bit word per
clock cycle for sequential reads
Secure State for supporting FlashVault technology
32-bit HSB interface for reads from flash and writes to page buffer
32-bit PB interface for issuing commands to and configuration of the controller
Flash memory is divided into 16 regions can be individually protected or unprotected
Additional protection of the Boot Loader pages
Supports reads and writes of general-purpose Non Volatile Memory (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
Overview
The Flash Controller (FLASHCDW) interfaces the on-chip flash memory with the 32-bit internal
HSB bus. The controller manages the reading, writing, erasing, locking, and unlocking
sequences.
8.3
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
8.3.1
Power Management
If the CPU enters a sleep mode that disables clocks used by the FLASHCDW, the FLASHCDW
will stop functioning and resume operation after the system wakes up from sleep mode.
8.3.2
Clocks
The FLASHCDW has two bus clocks connected: One High Speed Bus clock
(CLK_FLASHCDW_HSB) and one Peripheral Bus clock (CLK_FLASHCDW_PB). These clocks
are generated by the Power Manager. Both clocks are enabled at reset, and can be disabled by
writing to the Power Manager. The user has to ensure that CLK_FLASHCDW_HSB is not turned
off before reading the flash or writing the pagebuffer and that CLK_FLASHCDW_PB is not
turned off before accessing the FLASHCDW configuration and control registers. Failing to do so
may deadlock the bus.
8.3.3
Interrupts
The FLASHCDW interrupt request lines are connected to the interrupt controller. Using the
FLASHCDW interrupts requires the interrupt controller to be programmed first.
77
32099I–01/2012
AT32UC3L016/32/64
8.3.4
8.4
8.4.1
Debug Operation
When an external debugger forces the CPU into debug mode, the FLASHCDW continues normal operation. If the FLASHCDW is configured in a way that requires it to be periodically
serviced by the CPU through interrupts or similar, improper operation or data loss may result
during debugging.
Functional Description
Bus Interfaces
The FLASHCDW has two bus interfaces, one High Speed Bus (HSB) interface for reads from the
flash memory and writes to the page buffer, and one Peripheral Bus (PB) interface for issuing
commands and reading status from the controller.
8.4.2
Memory Organization
The flash memory is divided into a set of pages. A page is the basic unit addressed when programming the flash. A page consists of several words. The pages are grouped into 16 regions of
equal size. Each of these regions can be locked by a dedicated fuse bit, protecting it from accidental modification.
• p pages (FLASH_P)
• w bytes in each page and in the page buffer (FLASH_W)
• pw bytes in total (FLASH_PW)
• f general-purpose fuse bits (FLASH_F), used as region lock bits and for other device-specific
purposes
• 1 security fuse bit
• 1 User page
8.4.3
User Page
The User page is an additional page, outside the regular flash array, that can be used to store
various data, such as calibration data and serial numbers. This page is not erased by regular
chip erase. The User page can only be written and erased by a special set of commands. Read
accesses to the User page are performed just as any other read accesses to the flash. The
address map of the User page is given in Figure 8-1 on page 80.
8.4.4
Read Operations
The on-chip flash memory is typically used for storing instructions to be executed by the CPU.
The CPU will address instructions using the HSB bus, and the FLASHCDW will access the flash
memory and return the addressed 32-bit word.
In systems where the HSB clock period is slower than the access time of the flash memory, the
FLASHCDW can operate in 0 wait state mode, and output one 32-bit word on the bus per clock
cycle. If the clock frequency allows, the user should use 0 wait state mode, because this gives
the highest performance as no stall cycles are encountered.
The FLASHCDW can also operate in systems where the HSB bus clock period is faster than the
access speed of the flash memory. Wait state support and a read granularity of 64 bits ensure
efficiency in such systems.
Performance for systems with high clock frequency is increased since the internal read word
width of the flash memory is 64 bits. When a 32-bit word is to be addressed, the word itself and
78
32099I–01/2012
AT32UC3L016/32/64
also the other word in the same 64-bit location is read. The first word is output on the bus, and
the other word is put into an internal buffer. If a read to a sequential address is to be performed
in the next cycle, the buffered word is output on the bus, while the next 64-bit location is read
from the flash memory. Thus, latency in 1 wait state mode is hidden for sequential fetches.
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 memory.
In 0ws mode, no wait states are encountered on any flash read operations. In 1 ws mode, one
stall cycle is encountered on the first access in a single or burst transfer. In 1 ws mode, if the first
access in a burst access is to an address that is not 64-bit aligned, an additional stall cycle is
also encountered when reading the second word in the burst. All subsequent words in the burst
are accessed without any stall cycles.
The Flash Controller provides two sets of buffers that can be enabled in order to speed up
instruction fetching. These buffers can be enabled by writing a one to the FCR.SEQBUF and
FCR.BRBUF bits. The SEQBUF bit enables buffering hardware optimizing sequential instruction
fetches. The BRBUF bit enables buffering hardware optimizing tight inner loops. These buffers
are never used when the flash is in 0 wait state mode. Usually, both these buffers should be
enabled when operating in 1 wait state mode. Some users requiring absolute cycle determinism
may want to keep the buffers disabled.
The Flash Controller address space is displayed in Figure 8-1. The memory space between
address pw and the User page is reserved, and reading addresses in this space returns an
undefined result. The User page is permanently mapped to an offset of 0x00800000 from the
start address of the flash memory.
Table 8-1.
User Page Addresses
Memory type
Start address, byte sized
Size
Main array
0
pw bytes
User
0x00800000
w bytes
79
32099I–01/2012
AT32UC3L016/32/64
Figure 8-1.
Memory Map for the Flash Memories
Offset from
base address
Reserved
User Page
Reserved
0x0080 0000
Flash data array
pw
0
Flash base address
Flash with User Page
All addresses are byte addresses
8.4.5
High Speed Read Mode
The flash provides a High Speed Read Mode, offering slightly higher flash read speed at the
cost of higher power consumption. Two dedicated commands, High Speed Read Mode Enable
(HSEN) and High Speed Read Mode Disable (HSDIS) control the speed mode. The High Speed
Mode (HSMODE) bit in the Flash Status Register (FSR) shows which mode the flash is in. After
reset, the High Speed Mode is disabled, and must be manually enabled if the user wants to.
Refer to the Electrical Characteristics chapter at the end of this datasheet for details on the maximum clock frequencies in Normal and High Speed Read Mode.
80
32099I–01/2012
AT32UC3L016/32/64
Figure 8-2.
High Speed Mode
Frequency
1 wait state
0 wait state
Frequency limit
for 0 wait state
operation
Speed mode
H
al
m
h
ig
or
N
8.4.6
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.
8.4.7
Quick User Page Read
A dedicated command, Quick User Page Read (QPRUP), is provided to read all words in the
user 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
QPRUP command is useful to check that a page is in an erased state. The QPRUP instruction is
much faster than performing the erased-page check using a regular software subroutine.
8.4.8
Page Buffer Operations
The flash memory has a write and erase granularity of one page; data is written and erased in
chunks of one page. When programming a page, the user must first write the new data into the
Page Buffer. The contents of the entire Page Buffer is copied into the desired page in flash
memory when the user issues the Write Page command, Refer to Section 8.5.1 on page 83.
In order to program data into flash page Y, write the desired data to locations Y0 to Y31 in the
regular flash memory map. Writing to an address A in the flash memory map will not update the
flash memory, but will instead update location A%32 in the page buffer. The PAGEN field in the
Flash Command (FCMD) register will at the same time be updated with the value A/32.
81
32099I–01/2012
AT32UC3L016/32/64
Figure 8-3.
Mapping from Page Buffer to Flash
Flash
All locations are doubleword locations
Page Buffer
64-bit data
31
27
23
19
15
11
7
3
30
26
22
18
14
10
6
2
29
25
21
17
13
9
5
1
28
24
20
16
12
8
4
0
Z31
Z27
Z23
Z19
Z15
Z11
Z7
Z3
Y31
Y27
Y23
Y19
Y15
Y11
Y7
Y3
X31
X27
X23
X19
X15
X11
X7
X3
Z30
Z29
Z26
Z25
Z22
Z21
Z18
Z17
Page Z
Z14
Z13
Z10
Z9
Z6
Z5
Z2
Z1
Y30
Y29
Y26
Y25
Y22
Y21
Y18
Y17
Page Y
Y14
Y13
Y10
Y9
Y6
Y5
Y2
Y1
X30
X29
X26
X25
X22
X21
X18
X17
Page X
X14
X13
X10
X9
X6
X5
X2
X1
Z28
Z24
Z20
Z16
Z12
Z8
Z4
Z0
Y28
Y24
Y20
Y16
Y12
Y8
Y4
Y0
X28
X24
X20
X16
X12
X8
X4
X0
Internally, the flash memory stores data in 64-bit doublewords. Therefore, the native data size of
the Page Buffer is also a 64-bit doubleword. All locations shown in Figure 8-3 are therefore doubleword locations. Since the HSB bus only has a 32-bit data width, two 32-bit HSB transfers
must be performed to write a 64-bit doubleword into the Page Buffer. The FLASHCDW has logic
to combine two 32-bit HSB transfers into a 64-bit data before writing this 64-bit data into the
Page Buffer. This logic requires the word with the low address to be written to the HSB bus
before the word with the high address. To exemplify, to write a 64-bit value to doubleword X0
residing in page X, first write a 32-bit word to the byte address pointing to address X0, thereafter
write a word to the byte address pointing to address (X0+4).
The page buffer is word-addressable and should only be written with aligned word transfers,
never with byte or halfword transfers. The page buffer can not be read.
The page buffer is also used for writes to the User page.
Page buffer write operations are performed with 4 wait states. Any accesses attempted to the
FLASHCDW on the HSB bus during these cycles will be automatically stalled.
Writing to the page buffer can only change page buffer bits from one to zero, i.e. 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 erase the entire page buffer with
the Clear Page Buffer command.
82
32099I–01/2012
AT32UC3L016/32/64
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.
8.5
Flash Commands
The FLASHCDW offers a command set to manage programming of the flash memory, locking
and unlocking of regions, and full flash erasing. See Section 8.8.2 for a complete list of
commands.
To run a command, the CMD field in the Flash Command Register (FCMD) has to be written
with the command number. As soon as the FCMD register is written, the FRDY bit in the Flash
Status Register (FSR) is automatically cleared. Once the current command is complete, the
FSR.FRDY bit is automatically set. If an interrupt has been enabled by writing a one to
FCR.FRDY, the interrupt request line of the Flash Controller is activated. All flash commands
except for Quick Page Read (QPR) and Quick User Page Read (QPRUP) will generate an interrupt request upon completion if FCR.FRDY is one.
Any HSB bus transfers attempting to read flash memory when the FLASHCDW is busy executing a flash command will be stalled, and allowed to continue when the flash command is
complete.
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 FLASHCDW
is IDLE. The user must make sure that the access pattern to the FLASHCDW 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 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 bit is
set in the Flash Status Register (FSR). This bit 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 bit is set in the Flash Status Register (FSR). This bit 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 bit is
set in the FSR register. This bit is automatically cleared by a read access to the FSR register.
8.5.1
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.
83
32099I–01/2012
AT32UC3L016/32/64
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 is stored in an internal buffer called the 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 as described in Section 8.4.8 on page 81.
• 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 FRDY bit in the Flash Status Register (FSR) is set. If
an interrupt was enabled by writing FCR.FRDY to one, an interrupt request is generated.
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.
• Lock Error: Can have two different causes:
– The page to be programmed belongs to a locked region. A command must be
executed to unlock the corresponding region before programming can start.
– A bus master without secure status attempted to program a page requiring secure
privileges.
8.5.2
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 configured with a BOOTPROT 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 LOCKE bit in FCR is one, an interrupt request
is set generated.
When the command is complete, the FRDY bit in the Flash Status Register (FSR) is set. If an
interrupt has been enabled by writing FCR.FRDY to one, an interrupt request is generated. Two
errors can be detected in the FSR register after issuing the command:
84
32099I–01/2012
AT32UC3L016/32/64
• Programming Error: A bad keyword and/or an invalid command have been written in the
FCMD register.
• Lock Error: At least one lock region is protected, or BOOTPROT is different from 0. The erase
command has been aborted and no page has been erased. A “Unlock region containing
given page” (UP) command must be executed to unlock any locked regions.
8.5.3
Region Lock Bits
The flash memory 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 Section 8.6. 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 8.6.
8.6
General-purpose Fuse Bits
The flash memory 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
85
32099I–01/2012
AT32UC3L016/32/64
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 8-2.
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 or aWire 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.
This fuse has no effect in devices with no External Memory
Interface (EBI).
BOOTPROT
Used to select one of eight different bootloader sizes. Pages
included in the bootloader area can not be erased or
programmed except by a JTAG or aWire chip erase.
BOOTPROT can only be changed when the security bit is
cleared.
If the security bit is set, only an external JTAG or aWire 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.
SECURE
Used to configure secure state and secure state debug
capabilities. Decoded into SSE and SSDE signals as shown in
Table 8-5. Refer to the AVR32 Architecture Manual and the
AVR32UC Technical Reference Manual for more details on
SSE and SSDE.
UPROT
If programmed (i.e. “0”), the JTAG USER PROTECTION
feature is enabled. If this fuse is programmed some HSB
addresses will be accessible by JTAG access even if the flash
security fuse is programmed. Refer to the JTAG documentation
for more information on this functionality. This bit can only be
changed when the security bit is cleared.
16
19:17
21:20
22
The BOOTPROT fuses protects the following address space for the Boot Loader:
86
32099I–01/2012
AT32UC3L016/32/64
Table 8-4.
Boot Loader Area Specified by BOOTPROT
BOOTPROT
Pages protected by
BOOTPROT
Size of protected
memory
7
None
0
6
0-1
512 byte
5
0-3
1Kbyte
4
0-7
2Kbyte
3
0-15
4Kbyte
2
0-31
8Kbyte
1
0-63
16Kbyte
0
0-127
32Kbyte
The SECURE fuses have the following functionality:
Table 8-5.
Secure State Configuration
SECURE
Functionality
SSE
SSDE
00
Secure state disabled
0
0
01
Secure enabled, secure state debug enabled
1
1
10
Secure enabled, secure state debug disabled
1
0
11
Secure state disabled
0
0
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 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 the BOOTPROT or EPFL or UPROT fuse bits was attempted
while the flash is locked by the security bit.
– A write or erase of the SECURE fuse bits was attempted when SECURE mode was
enabled.
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 8.5.3.
87
32099I–01/2012
AT32UC3L016/32/64
8.7
Security Bit
The security bit allows the entire device to be locked from external JTAG, aWire, 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 or aWire 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.
88
32099I–01/2012
AT32UC3L016/32/64
8.8
User Interface
Table 8-6.
Note:
FLASHCDW Register Memory Map
Offset
Register
Register Name
Access
Reset
0x00
Flash Control Register
FCR
Read/Write
0x00000000
0x04
Flash Command Register
FCMD
Read/Write
0x00000000
0x08
Flash Status Register
FSR
Read-only
-(1)
0x0C
Flash Parameter Register
FPR
Read-only
-(3)
0x10
Flash Version Register
FVR
Read-only
-(3)
0x14
Flash General Purpose Fuse Register Hi
FGPFRHI
Read-only
-(2)
0x18
Flash General Purpose Fuse Register Lo
FGPFRLO
Read-only
-(2)
1. The value of the Lock bits depend on their programmed state. All other bits in FSR are 0.
2. All bits in FGPRHI/LO are dependent on the programmed state of the fuses they map to. Any bits in these registers not
mapped to a fuse read as 0.
3. The reset values for these registers are device specific. Please refer to the Module Configuration section at the end of this
chapter.
89
32099I–01/2012
AT32UC3L016/32/64
8.8.1
Name:
Flash Control Register
FCR
Access Type:
Read/Write
Offset:
0x00
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
-
-
-
-
-
BRBUF
SEQBUF
-
7
6
5
4
3
2
1
0
-
FWS
-
-
PROGE
LOCKE
-
FRDY
• BRBUF: Branch Target Instruction Buffer Enable
0: The Branch Target Instruction Buffer is disabled.
1: The Branch Target Instruction Buffer is enabled.
• SEQBUF: Sequential Instruction Fetch Buffer Enable
0: The Sequential Instruction Fetch Buffer is disabled.
1: The Sequential Instruction Fetch Buffer is enabled.
• FWS: Flash Wait State
0: The flash is read with 0 wait states.
1: The flash is read with 1 wait state.
• PROGE: Programming Error Interrupt Enable
0: Programming Error does not generate an interrupt request.
1: Programming Error generates an interrupt request.
• LOCKE: Lock Error Interrupt Enable
0: Lock Error does not generate an interrupt request.
1: Lock Error generates an interrupt request.
• FRDY: Flash Ready Interrupt Enable
0: Flash Ready does not generate an interrupt request.
1: Flash Ready generates an interrupt request.
90
32099I–01/2012
AT32UC3L016/32/64
8.8.2
Name:
Flash Command Register
FCMD
Access Type:
Read/Write
Offset:
0x04
Reset Value:
0x00000000
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 in FSR 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
• 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.
• 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 8-7.
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
91
32099I–01/2012
AT32UC3L016/32/64
Table 8-7.
Semantic of PAGEN field in different commands
Command
PAGEN description
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
High Speed Mode Enable
Not used
High Speed Mode Disable
Not used
• CMD: Command
This field defines the flash command. Issuing any unused command will cause the Programming Error bit in FSR to be set, and
the corresponding interrupt to be requested if the PROGE bit in FCR is one.
Table 8-8.
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
High Speed Mode Enable
16
HSEN
High Speed Mode Disable
17
HSDIS
RESERVED
16-31
92
32099I–01/2012
AT32UC3L016/32/64
8.8.3
Name:
Flash Status Register
FSR
Access Type:
Read-only
Offset:
0x08
Reset Value:
0x00000000
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
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
HSMODE
QPRR
SECURITY
PROGE
LOCKE
-
FRDY
• LOCKx: Lock Region x Lock Status
0: The corresponding lock region is not locked.
1: The corresponding lock region is locked.
• HSMODE: High-Speed Mode
0: High-speed mode disabled.
1: High-speed mode enabled.
• 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.
• SECURITY: Security Bit Status
0: The security bit is inactive.
1: The security bit is active.
• 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.
• 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.
• 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.
93
32099I–01/2012
AT32UC3L016/32/64
8.8.4
Name:
Flash Parameter Register
FPR
Access Type:
Read-only
Offset:
0x0C
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
-
-
-
-
PSZ
2
1
0
FSZ
• PSZ: Page Size
The size of each flash page.
Table 8-9.
Flash Page Size
PSZ
Page Size
0
32 Byte
1
64 Byte
2
128 Byte
3
256 Byte
4
512 Byte
5
1024 Byte
6
2048 Byte
7
4096 Byte
94
32099I–01/2012
AT32UC3L016/32/64
• FSZ: Flash Size
The size of the flash. Not all device families will provide all flash sizes indicated in the table.
Table 8-10.
Flash Size
FSZ
Flash Size
FSZ
Flash Size
0
4 Kbyte
8
192 Kbyte
1
8 Kbyte
9
256 Kbyte
2
16 Kbyte
10
384 Kbyte
3
32 Kbyte
11
512 Kbyte
4
48 Kbyte
12
768 Kbyte
5
64 Kbyte
13
1024 Kbyte
6
96 Kbyte
14
2048 Kbyte
7
128 Kbyte
15
Reserved
95
32099I–01/2012
AT32UC3L016/32/64
8.8.5
Name:
Flash Version Register
FVR
Access Type:
Read-only
Offset:
0x10
Reset Value:
0x00000000
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the module. No functionality associated.
96
32099I–01/2012
AT32UC3L016/32/64
8.8.6
Name:
Flash General Purpose Fuse Register High
FGPFRHI
Access Type:
Read-only
Offset:
0x14
Reset Value:
-
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.
97
32099I–01/2012
AT32UC3L016/32/64
8.8.7
Name:
Flash General Purpose Fuse Register Low
FGPFRLO
Access Type:
Read-only
Offset:
0x18
Reset Value:
-
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.
98
32099I–01/2012
AT32UC3L016/32/64
8.9
Fuse Settings
The flash contains 32 general purpose fuses. These 32 fuses can be found in the Flash General
Purpose Fuse Register Low (FGPFRLO). The Flash General Purpose Fuse Register High
(FGPFRHI) is not used. In addition to the general purpose fuses, parts of the flash user page
can have a defined meaning outside of the flash controller and will also be described in this
section.
Note that when writing to the user page the values do not get loaded by the other modules on
the device until a chip reset occurs.
The general purpose fuses are erased by a JTAG or aWire chip erase.
99
32099I–01/2012
AT32UC3L016/32/64
8.9.1
Flash General Purpose Fuse Register Low (FGPFRLO)
31
30
29
BODEN
28
27
BODHYST
23
22
21
BODLEVEL[0]
UPROT
15
14
25
24
17
16
BODLEVEL[5:1]
20
19
SECURE
13
26
18
BOOTPROT
12
EPFL
11
10
9
8
3
2
1
0
LOCK[15:8]
7
6
5
4
LOCK[7:0]
• BODEN: Brown Out Detector Enable
BODEN
Description
00
BOD disabled
01
BOD enabled, BOD reset enabled
10
BOD enabled, BOD reset disabled
11
BOD disabled
• BODHYST: Brown Out Detector Hysteresis
0: The Brown out detector hysteresis is disabled
1: The 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 ”Electrical Characteristics” on page 770.
• UPROT, SECURE, BOOTPROT, EPFL, LOCK
These are Flash Controller fuses and are described in the FLASHCDW section.
8.9.1.1
Default Fuse Value
The devices are shipped with the FGPFRLO register value:0xE075FFFF:
• BODEN fuses set to 11. BOD is disabled.
• BODHYST fuse set to 1. The BOD hysteresis is enabled.
• BODLEVEL fuses set to 000000. This is the minimum voltage trigger level for BOD. This level
is lower than the POR level, so when BOD is enabled, it will never trigger with this default
value.
• UPROT fuse set to 1.
• SECURE fuse set to 11.
• BOOTPROT fuses set to 010. The bootloader protected size is 8Kbytes.
• EPFL fuse set to 1. External privileged fetch is not locked.
• LOCK fuses set to 1111111111111111. No region locked.
After the JTAG or aWire chip erase command, the FGPFR register value is 0xFFFFFFFF.
100
32099I–01/2012
AT32UC3L016/32/64
8.9.2
First Word of the User Page (Address 0x80800000)
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
-
-
-
-
-
-
-
WDTAUTO
• WDTAUTO: WatchDog Timer Auto Enable at Startup
0: The WDT is automatically enabled at startup.
1: The WDT is not automatically enabled at startup.
Please refer to the WDT chapter for detail about timeout settings when the WDT is automatically enabled.
8.9.2.1
Default user page first word value
The devices are shipped with the user page erased (all bits 1):
• WDTAUTO set to 1, WDT disabled.
101
32099I–01/2012
AT32UC3L016/32/64
8.9.3
Second Word of the User Page (Address 0x80800004)
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
SSADRR[15:8]
23
22
21
20
SSADRR[7:0]
15
14
13
12
SSADRF[15:8]
7
6
5
4
SSADRF[7:0]
• SSADRR: Secure State End Address for the RAM
• SSADRF: Secure State End Address for the Flash
8.9.3.1
Default user page second word value
The devices are shipped with the User page erased (all bits 1).
102
32099I–01/2012
AT32UC3L016/32/64
8.10
Bootloader Configuration
The bootloader uses two words in the flash User page to store its configuration:
• Configuration word 1 at address 0x808000FC is read first at boot time to know if it should start
the ISP process inconditionally and whether it should use the configuration word 2 where
further configuration is stored.
• Configuration word 2 at address 0x808000F8 stores the I/O conditions that determine which of
the ISP and the application to start at the end of the boot process.
Please refer to the bootloader documentation for more information.
8.11
Serial Number
Each device has a unique 120 bits serial number readable from address 0x8080010C to
0x8080011A.
8.12
Module Configuration
The specific configuration for each FLASHCDW instance is listed in the following tables. The
module bus clocks listed here are connected to the system bus clocks. Please refer to the Power
Manager chapter for details.
Table 8-11.
Module Configuration
Feature
AT32UC3L064
AT32UC3L032
AT32UC3L016
Flash size
64Kbytes
32Kbytes
16Kbytes
Number of pages
256
128
64
Page size
256 bytes
256 bytes
256 bytes
Table 8-12.
Module Name
Module Clock Name
Clock Name
Description
CLK_FLASHCDW_HSB
Clock for the FLASHCDW HSB interface
CLK_FLASHCDW_PB
Clock for the FLASHCDW PB interface
FLASHCDW
Table 8-13.
Register Reset Values
Register
Reset Value
FVR
0x00000102
FPR
0x00000305
103
32099I–01/2012
AT32UC3L016/32/64
9. Secure Access Unit (SAU)
Rev: 1.1.0.3
9.1
Features
• Remaps registers in memory regions protected by the MPU to regions not protected by the MPU
• Programmable physical address for each channel
• Two modes of operation: Locked and Open
– In Locked Mode, access to a channel must be preceded by an unlock action
• An unlocked channel remains open only for a specific amount of time, if no access is
performed during this time, the channel is relocked
• Only one channel can be open at a time, opening a channel while another one is open
locks the first one
• Access to a locked channel is denied, a bus error and optionally an interrupt is returned
• If a channel is relocked due to an unlock timeout, an interrupt can optionally be
generated
– In Open Mode, all channels are permanently unlocked
9.2
Overview
In many systems, erroneous access to peripherals can lead to catastrophic failure. An example
of such a peripheral is the Pulse Width Modulator (PWM) used to control electric motors. The
PWM outputs a pulse train that controls the motor. If the control registers of the PWM module
are inadvertently updated with wrong values, the motor can start operating out of control, possibly causing damage to the application and the surrounding environment. However, sometimes
the PWM control registers must be updated with new values, for example when modifying the
pulse train to accelerate the motor. A mechanism must be used to protect the PWM control registers from inadvertent access caused by for example:
• Errors in the software code
• Transient errors in the CPU caused by for example electrical noise altering the execution path
of the program
To improve the security in a computer system, the AVR32UC implements a Memory Protection
Unit (MPU). The MPU can be set up to limit the accesses that can be performed to specific
memory addresses. The MPU divides the memory space into regions, and assigns a set of
access restrictions on each region. Access restrictions can for example be read/write if the CPU
is in supervisor mode, and read-only if the CPU is in application mode. The regions can be of different size, but each region is usually quite large, e.g. protecting 1 kilobyte of address space or
more. Furthermore, access to each region is often controlled by the execution state of the CPU,
i.e. supervisor or application mode. Such a simple control mechanism is often too inflexible (too
coarse-grained chunks) and with too much overhead (often requiring system calls to access protected memory locations) for simple or real-time systems such as embedded microcontrollers.
Usually, the Secure Access Unit (SAU) is used together with the MPU to provide the required
security and integrity. The MPU is set up to protect regions of memory, while the SAU is set up
to provide a secure channel into specific memory locations that are protected by the MPU.
These specific locations can be thought of as fine-grained overrides of the general coarsegrained protection provided by the MPU.
104
32099I–01/2012
AT32UC3L016/32/64
9.3
Block Diagram
Figure 9-1 presents the SAU integrated in an example system with a CPU, some memories,
some peripherals, and a bus system. The SAU is connected to both the Peripheral Bus (PB) and
the High Speed Bus (HSB). Configuration of the SAU is done via the PB, while memory
accesses are done via the HSB. The SAU receives an access on its HSB slave interface,
remaps it, checks that the channel is unlocked, and if so, initiates a transfer on its HSB master
interface to the remapped address.
The thin arrows in Figure 9-1 exemplifies control flow when using the SAU. The CPU wants to
read the RX Buffer in the USART. The MPU has been configured to protect all registers in the
USART from user mode access, while the SAU has been configured to remap the RX Buffer into
a memory space that is not protected by the MPU. This unprotected memory space is mapped
into the SAU HSB slave space. When the CPU reads the appropriate address in the SAU, the
SAU will perform an access to the desired RX buffer register in the USART, and thereafter return
the read results to the CPU. The return data flow will follow the opposite direction of the control
flow arrows in Figure 9-1.
Figure 9-1.
SAU Block Diagram
CPU
MPU
Flash
RAM
Bus master
Bus slave
Bus slave
High Speed Bus
Bus slave
Bus master
Bus slave
Bus bridge
SAU Channel
SAU
Interrupt
request
USART
PWM
Peripheral Bus
SAU Configuration
105
32099I–01/2012
AT32UC3L016/32/64
9.4
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
9.4.1
Power Management
If the CPU enters a sleep mode that disables clocks used by the SAU, the SAU will stop functioning and resume operation after the system wakes up from sleep mode.
9.4.2
Clocks
The SAU has two bus clocks connected: One High Speed Bus clock (CLK_SAU_HSB) and one
Peripheral Bus clock (CLK_SAU_PB). These clocks are generated by the Power Manager. Both
clocks are enabled at reset, and can be disabled by writing to the Power Manager. The user has
to ensure that CLK_SAU_HSB is not turned off before accessing the SAU. Likewise, the user
must ensure that no bus access is pending in the SAU before disabling CLK_SAU_HSB. Failing
to do so may deadlock the High Speed Bus.
9.4.3
Interrupt
The SAU interrupt request line is connected to the interrupt controller. Using the SAU interrupt
requires the interrupt controller to be programmed first.
9.4.4
9.5
Debug Operation
When an external debugger forces the CPU into debug mode, the SAU continues normal operation. If the SAU is configured in a way that requires it to be periodically serviced by the CPU
through interrupts or similar, improper operation or data loss may result during debugging.
Functional Description
9.5.1
Enabling the SAU
The SAU is enabled by writing a one to the Enable (EN) bit in the Control Register (CR). This will
set the SAU Enabled (EN) bit in the Status Register (SR).
9.5.2
Configuring the SAU Channels
The SAU has a set of channels, mapped in the HSB memory space. These channels can be
configured by a Remap Target Register (RTR), located at the same memory address. When the
SAU is in normal mode, the SAU channel is addressed, and when the SAU is in setup mode, the
RTR can be addressed.
Before the SAU can be used, the channels must be configured and enabled. To configure a
channel, the corresponding RTR must be programmed with the Remap Target Address. To do
this, make sure the SAU is in setup mode by writing a one to the Setup Mode Enable (SEN) bit
in CR. This makes sure that a write to the RTR address accesses the RTR, not the SAU channel. Thereafter, the RTR is written with the address to remap to, typically the address of a
specific PB register. When all channels have been configured, return to normal mode by writing
a one to the Setup Mode Disable (SDIS) in CR. The channels can now be enabled by writing
ones to the corresponding bits in the Channel Enable Registers (CERH/L).
The SAU is only able to remap addresses above 0xFFFC0000.
106
32099I–01/2012
AT32UC3L016/32/64
9.5.2.1
9.5.3
Protecting SAU configuration registers
In order to prevent the SAU configuration registers to be changed by malicious or runaway code,
they should be protected by the MPU as soon as they have been configured. Maximum security
is provided in the system if program memory does not contain any code to unprotect the configuration registers in the MPU. This guarantees that runaway code can not accidentally unprotect
and thereafter change the SAU configuration registers.
Lock Mechanism
The SAU can be configured to use two different access mechanisms: Open and Locked. In
Open Mode, SAU channels can be accessed freely after they have been configured and
enabled. In order to prevent accidental accesses to remapped addresses, it is possible to configure the SAU in Locked Mode. Writing a one to the Open Mode bit in the CONFIG register
(CONFIG.OPEN) will enable Open Mode. Writing a zero to CONFIG.OPEN will enable Locked
Mode.
When using Locked Mode, the lock mechanism must be configured by writing a user defined key
value to the Unlock Key (UKEY) field in the Configuration Register (CONFIG). The number of
CLK_SAU_HSB cycles the channel remains unlocked must be written to the Unlock Number of
Clock Cycles (UCYC) field in CONFIG.
Access control to the SAU channels is enabled by means of the Unlock Register (UR), which
resides in the same address space as the SAU channels. Before a channel can be accessed,
the unlock register must be written with th correct key and channel number (single write access).
Access to the channel is then permitted for the next CONFIG.UCYC clock cycles, or until a successful access to the unlocked channel has been made.
Only one channel can be unlocked at a time. If any other channel is unlocked at the time of writing UR, this channel will be automatically locked before the channel addressed by the UR write
is unlocked.
An attempted access to a locked channel will be aborted, and the Channel Access Unsuccessful
bit (SR.CAU) will be set.
Any pending errors bits in SR must be cleared before it is possible to access UR. The following
SR bits are defined as error bits: EXP, CAU, URREAD, URKEY, URES, MBERROR, RTRADR.
If any of these bits are set while writing to UR, the write is aborted and the Unlock Register Error
Status (URES) bit in SR is set.
9.5.4
Normal Operation
The following sequence must be used in order to access a SAU channel in normal operation
(CR.SEN=0):
1. If not in Open Mode, write the unlock key to UR.KEY and the channel number to
UR.CHANNEL.
2. Perform the read or write operation to the SAU channel. If not in Open Mode, this must
be done within CONFIG.UCYC clock cycles of unlocking the channel. The SAU will use
its HSB master interface to remap the access to the target address pointed to by the
corresponding RTR.
3. To confirm that the access was successful, wait for the IDLE transfer status bit
(SR.IDLE) to indicate the operation is completed. Then check SR for possible error conditions. The SAU can be configured to generate interrupt requests or a Bus Error
Exception if the access failed.
107
32099I–01/2012
AT32UC3L016/32/64
9.5.4.1
Operation example
Figure 9-2 shows a typical memory map, consisting of some memories, some simple peripherals, and a SAU with multiple channels and an Unlock Register (UR). Imagine that the MPU has
been set up to disallow all accesses from the CPU to the grey modules. Thus the CPU has no
way of accessing for example the Transmit Holding register in the UART, present on address X
on the bus. Note that the SAU RTRs are not protected by the MPU, thus the RTRs can be
accessed. If for example RTR0 is configured to point to address X, an access to RTR0 will be
remapped by the SAU to address X according to the algorithm presented above. By programming the SAU RTRs, specific addresses in modules that have generally been protected by the
MPU can be performed.
Figure 9-2.
Example Memory Map for a System with SAU
SAU
CONFIG
UART
Receive Holding
Transmit Holding
Baudrate
Control
Address X
SAU
CHANNEL
9.5.5
...
UR
RTR62
Channel
RTR1 1
RTR0
Address Z
Interrupts
The SAU can generate an interrupt request to signal different events. All events that can generate an interrupt request have dedicated bits in the Status Register (SR). An interrupt request will
be generated if the corresponding bit in the Interrupt Mask Register (IMR) is set. Bits in IMR are
set by writing a one to the corresponding bit in the Interrupt Enable Register (IER), and cleared
by writing a one to the corresponding bit in the Interrupt Disable Register (IDR). The interrupt
request remains active until the corresponding bit in SR is cleared by writing a one to the corresponding bit in the Interrupt Clear Register (ICR).
The following SR bits are used for signalling the result of SAU accesses:
• RTR Address Error (RTRADR) is set if an illegal address is written to the RTRs. Only
addresses in the range 0xFFFC0000-0xFFFFFFFF are allowed.
• Master Interface Bus Error (MBERROR) is set if any of the conditions listed in Section 9.5.7
occurred.
108
32099I–01/2012
AT32UC3L016/32/64
• Unlock Register Error Status (URES) is set if an attempt was made to unlock a channel by
writing to the Unlock Register while one or more error bits in SR were set (see Section 9.5.6).
The unlock operation was aborted.
• Unlock Register Key Error (URKEY) is set if the Unlock Register was attempted written with
an invalid key.
• Unlock Register Read (URREAD) is set if the Unlock Register was attempted read.
• Channel Access Unsuccessful (CAU) is set if the channel access was unsuccessful.
• Channel Access Successful (CAS) is set if the channel access was successful.
• Channel Unlock Expired (EXP) is set if the channel lock expired, with no channel being
accessed after the channel was unlocked.
9.5.6
Error bits
If error bits are set when attempting to unlock a channel, SR.URES will be set. The following SR
bits are considered error bits:
• EXP
• CAU
• URREAD
• URKEY
• URES
• MBERROR
• RTRADR
9.5.7
Bus Error Responses
By writing a one to the Bus Error Response Enable bit (CR.BERREN), serious access errors will
be configured to return a bus error to the CPU. This will cause the CPU to execute its Bus Error
Data Fetch exception routine.
The conditions that can generate a bus error response are:
• Reading the Unlock Register
• Trying to access a locked channel
• The SAU HSB master receiving a bus error response from its addressed slave
9.5.8
Disabling the SAU
To disable the SAU, the user must first ensure that no SAU bus operations are pending. This
can be done by checking that the SR.IDLE bit is set.
The SAU may then be disabled by writing a one to the Disable (DIS) bit in CR.
109
32099I–01/2012
AT32UC3L016/32/64
9.6
User Interface
The following addresses are used by SAU channel configuration registers. All offsets are relative to the SAU’s PB base
address.
Table 9-1.
SAU Configuration Register Memory Map
Offset
Register
Register Name
Access
Reset
0x00
Control Register
CR
Write-only
0x00000000
0x04
Configuration Register
CONFIG
Write-only
0x00000000
0x08
Channel Enable Register High
CERH
Read/Write
0x00000000
0x0C
Channel Enable Register Low
CERL
Read/Write
0x00000000
0x10
Status Register
SR
Read-only
0x00000400
0x14
Interrupt Enable Register
IER
Write-only
0x00000000
0x18
Interrupt Disable Register
IDR
Write-only
0x00000000
0x1C
Interrupt Mask Register
IMR
Read-only
0x00000000
0x20
Interrupt Clear Register
ICR
Write-only
0x00000000
0x24
Parameter Register
PARAMETER
Read-only
-(1)
0x28
Version Register
VERSION
Read-only
-(1)
1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter.
Note:
The following addresses are used by SAU channel registers. All offsets are relative to the SAU’s HSB base address. The
number of channels implemented is device specific, refer to the Module Configuration section at the end of this chapter.
Table 9-2.
SAU Channel Register Memory Map
Offset
Register
Register Name
Access
Reset
0x00
Remap Target Register 0
RTR0
Read/Write
N/A
0x04
Remap Target Register 1
RTR1
Read/Write
N/A
0x08
Remap Target Register 2
RTR2
Read/Write
N/A
...
...
...
...
...
0x04*n
Remap Target Register n
RTRn
Read/Write
N/A
0xFC
Unlock Register
UR
Write-only
N/A
110
32099I–01/2012
AT32UC3L016/32/64
9.6.1
Name:
Control Register
CR
Access Type:
Write-only
Offset:
0x00
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
-
-
BERRDIS
BERREN
SDIS
SEN
DIS
EN
• BERRDIS: Bus Error Response Disable
Writing a zero to this bit has no effect.
Writing a one to this bit disables Bus Error Response from the SAU.
• BERREN: Bus Error Response Enable
Writing a zero to this bit has no effect.
Writing a one to this bit enables Bus Error Response from the SAU.
• SDIS: Setup Mode Disable
Writing a zero to this bit has no effect.
Writing a one to this bit exits setup mode.
• SEN: Setup Mode Enable
Writing a zero to this bit has no effect.
Writing a one to this bit enters setup mode.
• DIS: SAU Disable
Writing a zero to this bit has no effect.
Writing a one to this bit disables the SAU.
• EN: SAU Enable
Writing a zero to this bit has no effect.
Writing a one to this bit enables the SAU.
111
32099I–01/2012
AT32UC3L016/32/64
9.6.2
Name:
Configuration Register
CONFIG
Access Type:
Write-only
Offset:
0x04
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
OPEN
15
14
13
12
11
10
9
8
3
2
1
0
UCYC
7
6
5
4
UKEY
• OPEN: Open Mode Enable
Writing a zero to this bit disables open mode.
Writing a one to this bit enables open mode.
• UCYC: Unlock Number of Clock Cycles
Once a channel has been unlocked, it remains unlocked for this amount of CLK_SAU_HSB clock cycles or until one access to a
channel has been made.
• UKEY: Unlock Key
The value in this field must be written to UR.KEY to unlock a channel.
112
32099I–01/2012
AT32UC3L016/32/64
9.6.3
Name:
Channel Enable Register High
CERH
Access Type:
Read/Write
Offset:
0x08
Reset Value:
0x00000000
31
30
29
28
-
23
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
CERH[30:24]
22
21
20
CERH[23:16]
15
14
13
12
CERH[15:8]
7
6
5
4
CERH[7:0]
• CERH[n]: Channel Enable Register High
0: Channel (n+32) is not enabled.
1: Channel (n+32) is enabled.
113
32099I–01/2012
AT32UC3L016/32/64
9.6.4
Name:
Channel Enable Register Low
CERL
Access Type:
Read/Write
Offset:
0x0C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
CERL[31:24]
23
22
21
20
CERL[23:16]
15
14
13
12
CERL[15:8]
7
6
5
4
CERL[7:0]
• CERL[n]: Channel Enable Register Low
0: Channel n is not enabled.
1: Channel n is enabled.
114
32099I–01/2012
AT32UC3L016/32/64
9.6.5
Name:
Status Register
SR
Access Type:
Read-only
Offset:
0x10
Reset Value:
0x00000400
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
IDLE
SEN
EN
7
6
5
4
3
2
1
0
RTRADR
MBERROR
URES
URKEY
URREAD
CAU
CAS
EXP
• IDLE
•
•
•
•
•
•
•
This bit is cleared when a read or write operation to the SAU channel is started.
This bit is set when the operation is completed and no SAU bus operations are pending.
SEN: SAU Setup Mode Enable
This bit is cleared when the SAU exits setup mode.
This bit is set when the SAU enters setup mode.
EN: SAU Enabled
This bit is cleared when the SAU is disabled.
This bit is set when the SAU is enabled.
RTRADR: RTR Address Error
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set if, in the configuration phase, an RTR was written with an illegal address, i.e. the upper 16 bits in the address were
different from 0xFFFC, 0xFFFD, 0xFFFE or 0xFFFF.
MBERROR: Master Interface Bus Error
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set if a channel access generated a transfer on the master interface that received a bus error response from the
addressed slave.
URES: Unlock Register Error Status
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set if an attempt was made to unlock a channel by writing to the Unlock Register while one or more error bits were set
in SR. The unlock operation was aborted.
URKEY: Unlock Register Key Error
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set if the Unlock Register was attempted written with an invalid key.
URREAD: Unlock Register Read
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set if the Unlock Register was read.
115
32099I–01/2012
AT32UC3L016/32/64
• CAU: Channel Access Unsuccessful
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set if channel access was unsuccessful, i.e. an access was attempted to a locked or disabled channel.
• CAS: Channel Access Successful
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set if channel access successful, i.e. one access was made after the channel was unlocked.
• EXP: Channel Unlock Expired
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set if channel unlock has expired, i.e. no access being made after the channel was unlocked.
116
32099I–01/2012
AT32UC3L016/32/64
9.6.6
Name:
Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x14
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
RTRADR
MBERROR
URES
URKEY
URREAD
CAU
CAS
EXP
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
117
32099I–01/2012
AT32UC3L016/32/64
9.6.7
Name:
Interrupt Disable Register
IDR
Access Type:
Write-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
RTRADR
MBERROR
URES
URKEY
URREAD
CAU
CAS
EXP
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
118
32099I–01/2012
AT32UC3L016/32/64
9.6.8
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
Offset:
0x1C
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
RTRADR
MBERROR
URES
URKEY
URREAD
CAU
CAS
EXP
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
A bit in this register is cleared when the corresponding bit in IDR is written to one.
A bit in this register is set when the corresponding bit in IER is written to one.
119
32099I–01/2012
AT32UC3L016/32/64
9.6.9
Name:
Interrupt Clear Register
ICR
Access Type:
Write-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
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
RTRADR
MBERROR
URES
URKEY
URREAD
CAU
CAS
EXP
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in SR and any corresponding interrupt request.
120
32099I–01/2012
AT32UC3L016/32/64
9.6.10
Name:
Parameter Register
PARAMETER
Access Type:
Read-only
Offset:
0x24
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
-
-
CHANNELS
• CHANNELS:
Number of channels implemented.
121
32099I–01/2012
AT32UC3L016/32/64
9.6.11
Name:
Version Register
VERSION
Access Type:
Write-only
Offset:
0x28
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]
3
2
VERSION[7:0]
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the module. No functionality associated.
122
32099I–01/2012
AT32UC3L016/32/64
9.6.12
Name:
Remap Target Register n
RTRn
Access Type:
Read/Write
Offset:
n*4
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
RTR[31:24]
23
22
21
20
RTR[23:16]
15
14
13
12
RTR[15:8]
7
6
5
4
RTR[7:0]
• RTR: Remap Target Address for Channel n
RTR[31:16] must have one of the following values, any other value will result in UNDEFINED behavior:
0xFFFC
0xFFFD
0xFFFE
0xFFFF
RTR[1:0] must be written to 00, any other value will result in UNDEFINED behavior.
123
32099I–01/2012
AT32UC3L016/32/64
9.6.13
Name:
Unlock Register
UR
Access Type :
Write-only
Offset:
0xFC
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
KEY
7
6
-
-
5
4
CHANNEL
• KEY: Unlock Key
The correct key must be written in order to unlock a channel. The key value written must correspond to the key value defined in
CONFIG.UKEY.
• CHANNEL: Channel Number
Number of the channel to unlock.
124
32099I–01/2012
AT32UC3L016/32/64
9.7
Module Configuration
The specific configuration for each SAU instance is listed in the following tables.The module bus
clocks listed here are connected to the system bus clocks. Please refer to the Power Manager
chapter for details.
Table 9-3.
Module configuration
Feature
SAU
SAU Channels
16
Table 9-4.
Module clock name
Module name
Clock name
Description
SAU
CLK_SAU_HSB
Clock for the SAU HSB interface
SAU
CLK_SAU_PB
Clock for the SAU PB interface
Table 9-5.
Register Reset Values
Register
Reset Value
VERSION
0x00000110
PARAMETER
0x00000010
125
32099I–01/2012
AT32UC3L016/32/64
10. HSB Bus Matrix (HMATRIXB)
Rev: 1.3.0.3
10.1
Features
•
•
•
•
•
•
•
•
•
10.2
User Interface on peripheral bus
Configurable number of masters (up to 16)
Configurable number of slaves (up to 16)
One decoder 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)
Overview
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.
10.3
Product Dependencies
In order to configure this module by accessing the user registers, other parts of the system must
be configured correctly, as described below.
10.3.1
Clocks
The clock for the HMATRIX bus interface (CLK_HMATRIX) is generated by the Power Manager.
This clock is enabled at reset, and can be disabled in the Power Manager.
10.4
10.4.1
Functional Description
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.
126
32099I–01/2012
AT32UC3L016/32/64
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.
10.4.1.1
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.
10.4.1.2
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.
10.4.1.3
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).
10.4.2
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 is selected by the ARBT field in 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. This is described in “Arbitration
Rules” .
10.4.2.1
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. This is described below.
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. This is described below.
127
32099I–01/2012
AT32UC3L016/32/64
• 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 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 ULBT field in the Master Configuration Registers
(MCFG).
• 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, halfword, or word
transfer.
10.4.2.2
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:
1. Round-Robin arbitration without default master
2. Round-Robin arbitration with last default master
3. Round-Robin arbitration with fixed default master
• 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.
• 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. At the end of the cur128
32099I–01/2012
AT32UC3L016/32/64
rent 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.
• 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.
10.4.2.3
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).
10.4.3
Slave and Master assignation
The index number assigned to Bus Matrix slaves and masters are described in the Module Configuration section at the end of this chapter.
129
32099I–01/2012
AT32UC3L016/32/64
10.5
User Interface
Table 10-1.
HMATRIX Register Memory Map
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
130
32099I–01/2012
AT32UC3L016/32/64
Table 10-1.
HMATRIX Register Memory Map (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
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
–
0x0124
Special Function Register 5
SFR5
Read/Write
–
0x0128
Special Function Register 6
SFR6
Read/Write
–
131
32099I–01/2012
AT32UC3L016/32/64
Table 10-1.
HMATRIX Register Memory Map (Continued)
Offset
Register
Name
Access
Reset Value
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
–
132
32099I–01/2012
AT32UC3L016/32/64
10.5.1
Name:
Master Configuration Registers
MCFG0...MCFG15
Access Type:
Read/Write
Offset:
0x00 - 0x3C
Reset Value:
0x00000002
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
Table 10-2.
Undefined Length Burst Type
ULBT
Undefined Length Burst Type
Description
000
Inifinite Length Burst
No predicted end of burst is generated and therefore INCR bursts coming from this
master cannot be broken.
001
Single-Access
The undefined length burst is treated as a succession of single accesses, allowing rearbitration at each beat of the INCR burst.
010
4 Beat Burst
The undefined length burst is split into a four-beat burst, allowing re-arbitration at each
four-beat burst end.
011
8 Beat Burst
The undefined length burst is split into an eight-beat burst, allowing re-arbitration at
each eight-beat burst end.
100
16 Beat Burst
The undefined length burst is split into a sixteen-beat burst, allowing re-arbitration at
each sixteen-beat burst end.
133
32099I–01/2012
AT32UC3L016/32/64
10.5.2
Name:
Slave Configuration Registers
SCFG0...SCFG15
Access Type:
Read/Write
Offset:
0x40 - 0x7C
Reset Value:
0x00000010
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
• ARBT: Arbitration Type
0: Round-Robin Arbitration
1: Fixed Priority Arbitration
• 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.
• 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.
• 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.
134
32099I–01/2012
AT32UC3L016/32/64
10.5.3
Bus Matrix Priority Registers A For Slaves
Register Name:
PRAS0...PRAS15
Access Type:
Read/Write
Offset:
-
Reset Value:
0x00000000
31
30
-
-
23
22
-
-
15
14
-
-
7
6
-
-
29
28
M7PR
21
20
M5PR
13
12
M3PR
5
4
M1PR
27
26
-
-
19
18
-
-
11
10
-
-
3
2
-
-
25
24
M6PR
17
16
M4PR
9
8
M2PR
1
0
M0PR
• MxPR: Master x Priority
Fixed priority of Master x for accessing the selected slave. The higher the number, the higher the priority.
135
32099I–01/2012
AT32UC3L016/32/64
10.5.4
Name:
Priority Registers B For Slaves
PRBS0...PRBS15
Access Type:
Read/Write
Offset:
-
Reset Value:
0x00000000
31
30
-
-
23
22
-
-
15
14
-
-
7
6
-
-
29
28
M15PR
21
20
M13PR
13
12
M11PR
5
4
M9PR
27
26
-
-
19
18
-
-
11
10
-
-
3
2
-
-
25
24
M14PR
17
16
M12PR
9
8
M10PR
1
0
M8PR
• MxPR: Master x Priority
Fixed priority of Master x for accessing the selected slave. The higher the number, the higher the priority.
136
32099I–01/2012
AT32UC3L016/32/64
10.5.5
Name:
Special Function Registers
SFR0...SFR15
Access Type:
Read/Write
Offset:
0x110 - 0x14C
Reset Value:
-
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
Those registers are not a HMATRIX specific register. The field of those will be defined where they are used.
137
32099I–01/2012
AT32UC3L016/32/64
10.6
Module Configuration
The specific configuration for each HMATRIX instance is listed in the following tables.The module bus clocks listed here are connected to the system bus clocks. Please refer to the Power
Manager chapter for details.
Table 10-3.
10.6.1
HMATRIX Clocks
Clock Name
Description
CLK_HMATRIX
Clock for the HMATRIX bus interface
Bus Matrix Connections
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, HMATRIX MCFG0
register is associated with the CPU Data master interface.
Table 10-4.
High Speed Bus Masters
Master 0
CPU Data
Master 1
CPU Instruction
Master 2
CPU SAB
Master 3
SAU
Master 4
PDCA
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.
Accesses to unused areas returns an error result to the master requesting such an access.
Table 10-5.
High Speed Bus Slaves
Slave 0
Internal Flash
Slave 1
HSB-PB Bridge A
Slave 2
HSB-PB Bridge B
Slave 3
Internal SRAM
Slave 4
SAU
138
32099I–01/2012
AT32UC3L016/32/64
Figure 10-1. HMatrix Master / Slave Connections
HMATRIX MASTERS
CPU Data
0
CPU
Instruction
1
CPU SAB
2
SAU
3
PDCA
4
Internal Flash
HSB-PB
Bridge A
HSB-PB
Bridge B
Internal SRAM
SAU
HMATRIX SLAVES
0
1
2
3
4
139
32099I–01/2012
AT32UC3L016/32/64
11. Interrupt Controller (INTC)
Rev: 1.0.2.5
11.1
Features
• Autovectored low latency interrupt service with programmable priority
– 4 priority levels for regular, maskable interrupts
– One Non-Maskable Interrupt
• Up to 64 groups of interrupts with up to 32 interrupt requests in each group
11.2
Overview
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 a
pending interrupt of the same level, the group with the lowest number takes priority.
11.3
Block Diagram
Figure 11-1 gives an overview of the INTC. The grey boxes represent registers that can be
accessed via the user interface. 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.
140
32099I–01/2012
AT32UC3L016/32/64
Figure 11-1. INTC Block Diagram
Interrupt Controller
CPU
NMIREQ
Masks
OR
IRRn
GrpReqN
IREQ63
IREQ34
IREQ33
IREQ32
OR
GrpReq1
INT_level,
offset
IPRn
.
.
.
Request
Masking ValReq1
INT_level,
offset
IPR1
.
.
.
INTLEVEL
Prioritizer
.
.
.
ValReqN
SREG
Masks
I[3-0]M
GM
AUTOVECTOR
IRR1
IREQ31
IREQ2
IREQ1
IREQ0
OR
GrpReq0
ValReq0
IPR0
INT_level,
offset
IRR0
IRR Registers
11.4
IPR Registers
ICR Registers
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
11.4.1
Power Management
If the CPU enters a sleep mode that disables CLK_SYNC, the INTC will stop functioning and
resume operation after the system wakes up from sleep mode.
11.4.2
Clocks
The clock for the INTC bus interface (CLK_INTC) is generated by the Power Manager. This
clock is enabled at reset, and can be disabled in the Power Manager.
The INTC sampling logic runs on a clock which is stopped in any of the sleep modes where the
system RC oscillator is not running. This clock is referred to as CLK_SYNC. This clock is
enabled at reset, and only turned off in sleep modes where the system RC oscillator is stopped.
11.4.3
11.5
Debug Operation
When an external debugger forces the CPU into debug mode, the INTC continues normal
operation.
Functional Description
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 are active, the interrupt service routine must
prioritize 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 Interrupt Level (INTLEVEL) field in the corresponding
141
32099I–01/2012
AT32UC3L016/32/64
Interrupt Priority Register (IPR). The GrpReq inputs are then masked by the 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.
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 GM or the corresponding interrupt
level mask bit is set.
The Prioritizer hardware uses the ValReq lines and the INTLEVEL field in the IPRs to select the
pending interrupt of the highest priority. If an NMI interrupt request is pending, it automatically
gets the highest priority of any pending interrupt. If several interrupt groups of the highest pending interrupt level have pending interrupts, the interrupt group with the lowest number is
selected.
The INTLEVEL and handler autovector offset (AUTOVECTOR) of the selected interrupt are
transmitted to the CPU for interrupt handling and context switching. The CPU does not 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
user interface registers for checking which interrupts of the group are actually active.
The delay through the INTC from the peripheral interrupt request is set until the interrupt request
to the CPU is set is three cycles of CLK_SYNC.
11.5.1
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.
11.5.2
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 an interrupt of level 3 is approved for handling, the interrupt mask bits I3M,
I2M, I1M, and I0M are set in status register. If an interrupt of level 1 is approved, the masking
bits I1M and I0M are set in status register. The handler address is calculated by logical OR of
the AUTOVECTOR to the CPU system register Exception Vector Base Address (EVBA). The
CPU will then jump to the calculated address and start executing the interrupt handler.
Setting the interrupt mask bits prevents the interrupts from the same and lower levels to be
passed through the interrupt controller. Setting of the same level mask bit prevents also multiple
requests 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.
11.5.3
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
142
32099I–01/2012
AT32UC3L016/32/64
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.
143
32099I–01/2012
AT32UC3L016/32/64
11.6
User Interface
Table 11-1.
INTC Register Memory Map
Offset
Register
Register Name
Access
Reset
0x000
Interrupt Priority Register 0
IPR0
Read/Write
0x00000000
0x004
Interrupt Priority Register 1
IPR1
Read/Write
0x00000000
...
...
...
...
...
0x0FC
Interrupt Priority Register 63
IPR63
Read/Write
0x00000000
0x100
Interrupt Request Register 0
IRR0
Read-only
N/A
0x104
Interrupt Request Register 1
IRR1
Read-only
N/A
...
...
...
...
...
0x1FC
Interrupt Request Register 63
IRR63
Read-only
N/A
0x200
Interrupt Cause Register 3
ICR3
Read-only
N/A
0x204
Interrupt Cause Register 2
ICR2
Read-only
N/A
0x208
Interrupt Cause Register 1
ICR1
Read-only
N/A
0x20C
Interrupt Cause Register 0
ICR0
Read-only
N/A
144
32099I–01/2012
AT32UC3L016/32/64
11.6.1
Name:
Interrupt Priority Registers
IPR0...IPR63
Access Type:
Read/Write
Offset:
0x000 - 0x0FC
Reset Value:
0x00000000
31
30
INTLEVEL
29
28
27
26
25
24
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
7
6
1
0
AUTOVECTOR[13:8]
5
4
3
2
AUTOVECTOR[7:0]
• INTLEVEL: Interrupt Level
Indicates the EVBA-relative offset of the interrupt handler of the corresponding group:
00: INT0: Lowest priority
01: INT1
10: INT2
11: INT3: Highest priority
• AUTOVECTOR: Autovector Address
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.
145
32099I–01/2012
AT32UC3L016/32/64
11.6.2
Name:
Interrupt Request Registers
IRR0...IRR63
Access Type:
Read-only
Offset:
0x0FF - 0x1FC
Reset Value:
N/A
31
30
29
28
27
26
25
24
IRR[32*x+31]
IRR[32*x+30]
IRR[32*x+29]
IRR[32*x+28]
IRR[32*x+27]
IRR[32*x+26]
IRR[32*x+25]
IRR[32*x+24]
23
22
21
20
19
18
17
16
IRR[32*x+23]
IRR[32*x+22]
IRR[32*x+21]
IRR[32*x+20]
IRR[32*x+19]
IRR[32*x+18]
IRR[32*x+17]
IRR[32*x+16]
15
14
13
12
11
10
9
8
IRR[32*x+15]
IRR[32*x+14]
IRR[32*x+13]
IRR[32*x+12]
IRR[32*x+11]
IRR[32*x+10]
IRR[32*x+9]
IRR[32*x+8]
7
6
5
4
3
2
1
0
IRR[32*x+7]
IRR[32*x+6]
IRR[32*x+5]
IRR[32*x+4]
IRR[32*x+3]
IRR[32*x+2]
IRR[32*x+1]
IRR[32*x+0]
• IRR: Interrupt Request line
This bit is cleared when no interrupt request is pending on this input request line.
This bit is set when an interrupt request is pending on this input request line.
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.
146
32099I–01/2012
AT32UC3L016/32/64
11.6.3
Name:
Interrupt Cause Registers
ICR0...ICR3
Access Type:
Read-only
Offset:
0x200 - 0x20C
Reset Value:
N/A
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. This value is only defined when at least
one interrupt of level n is pending.
147
32099I–01/2012
AT32UC3L016/32/64
11.7
Module Configuration
The specific configuration for each INTC instance is listed in the following tables.The module
bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details.
Table 11-2.
11.7.1
INTC Clock Name
Module Name
Clock Name
Description
INTC
CLK_INTC
Clock for the INTC bus interface
Interrupt Request Signal Map
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.
The table below shows how the interrupt request signals are connected to the INTC.
Table 11-3.
Interrupt Request Signal Map
Group
Line
Module
Signal
0
0
AVR32 UC3 Core
SYSREG COMPARE
0
AVR32 UC3 Core
OCD DCEMU_DIRTY
1
AVR32 UC3 Core
OCD DCCPU_READ
2
0
Flash Controller
3
0
Secure Access Unit
0
Peripheral DMA Controller
PDCA 0
1
Peripheral DMA Controller
PDCA 1
2
Peripheral DMA Controller
PDCA 2
3
Peripheral DMA Controller
PDCA 3
0
Peripheral DMA Controller
PDCA 4
1
Peripheral DMA Controller
PDCA 5
2
Peripheral DMA Controller
PDCA 6
3
Peripheral DMA Controller
PDCA 7
0
Peripheral DMA Controller
PDCA 8
1
Peripheral DMA Controller
PDCA 9
2
Peripheral DMA Controller
PDCA 10
3
Peripheral DMA Controller
PDCA 11
7
0
Power Manager
8
0
System Control Interface
9
0
Asynchronous Timer
1
FLASHCDW
SAU
4
5
6
PM
SCIF
AST ALARM
148
32099I–01/2012
AT32UC3L016/32/64
Table 11-3.
Interrupt Request Signal Map
0
Asynchronous Timer
AST PER
1
Asynchronous Timer
AST OVF
2
Asynchronous Timer
AST READY
3
Asynchronous Timer
AST CLKREADY
0
External Interrupt Controller
EIC 1
1
External Interrupt Controller
EIC 2
2
External Interrupt Controller
EIC 3
3
External Interrupt Controller
EIC 4
12
0
External Interrupt Controller
EIC 5
13
0
Frequency Meter
FREQM
0
General Purpose Input/Output Controller
GPIO 0
1
General Purpose Input/Output Controller
GPIO 1
2
General Purpose Input/Output Controller
GPIO 2
3
General Purpose Input/Output Controller
GPIO 3
4
General Purpose Input/Output Controller
GPIO 4
5
General Purpose Input/Output Controller
GPIO 5
15
0
Universal Synchronous/Asynchronous
Receiver/Transmitter
USART0
16
0
Universal Synchronous/Asynchronous
Receiver/Transmitter
USART1
17
0
Universal Synchronous/Asynchronous
Receiver/Transmitter
USART2
18
0
Universal Synchronous/Asynchronous
Receiver/Transmitter
USART3
19
0
Serial Peripheral Interface
SPI
20
0
Two-wire Master Interface
TWIM0
21
0
Two-wire Master Interface
TWIM1
22
0
Two-wire Slave Interface
TWIS0
23
0
Two-wire Slave Interface
TWIS1
24
0
Pulse Width Modulation Controller
PWMA
0
Timer/Counter
TC00
1
Timer/Counter
TC01
2
Timer/Counter
TC02
0
Timer/Counter
TC10
1
Timer/Counter
TC11
2
Timer/Counter
TC12
0
ADC Interface
ADCIFB
10
11
14
25
26
27
149
32099I–01/2012
AT32UC3L016/32/64
Table 11-3.
Interrupt Request Signal Map
28
0
Analog Comparator Interface
ACIFB
29
0
Capacitive Touch Module
CAT
30
0
aWire
AW
150
32099I–01/2012
AT32UC3L016/32/64
12. Power Manager (PM)
Rev: 4.1.1.4
12.1
Features
•
•
•
•
•
•
•
12.2
Generates clocks and resets for digital logic
On-the-fly frequency change of CPU, HSB and PBx clocks
Sleep modes allow simple disabling of logic clocks and clock sources
Module-level clock gating through maskable peripheral clocks
Wake-up from internal or external interrupts
Automatic identification of reset sources
Supports advanced Shutdown sleep mode
Overview
The Power Manager (PM) provides synchronous clocks used to clock the main digital logic in the
device, namely the CPU, and the modules and peripherals connected to the High Speed Bus
(HSB) and the Peripheral Buses (PBx).
The PM contains advanced power-saving features, allowing the user to optimize the power consumption for an application. The synchronous clocks are divided into a number of clock
domains, one for the CPU and HSB, and one for each PBx. The clocks can run at different
speeds, allowing the user to save power by running peripherals relatively slow, whilst maintaining high CPU performance. The clocks can be independently changed on-the-fly, without halting
any peripherals. The user may adjust CPU and memory speeds according to the dynamic application load, without disturbing or re-configuring active peripherals.
Each module has a separate clock, enabling the user to save power by switching off clocks to
inactive modules. Clocks and oscillators can be automatically switched off during idle periods by
the CPU sleep instruction. The system will return to normal operation when interrupts occur.
To achieve minimal power usage, a special sleep mode, called Shutdown is available, where
power on all internal logic (CPU, peripherals) and most of the I/O lines is removed, reducing current leakage. Only a small amount of logic, including the 32KHz crystal oscillator (OSC32K) and
the AST remain powered.
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.
151
32099I–01/2012
AT32UC3L016/32/64
12.3
Block Diagram
Figure 12-1. PM Block Diagram
Main Clock Sources
Synchronous
Clock Generator
Synchronous
clocks
CPU, HSB,
PBx
Interrupts
Sleep Controller
Sleep
Instruction
Reset Controller
Resets
Reset Sources
Power-on Reset
Detector(s)
External Reset Pin
12.4
I/O Lines Description
Table 12-1.
I/O Lines Description
Name
Description
Type
Active Level
RESET_N
Reset
Input
Low
12.5
12.5.1
Product Dependencies
Interrupt
The PM interrupt line is connected to one of the interrupt controllers internal sources. Using the
PM interrupt requires the interrupt controller to be configured first.
12.5.2
Clock Implementation
In AT32UC3L016/32/64, the HSB shares source clock with the CPU. Write attempts to the HSB
Clock Select register (HSBSEL) will be ignored, and it will always read the same as the CPU
Clock Select register (CPUSEL).
The PM bus interface clock (CLK_PM) is generated by the Power Manager. This clock is
enabled at reset, and can be disabled in the Power Manager. If disabled it can only be reenabled by a reset.
12.5.3
Power Considerations
The Shutdown mode is only available for the “3.3V supply mode, with 1.8V regulated I/O lines“
power configuration.
152
32099I–01/2012
AT32UC3L016/32/64
12.6
Functional Description
12.6.1
Synchronous Clocks
The System RC Oscillator (RCSYS) and a selection of other clock sources can provide the
source for the main clock, which is the origin for the synchronous CPU/HSB and PBx module
clocks. For details about the other main clock sources, please refer to the Main Clock Control
(MCCTRL) register description. The synchronous clocks can run of the main clock and all the 8bit prescaler settings as long as fCPU ≥ fPBx,. The synchronous clock source can be changed onthe fly, according to variations in application load. The clock domains can be shut down in sleep
mode, as described in Section 12.6.3. The module clocks in every synchronous clock domain
can be individually masked to minimize power consumption in inactive modules.
Figure 12-2. Synchronous Clock Generation
Sleep
Controller
Sleep
Instruction
0
Main Clock
Main Clock
Sources
1
Prescaler
CPUDIV
MCSEL
Mask
CPU Clocks
HSB Clocks
CPUMASK
PBx Clocks
CPUSEL
12.6.1.1
Selecting the main clock source
The common main clock can be connected to RCSYS or a selection of other clock sources. For
details about the other main clock sources, please refer to the MCCTRL register description. By
default, the main clock will be connected to RCSYS. The user can connect the main clock to
another source by writing to the Main Clock Select (MCCTRL.MCSEL) field. The user must first
assure that the source is enabled and ready in order to avoid a deadlock. Care should also be
taken so that the new synchronous clock frequencies do not exceed the maximum frequency for
each clock domain.
12.6.1.2
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 prescaler division for the CPU clock by writing a one to the CPU Division bit in the CPU Clock Select
register (CPUSEL.CPUDIV), and a value to the CPU Clock Select field (CPUSEL.CPUSEL),
resulting in a CPU clock frequency:
fCPU = fmain / 2(CPUSEL+1)
153
32099I–01/2012
AT32UC3L016/32/64
Similarly, the PBx clocks can be divided by writing their respective Clock Select (PBxSEL) registers to get the divided PBx frequency:
fPBx = fmain / 2(PBSEL+1)
The PBx clock frequency can not exceed the CPU clock frequency. The user must select a PBxSEL.PBSEL value greater than or equal to the CPUSEL.CPUSEL value, so that fCPU ≥ fPBx. If the
user selects division factors that will result in fCPU< fPBx, the Power Manager will automatically
change the PBxSEL.PBSEL/PBDIV values to ensure correct operation (fCPU ≥ fPBx).
The HSB clock will always be forced to the same division as the CPU clock.
To ensure correct operation, the frequencies must never exceed the specified maximum frequency for each clock domain.
For modules connected to the HSB bus, the PB clock frequency must be the same as the CPU
clock frequency.
12.6.1.3
Clock Ready flag
There is a slight delay from CPUSEL and PBxSEL being written to the new clock setting taking
effect. During this interval, the Clock Ready bit in the Status Register (SR.CKRDY) will read as
zero. When the clock settings change is completed, the bit will read as one. The Clock Select
registers (CPUSEL, PBxSEL) must not be written to while SR.CKRDY is zero, or the system
may become unstable or hang.
The Clock Ready bit in the Interrupt Status Register (ISR.CKRDY) is set on a SR.CKRDY zeroto-one transition. If the Clock Ready bit in the Interrupt Mask Register (IMR.CKRDY) is set, an
interrupt request is generated. IMR.CKRDY is set by writing a one to the corresponding bit in the
Interrupt Enable Register (IER.CKRDY).
12.6.2
Peripheral Clock Masking
By default, the clocks 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, or PBx clock
domain by writing a zero to the corresponding bit in the corresponding Clock Mask (CPUMASK/HSBMASK/PBxMASK) register. When a module is not clocked, it will cease operation,
and its registers cannot be read nor written. The module can be re-enabled later by writing a one
to the corresponding mask bit. A module may be connected to several clock domains, in which
case it will have several mask bits. The Maskable Module Clocks table in the Clock Mask register description contains a list of implemented maskable clocks.
12.6.2.1
12.6.3
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 Flash Controller will cause a problem if the CPU needs to read
from the flash. Switching off the clock to the Power Manager, which contains the mask registers,
or the corresponding PBx bridge, will make it impossible to write to the mask registers again. In
this case, they can only be re-enabled by a system reset.
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 it and other (optional) clock domains off
to save power. This is done by the sleep instruction, which takes the sleep mode index number
from Table 12-2 on page 155 as argument.
154
32099I–01/2012
AT32UC3L016/32/64
12.6.3.1
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 in the mask registers.
Clock sources can also be switched off to save power. Some of these 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.
12.6.3.2
Supported sleep modes
The following sleep modes are supported. These are detailed in Table 12-2 on page 155.
• Idle: The CPU is stopped, the rest of the device is operational.
• Frozen: The CPU and HSB modules are stopped, peripherals are operational.
• Standby: All synchronous clocks are stopped, and the clock sources are running, allowing for
a quick wake-up to normal mode.
• Stop: As Standby, but oscillators, and other clock sources are also stopped. 32KHz Oscillator
OSC32K(2), RCSYS, AST, and WDT will remain operational.
• DeepStop: All synchronous clocks and clock sources are stopped. Bandgap voltage
reference and BOD are turned off. OSC32K(2) and RCSYS remain operational.
• Static: All clock sources, including RCSYS are stopped. Bandgap voltage reference and BOD
are turned off. OSC32K(2) remains operational.
• Shutdown: All clock sources, including RCSYS are stopped. Bandgap voltage reference,
BOD detector, and Voltage regulator are turned off. OSC32K(2) remains operational. This
mode can only be used in the “3.3V supply mode, with 1.8V regulated I/O lines“
configuration (described in Power Considerations chapter). Refer to Section 12.6.4 for more
details.
Table 12-2.
Sleep Modes
Clock Sources(3),
SYSTIMER(4)
OSC32K
Sleep Mode
CPU
HSB
PBx,
GCLK
0
Idle
Stop
Run
Run
Run
Run
Run
On
Normal mode
1
Frozen
Stop
Stop
Run
Run
Run
Run
On
Normal mode
2
Standby
Stop
Stop
Stop
Run
Run
Run
On
Normal mode
3
Stop
Stop
Stop
Stop
Stop
Run
Run
On
Low power mode
4
DeepStop
Stop
Stop
Stop
Stop
Run
Run
Off
Low power mode
5
Static
Stop
Stop
Stop
Stop
Run
Stop
Off
Low power mode
6
Shutdown
Stop
Stop
Stop
Stop
Run
Stop
Off
Off
(1)
Index
Notes:
(2)
RCSYS
BOD &
Bandgap
Voltage
Regulator
1. The sleep mode index is used as argument for the sleep instruction.
2. OSC32K will only remain operational if pre-enabled.
3. Clock sources other than those specifically listed in the table.
4. SYSTIMER is the clock for the CPU COUNT and COMPARE registers.
The internal voltage regulator is also adjusted according to the sleep mode in order to reduce its
power consumption.
155
32099I–01/2012
AT32UC3L016/32/64
12.6.3.3
Waking from sleep modes
There are two types of wake-up sources from sleep mode, synchronous and asynchronous.
Synchronous wake-up sources are all non-masked interrupts. Asynchronous wake-up sources
are AST, WDT, external interrupts from EIC, external reset, external wake pin (WAKE_N), and
all asynchronous wake-ups enabled in the Asynchronous Wake Up Enable (AWEN) register. The
valid wake-up sources for each sleep mode are detailed in Table 12-3 on page 156.
In Shutdown the only wake-up sources are external reset, external wake-up pin or AST. See
Section 12.6.4.3 on page 158.
Table 12-3.
Index(1)
Wake-up Sources
Sleep Mode
Wake-up Sources
0
Idle
Synchronous, Asynchronous
1
Frozen
Synchronous(2), Asynchronous
2
Standby
Asynchronous
3
Stop
Asynchronous
4
DeepStop
Asynchronous
5
Static
Asynchronous(3)
6
Shutdown
External reset, External wake-up pin
Notes:
1. The sleep mode index is used as argument for the sleep instruction.
2. Only PB modules operational, as HSB module clocks are stopped.
3. WDT only available if clocked from pre-enabled OSC32K.
12.6.3.4
SleepWalking
In all sleep modes where the PBx clocks are stopped, except for Shutdown mode, the device
can partially wake up if a PBx module asynchronously discovers that it needs its clock. Only the
requested clocks and clock sources needed will be started, all other clocks will remain masked
to zero. E.g. if the main clock source is OSC0, only OSC0 will be started even if other clock
sources were enabled in normal mode. Generic clocks can also be started in a similar way. The
state where only requested clocks are running is referred to as SleepWalking.
The time spent to start the requested clock is mostly limited by the startup time of the given clock
source. This allows PBx modules to handle incoming requests, while still keeping the power consumption at a minimum.
When the device is SleepWalking any asynchronous wake-up can wake the device up at any
time without stopping the requested PBx clock.
All requests to start clocks can be masked by writing to the Peripheral Power Control Register
(PPCR), all requests are enabled at reset.
During SleepWalking the interrupt controller clock will be running. If an interrupt is pending when
entering SleepWalking, it will wake the whole device up.
12.6.3.5
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 will prevent erratic behavior caused by
entering or exiting sleep modes. Please refer to the relevant module documentation for recommended actions.
156
32099I–01/2012
AT32UC3L016/32/64
Communication between the synchronous clock domains is disturbed when entering and exiting
sleep modes. Bus transactions over clock domains affected by the sleep mode are therefore not
recommended. The system may hang if the bus clocks are stopped during 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. In order to let potential PBx write operations complete, the user should
let the CPU perform a PBx register read operation before issuing the sleep instruction. This will
stall the CPU until pending PBx operations have completed.
The Shutdown sleep mode requires extra care. Please refer to Section 12.6.4.
12.6.4
Shutdown Sleep Mode
12.6.4.1
Description
The Shutdown sleep mode is available only when the device is used in the “3.3V supply mode,
with 1.8V regulated I/O lines“ configuration (refer to the Power Considerations chapter). In this
configuration, the voltage regulator supplies both VDDCORE and VDDIO power supplies.
When the device enters Shutdown mode, the regulator is turned off and only the following logic
is kept powered by VDDIN:
– OSC32K using alternate pinout PA13/PA20
– AST core logic (internal counter and alarm detection logic)
– Backup Registers
– I/O lines PA11, PA13, PA20, PA21, PB04, PB05, and PB10
– RESET_N line
The table below lists I/O line functionality that remains operational during Shutdown sleep mode.
If no special function is used the I/O line will keep its setting when entering the sleep mode
Table 12-4.
I/O Lines Usage During Shutdown Mode
Pin
Possible Usage During Shutdown Sleep Mode
PA11
WAKE_N signal (active low wake-up)
PA13
XIN32_2 (OSC32K using alternate pinout)
PA20
XOUT32_2 (OSC32K using alternate pinout)
PA21
PB04
PB05
PB10
RESET_N
12.6.4.2
Reset pin
Entering Shutdown sleep mode
Before entering the Shutdown sleep mode, a few actions are required:
– All modules should normally be disabled before entering Shutdown sleep mode (see
Section 12.6.3.5)
157
32099I–01/2012
AT32UC3L016/32/64
– The POR33 must be masked to avoid spurious resets when the power is back. This
must also be done when POR33 is disabled, as POR33 will be enabled
automatically when the device wakes up from Shutdown mode. Disable the POR33
by writing a one to the POR33MASK bit in the SCIF.VREGCR register. Due to
internal synchronisation, this bit must be read as a one before the sleep instruction is
executed by the CPU. Refer to the System Control Interface (SCIF) chapter for more
details.
– The 32KHz RC oscillator (RC32K) must be running and stable. This is done by
writing a one to the EN bit in the SCIF.RC32KCR register. Due to internal
synchronisation, this bit must be read as a one to ensure that the oscillator is stable
before the sleep instruction is executed by the CPU.
As soon as the Shutdown sleep mode is entered, all CPU and peripherals are reset to ensure a
consistent state. POR33 and RC32K are automatically disabled to save extra power.
12.6.4.3
Leaving Shutdown sleep mode
Exiting Shutdown sleep mode can be done by the events described in Table 12-5.
Table 12-5.
Events That Can Wake up the Device from Shutdown Mode
Source
How
PA11 (WAKE_N)
Pulling-down PA11 will wake up the device
RESET_N
Pulling-down RESET_N pin will wake up the device
The device is kept under reset until RESET_N is tied high
again
AST
OSC32K must be set-up to use alternate pinout (XIN32_2
and XOUT32_2) Refer to the SCIF Chapter
AST must be configured to use the clock from OSC32K
AST must be configured to allow alarm, periodic, or
overflow wake-up
When a wake-up event occurs, the regulator is turned on and the device will wait for VDDCORE
to be valid before starting. The Sleep Reset bit in the Reset Cause register (RCAUSE.SLEEP) is
then set, allowing software running on the device to distinguish between the first power-up and a
wake-up from Shutdown mode.
12.6.4.4
12.6.5
Special consideration regarding waking up from Shutdown sleep mode using the WAKE_N pin
By default, the WAKE_N pin will only wake the device up if it is pulled low after entering Shutdown mode. If the WAKE_N is pulled low before the Shutdown mode is entered, it will not wake
the device from the Shutdown sleep mode. In order to wake the device by pulling WAKE_N low
before entering Shutdown mode, the user has to write a one to the bit corresponding to the
WAKEN wake-up source in the AWEN register. In this scenario, the CPU execution will proceed
with the next instruction, and the RCAUSE register content will not be altered.
Divided PB Clocks
The clock generator in the Power Manager provides divided PBx 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 directly maskable, and are stopped in sleep modes where the PBx
clocks are stopped.
158
32099I–01/2012
AT32UC3L016/32/64
12.6.6
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 Reset (POR) 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.
It is also possible to reset the device by pulling the RESET_N pin low. This pin has an internal
pull-up, and does not need to be driven externally during normal operation. Table 12-6 on page
159 lists these and other reset sources supported by the Reset Controller.
Figure 12-3. Reset Controller Block Diagram
RCAUSE
RESET_N
Power-on Reset
Detector(s)
CPU, HSB, PBx
Brown-out
Detector
Reset
Controller
OCD, AST, WDT,
Clock Generator
SM33 Detector
JTAG
AWIRE
OCD
Watchdog Reset
In addition to the listed reset types, the JTAG & aWire can keep parts of the device statically
reset. See JTAG and aWire documentation for details.
Table 12-6.
Reset Description
Reset Source
Description
Power-on Reset
Supply voltage below the Power-on Reset detector threshold
voltage VPOT
External Reset
RESET_N pin asserted
Brown-out Reset
VDDCORE supply voltage below the Brown-out detector
threshold voltage
159
32099I–01/2012
AT32UC3L016/32/64
Reset Source
Description
SM33 Reset
Internal regulator supply voltage below the SM33 threshold
voltage. This generates a Power-on Reset.
Watchdog Timer
See Watchdog Timer documentation
OCD
See On-Chip Debug documentation
Depending on the reset source, when a reset occurs, some parts of the device are not always
reset. Only the Power-on Reset (POR) will force a whole device reset. Refer to the table in the
Module Configuration section at the end of this chapter for further details. The latest reset cause
can be read in the RCAUSE register, and can be read during the applications boot sequence in
order to determine proper action.
12.6.6.1
Power-on Reset Detector
The Power-on Reset 1.8V (POR18) detector monitors the VDDCORE supply pin and generates
a Power-on Reset (POR) when the device is powered on. The POR is active until the
VDDCORE voltage is above the power-on threshold level (VPOT). The POR will be re-generated
if the voltage drops below the power-on threshold level. See Electrical Characteristics for parametric details.
The Power-on Reset 3.3V (POR33) detector monitors the internal regulator supply pin and generates a Power-on Reset (POR) when the device is powered on. The POR is active until the
internal regulator supply voltage is above the regulator power-on threshold level (VPOT). The
POR will be re-generated if the voltage drops below the regulator power-on threshold level. See
Electrical Characteristics for parametric details.
12.6.6.2
12.6.7
External Reset
The external reset detector monitors the RESET_N pin state. By default, a low level on this pin
will generate a reset.
Clock Failure Detector
This mechanism automatically switches the main clock source to the safe RCSYS clock when
the main clock source fails. This may happen when an external crystal is selected as a source
for the main clock and the crystal is not mounted on the board. The main clock is compared with
RCSYS, and if no rising edge of the main clock is detected during one RCSYS period, the clock
is considered to have failed.
The detector is enabled by writing a one to the Clock Failure Detection Enable bit in the Clock
Failure Detector Control Register (CFDCTRL.CFDEN). As soon as the detector is enabled, the
clock failure detector will monitor the divided main clock. Note that the detector does not monitor
the main clock if RCSYS is the source of the main clock, or if the main clock is temporarily not
available (startup-time after a wake-up, switching timing etc.), or in sleep mode where the main
clock is driven by the RCSYS (Stop and DeepStop mode). When a clock failure is detected, the
main clock automatically switches to the RCSYS clock and the Clock Failure Detected (CFD)
interrupt is generated if enabled. The MCCTRL register is also changed by hardware to indicate
that the main clock comes from RCSYS.
12.6.8
Interrupts
The PM has a number of interrupt sources:
• AE - Access Error,
160
32099I–01/2012
AT32UC3L016/32/64
– A lock protected register is written to without first being unlocked.
• CKRDY - Clock Ready:
– New Clock Select settings in the CPUSEL/PBxSEL registers have taken effect. (A
zero-to-one transition on SR.CKRDY is detected).
• CFD - Clock Failure Detected:
– The system detects that the main clock is not running.
The Interrupt Status Register contains one bit for each interrupt source. A bit in this register is
set on a zero-to-one transition of the corresponding bit in the Status Register (SR), and cleared
by writing a one to the corresponding bit in the Interrupt Clear Register (ICR). The interrupt
sources will generate an interrupt request if the corresponding bit in the Interrupt Mask Register
is set. The interrupt sources are ORed together to form one interrupt request. The Power Manager will generate an interrupt request if at least one of the bits in the Interrupt Mask Register
(IMR) is set. Bits in IMR are set by writing a one to the corresponding bit in the Interrupt Enable
Register (IER), and cleared by writing a one to the corresponding bit in the Interrupt Disable
Register (IDR). The interrupt request remains active until the corresponding bit in the Interrupt
Status Register (ISR) is cleared by writing a one to the corresponding bit in the Interrupt Clear
Register (ICR). Because all the interrupt sources are ORed together, the interrupt request from
the Power Manager will remain active until all the bits in ISR are cleared.
161
32099I–01/2012
AT32UC3L016/32/64
12.7
User Interface
Table 12-7.
PM Register Memory Map
Offset
Register
Register Name
Access
Reset
0x000
Main Clock Control
MCCTRL
Read/Write
0x00000000
0x004
CPU Clock Select
CPUSEL
Read/Write
0x00000000
0x008
HSB Clock Select
HSBSEL
Read-only
0x00000000
0x00C
PBA Clock Select
PBASEL
Read/Write
0x00000000
0x010
PBB Clock Select
PBBSEL
Read/Write
0x00000000
0x014 - 0x01C
Reserved
0x020
CPU Mask
CPUMASK
Read/Write
0x00010003
0x024
HSB Mask
HSBMASK
Read/Write
0x0000007F
0x028
PBA Mask
PBAMASK
Read/Write
0x03FFFFFF
0x02C
PBB Mask
PBBMASK
Read/Write
0x00000007
0x030- 0x03C
Reserved
0x040
PBA Divided Mask
PBADIVMASK
Read/Write
0x0000007F
0x044 - 0x050
Reserved
0x054
Clock Failure Detector Control
CFDCTRL
Read/Write
0x00000000
0x058
Unlock Register
UNLOCK
Write-only
0x00000000
0x05C - 0x0BC
Reserved
0x0C0
Interrupt Enable Register
IER
Write-only
0x00000000
0x0C4
Interrupt Disable Register
IDR
Write-only
0x00000000
0x0C8
Interrupt Mask Register
IMR
Read-only
0x00000000
0x0CC
Interrupt Status Register
ISR
Read-only
0x00000000
0x0D0
Interrupt Clear Register
ICR
Write-only
0x00000000
0x0D4
Status Register
SR
Read-only
0x00000020
0x0D8 - 0x15C
Reserved
0x160
Peripheral Power Control Register
PPCR
Read/Write
0x000001FA
0x164 - 0x17C
Reserved
0x180
Reset Cause Register
RCAUSE
Read-only
-(2)
0x184
Wake Cause Register
WCAUSE
Read-only
-(3)
0x188
Asynchronous Wake Up Enable Register
AWEN
Read/Write
0x00000000
0x18C - 0x3F4
Reserved
0x3F8
Configuration Register
CONFIG
Read-only
0x00000043
0x3FC
Version Register
VERSION
Read-only
-(1)
Note:
1. The reset value is device specific. Please refer to the Module Configuration section at the end of this chapter.
2. Latest Reset Source.
3. Latest Wake Source.
162
32099I–01/2012
AT32UC3L016/32/64
12.7.1
Main Clock Control
Name:
MCCTRL
Access Type:
Read/Write
Offset:
0x000
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
-
-
-
-
-
MCSEL
• MCSEL: Main Clock Select
Table 12-8.
Note:
Main clocks in AT32UC3L016/32/64
MCSEL[2:0]
Main clock source
0
System RC oscillator (RCSYS)
1
Oscillator0 (OSC0)
2
DFLL
3
120MHz RC oscillator
(RC120M)(1)
1. If the 120MHz RC oscillator is selected as main clock source, it must be divided by at least 4 before being used as clock
source for the CPU. This division is selected by writing to the CPUSEL and CPUDIV bits in the CPUSEL register, before
switching to RC120M as main clock source.
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
163
32099I–01/2012
AT32UC3L016/32/64
12.7.2
CPU Clock Select
Name:
CPUSEL
Access Type:
Read/Write
Offset:
0x004
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
CPUDIV
-
-
-
-
CPUSEL
• 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 CPUDIV is written to 0, CPUSEL should also be written to 0 to ensure correct operation.
Also note that writing this register clears SR.CKRDY. The register must not be re-written until CKRDY is set.
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
164
32099I–01/2012
AT32UC3L016/32/64
12.7.3
HSB Clock Select
Name:
HSBSEL
Access Type:
Read
Offset:
0x008
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
HSBDIV
-
-
-
-
HSBSEL
This register is read-only and its content is always equal to CPUSEL.
165
32099I–01/2012
AT32UC3L016/32/64
12.7.4
PBx Clock Select
Name:
PBxSEL
Access Type:
Read/Write
Offset:
0x00C-0x010
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
PBDIV
-
-
-
-
PBSEL
• PBDIV, PBSEL: PBx Division and Clock Select
PBDIV = 0: PBx clock equals main clock.
PBDIV = 1: PBx clock equals main clock divided by 2(PBSEL+1).
Note that if PBDIV is written to 0, PBSEL should also be written to 0 to ensure correct operation.
Also note that writing this register clears SR.CKRDY. The register must not be re-written until SR.CKRDY is set.
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
166
32099I–01/2012
AT32UC3L016/32/64
12.7.5
Clock Mask
Name:
CPUMASK/HSBMASK/PBAMASK/PBBMASK
Access Type:
Read/Write
Offset:
0x020-0x02C
Reset Value:
-
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 12-9.
Table 12-9.
Maskable Module Clocks in AT32UC3L016/32/64.
Bit
CPUMASK
HSBMASK
PBAMASK
PBBMASK
0
OCD
PDCA
PDCA
FLASHCDW
1
Reserved
FLASHCDW
INTC
HMATRIX
2
-
SAU
PM
SAU
3
-
PBB bridge
SCIF
-
4
-
PBA bridge
AST
-
5
-
Peripheral Event System
WDT
-
6
-
-
EIC
-
7
-
-
FREQM
-
8
-
-
GPIO
-
9
-
-
USART0
-
10
-
-
USART1
-
11
-
-
USART2
-
167
32099I–01/2012
AT32UC3L016/32/64
Table 12-9.
Maskable Module Clocks in AT32UC3L016/32/64.
Bit
CPUMASK
HSBMASK
PBAMASK
PBBMASK
12
-
-
USART3
-
13
-
-
SPI
-
14
-
-
TWIM0
-
15
-
-
TWIM1
-
16
SYSTIMER
-
TWIS0
-
17
-
-
TWIS1
-
18
-
-
PWMA
-
19
-
-
TC0
-
20
-
-
TC1
-
21
-
-
ADCIFB
-
22
-
-
ACIFB
-
23
-
-
CAT
-
24
-
-
GLOC
-
25
-
-
AW
-
31:26
-
-
-
-
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
168
32099I–01/2012
AT32UC3L016/32/64
12.7.6
PBA Divided Mask
Name:
PBADIVMASK
Access Type:
Read/Write
Offset:
0x040
Reset Value:
0x0000007F
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
-
MASK[6:0]
• MASK: Clock Mask
If bit n is written to zero, the clock divided by 2(n+1) is stopped. If bit n is written to one, the clock divided by 2(n+1) is enabled
according to the current power mode. Table 12-10 shows what clocks are affected by the different MASK bits.
Table 12-10. Divided Clock Mask
Bit
USART0
USART1
USART2
USART3
TC0
TC1
0
-
-
-
-
TIMER_CLOCK2
TIMER_CLOCK2
1
-
-
-
-
-
-
2
CLK_USART/
DIV
CLK_USART/
DIV
CLK_USART/
DIV
CLK_USART/
DIV
TIMER_CLOCK3
TIMER_CLOCK3
3
-
-
-
-
-
-
4
-
-
-
-
TIMER_CLOCK4
TIMER_CLOCK4
5
-
-
-
-
-
-
6
-
-
-
-
TIMER_CLOCK5
TIMER_CLOCK5
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
169
32099I–01/2012
AT32UC3L016/32/64
12.7.7
Clock Failure Detector Control Register
Name:
CFDCTRL
Access Type:
Read/Write
Offset:
0x054
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
SFV
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
CFDEN
• SFV: Store Final Value
0: The register is read/write
1: The register is read-only, to protect against further accidental writes.
• CFDEN: Clock Failure Detection Enable
0: Clock Failure Detector is disabled
1: Clock Failure Detector is enabled
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
170
32099I–01/2012
AT32UC3L016/32/64
12.7.8
Unlock Register
Name:
UNLOCK
Access Type:
Write-only
Offset:
0x058
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
KEY
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
7
6
5
4
3
2
ADDR[9:8]
1
0
ADDR[7:0]
To unlock a write protected register, first write to the UNLOCK register with the address of the register to unlock in the
ADDR field and 0xAA in the KEY field. Then, in the next PB access write to the register specified in the ADDR field.
• KEY: Unlock Key
Write this bit field to 0xAA to enable unlock.
• ADDR: Unlock Address
Write the address of the register to unlock to this field.
171
32099I–01/2012
AT32UC3L016/32/64
12.7.9
Interrupt Enable Register
Name:
IER
Access Type:
Write-only
Offset:
0x0C0
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
AE
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
CKRDY
-
-
-
-
CFD
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
172
32099I–01/2012
AT32UC3L016/32/64
12.7.10
Interrupt Disable Register
Name:
IDR
Access Type:
Write-only
Offset:
0x0C4
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
AE
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
CKRDY
-
-
-
-
CFD
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
173
32099I–01/2012
AT32UC3L016/32/64
12.7.11
Interrupt Mask Register
Name:
IMR
Access Type:
Read-only
Offset:
0x0C8
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
AE
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
CKRDY
-
-
-
-
CFD
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
This bit is cleared when the corresponding bit in IDR is written to one.
This bit is set when the corresponding bit in IER is written to one.
174
32099I–01/2012
AT32UC3L016/32/64
12.7.12
Interrupt Status Register
Name:
ISR
Access Type:
Read-only
Offset:
0x0CC
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
AE
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
CKRDY
-
-
-
-
CFD
0: The corresponding interrupt is cleared.
1: The corresponding interrupt is pending.
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set on a zero-to-one transition of the corresponding bit in the Status Register (SR).
175
32099I–01/2012
AT32UC3L016/32/64
12.7.13
Interrupt Clear Register
Name:
ICR
Access Type:
Write-only
Offset:
0x0D0
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
AE
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
CKRDY
-
-
-
-
CFD
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in ISR.
176
32099I–01/2012
AT32UC3L016/32/64
12.7.14
Status Register
Name:
SR
Access Type:
Read-only
Offset:
0x0D4
Reset Value:
0x00000020
31
30
29
28
27
26
25
24
AE
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
CKRDY
-
-
-
-
CFD
• AE: Access Error
0: No access error has occurred.
1: A write to lock protected register without unlocking it has occurred.
• CKRDY: Clock Ready
0: One of the CPUSEL/PBxSEL registers has been written, and the new clock setting is not yet effective.
1: The synchronous clocks have frequencies as indicated in the CPUSEL/PBxSEL registers.
• CFD: Clock Failure Detected
0: Main clock is running correctly.
1: Failure on main clock detected. Main clock is now running on RCSYS.
177
32099I–01/2012
AT32UC3L016/32/64
12.7.15
Peripheral Power Control Register
Name:
PPCR
Access Type:
Read/Write
Offset:
0x004
Reset Value:
0x000001FA
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
PPC[31:24]
23
22
21
20
PPC[23:16]
15
14
13
12
PPC[15:8]
7
6
5
4
PPC[7:0]
Table 12-11. Peripheral Power Control
Bit
Name
0
RSTPUN
1
FRC32
2
RSTTM
3
CATRCMASK
4
ACIFBCRCMASK
5
ADCIFBRCMASK
6
ASTRCMASK
7
TWIS0RCMASK
8
TWIS1RCMASK
31:9
-
• RSTTM: Reset test mode
0: External reset not in test mode
1: External reset in test mode
• FRC32: Force RC32 out
0: RC32 signal is not forced as output
1: RC32 signal is forced as output
• RSTPUN: Reset Pullup, active low
0: Pull-up for external reset on
1: Pull-up for external reset off
178
32099I–01/2012
AT32UC3L016/32/64
• CATRCMASK: CAT Request Clock Mask
0: CAT Request Clock is disabled
1: CAT Request Clock is enabled
• ACIFBRCMASK: ACIFB Request Clock Mask
0: ACIFB Request Clock is disabled
1: ACIFB Request Clock is enabled
• ADCIFBRCMASK: ADCIFB Request Clock Mask
0: ADCIFB Request Clock is disabled
1: ADCIFB Request Clock is enabled
• ASTRCMASK: AST Request Clock Mask
0: AST Request Clock is disabled
1: AST Request Clock is enabled
• TWIS0RCMASK: TWIS0 Request Clock Mask
0: TWIS0 Request Clock is disabled
1: TWIS0 Request Clock is enabled
• TWIS1RCMASK: TWIS1 Request Clock Mask
0: TWIS1 Request Clock is disabled
1: TWIS1 Request Clock is enabled
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
179
32099I–01/2012
AT32UC3L016/32/64
12.7.16
Reset Cause Register
Name:
RCAUSE
Access Type:
Read-only
Offset:
0x180
Reset Value:
Latest Reset Source
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
AWIRE
-
-
JTAG
OCDRST
7
6
5
4
3
2
1
0
-
SLEEP
-
-
WDT
EXT
BOD
POR
• AWIRE: aWire Reset
This bit is set when the last reset was caused by the aWire.
• JTAG: JTAG Reset
This bit is set when the last reset was caused by the JTAG.
• OCDRST: OCD Reset
This bit is set when the last reset was due to the RES bit in the OCD Development Control register having been written to one.
• SLEEP: Sleep Reset
This bit is set when the last reset was due to the device waking up from the Shutdown sleep mode.
• WDT: Watchdog Reset
This bit is set when the last reset was due to a watchdog time-out.
• EXT: External Reset Pin
This bit is set when the last reset was due to the RESET_N pin being pulled low.
• BOD: Brown-out Reset
This bit is set when the last reset was due to the core supply voltage being lower than the brown-out threshold level.
• POR: Power-on Reset
This bit is set when the last reset was due to the core supply voltage VDDCORE being lower than the power-on threshold level
(the reset is generated by the POR18 detector), or the internal regulator supply voltage being lower than the regulator power-on
threshold level (generated by the POR33 detector), or the internal regulator supply voltage being lower than the minimum
required input voltage (generated by the 3.3V supply monitor SM33).
180
32099I–01/2012
AT32UC3L016/32/64
12.7.17
Wake Cause Register
Name:
WCAUSE
Access Type:
Read-only
Offset:
0x184
Reset Value:
Latest Wake Source
31
30
29
28
27
26
25
24
18
17
16
10
9
8
2
1
0
WCAUSE[31:24]
23
22
21
20
19
WCAUSE[23:16]
15
14
13
12
11
WCAUSE[15:8]
7
6
5
4
3
WCAUSE[7:0]
A bit in this register is set on wake up caused by the peripheral referred to in Table 12-12 on page 181.
Table 12-12. Wake Cause
Bit
Wake Cause
0
CAT
1
ACIFB
2
ADCIFB
3
TWI Slave 0
4
TWI Slave 1
5
WAKE_N
6
ADCIFB Pen Detect
15:7
-
16
EIC
17
AST
31:18
-
181
32099I–01/2012
AT32UC3L016/32/64
12.7.18
Asynchronous Wake Up Enable Register
Name:
AWEN
Access Type:
Read/Write
Offset:
0x188
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
AWEN[31:24]
23
22
21
20
AWEN[23:16]
15
14
13
12
AWEN[15:8]
7
6
5
4
AWEN[7:0]
Each bit in this register corresponds to an asynchronous wake-up source, according to Table 12-13 on page 182.
0: The corresponding wake-up source is disabled.
1: The corresponding wake-up source is enabled.
Table 12-13. Asynchronous Wake-up Sources
Bit
Asynchronous Wake-up Source
0
CAT
1
ACIFB
2
ADCIFB
3
TWIS0
4
TWIS1
5
WAKEN
6
ADCIFBPD
31:7
-
182
32099I–01/2012
AT32UC3L016/32/64
12.7.19
Configuration Register
Name:
CONFIG
Access Type:
Read-Only
Offset:
0x3F8
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
HSBPEVC
-
-
-
PBD
PBC
PBB
PBA
This register shows the configuration of the PM.
• HSBPEVC:HSB PEVC Clock Implemented
0: HSBPEVC not implemented.
1: HSBPEVC implemented.
• PBD: PBD Implemented
0: PBD not implemented.
1: PBD implemented.
• PBC: PBC Implemented
0: PBC not implemented.
1: PBC implemented.
• PBB: PBB Implemented
0: PBB not implemented.
1: PBB implemented.
• PBA: PBA Implemented
0: PBA not implemented.
1: PBA implemented.
183
32099I–01/2012
AT32UC3L016/32/64
12.7.20
Version Register
Name:
VERSION
Access Type:
Read-Only
Offset:
0x3FC
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the module. No functionality associated.
184
32099I–01/2012
AT32UC3L016/32/64
12.8
Module Configuration
The specific configuration for each PM instance is listed in the following tables. The module bus
clocks listed here are connected to the system bus clocks. Please refer to the “Synchronous
Clocks”, “Peripheral Clock Masking” and “Sleep Modes” sections for details.
Table 12-14. Power Manager Clocks
Clock Name
Description
CLK_PM
Clock for the PM bus interface
Table 12-15. Register Reset Values
Register
Reset Value
VERSION
0x00000411
Table 12-16. Effect of the Different Reset Events
Power-On
Reset
External
Reset
Watchdog
Reset
BOD
Reset
SM33
Reset
CPU Error
Reset
OCD
Reset
JTAG
Reset
CPU/HSB/PBx
(excluding Power Manager)
Y
Y
Y
Y
Y
Y
Y
Y
32KHz oscillator
Y
N
N
N
N
N
N
N
RC Oscillator Calibration register
Y
N
N
N
N
N
N
N
Other oscillator control registers
Y
Y
Y
Y
Y
Y
Y
Y
AST register, except interrupt
registers
Y
N
N
N
N
N
N
N
Watchdog control register
Y
Y
N
Y
Y
Y
Y
Y
Voltage Calibration register
Y
N
N
N
N
N
N
N
SM33 control register
Y
Y
Y
Y
Y
Y
Y
Y
BOD control register
Y
Y
Y
N
Y
Y
Y
Y
Clock control registers
Y
Y
Y
Y
Y
Y
Y
Y
OCD system and OCD registers
Y
Y
N
Y
Y
Y
N
Y
185
32099I–01/2012
AT32UC3L016/32/64
13. System Control Interface (SCIF)
Rev: 1.0.2.2
13.1
Features
•
•
•
•
•
•
•
•
•
•
•
•
•
13.2
Supports crystal oscillator 0.45-16MHz (OSC0)
Supports Digital Frequency Locked Loop 40-150MHz (DFLL)
Supports 32KHz ultra-low-power oscillator (OSC32K)
Supports 32kHz RC oscillator (RC32K)
Integrated low-power RC oscillator (RCSYS)
Generic clocks (GCLK) with wide frequency range provided
Controls Brown-out detectors (BOD) and supply monitors
Controls Bandgap
Controls Voltage Regulator (VREG) behavior and calibration
Controls Temperature Sensor
Controls Supply Monitor 33 (SM33) operating modes and calibration
Controls 120MHz integrated RC Oscillator (RC120M)
Four 32-bit general-purpose backup registers
Overview
The System Control Interface (SCIF) controls the oscillators, Generic Clocks, BODs, Bandgap,
VREG, Temperature Sensor, and Backup Registers.
13.3
I/O Lines Description
Table 13-1.
13.4
I/O Lines Description
Pin Name
Pin Description
Type
RC32OUT
RC32 output at start-up
Output
XIN0
Crystal 0 input
Analog/Digital
XIN32
Crystal 32 input (primary location)
Analog/Digital
XIN32_2
Crystal 32 input (secondary location)
Analog/Digital
XOUT0
Crystal 0 output
Analog
XOUT32
Crystal 32 output (primary location)
Analog
XOUT32_2
Crystal 32 output (secondary location)
Analog
GCLK4-GCLK0
Generic clock output
Output
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
13.4.1
I/O Lines
The SCIF 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 SCIF are not used by the application,
they can be used for other purposes by the GPIO controller. Oscillator pins are also multiplexed
186
32099I–01/2012
AT32UC3L016/32/64
with GPIO. When oscillators are used, the related pins are controlled directly by the SCIF, overriding the GPIO settings.
RC32OUT will be output after reset, and the GPIO controller can assign this pin to other peripheral function after start-up.
13.4.2
Power Management
The BOD and the oscillators, except the 32 KHz oscillator (OSC32K), are turned off in some
sleep modes and turned automatically on when the device wakes up. The Voltage Regulator is
set in low power mode in some sleep modes and automatically set back in normal mode when
the device wakes up. Please refer to the Power Manager chapter for details.
The BOD control registers will not be reset by the Power Manager on a BOD reset.
13.4.3
Clocks
The SCIF controls all oscillators in the device. The oscillators can be used as source for the CPU
and peripherals. Selection of source is done in the Power Manager. The oscillators can also be
used as source for generic clocks.
13.4.4
Interrupts
The SCIF interrupt request line is connected to the interrupt controller. Using the SCIF interrupt
requires the interrupt controller to be programmed first.
13.4.5
13.5
13.5.1
Debug Operation
The SCIF does not interact with debug operations.
Functional Description
Oscillator (OSC) Operation
Rev: 1.0.0.0
The main oscillator (OSCn) is designed to be used with an external 0.450 to 16MHz crystal and
two biasing capacitors, as shown in the Electrical Characteristics chapter, or with an external
clock connected to the XIN. The oscillator can be used as source for the main clock in the
device, as described in the Power Manager chapter. The oscillator can be used as source for the
generic clocks, as described in the Generic Clocks section.
The oscillator is disabled by default after reset. When the oscillator is disabled, the XIN and
XOUT pins can be used as general purpose I/Os. When the oscillator is enabled, the XIN and
XOUT pins are controlled directly by the SCIF, overriding GPIO settings. When the oscillator is
configured to use an external clock, the clock must be applied to the XIN pin while the XOUT pin
can be used as general purpose I/O.
The oscillator is enabled by writing a one to the Oscillator Enable bit in the Oscillator Control register (OSCCTRLn.OSCEN). Operation mode (external clock or crystal) is selected by writing to
the Oscillator Mode bit in OSCCTRLn (OSCCTRLn.MODE). The oscillator is automatically disabled in certain sleep modes to reduce power consumption, as described in the Power Manager
chapter.
After a hard reset, or when waking up from a sleep mode where the oscillators were disabled,
the oscillator will need a certain amount of time to stabilize on the correct frequency. This startup time can be set in the OSCCTRLn register.
187
32099I–01/2012
AT32UC3L016/32/64
The SCIF masks the oscillator outputs during the start-up time, to ensure that no unstable clocks
propagate to the digital logic.
The OSCn Ready bit in the Power and Clock Status Register (PCLKSR.OSCnRDY) is set when
the oscillator is stable and ready to be used as clock source. An interrupt can be generated on a
zero-to-one transition on OSCnRDY if the OSCnRDY bit in the Interrupt Mask Register
(IMR.OSCnRDY) is set. This bit is set by writing a one to the corresponding bit in the Interrupt
Enable Register (IER.OSCnRDY).
13.5.2
32KHz Oscillator (OSC32K) Operation
Rev: 1.0.1.1
The 32KHz oscillator operates as described for the oscillator above. The 32KHz oscillator can
be used as source clock for the Asynchronous Timer (AST) and the Watchdog Timer (WDT).
The 32KHz oscillator can also be used as source for the generic clocks.
The oscillator is disabled by default after reset. When the oscillator is disabled, the XIN32 and
XOUT32 pins can be used as general-purpose I/Os. When the oscillator is enabled, the XIN32
and XOUT32 pins are controlled directly by the SCIF, overriding GPIO settings. When the oscillator is configured to use an external clock, the clock must be applied to the XIN32 pin while the
XOUT32 pin can be used as general-purpose I/O.
The oscillator is enabled writing a one to the OSC32 Enable bit in the 32KHz Oscillator Control
Register (OSCCTRL32OSC32EN). The oscillator is disabled by writing a zero to the OSC32EN
bit, while keeping the other bits unchanged. Writing to OSC32EN while also writing to other bits
may result in unpredictable behavior. Operation mode (external clock or crystal) is selected by
writing to the Oscillator Mode bit in OSCCTRL32 (OSCCTRL32.MODE). The oscillator is an
ultra-low-power design and remains enabled in all sleep modes.
The start-up time of the 32KHz oscillator is selected by writing to the Oscillator Start-up Time
field in the OSCCTRL32 register (OSCCTRL32.STARTUP). The SCIF masks the oscillator output during the start-up time, to ensure that no unstable clock cycles propagate to the digital logic.
The OSC32 Ready bit in the Power and Clock Status Register (PCLKSR.OSC32RDY) is set
when the oscillator is stable and ready to be used as clock source. An interrupt can be generated on a zero-to-one transition on PCLKSR.OSC32RDY if the OSC32RDY bit in the Interrupt
Mask Register (IMR.OSC32RDY) is set. This bit is set by writing a one to the corresponding bit
in the Interrupt Enable Register (IER.OSC32RDY).
.As a crystal oscillator usually requires a very long start-up time (up to 1 second), the 32 KHz
oscillator will keep running across resets, except a Power-on Reset (POR).
The 32KHz oscillator also has a 1KHz output. This is enabled by writing a one to the Enable
1KHz output bit in OSCCTRL32 register (OSCCTRL32.EN1K). If the 32KHz output clock is not
needed when 1K is enabled, this can be disabled by writing a zero to the Enable 32KHz output
bit in the OSCCTRL32 register (OSCCTRL32.EN32K). OSCCTRL32.EN32K is set after a POR.
The 32 KHz oscillator has two possible sets of pins. To select between them write to the Pin
Select bit in the OSCCTRL32 register (OSCCTRL32.PINSEL). If the 32KHz oscillator is to be
used in Shutdown mode, PINSEL must be written to one, and XIN32_2 and XOUT32_2 must be
used.
188
32099I–01/2012
AT32UC3L016/32/64
13.5.3
Digital Frequency Locked Loop (DFLL) Operation
Rev: 2.0.1.1
The DFLL is controlled by the Digital Frequency Locked Loop Interface (DFLLIF). The DFLL is
disabled by default, but can be enabled to provide a high-frequency source clock for synchronous and generic clocks.
Features:
• Internal oscillator with no external components
• 40-150MHz frequency in closed loop mode
• Can operate standalone as a high-frequency programmable oscillator in open loop mode
• Can operate as an accurate frequency multiplier against a known frequency in closed loop
mode
• Optional spread-spectrum clock generation
• Very high-frequency multiplication supported - can generate all frequencies from a 32KHz
clock
The DFLL can operate in both open loop mode and closed loop mode. In closed loop mode a
low frequency clock with high accuracy can be used as reference clock to get high accuracy on
the output clock (CLK_DFLL).
To prevent unexpected writes due to software bugs, write access to the configuration registers is
protected by a locking mechanism. For details please refer to the UNLOCK register description.
Figure 13-1. DFLLIF Block Diagram
COARSE
8
DFLL
CLK_DFLL
FINE
9
8+9
32
CSTEP
FSTEP
IMUL
FMUL
FREQUENCY
TUNER
CLK_DFLLIF_DITHER
CLK_DFLLIF_REF
DFLLLOCKF
DFLLLOCKLOSTF
DFLLLOCKC
DFLLLOCKA
DFLLLOCKLOSTC
DFLLLOCKLOSTA
13.5.3.1
Enabling the DFLL
The DFLL is enabled by writing a one to the Enable bit (EN) in the DFLLn Configuration Register
(DFLLnCONF). No other bits or fields in DFLLnCONF must be changed simultaneously, or
before the DFLL is enabled.
189
32099I–01/2012
AT32UC3L016/32/64
13.5.3.2
Internal synchronization
Due to multiple clock domains in the DFLLIF, values in the DFLLIF configuration registers need
to be synchronized to other clock domains. The status of this synchronization can be read from
the Power and Clocks Status Register (PCLKSR). Before writing to a DFLLIF configuration register, the user must check that the DFLLn Synchronization Ready bit (DFLLnRDY) in PCLKSR is
set. When this bit is set, the DFLL can be configured, and CLK_DFLL is ready to be used. Any
write to a DFLLIF configuration register while DFLLnRDY is cleared will be ignored.
Before reading the value in any of the DFLL configuration registers a one must be written to the
Synchronization bit (SYNC) in the DFLLn Synchronization Register (DFLLnSYNC). The DFLL
configuration registers are ready to be read when PCLKSR.DFLLnRDY is set.
13.5.3.3
Disabling the DFLL
The DFLL is disabled by writing a zero to DFLLnCONF.EN. No other bits or fields in DFLLnCONF must be changed simultaneously.
After disabling the DFLL, PCLKSR.DFLLnRDY will not be set. It is not required to wait for
PCLKSR.DFLLnRDY to be set before re-enabling the DFLL.
13.5.3.4
Open loop operation
After enabling the DFLL, open loop mode is selected by writing a zero to the Mode Selection bit
(MODE) in DFLLnCONF. When operating in open loop mode the output frequency of the DFLL
will be determined by the values written to the Coarse Calibration Value field (COARSE) and the
Fine Calibration Value field (FINE) in the DFLLnCONF register. When writing to COARSE and
FINE, be aware that the output frequency must not exceed the maximum frequency of the
device after the division in the clock generator. It is possible to change the value of COARSE
and FINE, and thereby the output frequency of the DFLL, while the DFLL is enabled and in use.
The DFLL clock is ready to be used when PCLKSR.DFLLnRDY is cleared after enabling the
DFLL.
The frequency range in open loop mode is 40-150MHz, but maximum frequency can be higher,
and the minimum frequency can be lower. The best way to start the DFLL at a specific frequency
in open loop mode is to first configure it for closed loop mode, see Section 13.5.3.5. When a lock
is achieved, read back the COARSE and FINE values and switch to open loop mode using these
values. An alternative approach is to use the Frequency Meter (FREQM) to monitor the DFLL
frequency and adjust the COARSE and FINE values based on measurement results form the
FREQM. Please refer to the FREQM chapter for more information on how to use it. Note that the
output frequency of the DFLL will drift when in open loop mode due to temperature and voltage
changes. Please refer to the Electrical Characteristics chapter for details.
13.5.3.5
Closed loop operation
The DFLL must be correctly configured before closed loop operation can be enabled. After
enabling the DFLL, enable and select a reference clock (CLK_DFLLIF_REF).
CLK_DFLLIF_REF is a generic clock, please refer to Generic Clocks section for details. Then
set the maximum step size allowed in finding the COARSE and FINE values by setting the
Coarse Maximum Step field (CSTEP) and Fine Maximum Step field (FSTEP) in the DFLLn Maximum Step Register (DFLLnSTEP). A small step size will ensure low overshoot on the output
frequency, but can typically result in longer lock times. A high value might give a big overshoot,
but can typically give faster locking. DFLLnSTEP.CSTEP and DFLLnSTEP.FSTEP must be
lower than 50% of the maximum value of DFLLnCONF.COARSE and DFLLnCONF.FINE
respectively. Then select the multiplication factor in the Integer Multiply Factor field (IMUL) and
190
32099I–01/2012
AT32UC3L016/32/64
the Fractional Multiply field (FMUL) in the DFLLn Multiplier Register (DFLLnMUL). Care must be
taken when choosing IMUL and FMUL so the output frequency does not exceed the maximum
frequency of the device. Start the closed loop mode by writing a one to DFLLnCONF.MODE bit.
The frequency of CLK_DFLL (fDFLL) is given by:
FMUL⎞
f DFLL = ⎛ IMUL + ---------------- f
16 ⎠ REF
⎝
2
where fREF is the frequency of CLK_DFLLIF_REF. COARSE and FINE in DFLLnCONF are readonly in closed loop mode, and are controlled by the DFLLIF to meet user specified frequency.
The values in COARSE when the closed loop mode is enabled is used by the frequency tuner as
a starting point for COARSE. Setting COARSE to a value close to the final value will reduce the
time needed to get a lock on COARSE.
Frequency locking
The locking of the frequency in closed loop mode is divided into three stages. In the COARSE
stage the control logic quickly finds the correct value for DFLLnCONF.COARSE and thereby
sets the output frequency to a value close to the correct frequency. The DFLLn Locked on
Coarse Value bit (DFLLnLOCKC) in PCLKSR will be set when this is done. In the FINE stage the
control logic tunes the value in DFLLnCONF.FINE so the output frequency will be very close to
the desired frequency. DFLLn Locked on Fine Value bit (DFLLnLOCKF) in PCLKSR will be set
when this is done. In the ACCURATE stage the DFLL frequency tuning mechanism uses dithering on the FINE bits to obtain an accurate average output frequency. DFLLn Locked on Accurate
Value bit (DFLLnLOCKA) in PCLKSR will be set when this is done. The ACCURATE stage will
only be executed if the Dithering Enable bit (DITHER) in DFLLnCONF has been written to a one.
If DITHER is written to a zero DFLLnLOCKA will never occur. If dithering is enabled, the frequency of the dithering is decided by a generic clock (CLK_DFLLIF_DITHER). This clock has to
be set up correctly before enabling dithering. Please refer to the Generic Clocks section for
details.
Figure 13-2. DFLL Closed loop State Diagram
Measure
fDFLLn
DFLLnLOCKC
1
DFLLnLOCKF
1
DITHER
1
DFLLnLOCKA
0
0
0
0
Calculate
new
COARSE
value
Calculate
new FINE
value
Compensate for
drift
Calculate
new
dithering
dutycycle
1
Compensate for
drift
When dithering is enabled the accuracy of the average output frequency of the DFLL will be
higher. However, the actual frequency will be alternating between two frequencies. If a fixed frequency is required, the dithering should not be enabled.
191
32099I–01/2012
AT32UC3L016/32/64
Figure 13-3. DFLL Locking in Closed loop
Target
frequency
Initial
frequency
DFLLnLOCKC DFLLnLOCKF DFLLnLOCKA
CLK_DFLL is ready to be used when the DFLLn Synchronization Ready bit (DFLLnRDY) in
PCLKSR is set after enabling the DFLL. However, the accuracy of the output frequency depends
on which locks are set.
For lock times, please refer to the Electrical Characteristics chapter.
Drift compensation
The frequency tuner will automatically compensate for drift in the fDFLL without losing either of the
locks. If the FINE value overflows or underflows, which should normally not happen, but could
occur due to large drift in temperature and voltage, all locks will be lost, and the COARSE and
FINE values will be recalibrated as described earlier. If any lock is lost the corresponding bit in
PCLKSR will be set, DFLLn Lock Lost on Coarse Value bit (DFLLnLOCKLOSTC) for lock lost on
COARSE value, DFLLn Lock Lost on Fine Value bit (DFLLnLOCKLOSTF) for lock lost on FINE
value and DFLLn Lock Lost on Accurate Value bit (DFLLnLOCKLOSTA) for lock lost on ACCURATE value. The corresponding lock status bit will be cleared when the lock lost bit is set, and
vice versa.
Reference clock stop detection
If CLK_DFLLIF_REF stops or is running at a very slow frequency, the DFLLn Reference Clock
Stopped bit (DFLLnRCS) in PCLKSR will be set. Note that the detection of the clock stop will
take a long time. The DFLLIF operate as if it was in open loop mode if it detects that the reference clock has stopped. This means that the COARSE and FINE values will be kept constant
while PCLKSR.DFLLnRCS is set. Closed loop mode operation will automatically resume if the
CLK_DFLLIF_REF is restarted, and compensate for any drift during the time CLK_DFLLIF_REF
was stopped. No locks will be lost.
Frequency error measurement
The ratio between CLK_DFLLIF_REF and CLK_DFLL is measured automatically by the DFLLIF.
The difference between this ratio and DFLLnMUL is stored in the Multiplication Ratio Difference
field (RATIODIFF) in the DFLLn Ratio Register (DFLLnRATIO). The relative error on CLK_DFLL
compared to the target frequency can be calculated as follows:
192
32099I–01/2012
AT32UC3L016/32/64
RATIODIFF ⋅ f REF
error = -----------------------------------------------2 NUMREF ⋅ f DFLL
where 2 NUMREF is the number of reference clock cycles the DFLLIF is using for calculating the
ratio.
13.5.3.6
Spread Spectrum Generator (SSG)
When the DFLL is used as the main clock source for the device, the EMI radiated from the
device will be synchronous to fDFLL. To provide better Electromagnetic Compatibility (EMC) the
DFLLIF can provide a clock with the energy spread in the frequency domain. This is done by
adding or subtracting values from the FINE value. SSG is enabled by writing a one to the Enable
bit (EN) in the DFLLn Spread Spectrum Generator Control Register (DFLLnSSG).
A generic clock sets the rate at which the SSG changes the frequency of the DFLL clock to generate a spread spectrum (CLK_DFLLIF_DITHER). This is the same clock used by the dithering
mechanism. The frequency of this clock should be higher than fREF to ensure that the DFLLIF
can lock. Please refer to the Generic clocks section for details.
Optionally, the clock ticks can be qualified by a Pseudo Random Binary Sequence (PRBS) if the
PRBS bit in DFLLnSSG is one. This reduces the modulation effect of CLK_DFLLIF_DITHER frequency onto fDFLL.
The amplitude of the frequency variation can be selected by setting the SSG Amplitude field
(AMPLITUDE) in DFLLnSSG. If AMPLITUDE is zero the SSG will toggle on the LSB of the FINE
value. If AMPLITUDE is one the SSG will add the sequence {1,-1, 0} to FINE.
The step size of the SSG is selected by writing to the SSG Step Size field (STEPSIZE) in
DFLLnSSG. STEPSIZE equal to zero or one will result in a step size equal to one. If the step
size is set to n, the output value from the SSG will be incremented/decremented by n on every
tick of the source clock.
The Spread Spectrum Generator is available in both open and closed loop mode.
When spread spectrum is enabled in closed loop mode, and the AMPLITUDE is high, an overflow/underflow in FINE is more likely to occur.
Figure 13-4. Spread Spectrum Generator Block Diagram.
FINE
CLK_DFLLIF_DITHER
Pseudorandom
Binary Sequence
To DFLL
1
Spread Spectrum
Generator
9
0
PRBS
13.5.3.7
AMPLITUDE,
STEPSIZE
Wake from sleep modes
When waking up from a sleep mode where the DFLL has been turned off, and CLK_DFLL was
the main clock before going to sleep, the DFLL will be re-enabled and start running with the
same configuration as before it was stopped even if the reference clock is not available. The
193
32099I–01/2012
AT32UC3L016/32/64
locks will not be lost. When the reference clock has restarted, the FINE tracking will quickly compensate for any frequency drift during sleep.
13.5.3.8
Accuracy
There are mainly three factors that decide the accuracy of the fDFLL. These can be tuned to
obtain maximum accuracy when fine lock is achieved.
• FINE resolution: The frequency step between two FINE values. This is relatively smaller for
high output frequencies.
• Resolution of the measurement: If the resolution of the measured fDFLL is low, i.e. the ratio
between CLK_DFLL frequency and CLK_DFLLIF_REF is small, then the DFLLIF might lock
at a frequency that is lower than the targeted frequency. It is recommended to use a
reference clock frequency of 32 KHz or lower to avoid this issue for low target frequencies.
• The accuracy of the reference clock.
13.5.3.9
Interrupts
A interrupt can be generated on a zero-to-one transaction on DFLLnLOCKC, DFLLnLOCKF,
DFLLnLOCKA, DFLLnLOCKLOSTC, DFLLnLOCKLOSTF, DFLLnLOCKLOSTA, DFLLnRDY or
DFLLnRCS.
13.5.4
Brown-Out Detection (BOD)
Rev: 1.0.1.0
The Brown-Out Detector monitors the VDDCORE supply pin and compares the supply voltage
to the brown-out detection level.
The BOD is disabled by default, and is enabled by writing to the BOD Control field in the BOD
Control Register (BOD.CTRL). This field can also be updated by flash fuses.
The BOD is powered by VDDIO and will not be powered during Shutdown sleep mode.
To prevent unexpected writes to the BOD register due to software bugs, write access to this register is protected by a locking mechanism. For details please refer to the UNLOCK register
description.
To prevent further modifications by software, the content of the BOD register can be set as readonly by writing a one to the Store Final Value bit (BOD.SFV). When this bit is one, software can
not change the BOD register content. This bit is cleared after flash calibration and after a reset
except after a BOD reset.
The brown-out detection level is selected by writing to the BOD Level field in BOD
(BOD.LEVEL). Please refer to the Electrical Characteristics chapter for parametric details.
If the BOD is enabled (BOD.CTRL is one or two) and the supply voltage goes below the detection level, the Brown-Out Detection bit in the Power and Clocks Status Register
(PCLKSR.BODDET) is set. This bit is cleared when the supply voltage goes above the detection
level. An interrupt request will be generated on a zero-to-one transition on PCLKSR.BODDET if
the Brown-Out Detection bit in the Interrupt Mask Register (IMR.BODDET) is set. This bit is set
by writing a one to the corresponding bit in the Interrupt Enable Register (IER.BODDET).
If BOD.CTRL is one, a BOD reset will be generated when the supply voltage goes below the
detection level. If BOD.CTRL is two, the device will not be reset.
Writing a one to the BOD Hysteresis bit in BOD (BOD.HYST) will add a hysteresis on the BOD
detection level.
194
32099I–01/2012
AT32UC3L016/32/64
Note that the BOD must be disabled before changing BOD.LEVEL, to avoid spurious reset or
interrupt. After enabling the BOD, the BOD output will be masked during one half of a RCSYS
clock cycle and two main clocks cycles to avoid false results.
When the JTAG or aWire is enabled, the BOD reset and interrupt are masked.
The CTRL, HYST, and LEVEL fields in the BOD Control Register are loaded factory defined calibration values from flash fuses after a reset. If the Flash Calibration Done bit in the BOD Control
Register (BOD.FCD) is zero, the flash calibration will be redone after any reset, and the
BOD.FCD bit will be set before program execution starts in the CPU. If BOD.FCD is one, the
flash calibration is redone after any reset except for a BOD reset. The BOD.FCD bit is cleared
after a reset, except for a BOD reset. BOD.FCD is set when these fields have been updated after
a flash calibration. It is possible to override the values in the BOD.CTRL, BOD.HYST, and
BOD.LEVEL fields after reset by writing to the BOD Control Register. Please refer to the Fuse
Settings chapter for more details about BOD fuses and how to program the fuses.
Figure 13-5. BOD Block Diagram
VDDCORE
BOD Hyst
BOD
Level
Interrupt
Reset
Reset
SCIF
POWER MANAGER(PM)
13.5.5
Enable
BOD
Bod
Detected
POR18
INTC
Bandgap
Rev: 1.0.1.0
The flash memory, the BOD, and the Temperature Sensor need a stable voltage reference to
operate. This reference voltage is provided by an internal Bandgap voltage reference. This refer-
195
32099I–01/2012
AT32UC3L016/32/64
ence is automatically turned on at start-up and turned off during some sleep modes to save
power. The Bandgap reference is powered by the internal regulator supply voltage and will not
be powered during Shutdown sleep mode. Please refer to the Power Manager chapter for
details.
13.5.6
System RC Oscillator (RCSYS)
Rev: 1.0.1.0
The system RC oscillator has a startup time of three cycles, and is always available except in
some sleep modes. Please refer to the Power Manager chapter for details. The system RC oscillator operates at a nominal frequency of 115kHz, and is calibrated using the Calibration Value
field (CALIB) in the RC Oscillator Calibration Register (RCCR). After a Power-on Reset (POR),
the RCCR.CALIB field is loaded with a factory defined value stored in the Flash fuses. Please
refer to the Fuse setting chapter for more details about RCCR fuses and how to program the
fuses.
If the Flash Calibration Done (FCD) bit in the RCCR is zero at any reset, the flash calibration will
be redone and the RCCR.FCD bit will be set before program execution starts in the CPU. If the
RCCR.FCD is one, the flash calibration will only be redone after a Power-on Reset.
To prevent unexpected writes to RCCR due to software bugs, write access to this register is protected by a locking mechanism. For details please refer to the UNLOCK register description.
Although it is not recommended to override default factory settings, it is still possible to override
the default values by writing to RCCR.CALIB.
13.5.7
Voltage Regulator (VREG)
Rev: 1.0.1.0
The embedded voltage regulator can be used to provide the VDDCORE voltage from the internal regulator supply voltage. It is controlled by the Voltage Regulator Calibration Register
(VREGCR). The voltage regulator is enabled by default at start-up but can be disabled by software if an external voltage is provided on the VDDCORE pin. The VREGCR also contains bits to
control the POR33 detector.
13.5.7.1
Register protection
To prevent unexpected writes to VREGCR due to software bugs, write access to this register is
protected by a locking mechanism. For details please refer to the UNLOCK register description.
To prevent further modifications by software, the content of the VREGCR register can be set as
read-only by writing a one to the Store Final Value bit (VREGCR.SFV). Once this bit is set, software can not change the VREGCR content until a Power-on Reset (POR) is applied.
13.5.7.2
Controlling voltage regulator output
The voltage regulator is always enabled at start-up, i.e. after a POR or when waking up from
Shutdown mode. It can be disabled by software by writing a zero to the Enable bit
(VREGCR.EN). This bit is set after a POR. Because of internal synchronization, the voltage regulator is not immediately enabled or disabled. The actual state of the voltage regulator can be
read from the ON bit (VREGCR.ON).
The voltage regulator output level is controlled by the Select VDD field (SELVDD) in VREGCR.
The default value of this field corresponds to a regulator output voltage of 1.8V. Other values of
this field are not defined, and it is not recommended to change the value of this field.
196
32099I–01/2012
AT32UC3L016/32/64
The Voltage Regulator OK bit (VREGCR.VREGOK) bit indicates when the voltage regulator output has reached the voltage threshold level.
13.5.7.3
Factory calibration
After a Power-on Reset (POR) the VREGCR.CALIB field is loaded with a factory defined calibration value. This value is chosen so that the normal output voltage of the regulator after a powerup is 1.8V.
Although it is not recommended to override default factory settings, it is still possible to override
these default values by writing to VREGCR.CALIB.
If the Flash Calibration Done bit in VREGCR (VREGCR.FCD) is zero, the flash calibration will be
redone after any reset, and the VREGCR.FCD bit will be set before program execution starts in
the CPU. If VREGCR.FCD is one, the flash calibration will only be redone after a POR.
13.5.7.4
POR33 control
VREGCR includes control bits for the Power-on Reset 3.3V (POR33) detector that monitors the
internal regulator supply voltage. The POR33 detector is enabled by default but can be disabled
by software to reduce power consumption. The 3.3V Supply Monitor (SM33) can then be used to
monitor the regulator power supply.
The POR33 detector is disabled by writing a zero to the POR33 Enable bit
(VREGCR.POR33EN). Because of internal synchronisation, the POR33 detector is not immediately enabled or disabled. The actual state of the POR33 detector can be read from the POR33
Status bit (VREGCR.POR33STATUS).
The 32kHz RC oscillator (RC32K) must be enabled before disabling the POR33 detector. Once
the POR33 detector has been disabled, the RC32K oscillator can be disabled again.
To avoid spurious resets, it is mandatory to mask the Power-on Reset when enabling or disabling the POR33 detector. The Power-on Reset generated by the POR33 detector can be
ignored by writing a one to the POR33 Mask bit (VREGCR.POR33MASK). Because of internal
synchronization, the masking is not immediately effective, so software should wait for the
VREGCR.POR33MASK to read as a one before assuming the masking is effective.
13.5.8
3.3 V Supply Monitor (SM33)
Rev: 1.0.0.0
The 3.3V supply monitor is a specific voltage detector for the internal regulator supply voltage. It
will indicate if the internal regulator supply voltage is above the minimum required input voltage
threshold. The user can choose to generate either a Power-on Reset (POR) and an interrupt
request, or only an interrupt request, when the internal regulator supply voltage drops below this
threshold.
Please refer to the Electrical Characteristics chapter for parametric details.
13.5.8.1
Register protection
To prevent unexpected writes to SM33 register due to software bugs, write access to this register is protected by a locking mechanism. For details please refer to the UNLOCK register
description.
197
32099I–01/2012
AT32UC3L016/32/64
To prevent further modifications by software, the content of the register can be set as read-only
by writing a one to the Store Final Value bit (SM33.SFV). When this bit is one, software can not
change the SM33 register content until the device is reset.
13.5.8.2
Operating modes
The SM33 is disabled by default and is enabled by writing to the Supply Monitor Control field in
the SM33 control register (SM33.CTRL). The current state of the SM33 can be read from the
Supply Monitor On Indicator bit in SM33 (SM33.ONSM). Enabling the SM33 will disable the
POR33 detector.
The SM33 can operate in continuous mode or in sampling mode. In sampling mode, the SM33 is
periodically enabled for a short period of time, just enough to make a a measurement, and then
disabled for a longer time to reduce power consumption.
By default, the SM33 operates in sampling mode during DeepStop and Static mode and in continuous mode for other sleep modes. Sampling mode can also be forced during sleep modes
other than DeepStop and Static, and during normal operation, by writing a one to the Force
Sampling Mode bit in the SM33 register (SM33.FS).
The user can select the sampling frequency by writing to the Sampling Frequency field in SM33
(SM33.SAMPFREQ). The sampling mode uses the 32 kHz RC oscillator (RC32K) as clock
source. The 32kHz RC oscillator is automatically enabled when the SM33 operates in sampling
mode.
13.5.8.3
Interrupt and reset generation
If the SM33 is enabled (SM33.CTRL is one or two) and the regulator supply voltage drops below
the SM33 threshold, the SM33DET bit in the Power and Clocks Status Register
(PCLKSR.SM33DET) is set. This bit is cleared when the supply voltage goes above the threshold. An interrupt request is generated on a zer-to-one transition of PCLKSR.SM33DET if the
Supply Monitor 3.3V Detection bit in the Interrupt Mask Register (IMR.SM33DET) is set. This bit
is set by writing a one to the corresponding bit in the Interrupt Enable Register (IER.SM33DET).
If SM33.CTRL is one, a POR will be generated when the voltage drops below the threshold. If
SM33.CTRL is two, the device will not be reset.
13.5.8.4
Factory calibration
After a reset the SM33.CALIB field is loaded with a factory defined value. This value is chosen
so that the nominal threshold value is 1.75V. The flash calibration is redone after any reset, and
the Flash Calibration Done bit in SM33 (SM33.FCD) is set before program execution starts in the
CPU.
Although it is not recommended to override default factory settings, it is still possible to override
the default value by writing to SM33.CALIB
13.5.9
Temperature Sensor
Rev: 1.0.0.0
The Temperature Sensor is connected to an ADC channel, please refer to the ADC chapter for
details. It is enabled by writing one to the Enable bit (EN) in the Temperature Sensor Configuration Register (TSENS). The Temperature Sensor can not be calibrated.
Please refer to the Electrical Characteristics chapter for more details.
198
32099I–01/2012
AT32UC3L016/32/64
13.5.10
120MHz RC Oscillator (RC120M)
Rev: 1.0.1.0
The 120MHz RC Oscillator can be used as source for the main clock in the device, as described
in the Power Manager chapter. The oscillator can also be used as source for the generic clocks,
as described in Generic Clock section. The RC120M must be enabled before it is used as a
source clock. To enable the clock, the user must write a one to the Enable bit in the 120MHz RC
Oscillator Control Register (RC120MCR.EN), and read back the RC120MCR register until the
EN bit reads one. The clock is disabled by writing a zero to RC120MCR.EN. The EN bit must be
read back as zero before the RC120M is re-enabled. If not, undefined behavior may occur.
The oscillator is automatically disabled in certain sleep modes to reduce power consumption, as
described in the Power Manager chapter.
13.5.11
Backup Registers (BR)
Rev: 1.0.0.1
Four 32-bit backup registers are available to store values when the device is in Shutdown
mode. These registers will keep their content even when the VDDCORE supply and the internal
regulator supply voltage supplies are removed. The backup registers can be accessed by reading from and writing to the BR0, BR1, BR2, and BR3 registers.
After writing to one of the backup registers the user must wait until the Backup Register Interface
Ready bit in tne Power and Clocks Status Register (PCLKSR.BRIFARDY) is set before writing to
another backup register. Writes to the backup register while PCLKSR.BRIFARDY is zero will be
discarded. An interrupt can be generated on a zero-to-one transition on PCLKSR.BRIFARDY if
the BRIFARDY bit in the Interrupt Mask Register (IMR.BRIFARDY) is set. This bit is set by writing a one to the corresponding bit in the Interrupt Enable Register (IER.BRIFARDY).
After powering up the device the Backup Register Interface Valid bit in PCLKSR (PCLKSR.BRIFAVALID) is cleared, indicating that the content of the backup registers has not been written and
contains the reset value. After writing to one of the backup registers the PCLKSR.BRIFAVALID
bit is set. During writes to the backup registers (when BRIFARDY is zero) BRIFAVALID will be
zero. If a reset occurs when BRIFARDY is zero, BRIFAVALID will be cleared after the reset, indicating that the content of the backup registers is not valid. If BRIFARDY is one when a reset
occurs, BRIFAVALID will be one and the content is the same as before the reset.
The user must ensure that BRIFAVALID and BRIFARDY are both set before reading the backup
register values.
13.5.12
32kHz RC Oscillator (RC32K)
Rev: 1.0.0.0
The RC32K can be used as source for the generic clocks, as described in The Generic Clocks
section.
The 32 kHz RC oscillator (RC32K) is forced on after reset, and output on PA20. The clock is
available on the pad until the PPCR.FRC32 bit in the Power Manager has been cleared or a different peripheral function has been chosen on PA20 (PA20 will start with peripheral function F
by default). Note that the forcing will only enable the clock output. To be able to use the RC32K
normally the oscillator must be enabled as described below.
199
32099I–01/2012
AT32UC3L016/32/64
The oscillator is enabled by writing a one to the Enable bit in the 32kHz RC Oscillator Configuration Register (RC32KCR.EN) and disabled by writing a zero to RC32KCR.EN. The oscillator is
also automatically enabled when the sampling mode is requested for the SM33. In this case,
writing a zero to RC32KCR.EN will not disable the RC32K until the sampling mode is no longer
requested.
13.5.13
Generic Clocks
Rev: 1.0.0.0
Timers, communication modules, and other modules connected to external circuitry may require
specific clock frequencies to operate correctly. The SCIF defines a number of generic clocks that
can provide a wide range of accurate clock frequencies.
Each generic clock runs from either clock source listed in the “Generic Clock Sources” table in
the SCIF Module Configuration section. 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 in the Power
Manager.
Figure 13-6. Generic Clock Generation
S leep C ontroller
G eneric
C lock
S ources
M ask
D ivider
OSCSEL
13.5.13.1
f G C LK
0
fS R C
D IV
G eneric C lock
1
D IV E N
CEN
Enabling a generic clock
A generic clock is enabled by writing a one to the Clock Enable bit (CEN) in the Generic Clock
Control Register (GCCTRL). Each generic clock can individually select a clock source by writing
to the Oscillator Select field (OSCSEL). The source clock can optionally be divided by writing a
one to the Divide Enable bit (DIVEN) and the Division Factor field (DIV), resulting in the output
frequency:
f SRC
f GCLK = ---------------------------2 ( DIV + 1 )
where fSRC is the frequency of the selected source clock, and fGCLK is the output frequency of the
generic clock.
13.5.13.2
Disabling a generic clock
A generic clock is disabled by writing a zero to CEN 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. After CEN has been written to zero, the bit will
still read as one until the next falling edge occurs, and the clock is actually switched off. When
200
32099I–01/2012
AT32UC3L016/32/64
writing a zero to CEN the other bits in GCCTRL should not be changed until CEN reads as zero,
to avoid glitches on the generic clock. The generic clocks will be automatically re-enabled when
waking from sleep.
13.5.13.3
Changing clock frequency
When changing the generic clock frequency by changing OSCSEL or DIV, the clock should be
disabled before being re-enabled with the new clock source or division setting. This prevents
glitches during the transition.
13.5.13.4
Generic clock allocation
The generic clocks are allocated to different functions as shown in the “Generic Clock Allocation”
table in the SCIF Module Configuration section.
13.5.14
Interrupts
The SCIF has the following interrupt sources:
• AE - Access Error:
– A protected SCIF register was accessed without first being correctly unlocked.
• BRIFARDY - Backup Register Interface Ready.
– A 0 to 1 transition on the PCLKSR.BRIFARDY bit is detected.
• DFLL0RCS - DFLL Reference Clock Stopped:
– A 0 to 1 transition on the PCLKSR.DFLLRCS bit is detected.
• DFLL0RDY - DFLL Ready:
– A 0 to 1 transition on the PCLKSR.DFLLRDY bit is detected.
• DFLL0LOCKLOSTA - DFLL lock lost on Accurate value:
– A 0 to 1 transition on the PCLKSR.DFLLLOCKLOSTA bit is detected.
• DFLL0LOCKLOSTF - DFLL lock lost on Fine value:
– A 0 to 1 transition on the PCLKSR.DFLLLOCKLOSTF bit is detected.
• DFLL0LOCKLOSTC - DFLL lock lost on Coarse value:
– A 0 to 1 transition on the PCLKSR.DFLLLOCKLOSTC bit is detected.
• DFLL0LOCKA - DFLL Locked on Accurate value:
– A 0 to 1 transition on the PCLKSR.DFLLLOCKA bit is detected.
• DFLL0LOCKF - DFLL Locked on Fine value:
– A 0 to 1 transition on the PCLKSR.DFLLLOCKF bit is detected.
• DFLL0LOCKC - DFLL Locked on Coarse value:
– A 0 to 1 transition on the PCLKSR.DFLLLOCKC bit is detected.
• BODDET - Brown out detection:
– A 0 to 1 transition on the PCLKSR.BODDET bit is detected.
• SM33DET - Supply Monitor 3.3V Detector:
– A 0 to 1 transition on the PCLKSR.SM33DET bit is detected.
• VREGOK - Voltage Regulator OK:
– A 0 to 1 transition on the PCLKSR.VREGOK bit is detected.
• OSC0RDY - Oscillator Ready:
– A 0 to 1 transition on the PCLKSR.OSC0RDY bit is detected.
201
32099I–01/2012
AT32UC3L016/32/64
• OSC32RDY - 32KHz Oscillator Ready:
– A 0 to 1 transition on the PCLKSR.OSC32RDY bit is detected.
The interrupt sources will generate an interrupt request if the corresponding bit in the Interrupt
Mask Register is set. The interrupt sources are ORed together to form one interrupt request. The
SCIF will generate an interrupt request if at least one of the bits in the Interrupt Mask Register
(IMR) is set. Bits in IMR are set by writing a one to the corresponding bit in the Interrupt Enable
Register (IER), and cleared by writing a one to the corresponding bit in the Interrupt Disable
Register (IDR). The interrupt request remains active until the corresponding bit in the Interrupt
Status Register (ISR) is cleared by writing a one to the corresponding bit in the Interrupt Clear
Register (ICR). Because all the interrupt sources are ORed together, the interrupt request from
the SCIF will remain active until all the bits in ISR are cleared.
202
32099I–01/2012
AT32UC3L016/32/64
13.6
User Interface
Table 13-2.
SCIF Register Memory Map
Offset
Register
Register Name
Access
Reset
0x0000
Interrupt Enable Register
IER
Write-only
0x00000000
0x0004
Interrupt Disable Register
IDR
Write-only
0x00000000
0x0008
Interrupt Mask Register
IMR
Read-only
0x00000000
0x000C
Interrupt Status Register
ISR
Read-only
0x00000000
0x0010
Interrupt Clear Register
ICR
Write-only
0x00000000
0x0014
Power and Clocks Status Register
PCLKSR
Read-only
0x00000000
0x0018
Unlock Register
UNLOCK
Write-only
0x00000000
0x001C
Oscillator 0 Control Register
OSCCTRL0
Read/Write
0x00000000
0x0020
Oscillator 32 Control Register
OSCCTRL32
Read/Write
0x00000004
0x0024
DFLL Config Register
DFLL0CONF
Read/Write
0x00000000
0x0028
DFLL Multiplier Register
DFLL0MUL
Write-only
0x00000000
0x002C
DFLL Step Register
DFLL0STEP
Read/Write
0x00000000
0x0030
DFLL Spread Spectrum Generator Control
Register
DFLL0SSG
Write-only
0x00000000
0x0034
DFLL Ratio Register
DFLL0RATIO
Read-only
0x00000000
0x0038
DFLL Synchronization Register
DFLL0SYNC
Write-only
0x00000000
0x003C
BOD Control Register
BOD
Read/Write
-(2)
0x0044
Voltage Regulator Calibration Register
VREGCR
Read/Write
-(2)
0x0048
System RC Oscillator Calibration Register
RCCR
Read/Write
-(2)
0x004C
Supply Monitor 33 Calibration Register
SM33
Read/Write
-(2)
0x0050
Temperature Sensor Configuration Register
TSENS
Read/Write
0x00000000
0x0058
120MHz RC Oscillator Control Register
RC120MCR
Read/Write
0x00000000
0x005C-0x0068
Backup Registers
BR
Read/Write
0x00000000
0x006C
32kHz RC Oscillator Control Register
RC32KCR
Read/Write
0x00000000
0x0070
Generic Clock Control0
GCCTRL0
Read/Write
0x00000000
0x0074
Generic Clock Control1
GCCTRL1
Read/Write
0x00000000
0x0078
Generic Clock Control2
GCCTRL2
Read/Write
0x00000000
0x007C
Generic Clock Control3
GCCTRL3
Read/Write
0x00000000
0x0080
Generic Clock Control4
GCCTR4
Read/Write
0x00000000
0x03C8
Oscillator 0 Version Register
OSC0VERSION
Read-only
-(1)
0x03CC
32 KHz Oscillator Version Register
OSC32VERSION
Read-only
-(1)
0x03D0
DFLL Version Register
DFLLIFVERSION
Read-only
-(1)
0x03D4
BOD Version Register
BODIFAVERSION
Read-only
-(1)
0x03D8
Voltage Regulator Version Register
VREGIFBVERSION
Read-only
-(1)
203
32099I–01/2012
AT32UC3L016/32/64
Table 13-2.
Note:
SCIF Register Memory Map
Offset
Register
Register Name
Access
Reset
0x03DC
System RC Oscillator Version Register
RCOSCIFAVERSION
Read-only
-(1)
0x03E0
3.3V Supply Monitor Version Register
SM33IFAVERSION
Read-only
-(1)
0x03E4
Temperature Sensor Version Register
TSENSIFAVERSION
Read-only
-(1)
0x03EC
120MHz RC Oscillator Version Register
RC120MIFAVERSION
Read-only
-(1)
0x03F0
Backup Register Interface Version Register
BRIFAVERSION
Read-only
-(1)
0x03F4
32kHz RC Oscillator Version Register
RC32KIFAVERSION
Read-only
-(1)
0x03F8
Generic Clock Version Register
GCLKVERSION
Read-only
-(1)
0x03FC
SCIF Version Register
VERSION
Read-only
-(1)
1. The reset value is device specific. Please refer to the Module Configuration section at the end of this chapter.
2. The reset value of this register depends on factory calibration.
204
32099I–01/2012
AT32UC3L016/32/64
13.6.1
Interrupt Enable Register
Name:
IER
Access Type:
Write-only
Offset:
0x0000
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
AE
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
BRIFARDY
15
14
13
12
11
10
9
8
DFLL0RCS
DFLL0RDY
DFLL0LOCK
LOSTA
DFLL0LOCK
LOSTF
DFLL0LOCK
LOSTC
DFLL0LOCK
A
DFLL0LOCK
F
DFLL0LOCK
C
7
6
5
4
3
2
1
0
BODDET
SM33DET
VREGOK
-
-
-
OSC0RDY
OSC32RDY
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
205
32099I–01/2012
AT32UC3L016/32/64
13.6.2
Interrupt Disable Register
Name:
IDR
Access Type:
Write-only
Offset:
0x0004
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
AE
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
BRIFARDY
15
14
13
12
11
10
9
8
DFLL0RCS
DFLL0RDY
DFLL0LOCK
LOSTA
DFLL0LOCK
LOSTF
DFLL0LOCK
LOSTC
DFLL0LOCK
A
DFLL0LOCK
F
DFLL0LOCK
C
7
6
5
4
3
2
1
0
BODDET
SM33DET
VREGOK
-
-
-
OSC0RDY
OSC32RDY
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
206
32099I–01/2012
AT32UC3L016/32/64
13.6.3
Interrupt Mask Register
Name:
IMR
Access Type:
Read-only
Offset:
0x0008
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
AE
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
BRIFARDY
15
14
13
12
11
10
9
8
DFLL0RCS
DFLL0RDY
DFLL0LOCK
LOSTA
DFLL0LOCK
LOSTF
DFLL0LOCK
LOSTC
DFLL0LOCK
A
DFLL0LOCK
F
DFLL0LOCK
C
7
6
5
4
3
2
1
0
BODDET
SM33DET
VREGOK
-
-
-
OSC0RDY
OSC32RDY
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
A bit in this register is cleared when the corresponding bit in IDR is written to one.
A bit in this register is set when the corresponding bit in IER is written to one.
207
32099I–01/2012
AT32UC3L016/32/64
13.6.4
Interrupt Status Register
Name:
ISR
Access Type:
Read-only
Offset:
0x000C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
AE
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
BRIFARDY
15
14
13
12
11
10
9
8
DFLL0RCS
DFLL0RDY
DFLL0LOCK
LOSTA
DFLL0LOCK
LOSTF
DFLL0LOCK
LOSTC
DFLL0LOCK
A
DFLL0LOCK
F
DFLL0LOCK
C
7
6
5
4
3
2
1
0
BODDET
SM33DET
VREGOK
-
-
-
OSC0RDY
OSC32RDY
0: The corresponding interrupt is cleared.
1: The corresponding interrupt is pending.
A bit in this register is cleared when the corresponding bit in ICR is written to one.
A bit in this register is set when the corresponding interrupt occurs.
208
32099I–01/2012
AT32UC3L016/32/64
13.6.5
Interrupt Clear Register
Name:
ICR
Access Type:
Write-only
Offset:
0x0010
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
AE
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
BRIFARDY
15
14
13
12
11
10
9
8
DFLL0RCS
DFLL0RDY
DFLL0LOCK
LOSTA
DFLL0LOCK
LOSTF
DFLL0LOCK
LOSTC
DFLL0LOCK
A
DFLL0LOCK
F
DFLL0LOCK
C
7
6
5
4
3
2
1
0
BODDET
SM33DET
VREGOK
-
-
-
OSC0RDY
OSC32RDY
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in ISR.
209
32099I–01/2012
AT32UC3L016/32/64
13.6.6
Power and Clocks Status Register
Name:
PCLKSR
Access Type:
Read-only
Offset:
0x0014
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
BRIFAVALID
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
BRIFARDY
15
14
13
12
11
10
9
8
DFLL0RCS
DFLL0RDY
DFLL0LOCK
LOSTA
DFLL0LOCK
LOSTF
DFLL0LOCK
LOSTC
DFLL0LOCK
A
DFLL0LOCK
F
DFLL0LOCK
C
7
6
5
4
3
2
1
0
BODDET
SM33DET
VREGOK
-
-
-
OSC0RDY
OSC32RDY
• BRIFAVALID: Backup Register Interface Valid
0: The values in the backup registers are not valid.
1: The values in the backup registers are valid.
• BRIFARDY: Backup Register Interface Ready
0: The backup register interface is busy updating the backup registers. Writes to BRn will be discarded.
1: The backup register interface is ready to accept new writes to the backup registers.
• DFLL0RCS: DFLL0 Reference Clock Stopped
0: The DFLL reference clock is running, or has never been enabled.
1: The DFLL reference clock has stopped or is too slow.
• DFLL0RDY: DFLL0 Synchronization Ready
0: Read or write to DFLL registers is invalid
1: Read or write to DFLL registers is valid
• DFLL0LOCKLOSTA: DFLL0 Lock Lost on Accurate Value
0: DFLL has not lost its Accurate lock or has never been enabled.
1: DFLL has lost its Accurate lock, either by disabling the DFLL or due to faulty operation.
• DFLL0LOCKLOSTF: DFLL0 Lock Lost on Fine Value
0: DFLL has not lost its Fine lock or has never been enabled.
1: DFLL has lost its Fine lock, either by disabling the DFLL or due to faulty operation.
• DFLL0LOCKLOSTC: DFLL0 Lock Lost on Coarse Value
0: DFLL has not lost its Coarse lock or has never been enabled.
1: DFLL has lost its Coarse lock, either by disabling the DFLL or due to faulty operation.
• DFLL0LOCKA: DFLL0 Locked on Accurate Value
0: DFLL is unlocked on Accurate value.
1: DFLL is locked on Accurate value, and is ready to be selected as clock source with an accurate output clock.
• DFLL0LOCKF: DFLL0 Locked on Fine Value
210
32099I–01/2012
AT32UC3L016/32/64
•
•
•
•
•
•
0: DFLL is unlocked on Fine value.
1: DFLL is locked on Fine value, and is ready to be selected as clock source with a high accuracy on the output clock.
DFLL0LOCKC: DFLL0 Locked on Coarse Value
0: DFLL is unlocked on Coarse value.
1: DFLL is locked on Coarse value, and is ready to be selected as clock source with medium accuracy on the output clock.
BODDET: Brown-Out Detection
0: No BOD Event.
1: BOD has detected that the supply voltage is below the BOD reference value.
SM33DET: Supply Monitor 3.3V Detector
0: SM33 not enabled or the supply voltage is above the SM33 threshold.
1: SM33 enabled and the supply voltage is below the SM33 threshold.
VREGOK: Voltage Regulator OK
0: Voltage regulator not enabled or not ready.
1: Voltage regulator has reached its output threshold value after being enabled.
OSC0RDY: OSC0 Ready
0: Oscillator not enabled or not ready.
1: Oscillator is stable and ready to be used as clock source.
OSC32RDY: 32 KHz oscillator Ready
0: OSC32K not enabled or not ready.
1: OSC32K is stable and ready to be used as clock source.
211
32099I–01/2012
AT32UC3L016/32/64
13.6.7
Unlock Register
Name:
UNLOCK
Access Type:
Write-only
Offset:
0x0018
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
KEY
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
7
6
5
4
3
2
ADDR[9:8]
1
0
ADDR[7:0]
To unlock a write protected register, first write to the UNLOCK register with the address of the register to unlock in the ADDR
field and 0xAA in the KEY field. Then, in the next PB access write to the register specified in the ADDR field.
• KEY: Unlock Key
Write this bit field to 0xAA to enable unlock.
• ADDR: Unlock Address
Write the address offset of the register to unlock to this field.
212
32099I–01/2012
AT32UC3L016/32/64
13.6.8
Oscillator Control Register
Name:
OSCCTRLn
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
OSCEN
15
14
13
12
11
10
9
8
-
-
-
-
7
6
5
4
3
1
0
-
-
-
-
AGC
STARTUP[3:0]
2
GAIN[1:0]
MODE
• OSCEN: Oscillator Enable
0: The oscillator is disabled.
1: The oscillator is enabled.
• STARTUP: Oscillator Start-up Time
Select start-up time for the oscillator. Please refer to the “Oscillator Startup Time” table in the SCIF Module Configuration
section for details.
• AGC: Automatic Gain Control
For test purposes.
• GAIN: Gain
Selects the gain for the oscillator. Please refer to the “Oscillator Gain Settings” table in the SCIF Module Configuration section
for details.
• MODE: Oscillator Mode
0: External clock connected on XIN. XOUT can be used as general-purpose I/O (no crystal).
1: Crystal is connected to XIN/XOUT.
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
213
32099I–01/2012
AT32UC3L016/32/64
13.6.9
32KHz Oscillator Control Register
Name:
OSCCTRL32
Access Type:
Read/Write
Reset Value:
0x00000004
31
30
29
28
27
26
25
24
RESERVED
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
15
14
13
12
11
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
EN1K
EN32K
PINSEL
OSC32EN
STARTUP[2:0]
10
9
8
MODE[2:0]
Note: This register is only reset by Power-On Reset
• RESERVED
This bit must always be written to zero.
• STARTUP: Oscillator Start-up Time
Select start-up time for 32 KHz oscillator
Table 13-3.
Start-up Time for 32 KHz Oscillator
Number of RCSYS 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
214
32099I–01/2012
AT32UC3L016/32/64
• MODE: Oscillator Mode
Table 13-4.
MODE
Operation Mode for 32 KHz Oscillator
Description
0
External clock connected to XIN32, XOUT32 can be used as general-purpose I/O (no crystal)
1
Crystal mode. Crystal is connected to XIN32/XOUT32.
2
Reserved
3
Reserved
4
Crystal and high current mode. Crystal is connected to XIN32/XOUT32.
5
Reserved
6
Reserved
7
Reserved
• EN1K: 1 KHz output Enable
0: The 1 KHz output is disabled.
1: The 1 KHz output is enabled.
• EN32K: 32 KHz output Enable
0: The 32 KHz output is disabled.
1: The 32 KHz output is enabled.
• PINSEL: Pins Select
0: Default pins used.
1: Alternate pins: XIN32_2 pin is used instead of XIN32 pin, XOUT32_2 pin is used instead of XOUT32.
• OSC32EN: 32 KHz Oscillator Enable
0: The 32 KHz Oscillator is disabled
1: The 32 KHz Oscillator is enabled
215
32099I–01/2012
AT32UC3L016/32/64
13.6.10
DFLLn Configuration Register
Name:
DFLLnCONF
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
COARSE[7:0]
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
FINE[8]
15
14
13
12
11
10
9
8
FINE[7:0]
7
6
5
4
3
2
1
0
-
-
-
-
-
DITHER
MODE
EN
• COARSE: Coarse Calibration Value
Set the value of the coarse calibration register. If in closed loop mode, this field is Read-only.
• FINE: FINE Calibration Value
Set the value of the fine calibration register. If in closed loop mode, this field is Read-only.
• DITHER: Enable Dithering
0: The fine LSB input to the VCO is constant.
1: The fine LSB input to the VCO is dithered to achieve sub-LSB approximation to the correct multiplication ratio.
• MODE: Mode Selection
0: The DFLL is in open loop operation.
1: The DFLL is in closed loop operation.
• EN: Enable
0: The DFLL is disabled.
1: The DFLL is enabled.
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
216
32099I–01/2012
AT32UC3L016/32/64
13.6.11
DFLLn Multiplier Register
Name:
DFLLnMUL
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
IMUL[15:8]
23
22
21
20
IMUL[7:0]
15
14
13
12
FMUL[15:8]
7
6
5
4
FMUL[7:0]
• IMUL: Integer Multiply Factor
This field, together with FMUL, determines the ratio between fDFLL and fREFthe DFLL. IMUL is the integer part, while the FMUL is
the fractional part.
In open loop mode, writing to this register has no effect.
• FMUL: Fractional Multiply Factor
This field, together with IMUL, determines the ratio between fDFLL and fREFthe DFLL. IMUL is the integer part, while the FMUL is
the fractional part.
In open loop mode, writing to this register has no effect.
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
217
32099I–01/2012
AT32UC3L016/32/64
13.6.12
DFLLn Maximum Step Register
Name:
DFLLnSTEP
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
FSTEP[8]
23
22
21
20
19
18
17
16
FSTEP[7:0]
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
CSTEP[7:0]
• FSTEP: Fine Maximum Step
This indicates the maximum step size during fine adjustment in closed-loop mode. When adjusting to a new frequency, the
expected overshoot of that frequency depends on this step size.
• CSTEP: Coarse Maximum Step
This indicates the maximum step size during coarse adjustment in closed-loop mode. When adjusting to a new frequency, the
expected overshoot of that frequency depends on this step size.
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
218
32099I–01/2012
AT32UC3L016/32/64
13.6.13
DFLLn Spread Spectrum Generator Control Register
Name:
DFLLnSSG
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
9
8
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
-
PRBS
EN
STEPSIZE[4:0]
12
11
10
AMPLITUDE[4:0]
• STEPSIZE: SSG Step Size
Sets the step size of the spread spectrum.
• AMPLITUDE: SSG Amplitude
Sets the amplitude of the spread spectrum.
• PRBS: Pseudo Random Bit Sequence
0: Each spread spectrum frequency is applied at constant intervals
1: Each spread spectrum frequency is applied at pseudo-random intervals
• EN: Enable
0: SSG is disabled.
1: SSG is enabled.
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
219
32099I–01/2012
AT32UC3L016/32/64
13.6.14
DFLLn Ratio Register
Name:
DFLLnRATIO
Access Type:
Read-only
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
18
17
16
RATIODIFF[15:8]
23
22
21
20
19
RATIODIFF[7:0]
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
NUMREF[4:0]
• RATIODIFF: Multiplication Ratio Difference
In closed-loop mode, this field indicates the error in the ratio between the VCO frequency and the target frequency.
• NUMREF: Numerical Reference
The number of reference clock cycles used to measure the VCO frequency equals 2^NUMREF.
220
32099I–01/2012
AT32UC3L016/32/64
13.6.15
DFLLn Synchronization Register
Name:
DFLLnSYNC
Access Type:
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
6
5
4
3
2
1
0
-
-
-
-
-
-
-
SYNC
• SYNC: Synchronization
To be able to read the current value of DFLLnCONF or DFLLnRATIO in closed-loop mode, this bit should be written to one. The
updated value is available in DFLLnCONF and DFLLnRATIO when PCLKSR.DFLLnRDY is set.
221
32099I–01/2012
AT32UC3L016/32/64
13.6.16
BOD Control Register
Name:
BOD
Access Type:
Read/Write
Reset Value:
-
31
30
29
28
27
26
25
24
SFV
-
-
-
-
-
-
-
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
• SFV: Store Final Value
0: The register is read/write
1: The register is read-only, to protect against further accidental writes.
This bit is cleared after any reset except for a BOD reset, and during flash calibration.
• FCD: Fuse Calibration Done
0: The flash calibration will be redone after any reset.
1: The flash calibration will be redone after any reset except for a BOD reset.
This bit is cleared after any reset, except for a BOD reset.
This bit is set when the CTRL, HYST and LEVEL fields have been updated by the flash fuses after a reset.
• CTRL: BOD Control
Table 13-5.
CTRL
Operation Mode for BOD
Description
0
BOD is disabled.
1
BOD is enabled and can reset the device. An interrupt request will be generated, if enabled in the IMR
register.
2
BOD is enabled but cannot reset the device. An interrupt request will be generated, if enabled in the IMR
register.
3
Reserved.
• 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.
222
32099I–01/2012
AT32UC3L016/32/64
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
223
32099I–01/2012
AT32UC3L016/32/64
13.6.17
Voltage Regulator Calibration Register
Name:
VREGCR
Access Type:
Read/Write
Reset Value:
-
31
30
29
28
27
26
25
24
SFV
INTPD
-
-
-
DBG-
-
-
23
22
21
20
19
18
17
16
-
-
-
POR33MASK
POR33STAT
US
POR33EN
DEEPDIS
FCD
15
14
13
12
11
10
9
8
-
-
-
-
7
6
5
4
3
1
0
ON
VREGOK
EN
-
-
CALIB
2
SELVDD
• SFV: Store Final Value
0: The register is read/write.
1: The register is read-only, to protect against further accidental writes.
This bit is cleared by a Power-on Reset.
• INTPD: Internal Pull-down
This bit is used for test purposes only.
0: The voltage regulator output is not pulled to ground.
1: The voltage regulator output has a pull-down to ground.
• POR33MASK: Power-on Reset 3.3V Output Mask
0: Power-on Reset 3.3V is not masked.
1: Power-on Reset 3.3V is masked.
• POR33STATUS: Power-on Reset 3.3V Status
0: Power-on Reset is disabled.
1: Power-on Reset is enabled.
This bit is read-only. Writing to this bit has no effect.
• POR33EN: Power-on Reset 3.3V Enable
0: Writing a zero to this bit disables the POR33 detector.
1: Writing a one to this bit enables the POR33 detector.
• DEEPDIS: Disable Regulator Deep Mode
0: Regulator will enter deep mode in low-power sleep modes for lower power consumption.
1: Regulator will stay in full-power mode in all sleep modes for shorter start-up time.
• FCD: Flash Calibration Done
0: The flash calibration will be redone after any reset.
1: The flash calibration will only be redone after a Power-on Reset.
This bit is cleared after a Power-on Reset.
This bit is set when the CALIB field has been updated by flash calibration after a reset.
• CALIB: Calibration Value
Calibration value for Voltage Regulator. This is calibrated during production and should not be changed.
224
32099I–01/2012
AT32UC3L016/32/64
• ON: Voltage Regulator On Status
0: The voltage regulator is currently disabled.
1: The voltage regulator is currently enabled.
This bit is read-only. Writing to this bit has no effect.
• VREGOK: Voltage Regulator OK Status
0: The voltage regulator is disabled or has not yet reached a stable output voltage.
1: The voltage regulator has reached the output voltage threshold level after being enabled.
This bit is read-only. Writing to this bit has no effect.
• EN: Enable
0: The voltage regulator is disabled.
1: The voltage regulator is enabled.
Note: This bit is set after a Power-on Reset (POR).
• SELVDD: Select VDD
Output voltage of the Voltage Regulator. The default value of this bit corresponds to an output voltage of 1.8V.
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
225
32099I–01/2012
AT32UC3L016/32/64
13.6.18
System RC Oscillator Calibration Register
Name:
RCCR
Access Type:
Read/Write
Reset Value:
-
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
FCD
15
14
13
12
11
10
9
8
-
-
-
-
-
-
7
6
5
4
3
2
CALIB[9:8]
1
0
CALIB[7:0]
• FCD: Flash Calibration Done
0: The flash calibration will be redone after any reset.
1: The flash calibration will only be redone after a Power-on Reset.
This bit is cleared after a POR.
This bit is set when the CALIB field has been updated by the flash fuses after a reset.
• CALIB: Calibration Value
Calibration Value for the System RC oscillator (RCSYS).
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
226
32099I–01/2012
AT32UC3L016/32/64
13.6.19
Supply Monitor 33 Calibration Register
Name:
SM33
Access Type:
Read/Write
Reset Value:
-
31
30
29
28
27
-
-
-
-
23
22
21
20
19
-
-
-
-
15
14
13
12
-
-
-
-
7
6
5
4
FS
-
-
-
26
25
24
18
17
16
-
ONSM
SFV
FCD
11
10
9
8
1
0
SAMPFREQ
CALIB
3
2
CTRL
• SAMPFREQ: Sampling Frequency
Selects the sampling mode frequency of the 3.3V supply monitor. In sampling mode, the SM33 performs a measurement every
2(SAMPFREQ+5) cycles of the internal 32kHz RC oscillator.
• ONSM: Supply Monitor On Indicator
0: The supply monitor is disabled.
1: The supply monitor is enabled.
This bit is read-only. Writing to this bit has no effect.
• SFV: Store Final Value
0: The register is read/write
1: The register is read-only, to protect against further accidental writes.
This bit is cleared after a reset.
• FCD: Flash Calibration Done
This bit is cleared after a reset.
This bit is set when CALIB field has been updated after a reset.
• CALIB: Calibration Value
Calibration Value for the SM33.
• FS: Force Sampling Mode
0: Sampling mode is enabled in DeepStop and Static mode only.
1: Sampling mode is always enabled.
• CTRL: Supply Monitor Control
227
32099I–01/2012
AT32UC3L016/32/64
Selects the operating mode for the SM33.
Table 13-6.
CTRL
Operation Mode for SM33
Description
0
SM33 is disabled.
1
SM33 is enabled and can reset the device. An interrupt request will be generated if the corresponding
interrupt is enabled in the IMR register.
2
SM33 is enabled and cannot reset the device. An interrupt request will be generated if the corresponding
interrupt is enabled in the IMR register.
3
SM33 is disabled
4-7
Reserved
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
228
32099I–01/2012
AT32UC3L016/32/64
13.6.20
Temperature Sensor Configuration Register
Name:
TSENS
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
-
-
-
-
-
EN
• EN: Temperature Sensor Enable
0: The Temperature Sensor is disabled.
1: The Temperature Sensor is enabled.
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
229
32099I–01/2012
AT32UC3L016/32/64
13.6.21
120MHz RC Oscillator Configuration Register
Name:
RC120MCR
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
-
-
-
-
-
-
-
EN
• EN: RC120M Enable
0: The 120 MHz RC oscillator is disabled.
1: The 120 MHz RC oscillator is enabled.
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
230
32099I–01/2012
AT32UC3L016/32/64
13.6.22
Backup Register n
Name:
BRn
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
DATA[31:24]
23
22
21
20
DATA[23:16]
15
14
13
12
DATA[15:8]
7
6
5
4
DATA[7:0]
This is a set of general-purpose read/write registers. Data stored in these registers is retained when the device is in Shutdown. Before writing to these registers the user must ensure that PCLKSR.BRIFARDY is not set.
Note that this registers are protected by a lock. To write to these registers the UNLOCK register has to be written first.
Please refer to the UNLOCK register description for details.
231
32099I–01/2012
AT32UC3L016/32/64
13.6.23
32kHz RC Oscillator Configuration Register
Name:
RC32KCR
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
-
-
-
-
-
-
-
EN
• EN: RC32K Enable
0: The 32 kHz RC oscillator is disabled.
1: The 32 kHz RC oscillator is enabled.
Note that this register is protected by a lock. To write to this register the UNLOCK register has to be written first. Please
refer to the UNLOCK register description for details.
232
32099I–01/2012
AT32UC3L016/32/64
13.6.24
Generic Clock Control
Name:
GCCTRL
Access Type:
Read/Write
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
DIV[15:8]
23
22
21
20
DIV[7:0]
15
14
13
12
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
-
DIVEN
CEN
OSCSEL[4:0]
There is one GCCTRL register per generic clock in the design.
• DIV: Division Factor
The number of DIV bits for each generic clock is as shown in the “Generic Clock number of DIV bits” table in the SCIF Module
Configuration section.
• OSCSEL: Oscillator Select
Selects the source clock for the generic clock. Please refer to the “Generic Clock Sources” table in the SCIF Module
Configuration section.
• 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: The generic clock is disabled.
1: The generic clock is enabled.
233
32099I–01/2012
AT32UC3L016/32/64
13.6.25
Oscillator 0 Version Register
Name:
OSC0VERSION
Access Type:
Read-only
Offset:
0x03C8
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
234
32099I–01/2012
AT32UC3L016/32/64
13.6.26
32KHz Oscillator Version Register
Name:
OSC32VERSION
Access Type:
Read-only
Offset:
0x03CC
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
235
32099I–01/2012
AT32UC3L016/32/64
13.6.27
Digital Frequency Locked Loop Version Register
Name:
DFLLIF VERSION
Access Type:
Read-only
Offset:
0x03D0
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
236
32099I–01/2012
AT32UC3L016/32/64
13.6.28
Brown-Out Detector Version Register
Name:
BODIFAVERSION
Access Type:
Read-only
Offset:
0x03D4
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
237
32099I–01/2012
AT32UC3L016/32/64
13.6.29
Voltage Regulator Version Register
Name:
VREGIFBVERSION
Access Type:
Read-only
Offset:
0x03D8
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
238
32099I–01/2012
AT32UC3L016/32/64
13.6.30
RC Oscillator Version Register
Name:
RCOSCIFAVERSION
Access Type:
Read-only
Offset:
0x03DC
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
239
32099I–01/2012
AT32UC3L016/32/64
13.6.31
3.3V Supply Monitor Version Register
Name:
SM33IFAVERSION
Access Type:
Read-only
Offset:
0x03E0
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
240
32099I–01/2012
AT32UC3L016/32/64
13.6.32
Temperature Sensor Version Register
Name:
TSENSIFAVERSION
Access Type:
Read-only
Offset:
0x03E4
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
241
32099I–01/2012
AT32UC3L016/32/64
13.6.33
120MHz RC Oscillator Version Register
Name:
RC120MIFAVERSION
Access Type:
Read-only
Offset:
0x03EC
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
242
32099I–01/2012
AT32UC3L016/32/64
13.6.34
Backup Register Interface Version Register
Name:
BRIFAVERSION
Access Type:
Read-only
Offset:
0x03F0
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
243
32099I–01/2012
AT32UC3L016/32/64
13.6.35
32kHz RC Oscillator Version Register
Name:
RC32KIFAVERSION
Access Type:
Read-only
Offset:
0x03F4
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
244
32099I–01/2012
AT32UC3L016/32/64
13.6.36
Generic Clock Version Register
Name:
GCLKVERSION
Access Type:
Read-only
Offset:
0x03F8
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
245
32099I–01/2012
AT32UC3L016/32/64
13.6.37
SCIF Version Register
Name:
VERSION
Access Type:
Read-only
Offset:
0x03FC
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
VARIANT
11
10
VERSION[11:0]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
246
32099I–01/2012
AT32UC3L016/32/64
13.7
Module Configuration
The specific configuration for each SCIF instance is listed in the following tables.The module bus
clocks listed here are connected to the system bus clocks. Please refer to the Power Manager
chapter for details.
Table 13-7.
SCIF Clocks
Module Name
Clock Name
Description
SCIF
CLK_SCIF
Clock for the SCIF bus interface
Table 13-8.
Oscillator Startup Time
Number of System RC Oscillator Clock
Cycles
Approximative Equivalent time
(RCSYS = 115kHz)
0
0
0
1
64
560µs
2
128
1.1ms
3
2048
18ms
4
4096
36ms
5
8192
71ms
6
16384
142ms
7
Reserved
Reserved
STARTUP
Table 13-9.
Oscillator Gain Settings
GAIN[1:0]
Gain Setting
0
Oscillator is used with gain G0 (XIN from 0.45 MHz to 12.0 MHz)
1
Oscillator is used with gain G1 (XIN from 12.0 MHz to 16.0 MHz)
2
Oscillator is used with gain G2 (XIN equals 16.0 MHz. Used for e.g. increasing S/N
ratio, better drive strength for high ESR crystals)
3
Oscillator is used with gain G3 (XIN equals 16.0 MHz. Used for e.g. increasing S/N
ratio, better drive strength for high ESR crystals)
In AT32UC3L016/32/64, there are six generic clocks. These are allocated to different functions
as shown in Table 13-10. Note that only GCLK4-0 are routed out.
Table 13-10. Generic Clock Allocation
Clock number
Function
0
DFLLIF main reference and GCLK0 pin
(CLK_DFLLIF_REF)
1
DFLLIF dithering and ssg reference and GCLK1 pin
(CLK_DFLLIF_DITHER)
2
AST and GCLK2 pin
247
32099I–01/2012
AT32UC3L016/32/64
Table 13-10. Generic Clock Allocation
Clock number
3
PWMA and GCLK3 pin
4
CAT, ACIFB and GCLK4 pin
(1)
GLOC
5
Note:
Function
1. GCLK5 is not routed out.
Table 13-11. Generic Clock Sources
OSCSEL
Clock/Oscillator
Description
0
RCSYS
System RC oscillator clock
1
OSC32K
Output clock from OSC32K
2
DFLL0
Output clock from DFLL0
3
OSC0
Output clock from Oscillator0
4
RC120M
Output from 120MHz RCOSC
5
CLK_CPU
The clock the CPU runs on
6
CLK_HSB
High Speed Bus clock
7
CLK_PBA
Peripheral Bus A clock
8
CLK_PBB
Peripheral Bus B clock
9
RC32K
Output from 32KHz RCOSC
10
Reserved
11
CLK_1K
12-31
Reserved
1KHz output clock from OSC32K
.
Table 13-12. Generic Clock number of DIV bits
Generic Clock
Number of DIV bits
0
8
1
8
2
8
3
8
4
8
5
8
248
32099I–01/2012
AT32UC3L016/32/64
Table 13-13. Register Reset Values
Register
Reset Value
OSC0VERSION
0x00000100
OSC32VERSION
0x00000101
DFLLIFVERSION
0x00000201
BODIFAVERSION
0x00000101
VREGIFBVERSION
0x00000101
RCOSCIFAVERSION
0x00000101
SM33IFAVERSION
0x00000100
TSENSEIFAVERSION
0x00000100
RC120MIFAVERSION
0x00000101
BRIFAVERSION
0x00000100
RC32KIFAVERSION
0x00000100
GCLKVERSION
0x00000100
VERSION
0x00000102
249
32099I–01/2012
AT32UC3L016/32/64
14. Asynchronous Timer (AST)
Rev: 3.0.0.1
14.1
Features
• 32-bit counter with 32-bit prescaler
• Clocked Source
•
•
•
•
•
14.2
– System RC oscillator (RCSYS)
– 32KHz crystal oscillator (OSC32K)
– PB clock
– Generic clock (GCLK)
– 1KHz clock from 32KHz oscillator
Operation and wakeup during shutdown
Optional calendar mode supported
Digital prescaler tuning for increased accuracy
Periodic interrupt(s) and peripheral event(s) supported
Alarm interrupt(s) and peripheral event(s) supported
– Optional clear on alarm
Overview
The Asynchronous Timer (AST) enables periodic interrupts and periodic peripheral events, as
well as interrupts and peripheral events at a specified time in the future. The AST consists of a
32-bit prescaler which feeds a 32-bit up-counter. The prescaler can be clocked from five different clock sources, including the low-power 32KHz oscillator, which allows the AST to be used as
a real-time timer with a maximum timeout of more than 100 years. Also, the PB clock or a
generic clock can be used for high-speed operation, allowing the AST to be used as a general
timer.
The AST can generate periodic interrupts and peripheral events from output from the prescaler,
as well as alarm interrupts and peripheral events, which can trigger at any counter value. Additionally, the timer can trigger an overflow interrupt and peripheral event, and be reset on the
occurrence of any alarm. This allows periodic interrupts and peripheral events at very long and
accurate intervals.
To keep track of time during shutdown the AST can run while the rest of the core is powered off.
This will reduce the power consumption when the system is idle. The AST can also wake up the
system from shutdown using either the alarm wakeup, periodic wakeup. or overflow wakeup
mechanisms.
The AST has been designed to meet the system tick and Real Time Clock requirements of most
embedded operating systems.
250
32099I–01/2012
AT32UC3L016/32/64
14.3
Block Diagram
Figure 14-1. Asynchronous Timer Block Diagram
CLK_AST
CONTROL
REGISTER
CLK_AST
CSSEL
RCSYS
32-bit
Prescaler
CLK_AST_PRSC
GCLK
CLK_AST_CNT
32-bit
Counter
others
DIGITAL
TUNER
REGISTER
14.4
Wake
Control
COUNTER
VALUE
Wake
EN PSEL
OSC32
PB clock
CLK_AST
WAKE ENABLE
REGISTER
Periodic
Interrupts
Alarm
Interrupts
PERIODIC
INTERVAL
REGISTER
ALARM
REGISTER
OVF
Interrupt
Status
and
Control
IRQs
Events
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
14.4.1
Power Management
When the AST is enabled, it will remain clocked as long as its selected clock source is running. It
can also wake the CPU from the currently active sleep mode. Refer to the Power Manager chapter for details on the different sleep modes.
14.4.2
Clocks
The clock for the AST bus interface (CLK_AST) is generated by the Power Manager. This clock
is turned on by default, and can be enabled and disabled in the Power Manager.
A number of clocks can be selected as source for the internal prescaler clock CLK_AST_PRSC.
The prescaler, counter, and interrupt will function as long as this selected clock source is active.
The selected clock must be enabled in the System Control Interface (SCIF).
The following clock sources are available:
• System RC oscillator (RCSYS). This oscillator is always enabled, except in some sleep
modes. Please refer to the Electrical Characteristics chapter for the characteristic frequency
of this oscillator.
• 32KHz crystal oscillator (OSC32K). This oscillator must be enabled before use.
• Peripheral Bus clock (PB clock). This is the clock of the peripheral bus the AST is connected
to.
251
32099I–01/2012
AT32UC3L016/32/64
• Generic clock (GCLK). One of the generic clocks is connected to the AST. This clock must be
enabled before use, and remains enabled in sleep modes when the PB clock is active.
• 1KHz clock from the 32KHz oscillator (CLK_1K). This clock is only available in crystal mode,
and must be enabled before use.
In Shutdown mode only the 32 KHz oscillator and the 1KHz clock are available, using certain
pins. Please refer to the Power Manager chapter for details.
14.4.3
Interrupts
The AST interrupt request lines are connected to the interrupt controller. Using the AST interrupts requires the interrupt controller to be programmed first.
14.4.4
Peripheral Events
The AST peripheral events are connected via the Peripheral Event System. Refer to the Peripheral Event System chapter for details.
14.4.5
Debug Operation
The AST prescaler and counter is frozen during debug operation, unless the Run In Debug bit in
the Development Control Register is set and the bit corresponding to the AST is set in the
Peripheral Debug Register (PDBG). Please refer to the On-Chip Debug chapter in the
AVR32UC Technical Reference Manual, and the OCD Module Configuration section, for details.
If the AST is configured in a way that requires it to be periodically serviced by the CPU through
interrupts or similar, improper operation or data loss may result during debugging.
14.5
Functional Description
14.5.1
Initialization
Before enabling the AST, the internal AST clock CLK_AST_PRSC must be enabled, following
the procedure specified in Section 14.5.1.1. The Clock Source Select field in the Clock register
(CLOCK.CSSEL) selects the source for this clock. The Clock Enable bit in the Clock register
(CLOCK.CEN) enables the CLK_AST_PRSC.
When CLK_AST_PRSC is enabled, the AST can be enabled by writing a one to the Enable bit in
the Control Register (CR.EN).
14.5.1.1
Enabling and disabling the AST clock
The Clock Source Selection field (CLOCK.CSSEL) and the Clock Enable bit (CLOCK.CEN) cannot be changed simultaneously. Special procedures must be followed for enabling and disabling
the CLK_AST_PRSC and for changing the source for this clock.
To enable CLK_AST_PRSC:
• Write the selected value to CLOCK.CSSEL
• Wait until SR.CLKBUSY reads as zero
• Write a one to CLOCK.CEN, without changing CLOCK.CSSEL
• Wait until SR.CLKBUSY reads as zero
To disable the clock:
• Write a zero to CLOCK.CEN to disable the clock, without changing CLOCK.CSSEL
• Wait until SR.CLKBUSY reads as zero
252
32099I–01/2012
AT32UC3L016/32/64
14.5.1.2
Changing the source clock
The CLK_AST_PRSC must be disabled before switching to another source clock. The Clock
Busy bit in the Status Register (SR.CLKBUSY) indicates whether the clock is busy or not. This
bit is set when the CEN bit in the CLOCK register is changed, and cleared when the CLOCK register can be changed.
To change the clock:
• Write a zero to CLOCK.CEN to disable the clock, without changing CLOCK.CSSEL
• Wait until SR.CLKBUSY reads as zero
• Write the selected value to CLOCK.CSSEL
• Wait until SR.CLKBUSY reads as zero
• Write a one to CLOCK.CEN to enable the clock, without changing the CLOCK.CSSEL
• Wait until SR.CLKBUSY reads as zero
14.5.2
Basic Operation
14.5.2.1
Prescaler
When the AST is enabled, the 32-bit prescaler will increment on the rising edge of
CLK_AST_PRSC. The prescaler value cannot be read or written, but it can be reset by writing a
one to the Prescaler Clear bit in the Control Register (CR.PCLR).
The Prescaler Select field in the Control Register (CR.PSEL) selects the prescaler bit PSEL as
source clock for the counter (CLK_AST_CNT). This results in a counter frequency of:
f PRSC
f CNT = ----------------------PSEL + 1
2
where fPRSC is the frequency of the internal prescaler clock CLK_AST_PRSC.
14.5.2.2
Counter operation
When enabled, the AST will increment on every 0-to-1 transition of the selected prescaler tapping. When the Calender bit in the Control Register (CR.CAL) is zero, the counter operates in
counter mode. It will increment until it reaches the top value of 0xFFFFFFFF, and then wrap to
0x00000000. This sets the status bit Overflow in the Status Register (SR.OVF). Optionally, the
counter can also be reset when an alarm occurs (see Section 14.5.3.2 on page 255. This will
also set the OVF bit.
The AST counter value can be read from or written to the Counter Value (CV) register. Note that
due to synchronization, continuous reading of the CV register with the lowest prescaler setting
will skip every third value. In addition, if CLK_AST_PRSC is as fast as, or faster than, the
CLK_AST, the prescaler value must be 3 or higher to be able to read the CV without skipping
values.
253
32099I–01/2012
AT32UC3L016/32/64
14.5.2.3
Calendar operation
When the CAL bit in the Control Register is one, the counter operates in calendar mode. Before
this mode is enabled, the prescaler should be set up to give a pulse every second. The date and
time can then be read from or written to the Calendar Value (CALV) register.
Time is reported as seconds, minutes, and hours according to the 24-hour clock format. Date is
the numeral date of month (starting on 1). Month is the numeral month of the year (1 = January,
2 = February, etc.). Year is a 6-bit field counting the offset from a software-defined leap year
(e.g. 2000). The date is automatically compensated for leap years, assuming every year divisible
by 4 is a leap year.
All peripheral events and interrupts work the same way in calendar mode as in counter mode.
However, the Alarm Register (ARn) must be written in time/date format for the alarm to trigger
correctly.
14.5.3
Interrupts
The AST can generate five separate interrupt requests:
• OVF: OVF
• PER: PER0, PER1
• ALARM: ALARM0, ALARM1
• CLKREADY
• READY
This allows the user to allocate separate handlers and priorities to the different interrupt types.
The generation of the PER interrupt is described in Section 14.5.3.1., and the generation of the
ALARM interrupt is described in Section 14.5.3.2. The OVF interrupt is generated when the
counter overflows, or when the alarm value is reached, if the Clear on Alarm bit in the Control
Register is one. The CLKREADY interrupt is generated when SR.CLKBUSY has a 1-to-0 transition, and indicates that the clock synchronization is completed. The READY interrupt is
generated when SR.BUSY has a 1-to-0 transition, and indicates that the synchronization
described in Section 14.5.8 is completed.
An interrupt request will be generated if the corresponding bit in the Interrupt Mask Register
(IMR) is set. Bits in IMR are set by writing a one to the corresponding bit in the Interrupt Enable
Register (IER), and cleared by writing a one to the corresponding bit in the Interrupt Disable
Register (IDR). The interrupt request remains active until the corresponding bit in SR is cleared
by writing a one to the corresponding bit in the Status Clear Register (SCR).
The AST interrupts can wake the CPU from any sleep mode where the source clock and the
interrupt controller is active.
14.5.3.1
Periodic interrupt
The AST can generate periodic interrupts. If the PERn bit in the Interrupt Mask Register (IMR) is
one, the AST will generate an interrupt request on the 0-to-1 transition of the selected bit in the
254
32099I–01/2012
AT32UC3L016/32/64
prescaler when the AST is enabled. The bit is selected by the Interval Select field in the corresponding Periodic Interval Register (PIRn.INSEL), resulting in a periodic interrupt frequency of
f CS
f PA = ------------------------INSEL + 1
2
where fCS is the frequency of the selected clock source.
The corresponding PERn bit in the Status Register (SR) will be set when the selected bit in the
prescaler has a 0-to-1 transition.
Because of synchronization, the transfer of the INSEL value will not happen immediately. When
changing/setting the INSEL value, the user must make sure that the prescaler bit number INSEL
will not have a 0-to-1 transition before the INSEL value is transferred to the register. In that case,
the first periodic interrupt after the change will not be triggered.
14.5.3.2
Alarm interrupt
The AST can also generate alarm interrupts. If the ALARMn bit in IMR is one, the AST will generate an interrupt request when the counter value matches the selected alarm value, when the
AST is enabled. The alarm value is selected by writing the value to the VALUE field in the corresponding Alarm Register (ARn.VALUE).
The corresponding ALARMn bit in SR will be set when the counter reaches the selected alarm
value.
Because of synchronization, the transfer of the alarm value will not happen immediately. When
changing/setting the alarm value, the user must make sure that the counter will not count the
selected alarm value before the value is transferred to the register. In that case, the first alarm
interrupt after the change will not be triggered.
If the Clear on Alarm bit in the Control Register (CR.CAn) is one, the corresponding alarm interrupt will clear the counter and set the OVF bit in the Status Register. This will generate an
overflow interrupt if the OVF bit in IMR is set.
14.5.4
Peripheral events
The AST can generate a number of peripheral events:
• OVF
• PER0
• PER1
• ALARM0
• ALARM1
The PERn peripheral event(s) is generated the same way as the PER interrupt, as described in
Section 14.5.3.1. The ALARMn peripheral event(s) is generated the same way as the ALARM
interrupt, as described in Section 14.5.3.2. The OVF peripheral event is generated the same
way as the OVF interrupt, as described in Section 14.5.3-
255
32099I–01/2012
AT32UC3L016/32/64
The peripheral event will be generated if the corresponding bit in the Event Mask (EVM) register
is set. Bits in EVM register are set by writing a one to the corresponding bit in the Event Enable
(EVE) register, and cleared by writing a one to the corresponding bit in the Event Disable (EVD)
register.
14.5.5
AST wakeup
The AST can wake up the CPU directly, without the need to trigger an interrupt. A wakeup can
be generated when the counter overflows, when the counter reaches the selected alarm value,
or when the selected prescaler bit has a 0-to-1 transition. In this case, the CPU will continue
executing from the instruction following the sleep instruction.
The AST wakeup is enabled by writing a one to the corresponding bit in the Wake Enable Register (WER). When the CPU wakes from sleep, the wake signal must be cleared by writing a one
to the corresponding bit in SCR to clear the internal wake signal to the sleep controller. If the
wake signal is not cleared after waking from sleep, the next sleep instruction will have no effect
because the CPU will wake immediately after this sleep instruction.
The AST wakeup can wake the CPU from any sleep mode where the source clock is active. The
AST wakeup can be configured independently of the interrupt masking.
14.5.6
Shutdown Mode
If the AST is configured to use a clock that is available in Shutdown mode, the AST can be used
to wake up the system from shutdown. Both the alarm wakeup, periodic wakeup, and overflow
wakeup mechanisms can be used in this mode.
When waking up from Shutdown mode all control registers will have the same value as before
the shutdown was entered, except the Interrupt Mask Register (IMR). IMR will be reset with all
interrupts turned off. The software must first reconfigure the interrupt controller and then enable
the interrupts in the AST to again receive interrupts from the AST.
The CV register will be updated with the current counter value directly after wakeup from shutdown. The SR will show the status of the AST, including the status bits set during shutdown
operation.
When waking up the system from shutdown the CPU will start executing code from the reset
start address.
14.5.7
Digital Tuner
The digital tuner adds the possibility to compensate for a too-slow or a too-fast input clock. The
ADD bit in the Digital Tuner Register (DTR.ADD) selects if the prescaler frequency should be
reduced or increased. If ADD is ‘0’, the prescaler frequency is reduced:
⎛
⎞
⎜
⎟
1
f TUNED = f 0 ⎜ 1 – --------------------------------------------------------------------------------⎟
256
⎜
roundup ⎛ --------------------⎞ ⋅ ( 2 EXP ) + 1⎟⎠
⎝
⎝ VALUE⎠
where fTUNED is the tuned frequency, f0 is the original prescaler frequency, and VALUE and EXP
are the corresponding fields to be programmed in DTR. Note that DTR.EXP must be greater
than zero. Frequency tuning is disabled by programming DTR.VALUE as zero.
256
32099I–01/2012
AT32UC3L016/32/64
If ADD is ‘1’, the prescaler frequency is increased:
⎛
⎞
⎜
⎟
1
f TUNED = f 0 ⎜ 1 + --------------------------------------------------------------------------------⎟
256
⎜
roundup ⎛ --------------------⎞ ⋅ ( 2 EXP ) – 1⎟⎠
⎝
⎝ VALUE⎠
Note that for these formulas to be within an error of 0.01%, it is recommended that the prescaler
bit that is used as the clock for the counter (selected by CR.PSEL) or to trigger the periodic interrupt (selected by PIRn.INSEL) be bit 6 or higher.
14.5.8
Synchronization
As the prescaler and counter operate asynchronously from the user interface, the AST needs a
few clock cycles to synchronize the values written to the CR, CV, SCR, WER, EVE, EVD, PIRn,
ARn, and DTR registers. The Busy bit in the Status Register (SR.BUSY) indicates that the synchronization is ongoing. During this time, writes to these registers will be discarded and reading
will return a zero value.
Note that synchronization takes place also if the prescaler is clocked from CLK_AST.
257
32099I–01/2012
AT32UC3L016/32/64
14.6
User Interface
Table 14-1.
Offset
Register
Register Name
Access
Reset
0x00
Control Register
CR
Read/Write
0x00000000
0x04
Counter Value
CV
Read/Write
0x00000000
0x08
Status Register
SR
Read-only
0x00000000
0x0C
Status Clear Register
SCR
Write-only
0x00000000
0x10
Interrupt Enable Register
IER
Write-only
0x00000000
0x14
Interrupt Disable Register
IDR
Write-only
0x00000000
0x18
Interrupt Mask Register
IMR
Read-only
0x00000000
0x1C
Wake Enable Register
WER
Read/write
0x00000000
AR0
Read/Write
0x00000000
0x20
Alarm Register 0
(2)
Alarm Register 1
(2)
AR1
Read/Write
0x00000000
0x30
(2)
Periodic Interval Register 0
PIR0
Read/Write
0x00000000
0x34
Periodic Interval Register 1(2)
PIR1
Read/Write
0x00000000
0x40
Clock Control Register
CLOCK
Read/Write
0x00000000
0x44
Digital Tuner Register
DTR
Read/Write
0x00000000
0x48
Event Enable
EVE
Write-only
0x00000000
0x4C
Event Disable
EVD
Write-only
0x00000000
0x50
Event Mask
EVM
Read-only
0x00000000
0x54
Calendar Value
CALV
Read/Write
0x00000000
0xF0
Parameter Register
PARAMETER
Read-only
-(1)
0xFC
Version Register
VERSION
Read-only
-(1)
0x24
Note:
AST Register Memory Map
1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter.
2. The number of Alarm and Periodic Interval registers are device specific. Please refer to the Module Configuration section at
the end of this chapter.
258
32099I–01/2012
AT32UC3L016/32/64
14.6.1
Name:
Control Register
CR
Access Type:
Read/Write
Offset:
0x00
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
-
-
-
-
-
-
CA1
CA0
7
6
5
4
3
2
1
0
-
-
-
-
-
CAL
PCLR
EN
PSEL
When the SR.BUSY bit is set, writes to this register will be discarded and this register will read as zero.
• PSEL: Prescaler Select
Selects prescaler bit PSEL as source clock for the counter.
• CAn: Clear on Alarm n
0: The corresponding alarm will not clear the counter.
1: The corresponding alarm will clear the counter.
• CAL: Calendar Mode
0: The AST operates in counter mode.
1: The AST operates in calendar mode.
• PCLR: Prescaler Clear
Writing a zero to this bit has no effect.
Writing a one to this bit clears the prescaler.
This bit always reads as zero.
• EN: Enable
0: The AST is disabled.
1: The AST is enabled.
259
32099I–01/2012
AT32UC3L016/32/64
14.6.2
Name:
Counter Value
CV
Access Type:
Read/Write
Offset:
0x04
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
18
17
16
11
10
9
8
3
2
1
0
VALUE[31:24]
23
22
21
20
19
VALUE[23:16]
15
14
13
12
VALUE[15:8]
7
6
5
4
VALUE[7:0]
When the SR.BUSY bit is set, writes to this register will be discarded and this register will read as zero.
• VALUE: AST Value
The current value of the AST counter.
260
32099I–01/2012
AT32UC3L016/32/64
14.6.3
Name:
Status Register
SR
Access Type:
Read-only
Offset:
0x08
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
CLKRDY
CLKBUSY
-
-
READY
BUSY
23
22
21
20
19
18
17
16
-
-
-
-
-
-
PER1
PER0
15
14
13
12
11
10
9
8
-
-
-
-
-
-
ALARM1
ALARM0
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
OVF
• CLKRDY: Clock Ready
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when the SR.CLKBUSY bit has a 1-to-0 transition.
• CLKBUSY: Clock Busy
0: The clock is ready and can be changed.
1: CLOCK.CEN has been written and the clock is busy.
• READY: AST Ready
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when the SR.BUSY bit has a 1-to-0 transition.
• BUSY: AST Busy
0: The AST accepts writes to CR, CV, SCR, WER, EVE, EVD, ARn, PIRn, and DTR.
1: The AST is busy and will discard writes to CR, CV, SCR, WER, EVE, EVD, ARn, PIRn, and DTR.
• PERn: Periodic n
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when the selected bit in the prescaler has a 0-to-1 transition.
• ALARMn: Alarm n
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when the counter reaches the selected alarm value.
• OVF: Overflow
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when an overflow has occurred.
261
32099I–01/2012
AT32UC3L016/32/64
14.6.4
Name:
Status Clear Register
SCR
Access Type:
Write-only
Offset:
0x0C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
CLKRDY
-
-
-
READY
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
PER1
PER0
15
14
13
12
11
10
9
8
-
-
-
-
-
-
ALARM1
ALARM0
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
OVF
When the SR.BUSY bit is set, writes to this register will be discarded.
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in SR and the corresponding interrupt request.
262
32099I–01/2012
AT32UC3L016/32/64
14.6.5
Name:
Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x10
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
CLKRDY
-
-
-
READY
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
PER1
PER0
15
14
13
12
11
10
9
8
-
-
-
-
-
-
ALARM1
ALARM0
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
OVF
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
263
32099I–01/2012
AT32UC3L016/32/64
14.6.6
Name:
Interrupt Disable Register
IDR
Access Type:
Write-only
Offset:
0x14
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
CLKRDY
-
-
-
READY
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
PER1
PER0
15
14
13
12
11
10
9
8
-
-
-
-
-
-
ALARM1
ALARM0
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
OVF
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
264
32099I–01/2012
AT32UC3L016/32/64
14.6.7
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
Offset:
0x18
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
CLKRDY
-
-
-
READY
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
PER1
PER0
15
14
13
12
11
10
9
8
-
-
-
-
-
-
ALARM1
ALARM0
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
OVF
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
A bit in this register is cleared when the corresponding bit in IDR is written to one.
A bit in this register is set when the corresponding bit in IER is written to one.
265
32099I–01/2012
AT32UC3L016/32/64
14.6.8
Name:
Wake Enable Register
WER
Access Type:
Read/Write
Offset:
0x1C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
PER1
PER0
15
14
13
12
11
10
9
8
-
-
-
-
-
-
ALARM1
ALARM0
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
OVF
When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero.
This register enables the wakeup signal from the AST.
• PERn: Periodic n
0: The CPU will not wake up from sleep mode when the selected bit in the prescaler has a 0-to-1 transition.
1: The CPU will wake up from sleep mode when the selected bit in the prescaler has a 0-to-1 transition.
• ALARMn: Alarm n
0: The CPU will not wake up from sleep mode when the counter reaches the selected alarm value.
1: The CPU will wake up from sleep mode when the counter reaches the selected alarm value.
• OVF: Overflow
0: A counter overflow will not wake up the CPU from sleep mode.
1: A counter overflow will wake up the CPU from sleep mode.
266
32099I–01/2012
AT32UC3L016/32/64
14.6.9
Name:
Alarm Register 0
AR0
Access Type:
Read/Write
Offset:
0x20
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
18
17
16
11
10
9
8
3
2
1
0
VALUE[31:24]
23
22
21
20
19
VALUE[23:16]
15
14
13
12
VALUE[15:8]
7
6
5
4
VALUE[7:0]
When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero.
• VALUE: Alarm Value
When the counter reaches this value, an alarm is generated.
267
32099I–01/2012
AT32UC3L016/32/64
14.6.10
Name:
Alarm Register 1
AR1
Access Type:
Read/Write
Offset:
0x24
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
18
17
16
11
10
9
8
3
2
1
0
VALUE[31:24]
23
22
21
20
19
VALUE[23:16]
15
14
13
12
VALUE[15:8]
7
6
5
4
VALUE[7:0]
When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero.
• VALUE: Alarm Value
When the counter reaches this value, an alarm is generated.
268
32099I–01/2012
AT32UC3L016/32/64
14.6.11
Name:
Periodic Interval Register 0
PIR0
Access Type:
Read/Write
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
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
INSEL
When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero.
• INSEL: Interval Select
The PER0 bit in SR will be set when the INSEL bit in the prescaler has a 0-to-1 transition.
269
32099I–01/2012
AT32UC3L016/32/64
14.6.12
Name:
Periodic Interval Register 1
PIR1
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
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
INSEL
When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero.
• INSEL: Interval Select
The PER1 bit in SR will be set when the INSEL bit in the prescaler has a 0-to-1 transition.
270
32099I–01/2012
AT32UC3L016/32/64
14.6.13
Name:
Clock Control Register
CLOCK
Access Type:
Read/Write
Offset:
0x40
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
-
-
-
-
-
-
-
CEN
CSSEL
When writing to this register, follow the sequence in Section 14.5.1 on page 252.
• CSSEL: Clock Source Selection
This field defines the clock source CLK_AST_PRSC for the prescaler:
Table 14-2.
CSSEL
Clock Source Selection
Clock Source
0
System RC oscillator (RCSYS)
1
32KHz oscillator (OSC32K)
2
PB clock
3
Generic clock (GCLK)
4
1KHz clock from 32KHz oscillator (CLK_1K)
• CEN: Clock Enable
0: CLK_AST_PRSC is disabled.
1: CLK_AST_PRSC is enabled.
271
32099I–01/2012
AT32UC3L016/32/64
14.6.14
Name:
Digital Tuner Register
DTR
Access Type:
Read/Write
Offset:
0x44
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
VALUE
7
6
5
-
-
ADD
4
EXP
When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero.
• VALUE:
0: The frequency is unchanged.
1-255: The frequency will be adjusted according to the formula below.
• ADD:
⎛
⎞
⎜
⎟
1
0: The resulting frequency is f0 ⎜ 1 – -------------------------------------------------------------------------------⎟
256
⎜
for VALUE > 0 .
roundup ⎛ --------------------⎞ ⋅ 2 ( EXP ) + 1⎟⎠
⎝
⎝ VALUE⎠
⎛
⎞
⎜
⎟
1
1: The resulting frequency is f0 ⎜ 1 + -------------------------------------------------------------------------------⎟
256
⎜
for VALUE > 0 .
roundup ⎛ --------------------⎞ ⋅ 2 ( EXP ) – 1⎟⎠
⎝
⎝ VALUE⎠
• EXP:
The frequency will be adjusted according to the formula above.
272
32099I–01/2012
AT32UC3L016/32/64
14.6.15
Name:
Event Enable Register
EVE
Access Type:
Write-only
Offset:
0x48
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
PER1
PER0
15
14
13
12
11
10
9
8
-
-
-
-
-
-
ALARM1
ALARM0
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
OVF
When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero.
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in EVM.
273
32099I–01/2012
AT32UC3L016/32/64
14.6.16
Name:
Event Disable Register
EVD
Access Type:
Write-only
Offset:
0x4C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
PER1
PER0
15
14
13
12
11
10
9
8
-
-
-
-
-
-
ALARM1
ALARM0
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
OVF
When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero.
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in EVM.
274
32099I–01/2012
AT32UC3L016/32/64
14.6.17
Name:
Event Mask Register
EVM
Access Type:
Read-only
Offset:
0x50
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
PER1
PER0
15
14
13
12
11
10
9
8
-
-
-
-
-
-
ALARM1
ALARM0
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
OVF
0: The corresponding peripheral event is disabled.
1: The corresponding peripheral event is enabled.
This bit is cleared when the corresponding bit in EVD is written to one.
This bit is set when the corresponding bit in EVE is written to one.
275
32099I–01/2012
AT32UC3L016/32/64
14.6.18
Name:
Calendar Value
CALV
Access Type:
Read/Write
Offset:
0x54
Reset Value:
0x00000000
31
30
29
28
27
26
25
YEAR
23
22
21
MONTH[3:2]
20
MONTH[1:0]
15
19
18
17
DAY
14
13
12
6
16
HOUR[4]
11
10
HOUR[3:0]
7
24
9
8
1
0
MIN[5:2]
5
4
3
MIN[1:0]
2
SEC
When the SR.BUSY bit is set writes to this register will be discarded and this register will read as zero.
• YEAR: Year
Current year. The year is considered a leap year if YEAR[1:0] = 0.
• MONTH: Month
1 = January
2 = February
...
12 = December
• DAY: Day
Day of month, starting with 1.
• HOUR: Hour
Hour of day, in 24-hour clock format.
Legal values are 0 through 23.
• MIN: Minute
Minutes, 0 through 59.
• SEC: Second
Seconds, 0 through 59.
276
32099I–01/2012
AT32UC3L016/32/64
14.6.19
Name:
Parameter Register
PARAMETER
Access Type:
Read-only
Offset:
0xF0
Reset Value:
-
31
30
29
-
-
-
23
22
21
-
-
-
15
14
13
12
11
10
PIR1WA
PIR0WA
-
NUMPIR
-
-
7
6
5
4
3
2
-
•
•
•
•
•
•
•
28
27
26
25
24
17
16
9
8
PER1VALUE
20
19
18
PER0VALUE
DTEXPVALUE
NUMAR
1
0
DTEXPWA
DT
This register gives the configuration used in the specific device. Also refer to the Module Configuration section.
DT: Digital Tuner
0: Digital tuner not implemented.
1: Digital tuner implemented.
DTREXPWA: Digital Tuner Exponent Writeable
0: Digital tuner exponent is a constant value. Writes to EXP field in DTR will be discarded.
1: Digital tuner exponent is chosen by writing to EXP field in DTR.
DTREXPVALUE: Digital Tuner Exponent Value
Digital tuner exponent value if DTEXPWA is zero.
NUMAR: Number of Alarm Comparators
0: Zero alarm comparators.
1: One alarm comparator.
2: Two alarm comparators.
NUMPIR: Number of Periodic Comparators
0: One periodic comparator.
1: Two periodic comparator.
PIRnWA: Periodic Interval n Writeable
0: Periodic interval n prescaler tapping is a constant value. Writes to INSEL field in PIRn register will be discarded.
1: Periodic interval n prescaler tapping is chosen by writing to INSEL field in PIRn register.
PERnVALUE: Periodic Interval n Value
Periodic interval prescaler n tapping if PIRnWA is zero.
277
32099I–01/2012
AT32UC3L016/32/64
14.6.20
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the module. No functionality associated.
278
32099I–01/2012
AT32UC3L016/32/64
14.7
Module Configuration
The specific configuration for each AST instance is listed in the following tables.The module bus
clocks listed here are connected to the system bus clocks. Please refer to the Power Manager
chapter for details.
Table 14-3.
AST Configuration
Feature
AST
Number of alarm comparators
1
Number of periodic comparators
1
Digital tuner
On
Table 14-4.
AST Clocks
Clock Name
Description
CLK_AST
Clock for the AST bus interface
GCLK
The generic clock used for the AST is GCLK2
PB clock
Peripheral Bus clock from the PBA clock domain
Table 14-5.
Register Reset Values
Register
Reset Value
VERSION
0x00000300
PARAMETER
0x00004103
279
32099I–01/2012
AT32UC3L016/32/64
15. Watchdog Timer (WDT)
Rev: 4.0.2.0
15.1
Features
•
•
•
•
•
15.2
Watchdog Timer counter with 32-bit counter
Timing window watchdog
Clocked from system RC oscillator or the 32 KHz crystal oscillator
Configuration lock
WDT may be enabled at reset by a fuse
Overview
The Watchdog Timer (WDT) will reset the device unless it is periodically serviced by the software. This allows the device to recover from a condition that has caused the system to be
unstable.
The WDT has an internal counter clocked from the system RC oscillator or the 32 KHz crystal
oscillator.
The WDT counter must be periodically cleared by software to avoid a watchdog reset. If the
WDT timer is not cleared correctly, the device will reset and start executing from the boot vector.
15.3
Block Diagram
Figure 15-1. WDT Block Diagram
PB
PB Clock Domain
CLR
SR
CTRL
WDTCLR
WINDOW,
CLEARED
EN, MODE,
PSEL, TBAN
SYNC
RCSYS
0
OSC32K
1
CLK_CNT
32-bit Counter
Watchdog
Detector
Watchdog
Reset
CEN
CLK_CNT Domain
CSSEL
15.4
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
280
32099I–01/2012
AT32UC3L016/32/64
15.4.1
Power Management
When the WDT is enabled, the WDT remains clocked in all sleep modes. It is not possible to
enter sleep modes where the source clock of CLK_CNT is stopped. Attempting to do so will
result in the device entering the lowest sleep mode where the source clock is running, leaving
the WDT operational. Please refer to the Power Manager chapter for details about sleep modes.
After a watchdog reset the WDT bit in the Reset Cause Register (RCAUSE) in the Power Manager will be set.
15.4.2
Clocks
The clock for the WDT bus interface (CLK_WDT) is generated by the Power Manager. This
clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the WDT before disabling the clock, to avoid freezing the WDT in an undefined state.
There are two possible clock sources for the Watchdog Timer (CLK_CNT):
• System RC oscillator (RCSYS): This oscillator is always enabled when selected as clock
source for the WDT. Please refer to the Power Manager chapter for details about the RCSYS
and sleep modes. Please refer to the Electrical Characteristics chapter for the characteristic
frequency of this oscillator.
• 32 KHz crystal oscillator (OSC32K): This oscillator has to be enabled in the System Control
Interface before using it as clock source for the WDT. The WDT will not be able to detect if
this clock is stopped.
15.4.3
Debug Operation
The WDT counter is frozen during debug operation, unless the Run In Debug bit in the Development Control Register is set and the bit corresponding to the WDT is set in the Peripheral Debug
Register (PDBG). Please refer to the On-Chip Debug chapter in the AVR32UC Technical Reference Manual, and the OCD Module Configuration section, for details. If the WDT counter is not
frozen during debug operation it will need periodically clearing to avoid a watchdog reset.
15.4.4
Fuses
The WDT can be enabled at reset. This is controlled by the WDTAUTO fuse, see Section 15.5.4
for details. Please refer to the Fuse Settings section in the Flash Controller chapter for details
about WDTAUTO and how to program the fuses.
15.5
Functional Description
15.5.1
Basic Mode
15.5.1.1
WDT Control Register Access
To avoid accidental disabling of the watchdog, the Control Register (CTRL) must be written
twice, first with the KEY field set to 0x55, then 0xAA without changing the other bits. Failure to
do so will cause the write operation to be ignored, and the value in the CTRL Register will not be
changed.
15.5.1.2
Changing CLK_CNT Clock Source
After any reset, except for watchdog reset, CLK_CNT will be enabled with the RCSYS as
source.
281
32099I–01/2012
AT32UC3L016/32/64
To change the clock for the WDT the following steps need to be taken. Note that the WDT
should always be disabled before changing the CLK_CNT source:
1. Write a zero to the Clock Enable (CEN) bit in the CTRL Register, leaving the other bits as they
are in the CTRL Register. This will stop CLK_CNT.
2. Read back the CTRL Register until the CEN bit reads zero. The clock has now been stopped.
3. Modify the Clock Source Select (CSSEL) bit in the CTRL Register with your new clock selection and write it to the CTRL Register.
4. Write a one to the CEN bit, leaving the other bits as they are in the CTRL Register. This will
enable the clock.
5. Read back the CTRL Register until the CEN bit reads one. The clock has now been enabled.
15.5.1.3
Configuring the WDT
If the MODE bit in the CTRL Register is zero, the WDT is in basic mode. The Time Out Prescale
Select (PSEL) field in the CTRL Register selects the WDT timeout period:
Ttimeout = Tpsel = 2(PSEL+1) / fclk_cnt
15.5.1.4
Enabling the WDT
To enable the WDT write a one to the Enable (EN) bit in the CTRL Register. Due to internal synchronization, it will take some time for the CTRL.EN bit to read back as one.
15.5.1.5
Clearing the WDT Counter
The WDT counter is cleared by writing a one to the Watchdog Clear (WDTCLR) bit in the Clear
(CLR) Register, at any correct write to the CTRL Register, or when the counter reaches Ttimeout
and the device is reset. In basic mode the CLR.WDTCLR can be written at any time when the
WDT Counter Cleared (CLEARED) bit in the Status Register (SR) is one. Due to internal synchronization, clearing the WDT counter takes some time. The SR.CLEARED bit is cleared when
writing to CLR.WDTCLR bit and set when the clearing is done. Any write to the CLR.WDTCLR
bit while SR.CLEARED is zero will not clear the counter.
Writing to the CLR.WDTCLR bit has to be done in a particular sequence to be valid. The CLR
Register must be written twice, first with the KEY field set to 0x55 and WDTCLR set to one, then
a second write with the KEY set to 0xAA without changing the WDTCLR bit. Writing to the CLR
Register without the correct sequence has no effect.
If the WDT counter is periodically cleared within Tpsel no watchdog reset will be issued, see Figure 15-2 on page 283.
282
32099I–01/2012
AT32UC3L016/32/64
Figure 15-2. Basic Mode WDT Timing Diagram, normal operation.
t= t 0
T psel
T im e o u t
W rite o n e to
C L R .W D T C L R
W a tc h d o g re s e t
If the WDT counter is not cleared within Tpsel a watchdog reset will be issued at the end of Tpsel,
see Figure 15-3 on page 283.
Figure 15-3. Basic Mode WDT Timing Diagram, no clear within Tpsel.
t= t 0
T psel
T im e o u t
W rite o ne to
C L R .W D T C L R
W a tch d o g re se t
15.5.1.6
Watchdog Reset
A watchdog reset will result in a reset and the code will start executing from the boot vector,
please refer to the Power Manager chapter for details. If the Disable After Reset (DAR) bit in the
CTRL Register is zero, the WDT counter will restart counting from zero when the watchdog reset
is released.
If the CTRL.DAR bit is one the WDT will be disabled after a watchdog reset. Only the CTRL.EN
bit will be changed after the watchdog reset. However, if WDTAUTO fuse is configured to enable
the WDT after a watchdog reset, and the CTRL.FCD bit is zero, writing a one to the CTRL.DAR
bit will have no effect.
15.5.2
Window Mode
The window mode can protect against tight loops of runaway code. This is obtained by adding a
ban period to timeout period. During the ban period clearing the WDT counter is not allowed.
If the WDT Mode (MODE) bit in the CTRL Register is one, the WDT is in window mode. Note
that the CTRL.MODE bit can only be changed when the WDT is disabled (CTRL.EN=0).
283
32099I–01/2012
AT32UC3L016/32/64
The PSEL and Time Ban Prescale Select (TBAN) fields in the CTRL Register selects the WDT
timeout period
Ttimeout = Ttban + Tpsel = (2(TBAN+1) + 2(PSEL+1)) / fclk_cnt
where Ttban sets the time period when clearing the WDT counter by writing to the CLR.WDTCLR
bit is not allowed. Doing so will result in a watchdog reset, the device will receive a reset and the
code will start executing form the boot vector, see Figure 15-5 on page 284. The WDT counter
will be cleared.
Writing a one to the CLR.WDTCLR bit within the Tpsel period will clear the WDT counter and the
counter starts counting from zero (t=t0), entering Ttban, see Figure 15-4 on page 284.
If the value in the CTRL Register is changed, the WDT counter will be cleared without a watchdog reset, regardless of if the value in the WDT counter and the TBAN value.
If the WDT counter reaches Ttimeout, the counter will be cleared, the device will receive a reset
and the code will start executing form the boot vector.
Figure 15-4. Window Mode WDT Timing Diagram
t= t 0
T tb a n
T pse l
T im e o u t
W rite o n e to
C L R .W D T C L R
W a tch d o g re se t
Figure 15-5. Window Mode WDT Timing Diagram, clearing within Ttban, resulting in watchdog reset.
t= t 0
T tb a n
T psel
T im e o u t
W rite o n e to
C L R .W D T C L R
W a tc h d o g re s e t
284
32099I–01/2012
AT32UC3L016/32/64
15.5.3
Disabling the WDT
The WDT is disabled by writing a zero to the CTRL.EN bit. When disabling the WDT no other
bits in the CTRL Register should be changed until the CTRL.EN bit reads back as zero. If the
CTRL.CEN bit is written to zero, the CTRL.EN bit will never read back as zero if changing the
value from one to zero.
15.5.4
Flash Calibration
The WDT can be enabled at reset. This is controlled by the WDTAUTO fuse. The WDT will be
set in basic mode, RCSYS is set as source for CLK_CNT, and PSEL will be set to a value giving
Tpsel above 100 ms. Please refer to the Fuse Settings chapter for details about WDTAUTO and
how to program the fuses.
If the Flash Calibration Done (FCD) bit in the CTRL Register is zero at a watchdog reset the
flash calibration will be redone, and the CTRL.FCD bit will be set when the calibration is done. If
CTRL.FCD is one at a watchdog reset, the configuration of the WDT will not be changed during
flash calibration. After any other reset the flash calibration will always be done, and the
CTRL.FCD bit will be set when the calibration is done.
15.5.5
Special Considerations
Care must be taken when selecting the PSEL/TBAN values so that the timeout period is greater
than the startup time of the device. Otherwise a watchdog reset will reset the device before any
code has been run. This can also be avoided by writing the CTRL.DAR bit to one when configuring the WDT.
If the Store Final Value (SFV) bit in the CTRL Register is one, the CTRL Register is locked for
further write accesses. All writes to the CTRL Register will be ignored. Once the CTRL Register
is locked, it can only be unlocked by a reset (e.g. POR, OCD, and WDT).
The CTRL.MODE bit can only be changed when the WDT is disabled (CTRL.EN=0).
285
32099I–01/2012
AT32UC3L016/32/64
15.6
User Interface
Table 15-1.
Note:
WDT Register Memory Map
Offset
Register
Register Name
Access
Reset
0x000
Control Register
CTRL
Read/Write
0x00010080
0x004
Clear Register
CLR
Write-only
0x00000000
0x008
Status Register
SR
Read-only
0x00000003
0x3FC
Version Register
VERSION
Read-only
-(1)
1. The reset value for this register is device specific. Please refer to the Module Configuration section at the end of this chapter.
286
32099I–01/2012
AT32UC3L016/32/64
15.6.1
Name:
Control Register
CTRL
Access Type:
Read/Write
Offset:
0x000
Reset Value:
0x00010080
31
30
29
28
27
26
25
24
19
18
17
16
CSSEL
CEN
9
8
KEY
23
22
21
-
20
TBAN
15
14
13
12
11
10
-
-
-
7
6
5
4
3
2
1
0
FCD
-
-
-
SFV
MODE
DAR
EN
PSEL
• KEY
•
•
•
•
•
•
•
This field must be written twice, first with key value 0x55, then 0xAA, for a write operation to be effective. This field always reads
as zero.
TBAN: Time Ban Prescale Select
Counter bit TBAN is used as watchdog “banned” time frame. In this time frame clearing the WDT timer is forbidden, otherwise a
watchdog reset is generated and the WDT timer is cleared.
CSSEL: Clock Source Select
0: Select the system RC oscillator (RCSYS) as clock source.
1: Select the 32KHz crystal oscillator (OSC32K) as clock source.
CEN: Clock Enable
0: The WDT clock is disabled.
1: The WDT clock is enabled.
PSEL: Time Out Prescale Select
Counter bit PSEL is used as watchdog timeout period.
FCD: Flash Calibration Done
This bit is set after any reset.
0: The flash calibration will be redone after a watchdog reset.
1: The flash calibration will not be redone after a watchdog reset.
SFV: WDT Control Register Store Final Value
0: WDT Control Register is not locked.
1: WDT Control Register is locked.
Once locked, the Control Register can not be re-written, only a reset unlocks the SFV bit.
MODE: WDT Mode
0: The WDT is in basic mode, only PSEL time is used.
1: The WDT is in window mode. Total timeout period is now TBAN+PSEL.
Writing to this bit when the WDT is enabled has no effect.
287
32099I–01/2012
AT32UC3L016/32/64
• DAR: WDT Disable After Reset
0: After a watchdog reset, the WDT will still be enabled.
1: After a watchdog reset, the WDT will be disabled.
• EN: WDT Enable
0: WDT is disabled.
1: WDT is enabled.
After writing to this bit the read back value will not change until the WDT is enabled/disabled. This due to internal
synchronization.
288
32099I–01/2012
AT32UC3L016/32/64
15.6.2
Name:
Clear Register
CLR
Access Type:
Write-only
Offset:
0x004
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
KEY
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
WDTCLR
When the Watchdog Timer is enabled, this Register must be periodically written within the window time frame or within the
watchdog timeout period, to prevent a watchdog reset.
• KEY
This field must be written twice, first with key value 0x55, then 0xAA, for a write operation to be effective.
• WDTCLR: Watchdog Clear
Writing a zero to this bit has no effect.
Writing a one to this bit clears the WDT counter.
289
32099I–01/2012
AT32UC3L016/32/64
15.6.3
Name:
Status Register
SR
Access Type:
Read-only
Offset:
0x008
Reset Value:
0x00000003
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
-
-
-
-
-
-
CLEARED
WINDOW
• CLEARED: WDT Counter Cleared
This bit is cleared when writing a one to the CLR.WDTCLR bit.
This bit is set when clearing the WDT counter is done.
• WINDOW: Within Window
This bit is cleared when the WDT counter is inside the TBAN period.
This bit is set when the WDT counter is inside the PSEL period.
290
32099I–01/2012
AT32UC3L016/32/64
15.6.4
Name:
Version Register
VERSION
Access Type:
Read-only
Offset:
0x3FC
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
291
32099I–01/2012
AT32UC3L016/32/64
15.7
Module Configuration
The specific configuration for each WDT instance is listed in the following tables.The module bus
clocks listed here are connected to the system bus clocks. Please refer to the Power Manager
chapter for details.
Table 15-2.
WDT Clocks
Clock Name
Description
CLK_WDT
Clock for the WDT bus interface
Table 15-3.
Register Reset Values
Register
Reset Value
VERSION
0x00000402
292
32099I–01/2012
AT32UC3L016/32/64
16. External Interrupt Controller (EIC)
Rev: 3.0.1.0
16.1
Features
•
•
•
•
•
•
•
16.2
Dedicated interrupt request 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
Non-Maskable NMI interrupt
Overview
The External Interrupt Controller (EIC) allows pins to be configured as external interrupts. Each
external interrupt has its own interrupt request and can be individually masked. Each external
interrupt can generate an interrupt on rising or falling edge, or high or low level. Every interrupt
input has a configurable filter to remove spikes from the interrupt source. Every interrupt pin can
also be configured to be asynchronous in order to wake up the part from sleep modes where the
CLK_SYNC 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 EIC 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.
16.3
Block Diagram
Figure 16-1. EIC Block Diagram
EN
D IS
E n a b le
LEVEL
MODE
EDGE
ASYNC
P o la rity
co n tro l
A syn ch ro n u s
d e te cto r
F IL T E R
LEVEL
MODE
EDGE
F ilte r
E d g e/L e ve l
D e te cto r
E X T IN T n
NMI
CTRL
CLK_SYNC
IC R
CTRL
IE R
ID R
IN T n
M a sk
IS R
IM R
W a ke
d e te ct
IR Q n
E IC _ W A K E
293
32099I–01/2012
AT32UC3L016/32/64
16.4
I/O Lines Description
Table 16-1.
16.5
I/O Lines Description
Pin Name
Pin Description
Type
NMI
Non-Maskable Interrupt
Input
EXTINTn
External Interrupt
Input
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
16.5.1
I/O Lines
The external interrupt pins (EXTINTn and NMI) may be multiplexed with I/O Controller lines. The
programmer must first program the I/O Controller to assign the desired EIC pins to their peripheral function. If I/O lines of the EIC are not used by the application, they can be used for other
purposes by the I/O Controller.
It is only required to enable the EIC inputs actually in use. If an application requires two external
interrupts, then only two I/O lines will be assigned to EIC inputs.
16.5.2
Power Management
All interrupts are available in all sleep modes as long as the EIC module is powered. However, in
sleep modes where CLK_SYNC is stopped, the interrupt must be configured to asynchronous
mode.
16.5.3
Clocks
The clock for the EIC bus interface (CLK_EIC) is generated by the Power Manager. This clock is
enabled at reset, and can be disabled in the Power Manager.
The filter and synchronous edge/level detector runs on a clock which is stopped in any of the
sleep modes where the system RC oscillator (RCSYS) is not running. This clock is referred to as
CLK_SYNC.
16.5.4
Interrupts
The external interrupt request lines are connected to the interrupt controller. Using the external
interrupts requires the interrupt controller to be programmed first.
Using the Non-Maskable Interrupt does not require the interrupt controller to be programmed.
16.5.5
Debug Operation
When an external debugger forces the CPU into debug mode, the EIC continues normal operation. If the EIC is configured in a way that requires it to be periodically serviced by the CPU
through interrupts or similar, improper operation or data loss may result during debugging.
16.6
16.6.1
Functional Description
External Interrupts
The external interrupts are not enabled by default, allowing the proper interrupt vectors to be set
up by the CPU before the interrupts are enabled.
294
32099I–01/2012
AT32UC3L016/32/64
Each external interrupt INTn 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 has a bit INTn in each of these registers. Writing a zero to the INTn bit in the
MODE register enables edge triggered interrupts, while writing a one to the bit enables level triggered interrupts.
If INTn is configured as an edge triggered interrupt, writing a zero to the INTn bit in the EDGE
register will cause the interrupt to be triggered on a falling edge on EXTINTn, while writing a one
to the bit will cause the interrupt to be triggered on a rising edge on EXTINTn.
If INTn is configured as a level triggered interrupt, writing a zero to the INTn bit in the LEVEL
register will cause the interrupt to be triggered on a low level on EXTINTn, while writing a one to
the bit will cause the interrupt to be triggered on a high level on EXTINTn.
Each interrupt has a corresponding bit in each of the interrupt control and status registers. Writing a one to the INTn bit in the Interrupt Enable Register (IER) enables the external interrupt
from pin EXTINTn to propagate from the EIC to the interrupt controller, while writing a one to
INTn bit in the Interrupt Disable Register (IDR) disables this propagation. The Interrupt Mask
Register (IMR) can be read to check which interrupts are enabled. When an interrupt triggers,
the corresponding bit in the Interrupt Status Register (ISR) will be set. This bit remains set until a
one is written to the corresponding bit in the Interrupt Clear Register (ICR) or the interrupt is
disabled.
Writing a one to the INTn bit in the Enable Register (EN) enables the external interrupt on pin
EXTINTn, while writing a one to INTn bit in the Disable Register (DIS) disables the external interrupt. The Control Register (CTRL) can be read to check which interrupts are enabled. If a bit in
the CTRL register is set, but the corresponding bit in IMR is not set, an interrupt will not propagate to the interrupt controller. However, the corresponding bit in ISR will be set, and
EIC_WAKE will be set. Note that an external interrupt should not be enabled before it has been
configured correctly.
If the CTRL.INTn bit is zero, the corresponding bit in ISR will always be zero. Disabling an external interrupt by writing a one to the DIS.INTn bit will clear the corresponding bit in ISR.
Please refer to the Module Configuration section for the number of external interrupts.
16.6.2
Synchronization and Filtering of External Interrupts
In synchronous mode the pin value of the EXTINTn pin is synchronized to CLK_SYNC, so
spikes shorter than one CLK_SYNC cycle are not guaranteed to produce an interrupt. The synchronization of the EXTINTn to CLK_SYNC will delay the propagation of the interrupt to the
interrupt controller by two cycles of CLK_SYNC, see Figure 16-2 and Figure 16-3 for examples
(FILTER off).
It is also possible to apply a filter on EXTINTn by writing a one to the INTn bit in the FILTER register. This filter is a majority voter, if the condition for an interrupt is true for more than one of the
latest three cycles of CLK_SYNC the interrupt will be set. This will additionally delay the propagation of the interrupt to the interrupt controller by one or two cycles of CLK_SYNC, see Figure
16-2 and Figure 16-3 for examples (FILTER on).
295
32099I–01/2012
AT32UC3L016/32/64
Figure 16-2. Timing Diagram, Synchronous Interrupts, High Level or Rising Edge
CLK_SYNC
EXTINTn/NMI
ISR.INTn:
FILTER off
ISR.INTn:
FILTER on
Figure 16-3. Timing Diagram, Synchronous Interrupts, Low Level or Falling Edge
CLK_SYNC
EXTINTn/NMI
ISR.INTn:
FILTER off
ISR.INTn:
FILTER on
16.6.3
Non-Maskable Interrupt
The NMI supports the same features as the external interrupts, and is accessed through the
same registers. The description in Section 16.6.1 should be followed, accessing the NMI bit
instead of the INTn bits.
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 can be enabled and disabled
by accessing the registers in the EIC.
16.6.4
Asynchronous Interrupts
Each external interrupt can be made asynchronous by writing a one 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. If an interrupt is configured as edge triggered interrupt in asynchronous mode, a zero in EDGE.INTn will be interpreted
as low level, and a one in EDGE.INTn will be interpreted as high level.
EIC_WAKE will be set immediately after the source triggers the interrupt, while the corresponding bit in ISR and the interrupt to the interrupt controller will be set on the next rising edge of
CLK_SYNC. Please refer to Figure 16-4 on page 297 for details.
296
32099I–01/2012
AT32UC3L016/32/64
When CLK_SYNC is stopped only asynchronous interrupts remain active, and any short spike
on this interrupt will wake up the device. EIC_WAKE will restart CLK_SYNC and ISR will be
updated on the first rising edge of CLK_SYNC.
Figure 16-4. Timing Diagram, Asynchronous Interrupts
C LK _SYN C
CLK_SYNC
E X T IN T n /N M I
16.6.5
E X T IN T n /N M I
IS R .IN T n :
ris in g E D G E o r h ig h
LEVEL
IS R .IN T n :
ris in g E D G E o r h ig h
LEVEL
E IC _ W A K E :
ris in g E D G E o r h ig h
LEVEL
E IC _ W A K E :
ris in g E D G E o r h ig h
LEVEL
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 one, then the execution starts at the
interrupt handler for this interrupt. If the bit in IMR is zero, then the execution starts from the next
instruction after the sleep instruction.
297
32099I–01/2012
AT32UC3L016/32/64
16.7
User Interface
Table 16-2.
Note:
EIC Register Memory Map
Offset
Register
Register Name
Access
Reset
0x000
Interrupt Enable Register
IER
Write-only
0x00000000
0x004
Interrupt Disable Register
IDR
Write-only
0x00000000
0x008
Interrupt Mask Register
IMR
Read-only
0x00000000
0x00C
Interrupt Status Register
ISR
Read-only
0x00000000
0x010
Interrupt Clear Register
ICR
Write-only
0x00000000
0x014
Mode Register
MODE
Read/Write
0x00000000
0x018
Edge Register
EDGE
Read/Write
0x00000000
0x01C
Level Register
LEVEL
Read/Write
0x00000000
0x020
Filter Register
FILTER
Read/Write
0x00000000
0x024
Test Register
TEST
Read/Write
0x00000000
0x028
Asynchronous Register
ASYNC
Read/Write
0x00000000
0x030
Enable Register
EN
Write-only
0x00000000
0x034
Disable Register
DIS
Write-only
0x00000000
0x038
Control Register
CTRL
Read-only
0x00000000
0x3FC
Version Register
VERSION
Read-only
- (1)
1. The reset value is device specific. Please refer to the Module Configuration section at the end of this chapter.
298
32099I–01/2012
AT32UC3L016/32/64
16.7.1
Name:
Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x000
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
Writing a zero to this bit has no effect.
Writing a one to this bit will set the corresponding bit in IMR.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
Writing a zero to this bit has no effect.
Wrting a one to this bit will set the corresponding bit in IMR.
299
32099I–01/2012
AT32UC3L016/32/64
16.7.2
Name:
Interrupt Disable Register
IDR
Access Type:
Write-only
Offset:
0x004
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
Writing a zero to this bit has no effect.
Writing a one to this bit will clear the corresponding bit in IMR.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
Writing a zero to this bit has no effect.
Writing a one to this bit will clear the corresponding bit in IMR.
300
32099I–01/2012
AT32UC3L016/32/64
16.7.3
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
Offset:
0x008
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
This bit is cleared when the corresponding bit in IDR is written to one.
This bit is set when the corresponding bit in IER is written to one.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
0: The Non-Maskable Interrupt is disabled.
1: The Non-Maskable Interrupt is enabled.
This bit is cleared when the corresponding bit in IDR is written to one.
This bit is set when the corresponding bit in IER is written to one.
301
32099I–01/2012
AT32UC3L016/32/64
16.7.4
Name:
Interrupt Status Register
ISR
Access Type:
Read-only
Offset:
0x00C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
0: An interrupt event has not occurred.
1: An interrupt event has occurred.
This bit is cleared by writing a one to the corresponding bit in ICR.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
0: An interrupt event has not occurred.
1: An interrupt event has occurred.
This bit is cleared by writing a one to the corresponding bit in ICR.
302
32099I–01/2012
AT32UC3L016/32/64
16.7.5
Name:
Interrupt Clear Register
ICR
Access Type:
Write-only
Offset:
0x010
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
Writing a zero to this bit has no effect.
Writing a one to this bit will clear the corresponding bit in ISR.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
Writing a zero to this bit has no effect.
Writing a one to this bit will clear the corresponding bit in ISR.
303
32099I–01/2012
AT32UC3L016/32/64
16.7.6
Name:
Mode Register
MODE
Access Type:
Read/Write
Offset:
0x014
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
0: The external interrupt is edge triggered.
1: The external interrupt is level triggered.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
0: The Non-Maskable Interrupt is edge triggered.
1: The Non-Maskable Interrupt is level triggered.
304
32099I–01/2012
AT32UC3L016/32/64
16.7.7
Name:
Edge Register
EDGE
Access Type:
Read/Write
Offset:
0x018
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
0: The external interrupt triggers on falling edge.
1: The external interrupt triggers on rising edge.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
0: The Non-Maskable Interrupt triggers on falling edge.
1: The Non-Maskable Interrupt triggers on rising edge.
305
32099I–01/2012
AT32UC3L016/32/64
16.7.8
Name:
Level Register
LEVEL
Access Type:
Read/Write
Offset:
0x01C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
0: The external interrupt triggers on low level.
1: The external interrupt triggers on high level.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
0: The Non-Maskable Interrupt triggers on low level.
1: The Non-Maskable Interrupt triggers on high level.
306
32099I–01/2012
AT32UC3L016/32/64
16.7.9
Filter Register
Name:
FILTER
Access Type:
Read/Write
Offset:
0x020
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
0: The external interrupt is not filtered.
1: The external interrupt is filtered.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
0: The Non-Maskable Interrupt is not filtered.
1: The Non-Maskable Interrupt is filtered.
307
32099I–01/2012
AT32UC3L016/32/64
16.7.10
Test Register
Name:
TEST
Access Type:
Read/Write
Offset:
0x024
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
TESTEN
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• TESTEN: Test Enable
0: This bit disables external interrupt test mode.
1: This bit enables external interrupt test mode.
• INTn: External Interrupt n
Writing a zero to this bit will set the input value to INTn to zero, if test mode is enabled.
Writing a one to this bit will set the input value to INTn to one, if test mode is enabled.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
Writing a zero to this bit will set the input value to NMI to zero, if test mode is enabled.
Writing a one to this bit will set the input value to NMI to one, if test mode is enabled.
If TESTEN is 1, the value written to this bit will be the value to the interrupt detector and the value on the pad will be ignored.
308
32099I–01/2012
AT32UC3L016/32/64
16.7.11
Asynchronous Register
Name:
ASYNC
Access Type:
Read/Write
Offset:
0x028
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
0: The external interrupt is synchronized to CLK_SYNC.
1: The external interrupt is asynchronous.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
0: The Non-Maskable Interrupt is synchronized to CLK_SYNC.
1: The Non-Maskable Interrupt is asynchronous.
309
32099I–01/2012
AT32UC3L016/32/64
16.7.12
Enable Register
Name:
EN
Access Type:
Write-only
Offset:
0x030
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
Writing a zero to this bit has no effect.
Writing a one to this bit will enable the corresponding external interrupt.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
Writing a zero to this bit has no effect.
Writing a one to this bit will enable the Non-Maskable Interrupt.
310
32099I–01/2012
AT32UC3L016/32/64
16.7.13
Disable Register
Name:
DIS
Access Type:
Write-only
Offset:
0x034
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
Writing a zero to this bit has no effect.
Writing a one to this bit will disable the corresponding external interrupt.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
Writing a zero to this bit has no effect.
Writing a one to this bit will disable the Non-Maskable Interrupt.
311
32099I–01/2012
AT32UC3L016/32/64
16.7.14
Control Register
Name:
CTRL
Access Type:
Read-only
Offset:
0x038
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
INT30
INT29
INT28
INT27
INT26
INT25
INT24
23
22
21
20
19
18
17
16
INT23
INT22
INT21
INT20
INT19
INT18
INT17
INT16
15
14
13
12
11
10
9
8
INT15
INT14
INT13
INT12
INT11
INT10
INT9
INT8
7
6
5
4
3
2
1
0
INT7
INT6
INT5
INT4
INT3
INT2
INT1
NMI
• INTn: External Interrupt n
0: The corresponding external interrupt is disabled.
1: The corresponding external interrupt is enabled.
Please refer to the Module Configuration section for the number of external interrupts.
• NMI: Non-Maskable Interrupt
0: The Non-Maskable Interrupt is disabled.
1: The Non-Maskable Interrupt is enabled.
312
32099I–01/2012
AT32UC3L016/32/64
16.7.15
Name:
Version Register
VERSION
Access Type:
Read-only
Offset:
0x3FC
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
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VERSION: Version number
Version number of the module. No functionality associated.
313
32099I–01/2012
AT32UC3L016/32/64
16.8
Module Configuration
The specific configuration for each EIC instance is listed in the following tables.The module bus
clocks listed here are connected to the system bus clocks. Please refer to the Power Manager
chapter for details.
Table 16-3.
Module Configuration
Feature
EIC
Number of external interrupts, including NMI
6
Table 16-4.
EIC Clocks
Clock Name
Description
CLK_EIC
Clock for the EIC bus interface
Table 16-5.
Register Reset Values
Register
Reset Value
VERSION
0x00000301
314
32099I–01/2012
AT32UC3L016/32/64
17. Frequency Meter (FREQM)
Rev: 3.0.1.1
17.1
Features
•
•
•
•
17.2
Accurately measures a clock frequency
Selectable reference clock
A selectable clock can be measured
Ratio can be measured with 24-bit accuracy
Overview
The Frequency Meter (FREQM) can be used to accurately measure the frequency of a clock by
comparing it to a known reference clock.
17.3
Block Diagram
Figure 17-1. Frequency Meter Block Diagram
CLKSEL
START
CLK_MSR
Counter
VALUE
CLK_REF
Timer
Trigger
REFSEL
17.4
ISR
REFNUM,
START
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
17.4.1
Power Management
The device can enter a sleep mode while a measurement is ongoing. However, make sure that
neither CLK_MSR nor CLK_REF is stopped in the actual sleep mode. FREQM interrupts can
wake up the device from sleep modes when the measurement is done, but only from sleep
modes where CLK_FREQM is running. Please refer to the Power Manager chapter for details.
315
32099I–01/2012
AT32UC3L016/32/64
17.4.2
Clocks
The clock for the FREQM bus interface (CLK_FREQM) is generated by the Power Manager.
This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to
disable the FREQM before disabling the clock, to avoid freezing the FREQM ia an undefined
state.
A set of clocks can be selected as reference (CLK_REF) and another set of clocks can be
selected for measurement (CLK_MSR). Please refer to the CLKSEL and REFSEL tables in the
Module Configuration section for details.
17.4.3
Debug Operation
When an external debugger forces the CPU into debug mode, the FREQM continues normal
operation. If the FREQM is configured in a way that requires it to be periodically serviced by the
CPU through interrupts or similar, improper operation or data loss may result during debugging.
17.4.4
Interrupts
The FREQM interrupt request line is connected to the internal source of the interrupt controller.
Using the FREQM interrupt requires the interrupt controller to be programmed first.
17.5
Functional Description
The FREQM accuratly measures the frequency of a clock by comparing the frequency to a
known frequency:
fCLK_MSR = (VALUE/REFNUM)*fCLK_REF
17.5.1
Reference Clock
The Reference Clock Selection (REFSEL) field in the Mode Register (MODE) selects the clock
source for CLK_REF. The reference clock is enabled by writing a one to the Reference Clock
Enable (REFCEN) bit in the Mode Register. This clock should have a known frequency.
CLK_REF needs to be disabled before switching to another clock. The RCLKBUSY bit in the
Status Register (SR) indicates whether the clock is busy or not. This bit is set when the
MODE.REFCEN bit is written.
To change CLK_REF:
• Write a zero to the MODE.REFCEN bit to disable the clock, without changing the other
bits/fields in the Mode Register.
• Wait until the SR.RCLKBUSY bit reads as zero.
• Change the MODE.REFSEL field.
• Write a one to the MODE.REFCEN bit to enable the clock, without changing the other
bits/fields in the Mode Register.
• Wait until the SR.RCLKBUSY bit reads as zero.
To enable CLK_REF:
• Write the correct value to the MODE.REFSEL field.
• Write a one to the MODE.REFCEN to enable the clock, without changing the other bits/fields
in the Mode Register.
• Wait until the SR.RCLKBUSY bit reads as zero.
To disable CLK_REF:
316
32099I–01/2012
AT32UC3L016/32/64
• Write a zero to the MODE.REFCEN to disable he clock, without changing the other bits/fields
in the Mode register.
• Wait until the SR.RCLKBUSY bit reads as zero.
17.5.1.1
17.5.2
Cautionary note
Note that if clock selected as source for CLK_REF is stopped during a measurement, this will
not be detected by the FREQM. The BUSY bit in the STATUS register will never be cleared, and
the DONE interrupt will never be triggered. If the clock selected as soruce for CLK_REF is
stopped, it will not be possible to change the source for the reference clock as long as the
selected source is not running.
Measurement
In the Mode Register the Clock Source Selection (CLKSEL) field selects CLK_MSR and the
Number of Reference Clock Cycles (REFNUM) field selects the duration of the measurement.
The duration is given in number of CLK_REF periodes.
Writing a one to the START bit in the Control Register (CTRL) starts the measurement. The
BUSY bit in SR is cleared when the measurement is done.
The result of the measurement can be read from the Value Register (VALUE). The frequency of
the measured clock CLK_MSR is then:
fCLK_MSR = (VALUE/REFNUM)*fCLK_REF
17.5.3
Interrupts
The FREQM has two interrupt sources:
• DONE: A frequency measurement is done
• RCLKRDY: The reference clock is ready
These will generate an interrupt request if the corresponding bit in the Interrupt Mask Register
(IMR) is set. The interrupt sources are ORed together to form one interrupt request. The FREQM
will generate an interrupt request if at least one of the bits in the Interrupt Mask Register (IMR) is
set. Bits in IMR are set by writing a one to the corresponding bit in the Interrupt Enable Register
(IER) and cleared by writing a one to this bit in the Interrupt Disable Register (IDR). The interrupt
request remains active until the corresponding bit in the Interrupt Status Register (ISR) is
cleared by writing a one to this bit in the Interrupt Clear Register (ICR). Because all the interrupt
sources are ORed together, the interrupt request from the FREQM will remain active until all the
bits in ISR are cleared.
317
32099I–01/2012
AT32UC3L016/32/64
17.6
User Interface
Table 17-1.
Note:
FREQM Register Memory Map
Offset
Register
Register Name
Access
Reset
0x000
Control Register
CTRL
Write-only
0x00000000
0x004
Mode Register
MODE
Read/Write
0x00000000
0x008
Status Register
STATUS
Read-only
0x00000000
0x00C
Value Register
VALUE
Read-only
0x00000000
0x010
Interrupt Enable Register
IER
Write-only
0x00000000
0x014
Interrupt Disable Register
IDR
Write-only
0x00000000
0x018
Interrupt Mask Register
IMR
Read-only
0x00000000
0x01C
Interrupt Status Register
ISR
Read-only
0x00000000
0x020
Interrupt Clear Register
ICR
Write-only
0x00000000
0x3FC
Version Register
VERSION
Read-only
-(1)
1. The reset value for this register is device specific. Please refer to the Module Configuration section at the end of this chapter.
318
32099I–01/2012
AT32UC3L016/32/64
17.6.1
Name:
Control Register
CTRL
Access Type:
Write-only
Offset:
0x000
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
-
-
-
-
-
-
-
START
• START
Writing a zero to this bit has no effect.
Writing a one to this bit will start a measurement.
319
32099I–01/2012
AT32UC3L016/32/64
17.6.2
Name:
Mode Register
MODE
Access Type:
Read/Write
Offset:
0x004
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
REFCEN
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
15
14
13
CLKSEL
12
11
10
9
8
2
1
0
REFNUM
7
6
5
4
3
-
-
-
-
-
REFSEL
• REFCEN: Reference Clock Enable
0: The reference clock is disabled
1: The reference clock is enabled
• CLKSEL: Clock Source Selection
Selects the source for CLK_MSR. See table in Module Configuration chapter for details.
• REFNUM: Number of Reference Clock Cycles
Selects the duration of a measurement, given in number of CLK_REF cycles.
• REFSEL: Reference Clock Selection
Selects the source for CLK_REF. See table in Module Configuration chapter for details.
320
32099I–01/2012
AT32UC3L016/32/64
17.6.3
Status Register
Name:
STATUS
Access Type:
Read-only
Offset:
0x008
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
-
-
-
-
-
-
RCLKBUSY
BUSY
• RCLKBUSY: FREQM Reference Clock Status
0: The FREQM ref clk is ready, so a measurement can start.
1: The FREQM ref clk is not ready, so a measurement should not be started.
• BUSY: FREQM Status
0: The Frequency Meter is idle.
1: Frequency measurement is on-going.
321
32099I–01/2012
AT32UC3L016/32/64
17.6.4
Value Register
Name:
VALUE
Access Type:
Read-only
Offset:
0x00C
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
VALUE[23:16]
15
14
13
12
VALUE[15:8]
7
6
5
4
VALUE[7:0]
• VALUE:
Result from measurement.
322
32099I–01/2012
AT32UC3L016/32/64
17.6.5
Interrupt Enable Register
Name:
IER
Access Type:
Write-only
Offset:
0x010
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
-
-
-
-
-
-
RCLKRDY
DONE
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
323
32099I–01/2012
AT32UC3L016/32/64
17.6.6
Interrupt Disable Register
Name:
IDR
Access Type:
Write-only
Offset:
0x014
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
-
-
-
-
-
-
RCLKRDY
DONE
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
324
32099I–01/2012
AT32UC3L016/32/64
17.6.7
Interrupt Mask Register
Name:
IMR
Access Type:
Read-only
Offset:
0x018
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
-
-
-
-
-
-
RCLKRDY
DONE
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
A bit in this register is cleared when the corresponding bit in IDR is written to one.
A bit in this register is set when the corresponding bit in IER is written to one.
325
32099I–01/2012
AT32UC3L016/32/64
17.6.8
Interrupt Status Register
Name:
ISR
Access Type:
Read-only
Offset:
0x01C
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
-
-
-
-
-
-
RCLKRDY
DONE
0: The corresponding interrupt is cleared.
1: The corresponding interrupt is pending.
A bit in this register is set when the corresponding bit in STATUS has a one to zero transition.
A bit in this register is cleared when the corresponding bit in ICR is written to one.
326
32099I–01/2012
AT32UC3L016/32/64
17.6.9
Interrupt Clear Register
Name:
ICR
Access Type:
Write-only
Offset:
0x020
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
-
-
-
-
-
-
RCLKRDY
DONE
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in ISR and the corresponding interrupt request.
327
32099I–01/2012
AT32UC3L016/32/64
17.6.10
Name:
Version Register
VERSION
Access Type:
Read-only
Offset:
0x3FC
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
328
32099I–01/2012
AT32UC3L016/32/64
17.7
Module Configuration
The specific configuration for each FREQM instance is listed in the following tables. The module
bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details.
Table 17-2.
FREQM Clock Name
Module Name
FREQM
Table 17-3.
Clock Name
Description
CLK_FREQM
Bus interface clock
CLK_MSR
Measured clock
CLK_REF
Reference clock
Register Reset Values
Register
Reset Value
VERSION
0x00000301
Table 17-4.
Clock Sources for CLK_MSR
CLKSEL
Clock/Oscillator
Description
0
CLK_CPU
The clock the CPU runs on
1
CLK_HSB
High Speed Bus clock
2
CLK_PBA
Peripheral Bus A clock
3
CLK_PBB
Peripheral Bus B clock
4
OSC0
Output clock from Oscillator 0
5
OSC32K
Output clock from OSC32K
6
RCSYS
Output clock from RCSYS Oscillator
7
DFLL0
Output clock from DFLL0
8
Reserved
9-14
GCLK0-5
Generic clock 0 through 5
15
RC120M AW clock
Output clock from RC120M to AW
16
RC120M
Output clock from RC120M to main clock mux
17
RC32K
Output clock from RC32K
18-31
Reserved
Table 17-5.
Clock Sources for CLK_REF
REFSEL
Clock/Oscillator
Description
0
RCSYS
System RC oscillator clock
1
OSC32K
Output clock form OSC32K
2-7
Reserved
329
32099I–01/2012
AT32UC3L016/32/64
330
32099I–01/2012
AT32UC3L016/32/64
18. General-Purpose Input/Output Controller (GPIO)
Rev: 2.1.1.5
18.1
Features
•
•
•
•
•
•
•
18.2
Configurable pin-change, rising-edge, or falling-edge interrupt
Configurable peripheral event generator
Glitch filter providing rejection of pulses shorter than one clock cycle
Input visibility and output control
Multiplexing of peripheral functions on I/O pins
Programmable internal pull-up resistor
Optional locking of configuration to avoid accidental reconfiguration
Overview
The General Purpose Input/Output Controller (GPIO) controls the I/O pins of the microcontroller.
Each GPIO pin may be used as a general-purpose I/O or be assigned to a function of an embedded peripheral.
The GPIO is configured using the Peripheral Bus (PB). Some registers can also be configured
using the low latency CPU Local Bus. See Section 18.6.2.7 for details.
18.3
Block Diagram
Figure 18-1. GPIO Block Diagram
Configuration
Interface
Interrupt
Controller
GPIO Interrupt
Request
PIN
General Purpose
Input/Output - GPIO
Power Manager
CLK_GPIO
PIN
PIN
PIN
MCU
I/O
Pins
PIN
Embedded
Peripheral
Pin Control
Signals
331
32099I–01/2012
AT32UC3L016/32/64
18.4
I/O Lines Description
Pin Name
Description
Type
GPIOn
GPIO pin n
Digital
18.5
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
18.5.1
Power Management
If the CPU enters a sleep mode that disables clocks used by the GPIO, the GPIO will stop functioning and resume operation after the system wakes up from sleep mode.
If a peripheral function is configured for a GPIO pin, the peripheral will be able to control the
GPIO pin even if the GPIO clock is stopped.
18.5.2
Clocks
The GPIO is connected to a Peripheral Bus clock (CLK_GPIO). This clock is generated by the
Power Manager. CLK_GPIO is enabled at reset, and can be disabled by writing to the Power
Manager. CLK_GPIO must be enabled in order to access the configuration registers of the GPIO
or to use the GPIO interrupts. After configuring the GPIO, the CLK_GPIO can be disabled by
writing to the Power Manager if interrupts are not used.
If the CPU Local Bus is used to access the configuration interface of the GPIO, the CLK_GPIO
must be equal to the CPU clock to avoid data loss.
18.5.3
Interrupts
The GPIO interrupt request lines are connected to the interrupt controller. Using the GPIO interrupts requires the interrupt controller to be programmed first.
18.5.4
Peripheral Events
The GPIO peripheral events are connected via the Peripheral Event System. Refer to the
Peripheral Event System chapter for details.
18.5.5
Debug Operation
When an external debugger forces the CPU into debug mode, the GPIO continues normal operation. If the GPIO is configured in a way that requires it to be periodically serviced by the CPU
through interrupts or similar, improper operation or data loss may result during debugging.
332
32099I–01/2012
AT32UC3L016/32/64
18.6
Functional Description
The GPIO controls the I/O pins of the microcontroller. The control logic associated with each pin
is shown in the figure below.
Figure 18-2. Overview of the GPIO
PUER*
ODER
1
0
Periph. Func. A
Output
Pullup
0
Periph.Func. B
Periph. Func. C
1
....
GPER
PMRn
Output
Enable
0
0
1
OVR
PIN
1
Input
PVR
IER
0
Edge Detector
1
1
Glitch Filter
IFR
IMR1
GFER
0
Interrupt Request
IMR0
*) Register value is overrided if a peripheral function that
support this function is enabled
333
32099I–01/2012
AT32UC3L016/32/64
18.6.1
Basic Operation
18.6.1.1
Module Configuration
The GPIO user interface registers are organized into ports and each port controls 32 different
GPIO pins. Most of the registers supports bit wise access operations such as set, clear and toggle in addition to the standard word access. For details regarding interface registers, refer to
Section 18.7.
18.6.1.2
Available Features
The GPIO features implemented are device dependent, and not all functions are implemented
on all pins. The user must refer to the Module Configuration section and the GPIO Function Multiplexing section in the Package and Pinout chapter for the device specific settings used in the
AT32UC3L016/32/64.
Device specific settings includes:
• Number of GPIO pins
• Functions implemented on each pin
• Peripheral function(s) multiplexed on each GPIO pin
• Reset state of registers
18.6.1.3
Inputs
The level on each GPIO pin can be read through the Pin Value Register (PVR). This register
indicates the level of the GPIO pins regardless of the pins being driven by the GPIO or by an
external component. Note that due to power saving measures, the PVR register will only be
updated when the corresponding bit in GPER is one or if an interrupt is enabled for the pin, i.e.
IER is one for the corresponding pin.
18.6.1.4
Output Control
When the GPIO pin is assigned to a peripheral function, i.e. the corresponding bit in GPER is
zero, the peripheral determines whether the pin is driven or not.
When the GPIO pin is controlled by the GPIO, the value of Output Driver Enable Register
(ODER) determines whether the pin is driven or not. When a bit in this register is one, the corresponding GPIO pin is driven by the GPIO. When the bit is zero, the GPIO does not drive the pin.
The level driven on a GPIO pin can be determined by writing the value to the corresponding bit
in the Output Value Register (OVR).
18.6.1.5
Peripheral Muxing
The GPIO allows a single GPIO pin to be shared by multiple peripheral pins and the GPIO itself.
Peripheral pins sharing the same GPIO pin are arranged into peripheral functions that can be
selected one at a time. Peripheral functions are configured by writing the selected function value
to the Peripheral Mux Registers (PMRn). To allow a peripheral pin access to the shared GPIO
pin, GPIO control must be disabled for that pin, i.e. the corresponding bit in GPER must read
zero.
A peripheral function value is set by writing bit zero to PMR0 and bit one to the same index position in PMR1 and so on. In a system with 4 peripheral functions A,B,C, and D, peripheral
function C for GPIO pin four is selected by writing a zero to bit four in PMR0 and a one to the
same bit index in PMR1. Refer to the GPIO Function Multiplexing chapter for details regarding
pin function configuration for each GPIO pin.
334
32099I–01/2012
AT32UC3L016/32/64
18.6.2
Advanced Operation
18.6.2.1
Peripheral I/O Pin Control
When a GPIO pin is assigned to a peripheral function, i.e. the corresponding bit in GPER is zero,
output and output enable is controlled by the selected peripheral pin. In addition the peripheral
may control some or all of the other GPIO pin functions listed in Table 18-1, if the peripheral supports those features. All pin features not controlled by the selected peripheral is controlled by the
GPIO.
Refer to the Module Configuration section for details regarding implemented GPIO pin functions
and to the Peripheral chapter for details regarding I/O pin function control.
Table 18-1.
18.6.2.2
I/O Pin function Control
Function name
GPIO mode
Peripheral mode
Output
OVR
Peripheral
Output enable
ODER
Peripheral
Pull-up
PUER
Peripheral if supported, else GPIO
Pull-up Resistor Control
Pull-up can be configured for each GPIO pin. Pull-up allows the pin and any connected net to be
pulled up to VDD if the net is not driven.
Pull-up is useful for detecting if a pin is unconnected or if a mechanical button is pressed, for various communication protocols and to keep unconnected pins from floating.
Pull-up can be enabled and disabled by writing a one and a zero respectively to the corresponding bit in the Pull-up Enable Register (PUER).
18.6.2.3
Output Pin Timings
Figure 18-3 shows the timing of the GPIO pin when writing to the Output Value Register (OVR).
The same timing applies when performing a ‘set’ or ‘clear’ access, i.e. writing to OVRS or
OVRC. The timing of PVR is also shown.
Figure 18-3. Output Pin Timings
CLK_GPIO
Write OVR to 1
Write OVR to 0
PB Access
PB Access
OVR / I/O Line
PVR
18.6.2.4
Interrupts
The GPIO can be configured to generate an interrupt when it detects a change on a GPIO pin.
Interrupts on a pin are enabled by writing a one to the corresponding bit in the Interrupt Enable
335
32099I–01/2012
AT32UC3L016/32/64
Register (IER). The module can be configured to generate an interrupt whenever a pin changes
value, or only on rising or falling edges. This is controlled by the Interrupt Mode Registers
(IMRn). Interrupts on a pin can be enabled regardless of the GPIO pin being controlled by the
GPIO or assigned to a peripheral function.
An interrupt can be generated on each GPIO pin. These interrupt generators are further grouped
into groups of eight and connected to the interrupt controller. An interrupt request from any of the
GPIO pin generators in the group will result in an interrupt request from that group to the interrupt controller if the corresponding bit for the GPIO pin in the IER is set. By grouping interrupt
generators into groups of eight, four different interrupt handlers can be installed for each GPIO
port.
The Interrupt Flag Register (IFR) can be read by software to determine which pin(s) caused the
interrupt. The interrupt flag must be manually cleared by writing a zero to the corresponding bit
in IFR.
GPIO interrupts will only be generated when CLK_GPIO is enabled.
18.6.2.5
Input Glitch Filter
Input glitch filters can be enabled on each GPIO pin. When the glitch filter is enabled, a glitch
with duration of less than 1 CLK_GPIO cycle is automatically rejected, while a pulse with duration of 2 CLK_GPIO cycles or more is accepted. For pulse durations between 1 and 2
CLK_GPIO 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 CLK_GPIO
cycles, whereas for a glitch to be reliably filtered out, its duration must not exceed 1 CLK_GPIO
cycle. The filter introduces 2 clock cycles latency.
The glitch filters are controlled by the Glitch Filter Enable Register (GFER). When a bit in GFER
is one, 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.
18.6.2.6
Interrupt Timings
Figure 18-4 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 sampled on a rising edge
and will trigger an interrupt request.
Figure 18-4. Interrupt Timing with Glitch Filter Disabled
CLK_GPIO
Pin Level
IFR
Figure 18-5 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.
336
32099I–01/2012
AT32UC3L016/32/64
Figure 18-5. Interrupt Timing with Glitch Filter Enabled
CLK_GPIO
Pin Level
IFR
18.6.2.7
CPU Local Bus
The CPU Local Bus can be used for application where low latency read and write access to the
Output Value Register (OVR) and Output Drive Enable Register (ODER) is required. The CPU
Local Bus allows the CPU to configure the mentioned GPIO registers directly, bypassing the
shared Peripheral Bus (PB).
To avoid data loss when using the CPU Local Bus, the CLK_GPIO must run at the same frequency as the CLK_CPU. See Section 18.5.2 for details.
The CPU Local Bus is mapped to a different base address than the GPIO but the OVER and
ODER offsets are the same. See the CPU Local Bus Mapping section in the Memories chapter
for details.
18.6.2.8
Peripheral Events
Peripheral events allow direct peripheral to peripheral communication of specified events. See
the Peripheral Event System chapter for more information.
The GPIO can be programmed to output peripheral events whenever an interrupt condition is
detected. The peripheral events configuration depends on the interrupt configuration. An event
will be generated on the same condition as the interrupt (pin change, rising edge, or falling
edge). The interrupt configuration is controlled by the IMR register. Peripheral event on a pin is
enabled by writing a one to the corresponding bit in the Event Enable Register (EVER). The
Peripheral Event trigger mode is shared with the interrupt trigger and is configured by writing to
the IMR0 and IMR1 registers. Interrupt does not need to be enabled on a pin when peripheral
events are enabled. Peripheral Events are also affected by the Input Glitch Filter settings. See
Section 18.6.2.5 for more information.
A peripheral event can be generated on each GPIO pin. Each port can then have up to 32
peripheral event generators. Groups of eight peripheral event generators in each port are ORed
together to form a peripheral event line, so that each port has four peripheral event lines connected to the Peripheral Event System.
337
32099I–01/2012
AT32UC3L016/32/64
18.7
User Interface
The GPIO controller manages all the GPIO pins on the 32-bit AVR microcontroller. The pins are
managed as 32-bit ports that are configurable through a Peripheral Bus (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.
Figure 18-6. Port Configuration Registers
0x0000
Port 0 Configuration Registers
0x0200
Port 1 Configuration Registers
0x0400
….
Port 2 Configuration Registers
n*0x200
Port n Configuration Registers
In the peripheral muxing table in the Package and Pinout chapter each GPIO pin has a unique
number. Note that the PA, PB, PC, and PX ports do not necessarily 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 % 32, example: 36 % 32 = 4
Table 18-2 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 a GPIO pin.
18.7.1
Access Types
Most 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 one will be set. Bits written to zero will be unchanged by the
operation. Performing a “clear” access, all bits written to one will be cleared. Bits written to zero
will be unchanged by the operation. Finally, a toggle access will toggle the value of all bits writ-
338
32099I–01/2012
AT32UC3L016/32/64
ten to one. Again all bits written to zero 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.
18.7.2
Configuration Protection
In order to protect the configuration of individual GPIO pins from software failure, configuration
bits for individual GPIO pins may be locked by writing a one to the corresponding bit in the LOCK
register. While this bit is one, any write to the same bit position in any lockable GPIO register
using the Peripheral Bus (PB) will not have an effect. The CPU Local Bus is not checked and
thus allowed to write to all bits in a CPU Local Bus mapped register no mather the LOCK value.
The registers required to clear bits in the LOCK register are protected by the access protection
mechanism described in Section 18.7.3, ensuring the LOCK mechanism itself is robust against
software failure.
18.7.3
Access Protection
In order to protect critical registers from software failure, some registers are protected by a key
protection mechanism. These registers can only be changed by first writing the UNLOCK register, then the protected register. Protected registers are indicated in Table 18-2. The UNLOCK
register contains a key field which must always be written to 0xAA, and an OFFSET field corresponding to the offset of the register to be modified.
The next write operation resets the UNLOCK register, so if the register is to be modified again,
the UNLOCK register must be written again.
Attempting to write to a protected register without first writing the UNLOCK register results in the
write operation being discarded, and the Access Error bit in the Access Status Register
(ASR.AE) will be set.
Table 18-2.
GPIO Register Memory Map
Offset
Register
Function
Register Name
Access
Reset
Config.
Protection
Access
Protection
0x000
GPIO Enable Register
Read/Write
GPER
Read/Write
-(1)
Y
N
0x004
GPIO Enable Register
Set
GPERS
Write-only
Y
N
0x008
GPIO Enable Register
Clear
GPERC
Write-only
Y
N
0x00C
GPIO Enable Register
Toggle
GPERT
Write-only
Y
N
Y
N
-
(1)
0x010
Peripheral Mux Register 0
Read/Write
PMR0
Read/Write
0x014
Peripheral Mux Register 0
Set
PMR0S
Write-only
Y
N
0x018
Peripheral Mux Register 0
Clear
PMR0C
Write-only
Y
N
0x01C
Peripheral Mux Register 0
Toggle
PMR0T
Write-only
Y
N
Y
N
(1)
0x020
Peripheral Mux Register 1
Read/Write
PMR1
Read/Write
-
0x024
Peripheral Mux Register 1
Set
PMR1S
Write-only
Y
N
0x028
Peripheral Mux Register 1
Clear
PMR1C
Write-only
Y
N
339
32099I–01/2012
AT32UC3L016/32/64
Table 18-2.
GPIO Register Memory Map
Offset
Register
Function
Register Name
Access
0x02C
Peripheral Mux Register 1
Toggle
PMR1T
Write-only
Reset
-
(1)
Config.
Protection
Access
Protection
Y
N
Y
N
0x030
Peripheral Mux Register 2
Read/Write
PMR2
Read/Write
0x034
Peripheral Mux Register 2
Set
PMR2S
Write-only
Y
N
0x038
Peripheral Mux Register 2
Clear
PMR2C
Write-only
Y
N
0x03C
Peripheral Mux Register 2
Toggle
PMR2T
Write-only
Y
N
Y
N
(1)
0x040
Output Driver Enable Register
Read/Write
ODER
Read/Write
-
0x044
Output Driver Enable Register
Set
ODERS
Write-only
Y
N
0x048
Output Driver Enable Register
Clear
ODERC
Write-only
Y
N
0x04C
Output Driver Enable Register
Toggle
ODERT
Write-only
Y
N
N
N
-
(1)
0x050
Output Value Register
Read/Write
OVR
Read/Write
0x054
Output Value Register
Set
OVRS
Write-only
N
N
0x058
Output Value Register
Clear
OVRC
Write-only
N
N
0x05c
Output Value Register
Toggle
OVRT
Write-only
N
N
N
N
Depe
nding
on pin
states
0x060
Pin Value Register
Read
PVR
Read-only
0x064
Pin Value Register
-
-
-
N
N
0x068
Pin Value Register
-
-
-
N
N
0x06c
Pin Value Register
-
-
-
N
N
0x070
Pull-up Enable Register
Read/Write
PUER
Read/Write
Y
N
0x074
Pull-up Enable Register
Set
PUERS
Write-only
Y
N
0x078
Pull-up Enable Register
Clear
PUERC
Write-only
Y
N
0x07C
Pull-up Enable Register
Toggle
PUERT
Write-only
Y
N
0x090
Interrupt Enable Register
Read/Write
IER
Read/Write
N
N
0x094
Interrupt Enable Register
Set
IERS
Write-only
N
N
0x098
Interrupt Enable Register
Clear
IERC
Write-only
N
N
0x09C
Interrupt Enable Register
Toggle
IERT
Write-only
N
N
0x0A0
Interrupt Mode Register 0
Read/Write
IMR0
Read/Write
N
N
0x0A4
Interrupt Mode Register 0
Set
IMR0S
Write-only
N
N
0x0A8
Interrupt Mode Register 0
Clear
IMR0C
Write-only
N
N
0x0AC
Interrupt Mode Register 0
Toggle
IMR0T
Write-only
N
N
0x0B0
Interrupt Mode Register 1
Read/Write
IMR1
Read/Write
N
N
0x0B4
Interrupt Mode Register 1
Set
IMR1S
Write-only
N
N
0x0B8
Interrupt Mode Register 1
Clear
IMR1C
Write-only
N
N
0x0BC
Interrupt Mode Register 1
Toggle
IMR1T
Write-only
N
N
-(1)
-(1)
-(1)
-(1)
340
32099I–01/2012
AT32UC3L016/32/64
Table 18-2.
GPIO Register Memory Map
Offset
Register
Function
Register Name
Access
Reset
Config.
Protection
Access
Protection
0x0C0
Glitch Filter Enable Register
Read/Write
GFER
Read/Write
-(1)
N
N
0x0C4
Glitch Filter Enable Register
Set
GFERS
Write-only
N
N
0x0C8
Glitch Filter Enable Register
Clear
GFERC
Write-only
N
N
0x0CC
Glitch Filter Enable Register
Toggle
GFERT
Write-only
N
N
N
N
-
(1)
0x0D0
Interrupt Flag Register
Read
IFR
Read-only
0x0D4
Interrupt Flag Register
-
-
-
N
N
0x0D8
Interrupt Flag Register
Clear
IFRC
Write-only
N
N
0x0DC
Interrupt Flag Register
-
-
-
N
N
N
N
-
(1)
0x180
Event Enable Register
Read
EVER
Read/Write
0x184
Event Enable Register
Set
EVERS
Write-only
N
N
0x188
Event Enable Register
Clear
EVERC
Write-only
N
N
0x18C
Event Enable Register
Toggle
EVERT
Write-only
N
N
N
Y
0x1A0
Lock Register
Read/Write
LOCK
Read/Write
0x1A4
Lock Register
Set
LOCKS
Write-only
N
N
0x1A8
Lock Register
Clear
LOCKC
Write-only
N
Y
0x1AC
Lock Register
Toggle
LOCKT
Write-only
N
Y
0x1E0
Unlock Register
Read/Write
UNLOCK
Write-only
N
N
0x1E4
Access Status Register
Read/Write
ASR
Read/Write
0x1F8
Parameter Register
Read
PARAMETER
Read-only
-(1)
N
N
Read-only
(1)
N
N
0x1FC
Note:
Version Register
Read
VERSION
-
(1)
N
-
1. The reset values for these registers are device specific. Please refer to the Module Configuration section at the end of this
chapter.
341
32099I–01/2012
AT32UC3L016/32/64
18.7.4
Name:
GPIO Enable Register
GPER
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x000, 0x004, 0x008, 0x00C
Reset Value:
-
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.
342
32099I–01/2012
AT32UC3L016/32/64
18.7.5
Name:
Peripheral Mux Register 0
PMR0
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x010, 0x014, 0x018, 0x01C
Reset Value:
-
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
343
32099I–01/2012
AT32UC3L016/32/64
18.7.6
Name:
Peripheral Mux Register 1
PMR1
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x020, 0x024, 0x028, 0x02C
Reset Value:
-
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
344
32099I–01/2012
AT32UC3L016/32/64
18.7.7
Name:
Peripheral Mux Register 2
PMR2
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x030, 0x034, 0x038, 0x03C
Reset Value:
-
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 2
{PMR2, PMR1, PMR0}
000
001
010
011
100
101
110
111
Selected Peripheral Function
A
B
C
D
E
F
G
H
345
32099I–01/2012
AT32UC3L016/32/64
18.7.8
Name:
Output Driver Enable Register
ODER
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x040, 0x044, 0x048, 0x04C
Reset Value:
-
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.
346
32099I–01/2012
AT32UC3L016/32/64
18.7.9
Name:
Output Value Register
OVR
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x050, 0x054, 0x058, 0x05C
Reset Value:
-
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 GPIO pin is 0.
1: The value to be driven on the GPIO pin is 1.
347
32099I–01/2012
AT32UC3L016/32/64
18.7.10
Name:
Pin Value Register
PVR
Access:
Read-only
Offset:
0x060, 0x064, 0x068, 0x06C
Reset Value:
Depending on pin states
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 GPIO pin is at level zero.
1: The GPIO pin is at level one.
Note that the level of a pin can only be read when the corresponding pin in GPER is one or interrupt is enabled for the pin.
348
32099I–01/2012
AT32UC3L016/32/64
18.7.11
Name:
Pull-up Enable Register
PUER
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x070, 0x074, 0x078, 0x07C
Reset Value:
-
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
Writing a zero to a bit in this register will disable pull-up on the corresponding pin.
Writing a one to a bit in this register will enable pull-up on the corresponding pin.
349
32099I–01/2012
AT32UC3L016/32/64
18.7.12
Name:
Interrupt Enable Register
IER
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x090, 0x094, 0x098, 0x09C
Reset Value:
-
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.
350
32099I–01/2012
AT32UC3L016/32/64
18.7.13
Name:
Interrupt Mode Register 0
IMR0
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x0A0, 0x0A4, 0x0A8, 0x0AC
Reset Value:
-
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
351
32099I–01/2012
AT32UC3L016/32/64
18.7.14
Name:
Interrupt Mode Register 1
IMR1
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x0B0, 0x0B4, 0x0B8, 0x0BC
Reset Value:
-
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
352
32099I–01/2012
AT32UC3L016/32/64
18.7.15
Name:
Glitch Filter Enable Register
GFER
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x0C0, 0x0C4, 0x0C8, 0x0CC
Reset Value:
-
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 the corresponding bit in IER is zero. Updating GFER while
interrupt on the corresponding pin is enabled can cause an unintentional interrupt to be triggered.
353
32099I–01/2012
AT32UC3L016/32/64
18.7.16
Name:
Interrupt Flag Register
IFR
Access:
Read, Clear
Offset:
0x0D0, 0x0D8
Reset Value:
-
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: No interrupt condition has been detected on the corresponding pin.
1: An interrupt condition has been detected on the corresponding pin.
The number of interrupt request lines depends on the number of GPIO 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 one.
354
32099I–01/2012
AT32UC3L016/32/64
18.7.17
Name:
Event Enable Register
EVER
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x180, 0x184, 0x188, 0x18C
Reset Value:
-
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: Event Enable
0: Peripheral Event is disabled for the corresponding pin.
1: Peripheral Event is enabled for the corresponding pin.
355
32099I–01/2012
AT32UC3L016/32/64
18.7.18
Name:
Lock Register
LOCK
Access:
Read/Write, Set, Clear, Toggle
Offset:
0x1A0, 0x1A4, 0x1A8, 0x1AC
Reset Value:
-
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: Lock State
0: Pin is unlocked. The corresponding bit can be changed in any GPIO register for this port.
1: Pin is locked. The corresponding bit can not be changed in any GPIO register for this port.
The value of LOCK determines which bits are locked in the lockable registers.
The LOCK, LOCKC, and LOCKT registers are protected, which means they can only be written immediately after a write to the
UNLOCK register with the proper KEY and OFFSET.
LOCKS is not protected, and can be written at any time.
356
32099I–01/2012
AT32UC3L016/32/64
18.7.19
Name:
Unlock Register
UNLOCK
Access:
Write-only
Offset:
0x1E0
Reset Value:
-
31
30
29
28
27
26
25
24
KEY
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
-
-
-
-
-
-
7
6
5
4
3
2
8
OFFSET
1
0
OFFSET
• OFFSET: Register Offset
This field must be written with the offset value of the LOCK, LOCKC or LOCKT register to unlock. This offset must also include
the port offset for the register to unlock. LOCKS can not be locked so no unlock is required before writing to this register.
• KEY: Unlocking Key
This bitfield must be written to 0xAA for a write to this register to have an effect.
This register always reads as zero.
357
32099I–01/2012
AT32UC3L016/32/64
18.7.20
Name:
Access Status Register
ASR
Access:
Read/Write
Offset:
0x1E4
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
-
-
-
-
-
-
-
AE
• AE: Access Error
This bit is set when a write to a locked register occurs.
This bit can be written to 0 by software.
358
32099I–01/2012
AT32UC3L016/32/64
18.7.21
Name:
Parameter Register
PARAMETER
Access Type:
Read-only
Offset:
0x1F8
Reset Value:
-
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
PARAMETER
23
22
21
20
PARAMETER
15
14
13
12
PARAMETER
7
6
5
4
PARAMETER
• PARAMETER:
0: The corresponding pin is not implemented in this GPIO port.
1: The corresponding pin is implemented in this GPIO port.
There is one PARAMETER register per GPIO port. Each bit in the Parameter Register indicates whether the corresponding
GPER bit is implemented.
359
32099I–01/2012
AT32UC3L016/32/64
18.7.22
Name:
Version Register
VERSION
Access Type:
Read-only
Offset:
0x1FC
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]
3
2
VERSION[7:0]
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the module. No functionality associated.
360
32099I–01/2012
AT32UC3L016/32/64
18.8
Module Configuration
The specific configuration for each GPIO instance is listed in the following tables. The module
bus clocks listed here are connected to the system bus clocks. Refer to the Power Manager
chapter for details.
Table 18-3.
GPIO Configuration
Feature
GPIO
Number of GPIO ports
2
Number of peripheral functions
8
Table 18-4.
Implemented Pin Functions
Pin Function
Implemented
Notes
Pull-up
On all pins
Controlled by PUER or peripheral
Table 18-5.
GPIO Clocks
Module Name
Clock Name
Description
GPIO
CLK_GPIO
Clock for the GPIO bus interface
The reset values for all GPIO registers are zero, with the following exceptions:
Table 18-6.
Register Reset Values
Port
Register
Reset Value
0
GPER
0x004DFF5F
0
PMR0
0x00320020
0
PMR1
0x00020080
0
PMR2
0x00100800
0
PUER
0x00000001
0
GFER
0x007FFFFF
0
PARAMETER
0x007FFFFF
0
VERSION
0x00000211
1
GPER
0x00001FCF
1
PMR0
0x00000030
1
PMR1
0x00000030
1
GFER
0x00001FFF
1
PARAMETER
0x00001FFF
1
VERSION
0x00000211
361
32099I–01/2012
AT32UC3L016/32/64
19. Universal Synchronous Asynchronous Receiver Transmitter (USART)
Rev: 4.4.0.6
19.1
Features
• Configurable baud rate generator
• 5- to 9-bit full-duplex, synchronous and asynchronous, serial communication
•
•
•
•
19.2
– 1, 1.5, or 2 stop bits in asynchronous mode, and 1 or 2 in synchronous mode
– Parity generation and error detection
– Framing- and overrun error detection
– MSB- or LSB-first
– Optional break generation and detection
– Receiver frequency over-sampling by 8 or 16 times
– Optional RTS-CTS hardware handshaking
– Receiver Time-out and transmitter Timeguard
– Optional Multidrop mode with address generation and detection
SPI Mode
– Master or slave
– Configurable serial clock phase and polarity
– CLK SPI serial clock frequency up to a quarter of the CLK_USART internal clock frequency
LIN Mode
– Compliant with LIN 1.3 and LIN 2.0 specifications
– Master or slave
– Processing of Frames with up to 256 data bytes
– Configurable response data length, optionally defined automatically by the Identifier
– Self synchronization in slave node configuration
– Automatic processing and verification of the “Break Field” and “Sync Field”
– The “Break Field” is detected even if it is partially superimposed with a data byte
– Optional, automatic identifier parity management
– Optional, automatic checksum management
– Supports both “Classic” and “Enhanced” checksum types
– Full LIN error checking and reporting
– Frame Slot Mode: the master allocates slots to scheduled frames automatically.
– Wakeup signal generation
Test Modes
– Automatic echo, remote- and local loopback
Supports two Peripheral DMA Controller channels
– Buffer transfers without processor intervention
Overview
The Universal Synchronous Asynchronous Receiver Transceiver (USART) provides a full
duplex, universal, synchronous/asynchronous serial link. Data frame format is widely configurable, including basic length, parity, and stop bit settings, maximizing standards support. The
receiver implements parity-, framing-, and overrun error detection, and can handle un-fixed
frame lengths with the time-out feature. The USART supports several operating modes, providing an interface to, LIN, and SPI buses and infrared transceivers. Communication with slow and
remote devices is eased by the timeguard. Duplex multidrop communication is supported by
address and data differentiation through the parity bit. The hardware handshaking feature
enables an out-of-band flow control, automatically managing RTS and CTS pins. The Peripheral
DMA Controller connection enables memory transactions, and the USART supports chained
362
32099I–01/2012
AT32UC3L016/32/64
buffer management without processor intervention. Automatic echo, remote-, and local loopback
-test modes are also supported.
19.3
Block Diagram
Figure 19-1. USART Block Diagram
Peripheral DMA
Controller
Channel
Channel
USART
I/O
Controller
RXD
Receiver
RTS
Interrupt
Controller
USART
Interrupt
TXD
Transmitter
CTS
CLK_USART
Power
Manager
DIV
BaudRate
Generator
CLK_USART/DIV
CLK
User
Interface
Peripheral bus
Table 19-1.
SPI Operating Mode
PIN
USART
SPI Slave
SPI Master
RXD
RXD
MOSI
MISO
TXD
TXD
MISO
MOSI
RTS
RTS
–
CS
CTS
CTS
CS
–
363
32099I–01/2012
AT32UC3L016/32/64
19.4
I/O Lines Description
Table 19-2.
I/O Lines Description
Name
Description
Type
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
Output
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
19.5
19.5.1
Active Level
Product Dependencies
I/O Lines
The USART pins may be multiplexed with the I/O Controller lines. The user must first configure
the I/O Controller to assign these pins to their peripheral functions. Unused I/O lines may be
used for other purposes.
To prevent the TXD line from falling when the USART is disabled, the use of an internal pull up
is required. If the hardware handshaking feature or modem mode is used, the internal pull up on
TXD must also be enabled.
19.5.2
Clocks
The clock for the USART bus interface (CLK_USART) is generated by the Power Manager. This
clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the USART before disabling the clock, to avoid freezing the USART in an undefined state.
19.5.3
Interrupts
The USART interrupt request line is connected to the interrupt controller. Using the USART
interrupt requires the interrupt controller to be programmed first.
364
32099I–01/2012
AT32UC3L016/32/64
19.6
Functional Description
19.6.1
Baud Rate Generator
The baud rate generator provides the bit period clock named the Baud Rate Clock to both
receiver and transmitter. It is based on a 16-bit divider, which is specified in the Clock Divider
field in the Baud Rate Generator Register (BRGR.CD). A non-zero value enables the generator,
and if CD is one, the divider is bypassed and inactive. The Clock Selection field in the Mode
Register (MR.USCLKS) selects clock source between:
• CLK_USART (internal clock)
• CLK_USART/DIV (a divided CLK_USART, refer to Module Configuration section)
• CLK (external clock, available on the CLK pin)
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 at least 4.5 times longer than those provided by CLK_USART.
Figure 19-2. Baud Rate Generator
USCLKS
CLK_USART
CLK_USART/DIV
CLK
Reserved
CD
CD
0
1
2
CLK
16-bit Counter
FIDI
>1
3
1
0
SYNC
OVER
0
Sampling
Divider
0
0
BaudRate
Clock
1
1
SYNC
USCLKS= 3
19.6.1.1
Sampling
Clock
Baud Rate in Asynchronous Mode
If the USART is configured to operate in an asynchronous mode, the selected clock is divided by
the CD value before it is provided to the receiver as a sampling clock. Depending on the Oversampling Mode bit (MR.OVER) value, the clock is then divided by either 8 (OVER=1), or 16
(OVER=0). The baud rate is calculated with the following formula:
SelectedClock
BaudRate = -----------------------------------------------( 8 ( 2 – OVER )CD )
This gives a maximum baud rate of CLK_USART divided by 8, assuming that CLK_USART is
the fastest clock possible, and that OVER is one.
19.6.1.2
Baud Rate Calculation Example
Table 19-3 shows calculations based on the CD field to obtain 38400 baud from different source
clock frequencies. This table also shows the actual resulting baud rate and error.
365
32099I–01/2012
AT32UC3L016/32/64
Table 19-3.
Baud Rate Example (OVER=0)
Source Clock (Hz)
Expected Baud
Rate (bit/s)
Calculation Result
CD
Actual Baud Rate (bit/s)
Error
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%
The baud rate is calculated with the following formula (OVER=0):
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 ⎠
19.6.1.3
Fractional Baud Rate in Asynchronous Mode
The baud rate generator has a limitation: the source frequency is always a multiple of the baud
rate. An approach to this problem is to integrate a high resolution fractional N clock generator,
outputting fractional multiples of the reference source clock. This fractional part is selected with
the Fractional Part field (BRGR.FP), and is activated by giving it a non-zero value. The resolution is one eighth of CD. The resulting baud rate is calculated using the following formula:
SelectedClock
BaudRate = -------------------------------------------------------------------⎛ 8 ( 2 – OVER ) ⎛ CD + FP
-------⎞⎠ ⎞⎠
⎝
⎝
8
The modified architecture is presented below:
366
32099I–01/2012
AT32UC3L016/32/64
Figure 19-3. Fractional Baud Rate Generator
FP
USCLKS
CLK_USART
CLK_USART/DIV
CLK
Reserved
0
1
2
3
CD
Modulus
Control
FP
CD
16-bit Counter
glitch-free
logic
FIDI
>1
1
0
CLK
0
OVER
Sampling
Divider
0
SYNC
0
BaudRate
Clock
1
1
SYNC
USCLKS = 3
19.6.1.4
Sampling
Clock
Baud Rate in Synchronous and SPI Mode
If the USART is configured to operate in synchronous mode, the selected clock is divided by the
BRGR.CD field. This does not apply when CLK is selected.
SelectedClock
BaudRate = -------------------------------------CD
When CLK is selected the external frequency must be at least 4.5 times lower than the system
clock, and when either CLK or CLK_USART/DIV are selected, CD must be even to ensure a
50/50 duty cycle. If CLK_USART is selected, the generator ensures this regardless of value.
19.6.2
Receiver and Transmitter Control
After a reset, the transceiver is disabled. The receiver/transmitter is enabled by writing a one to
either the Receiver Enable, or Transmitter Enable bit in the Control Register (CR.RXEN, or
CR.TXEN). They may be enabled together and can be configured both before and after they
have been enabled. The user can reset the USART receiver/transmitter at any time by writing a
one to either the Reset Receiver (CR.RSTRX), or Reset Transmitter (CR.RSTTX) bit. This software reset clears status bits and resets internal state machines, immediately halting any
communication. The user interface configuration registers will retain their values.
The user can disable the receiver/transmitter by writing a one to either the Receiver Disable, or
Transmitter Disable bit (CR.RXDIS, or CR.TXDIS). If the receiver is disabled during a character
reception, the USART will wait for the current character to be received before disabling. If the
transmitter is disabled during transmission, the USART will wait until both the current character
and the character stored in the Transmitter Holding Register (THR) are transmitted before disabling. If a timeguard has been implemented it will remain functional during the transaction.
367
32099I–01/2012
AT32UC3L016/32/64
19.6.3
Synchronous and Asynchronous Modes
19.6.3.1
Transmitter Operations
The transmitter performs equally in both synchronous and asynchronous operating modes
(MR.SYNC). One start bit, up to 9 data bits, an optional parity bit, and up to two stop bits are
successively shifted out on the TXD pin at each falling edge of the serial clock. The number of
data bits is selected by the Character Length field (MR.CHRL) and the MR.MODE9 bit. Nine bits
are selected by writing a one to MODE9, overriding any value in CHRL. The parity bit configuration is selected in the MR.PAR field. The Most Significant Bit First bit (MR.MSBF) selects which
data bit to send first. The number of stop bits is selected by the MR.NBSTOP field. The 1.5 stop
bit configuration is only supported in asynchronous mode.
Figure 19-4. Character Transmit
Example: 8-bit, Parity Enabled One Stop
Baud Rate
Clock
TXD
Start
Bit
D0
D1
D2
D3
D4
D5
D6
D7
Parity
Bit
Stop
Bit
The characters are sent by writing to the Character to be Transmitted field (THR.TXCHR). The
transmitter reports status with the Transmitter Ready (TXRDY) and Transmitter Empty
(TXEMPTY) bits in the Channel Status Register (CSR). TXRDY is set when THR is empty.
TXEMPTY is set when both THR and the transmit shift register are empty (transmission complete). Both TXRDY and TXEMPTY are cleared when the transmitter is disabled. Writing a
character to THR while TXRDY is zero has no effect and the written character will be lost.
Figure 19-5. 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
THR
TXRDY
TXEMPTY
19.6.3.2
Asynchronous Receiver
If the USART is configured in an asynchronous operating mode (MR.SYNC = 0), the receiver will
oversample the RXD input line by either 8 or 16 times the baud rate clock, as selected by the
Oversampling Mode bit (MR.OVER). If the line is zero for half a bit period (four or eight consecutive samples, respectively), a start bit will be assumed, and the following 8th or 16th sample will
determine the logical value on the line, in effect resulting in bit values being determined at the
middle of the bit period.
368
32099I–01/2012
AT32UC3L016/32/64
The number of data bits, endianess, parity mode, and stop bits are selected by the same bits
and fields as for the transmitter (MR.CHRL, MODE9, MSBF, PAR, and NBSTOP). The synchronization mechanism will only consider one stop bit, regardless of the used protocol, and when
the first stop bit has been sampled, the receiver will automatically begin looking for a new start
bit, enabling resynchronization even if there is a protocol miss-match. Figure 19-6 and Figure
19-7 illustrate start bit detection and character reception in asynchronous mode.
Figure 19-6. Asynchronous Start Bit 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 19-7. 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
19.6.3.3
D1
D2
D3
D4
D5
D6
D7
Parity
Bit
Stop
Bit
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 bit. Configuration bits
and fields are the same as in asynchronous mode.
369
32099I–01/2012
AT32UC3L016/32/64
Figure 19-8. Synchronous Mode Character Reception
Example: 8-bit, Parity Enabled 1 Stop
Baud Rate
Clock
RXD
Sampling
Start
D0
D1
D2
D3
D4
D5
D6
Stop Bit
D7
Parity Bit
19.6.3.4
Receiver Operations
When a character reception is completed, it is transferred to the Received Character field in the
Receive Holding Register (RHR.RXCHR), and the Receiver Ready bit in the Channel Status
Register (CSR.RXRDY) is set. If RXRDY is already set, RHR will be overwritten and the Overrun
Error bit (CSR.OVRE) is set. Reading RHR will clear RXRDY, and writing a one to the Reset
Status bit in the Control Register (CR.RSTSTA) will clear OVRE.
Figure 19-9. 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
CR
Read
RHR
RXRDY
OVRE
19.6.3.5
Parity
The USART supports five parity modes selected by MR.PAR. The PAR field also enables the
Multidrop mode, see ”Multidrop Mode” on page 371. If even parity is selected, the parity bit will
be a zero if there is an even number of ones in the data character, and if there is an odd number
it will be a one. For odd parity the reverse applies. If space or mark parity is chosen, the parity bit
will always be a zero or one, respectively. See Table 19-4.
Table 19-4.
Parity Bit Examples
Parity Mode
Alphanum
Character
Hex
Bin
Odd
Even
Mark
Space
None
A
0x41
0100 0001
1
0
1
0
-
V
0x56
0101 0110
1
0
1
0
-
R
0x52
0101 0010
0
1
1
0
-
370
32099I–01/2012
AT32UC3L016/32/64
The receiver will report parity errors in CSR.PARE, unless parity is disabled. Writing a one to
CR.RSTSTA will clear PARE. See Figure 19-10
Figure 19-10. Parity Error
Baud Rate
Clock
RXD
Start
D0
Bit
D1
D2
D3
D4
D5
D6
D7
Bad Stop
Parity Bit
Bit
RSTSTA = 1
Write
CR
PARE
RXRDY
19.6.3.6
Multidrop Mode
If PAR is either 0x6 or 0x7, the USART runs in Multidrop mode. This mode differentiates data
and address characters. Data has the parity bit zero and addresses have a one. By writing a one
to the Send Address bit (CR.SENDA) the user will cause the next character written to THR to be
transmitted as an address. Receiving a character with a one as parity bit will set PARE.
19.6.3.7
Transmitter Timeguard
The timeguard feature enables the USART to interface slow devices by inserting an idle state on
the TXD line in between two characters. This idle state corresponds to a long stop bit, whose
duration is selected by the Timeguard Value field in the Transmitter Timeguard Register
(TTGR.TG). The transmitter will hold the TXD line high for TG bit periods, in addition to the number of stop bits. As illustrated in Figure 19-11, the behavior of TXRDY and TXEMPTY is modified
when TG has a non-zero value. If a pending character has been written to THR, the TXRDY bit
will not be set until this characters start bit has been sent. TXEMPTY will remain low until the
timeguard transmission has completed.
Figure 19-11. Timeguard Operation
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
THR
TXRDY
TXEMPTY
371
32099I–01/2012
AT32UC3L016/32/64
Table 19-5.
19.6.3.8
Maximum Baud Rate Dependent Timeguard Durations
Baud Rate (bit/sec)
Bit time (µs)
Timeguard (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 Time-out Value field in the Receiver Time-out Register (RTOR.TO) enables handling of variable-length frames by detection of selectable idle durations on the RXD line. The value written to
TO is loaded to a decremental counter, and unless it is zero, a time-out will occur when the
amount of inactive bit periods match the initial counter value. If a time-out has not occurred, the
counter will reload and restart every time a new character arrives. A time-out sets the TIMEOUT
bit in CSR. Clearing TIMEOUT can be done in two ways:
• Writing a one to the Start Time-out bit (CR.STTTO). This also aborts count down until the
next character has been received.
• Writing a one to the Reload and Start Time-out bit (CR.RETTO). This also reloads the
counter and restarts count down immediately.
Figure 19-12. Receiver Time-out Block Diagram
TO
Baud Rate
Clock
1
D
Q
Clock
16-bit Time-out
Counter
16-bit
Value
=
STTTO
Character
Received
Load
Clear
TIMEOUT
0
RETTO
Table 19-6.
Maximum Time-out Period
Baud Rate (bit/sec)
Bit Time (µs)
Time-out (ms)
600
1 667
109 225
1 200
833
54 613
2 400
417
27 306
4 800
208
13 653
372
32099I–01/2012
AT32UC3L016/32/64
19.6.3.9
Baud Rate (bit/sec)
Bit Time (µs)
Time-out (ms)
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
Framing Error
The receiver is capable of detecting framing errors. A framing error has occurred if a stop bit
reads as zero. This can occur if the transmitter and receiver are not synchronized. A framing
error is reported by CSR.FRAME as soon as the error is detected, at the middle of the stop bit.
Figure 19-13. Framing Error Status
Baud Rate
Clock
RXD
Start
D0
Bit
D1
D2
D3
D4
D5
D6
D7
Parity Stop
Bit Bit
RSTSTA = 1
Write
CR
FRAME
RXRDY
19.6.3.10
Transmit Break
When TXRDY is set, the user can request the transmitter to generate a break condition on the
TXD line by writing a one to The Start Break bit (CR.STTBRK). The break is treated as a normal
0x00 character transmission, clearing TXRDY and TXEMPTY, but with zeroes for preambles,
start, parity, stop, and time guard bits. Writing a one to the Stop Break bit (CR.STBRK) will stop
the generation of new break characters, and send ones for TG duration or at least 12 bit periods,
ensuring that the receiver detects end of break, before resuming normal operation. Figure 19-14
illustrates STTBRK and STPBRK effect on the TXD line.
Writing to STTBRK and STPBRK simultaneously can lead to unpredictable results. Writes to
THR before a pending break has started will be ignored.
373
32099I–01/2012
AT32UC3L016/32/64
Figure 19-14. 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
CR
TXRDY
TXEMPTY
19.6.3.11
Receive Break
A break condition is assumed when incoming data, parity, and stop bits are zero. This corresponds to a framing error, but FRAME will remain zero while the Break Received/End Of Break
bit (CSR.RXBRK) is set. Writing a one to CR.RSTSTA will clear RXBRK. An end of break will
also set RXBRK, and is assumed when TX is high for at least 2/16 of a bit period in asynchronous mode, or when a high level is sampled in synchronous mode.
19.6.3.12
Hardware Handshaking
The USART features an out-of-band hardware handshaking flow control mechanism, implementable by connecting the RTS and CTS pins with the remote device, as shown in Figure 1915.
Figure 19-15. Connection with a Remote Device for Hardware Handshaking
USART
Remote
Device
TXD
RXD
RXD
TXD
CTS
RTS
RTS
CTS
Writing 0x2 to the MR.MODE field configures the USART to operate in this mode. The receiver
will drive its RTS pin high when disabled or when the Reception Buffer Full bit (CSR.RXBUFF) is
set by the Buffer Full signal from the Peripheral DMA controller. If the receivers RTS pin is high,
the transmitters CTS pin will also be high and only the active character transactions will be completed. Allocating a new buffer to the DMA controller by clearing RXBUFF, will drive the RTS pin
low, allowing the transmitter to resume transmission. Detected level changes on the CTS pin
can trigger interrupts, and are reported by the CTS Input Change bit in the Channel Status Register (CSR.CTSIC).
Figure 19-16 illustrates receiver functionality, and Figure 19-17 illustrates transmitter
functionality.
374
32099I–01/2012
AT32UC3L016/32/64
Figure 19-16. Receiver Behavior when Operating with Hardware Handshaking
RXD
RXEN = 1
RXDIS = 1
Write
CR
RTS
RXBUFF
Figure 19-17. Transmitter Behavior when Operating with Hardware Handshaking
CTS
TXD
Figure 19-18.
19.6.4
SPI Mode
The USART features a Serial Peripheral Interface (SPI) link compliant mode, supporting synchronous, full-duplex communication, in both master and slave mode. Writing 0xE (master) or
0xF (slave) to MR.MODE will enable this mode. A SPI in master mode controls the data flow to
and from the other SPI devices, who are in slave mode. It is possible to let devices take turns
being masters (aka multi-master protocol), and one master may shift data simultaneously into
several slaves, but only one slave may respond at a time. A slave is selected when its slave
select (NSS) signal has been raised by the master. The USART can only generate one NSS signal, and it is possible to use standard I/O lines to address more than one slave.
19.6.4.1
Modes of Operation
The SPI system consists of two data lines and two control lines:
• Master Out Slave In (MOSI): This line supplies the data shifted from master to slave. In
master mode this is connected to TXD, and in slave mode to RXD.
• Master In Slave Out (MISO): This line supplies the data shifted from slave to master. In
master mode this is connected to RXD, and in slave mode to TXD.
• Serial Clock (CLK): This is controlled by the master. One period per bit transmission. In both
modes this is connected to CLK.
• Slave Select (NSS): This control line allows the master to select or deselect a slave. In
master mode this is connected to RTS, and in slave mode to CTS.
Changing SPI mode after initial configuration has to be followed by a transceiver software reset
in order to avoid unpredictable behavior.
19.6.4.2
Baud Rate
The baud rate generator operates as described in ”Baud Rate in Synchronous and SPI Mode”
on page 367, with the following requirements:
In SPI Master Mode:
375
32099I–01/2012
AT32UC3L016/32/64
• The Clock Selection field (MR.USCLKS) must not equal 0x3 (external clock, CLK).
• The Clock Output Select bit (MR.CLKO) must be one.
• The BRGR.CD field must be at least 0x4.
• If USCLKS is one (internal divided clock, CLK_USART/DIV), the value in CD has to be even,
ensuring a 50:50 duty cycle. CD can be odd if USCLKS is zero (internal clock, CLK_USART).
In SPI Slave Mode:
• CLK frequency must be at least four times lower than the system clock.
19.6.4.3
Data Transfer
• Up to nine data bits are successively shifted out on the TXD pin at each edge. There are no
start, parity, or stop bits, and MSB is always sent first. The SPI Clock Polarity (MR.CPOL),
and SPI Clock Phase (MR.CPHA) bits configure CLK by selecting the edges upon which bits
are shifted and sampled, resulting in four non-interoperable protocol modes see Table 19-7.
A master/slave pair must use the same configuration, and the master must be reconfigured if
it is to communicate with slaves using different configurations. See Figures 19-19 and 19-20.
Table 19-7.
SPI Bus Protocol Modes
SPI Bus Protocol Mode
CPOL
CPHA
0
0
1
1
0
0
2
1
1
3
1
0
Figure 19-19. SPI Transfer Format (CPHA=1, 8 bits per transfer)
CLK cycle (for reference)
1
2
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
376
32099I–01/2012
AT32UC3L016/32/64
Figure 19-20. 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
19.6.4.4
Receiver and Transmitter Control
See ”Transmitter Operations” on page 368, and ”Receiver Operations” on page 370.
19.6.4.5
Character Transmission and Reception
In SPI master mode, the slave select line (NSS) is asserted low one bit period before the start of
transmission, and released high one bit period after every character transmission. A delay for at
least three bit periods is always inserted in between characters. In order to address slave
devices supporting the Chip Select Active After Transfer (CSAAT) mode, NSS can be forced low
by writing a one to the Force SPI Chip Select bit (CR.RTSEN/FCS). Releasing NSS when FCS
is one, is only possible by writing a one to the Release SPI Chip Select bit (CR.RTSDIS/RCS).
In SPI slave mode, a low level on NSS for at least one bit period will allow the slave to initiate a
transmission or reception. The Underrun Error bit (CSR.UNRE) is set if a character must be sent
while THR is empty, and TXD will be high during character transmission, as if 0xFF was being
sent. If a new character is written to THR it will be sent correctly during the next transmission
slot. Writing a one to CR.RSTSTA will clear UNRE. To ensure correct behavior of the receiver in
SPI slave mode, the master device sending the frame must ensure a minimum delay of one bit
period in between each character transmission.
19.6.4.6
19.6.5
Receiver Time-out
Receiver Time-out’s are not possible in SPI mode as the baud rate clock is only active during
data transfers.
LIN Mode
The USART features a LIN (Local Interconnect Network) 1.3 and 2.0 compliant mode, embedding full error checking and reporting, automatic frame processing with up to 256 data bytes,
377
32099I–01/2012
AT32UC3L016/32/64
customizable response data lengths, and requires minimal CPU resources. Writing 0xA (master)
or 0xB (slave) to MR.MODE enables this mode.
19.6.5.1
Modes of operation
Changing LIN mode after initial configuration has to be followed by a transceiver software reset
in order to avoid unpredictable behavior.
19.6.5.2
Receiver and Transmitter Control
See Section “19.6.2” on page 367.
19.6.5.3
Baud Rate Configuration
The LIN nodes baud rate is configured in the Baud Rate Generator Register (BRGR), See Section “19.6.1.1” on page 365.
19.6.5.4
Character Transmission and Reception
See ”Transmitter Operations” on page 368, and ”Receiver Operations” on page 370.
19.6.5.5
Header Transmission (Master Node Configuration)
All LIN frames start with a header sent by the master. As soon as the identifier has been written
to the Identifier Character field in the LIN Identifier Register (LINIR.IDCHR), TXRDY is cleared
and the header is sent. The header consists of a Break, Sync, and Identifier field. TXRDY is set
when the identifier has been transferred into the transmitters shift register.
The Break field consists of 13 dominant bits, the break, and one recessive bit, the break delimiter. The Sync field is the character 0x55. The Identifier field contains the Identifier as written to
IDCHR. The identifier parity bits can be generated automatically (see Section 19.6.5.8).
Figure 19-21. Header Transmission
Baud Rate
Clock
TXD
Break Field
13 dominant bits (at 0)
Write
LINIR
LINIR
Break
Delimiter
1 recessive bit
(at 1)
Start
1
Bit
0
1
0
1
0
Synch Byte = 0x55
1
0
Stop Start
Stop
ID0 ID1 ID2 ID3 ID4 ID5 ID6 ID7
Bit Bit
Bit
ID
TXRDY
19.6.5.6
Header Reception (Slave Node Configuration)
The USART stays idle until it detects a break field, consisting of at least 11 consecutive dominant bits (zeroes) on the bus. A received break will set the Lin Break bit (CSR.LINBK). The Sync
field is used to synchronize the baud rate (see Section 19.6.5.7). IDCHR is updated and the LIN
Identifier bit (CSR.LINID) is set when the Identifier has been received. The Identifier parity bits
can be automatically checked (see Section 19.6.5.8). Writing a one to RSTSTA will clear LINBK
and LINID.
378
32099I–01/2012
AT32UC3L016/32/64
Figure 19-22. Header Reception
Baud Rate
Clock
RXD
Break Field
13 dominant bits (at 0)
LINID
Break
Delimiter
1 recessive bit
(at 1)
Start
1
Bit
0
1
0
1
0
1
0
Synch Byte = 0x55
Stop
Stop Start
ID0 ID1 ID2 ID3 ID4 ID5 ID6 ID7
Bit
Bit Bit
US_LINIR
Write US_CR
With RSTSTA=1
19.6.5.7
Slave Node Synchronization
Synchronization is only done by the slave. If the Sync field is not 0x55, an Inconsistent Sync
Field error (CSR.LINISFE) is generated. The time between falling edges is measured by a 19-bit
counter, driven by the sampling clock (see Section 19.6.1).
Figure 19-23. Sync Field
Synch Field
8 Tbit
2 Tbit
2 Tbit
2 Tbit
2 Tbit
Start
bit
Stop
bit
The counter starts when the Sync field start bit is detected, and continues for eight bit periods.
The 16 most significant bits (counter value divided by 8) becomes the new clock divider
(BRGR.CD), and the three least significant bits (the remainder) becomes the new fractional part
(BRGR.FP).
Figure 19-24. Slave Node Synchronization
Baud Rate
Clock
RXD
Break Field
13 dominant bits (at 0)
Break
Delimiter
1 recessive bit
(at 1)
Start
1
Bit
0
1
0
1
0
Synch Byte = 0x55
1
0
Stop Start
Stop
ID0 ID1 ID2 ID3 ID4 ID5 ID6 ID7
Bit Bit
Bit
LINIDRX
Reset
Synchro Counter
000_0011_0001_0110_1101
BRGR
Clcok Divider (CD)
Initial CD
0000_0110_0010_1101
BRGR
Fractional Part (FP)
Initial FP
101
The synchronization accuracy depends on:
• The theoretical slave node clock frequency; nominal clock frequency (FNom)
• The baud rate
379
32099I–01/2012
AT32UC3L016/32/64
• The oversampling mode (OVER=0 => 16x, or OVER=1 => 8x)
The following formula is used to calculate synchronization deviation, where FSLAVE is the real
slave node clock frequency, and FTOL_UNSYNC is the difference between FNom and FSLAVE According to the LIN specification, FTOL_UNSYNCH may not exceed ±15%, and the bit rates between two
nodes must be within ±2% of each other, resulting in a maximal BaudRate_deviation of ±1%.
[ α × 8 × ( 2 – OVER ) + β ] × BaudRate
BaudRate_deviation = ⎛ 100 × ---------------------------------------------------------------------------------------------------⎞ %
⎝
⎠
8 × F SLAVE
⎛
⎞
⎜
[ α × 8 × ( 2 – OVER ) + β ] × BaudRate⎟
BaudRate_deviation = ⎜ 100 × ---------------------------------------------------------------------------------------------------⎟ %
F TOL_UNSYNC
⎜
⎟
8 × ⎛ ------------------------------------⎞ xF Nom
⎝
⎠
⎝
⎠
100
– 0,5 ≤ α ≤ +0,5
-1 < β < +1
Minimum nominal clock frequency with a fractional part:
⎛
⎞
⎜
[ 0,5 × 8 × ( 2 – OVER ) + 1 ] × BaudRate⎟
F Nom ( min ) = ⎜ 100 × -------------------------------------------------------------------------------------------------------⎟ Hz
– 15
⎜
⎟
8 × ⎛ ---------- + 1⎞ × 1%
⎝
⎠
⎝ 100
⎠
Examples:
• Baud rate = 20 kbit/s, OVER=0 (Oversampling 16x) => FNom(min) = 2.64 MHz
• Baud rate = 20 kbit/s, OVER=1 (Oversampling 8x) => FNom(min) = 1.47 MHz
• Baud rate = 1 kbit/s, OVER=0 (Oversampling 16x) => FNom(min) = 132 kHz
• Baud rate = 1 kbit/s, OVER=1 (Oversampling 8x) => FNom(min) = 74 kHz
If the fractional part is not used, the synchronization accuracy is much lower. The 16 most significant bits, added with the first least significant bit, becomes the new clock divider (CD). The
equation of the baud rate deviation is the same as above, but the constants are:
– 4 ≤ α ≤ +4
-1 < β < +1
Minimum nominal clock frequency without a fractional part:
⎛
⎞
⎜
[ 4 × 8 × ( 2 – OVER ) + 1 ] × Baudrate⎟
F Nom ( min ) = ⎜ 100 × -----------------------------------------------------------------------------------------------⎟ Hz
– 15
⎜
⎟
8 × ⎛ ---------- + 1⎞ × 1%
⎝
⎠
⎝ 100
⎠
Examples:
• Baud rate = 20 kbit/s, OVER=0 (Oversampling 16x) => FNom(min) = 19.12 MHz
• Baud rate = 20 kbit/s, OVER=1 (Oversampling 8x) => FNom(min) = 9.71 MHz
• Baud rate = 1 kbit/s, OVER=0 (Oversampling 16x) => FNom(min) = 956 kHz
• Baud rate = 1 kbit/s, OVER=1 (Oversampling 8x) => FNom(min) = 485 kHz
19.6.5.8
Identifier Parity
An identifier field consists of two sub-fields; the identifier and its parity. Bits 0 to 5 are assigned
to the identifier, while bits 6 and 7 are assigned to parity. Automatic parity management is disabled by writing a one to the Parity Disable bit in the LIN Mode register (LINMR.PARDIS).
380
32099I–01/2012
AT32UC3L016/32/64
• PARDIS=0: During header transmission, the parity bits are computed and in the shift register
they replace bits six and seven from IDCHR. During header reception, the parity bits are
checked and can generate a LIN Identifier Parity Error (see Section 19.6.6). Bits six and
seven in IDCHR read as zero when receiving.
• PARDIS=1: During header transmission, all the bits in IDCHR are sent on the bus. During
header reception, all the bits in IDCHR are updated with the received Identifier.
19.6.5.9
Node Action
After an identifier transaction, a LIN response mode has to be selected. This is done in the Node
Action field (LINMR.NACT). Below are some response modes exemplified in a small LIN cluster:
• Response, from master to slave1:
Master: NACT=PUBLISH
Slave1: NACT=SUBSCRIBE
Slave2: NACT=IGNORE
• Response, from slave1 to master:
Master: NACT=SUBSCRIBE
Slave1: NACT=PUBLISH
Slave2: NACT=IGNORE
• Response, from slave1 to slave2:
Master: NACT=IGNORE
Slave1: NACT=PUBLISH
Slave2: NACT=SUBSCRIBE
19.6.5.10
LIN Response Data Length
The response data length is the number of data fields (bytes), excluding the checksum.
Figure 19-25. Response Data Length
User configuration: 1 - 256 data fields (DLC+1)
Identifier configuration: 2/4/8 data fields
Sync
Break
Sync
Field
Identifier
Field
Data
Field
Data
Field
Data
Field
Data
Field
Checksum
Field
The response data length can be configured, either by the user, or automatically by bits 4 and 5
in the Identifier (IDCHR), in accordance to LIN 1.1. The user selects mode by writing to the Data
Length Mode bit (LINMR.DML):
• DLM=0: the response data length is configured by the user by writing to the 8-bit Data Length
Control field (LINMR.DLC). The response data length equals DLC + 1 bytes.
381
32099I–01/2012
AT32UC3L016/32/64
• DLM=1: the response data length is defined by the Identifier bits according to the table below.
Table 19-8.
19.6.5.11
Response Data Length if DLM = 1
IDCHR[5]
IDCHR[4]
Response Data Length [bytes]
0
0
2
0
1
2
1
0
4
1
1
8
Checksum
The last frame field is the checksum. It is configured by the Checksum Type (LINMR.CHKTYP),
and the Checksum Disable (LINMR.CHKDIS) bits. TXRDY will not be set after the last THR data
write if enabled. Writing a one to CHKDIS will disable the automatic checksum generation/checking, and the user may send/check this last byte manually, disguised as a normal data.
The checksum is an inverted 8-bit sum with carry, either:
• over all data bytes, called a classic checksum. This is used for LIN 1.3 compliant slaves, and
automatically managed when CHKDIS=0, and CHKTYP=1.
• over all data bytes and the protected identifier, called an enhanced checksum. This is used
for LIN 2.0 compliant slaves, and automatically managed when CHKDIS=0, and CHKTYP=0.
19.6.5.12
Frame Slot Mode
A LIN master can be configured to use frame slots with a pre-defined minimum length. Writing a
one to the Frame Slot Mode Disable bit (LINMR.FSDIS) disables this mode. This mode will not
allow TXRDY to be set after a frame transfer until the entire frame slot duration has elapsed, in
effect preventing the master from sending a new header. The LIN Transfer Complete bit
(CSR.LINTC) will still be set after the checksum has been sent. Writing a one to CR.RSTST
clears LINTC.
Figure 19-26. Frame Slot Mode with Automatic Checksum
Frame slot = TFrame_Maximum
Frame
Header
Break
Synch
Data3
Interframe
space
Response
space
Protected
Identifier
Response
Data 1
Data N-1
Data N
Checksum
TXRDY
Frame Slot Mode Frame Slot Mode
Disabled
Enabled
Write
LINID
Write
THR
Data 1
Data 2
Data 3
Data N
LINTC
The minimum frame slot size is determined by TFrame_Maximum, and calculated below (all values in bit periods):
• THeader_Nominal = 34
382
32099I–01/2012
AT32UC3L016/32/64
• TFrame_Maximum = 1.4 x (THeader_Nominal + TResponse_Nominal + 1)(Note:)
Note:
The term “+1” leads to an integer result for TFrame_Max (LIN Specification 1.3)
If the Checksum is sent (CHKDIS=0):
• TResponse_Nominal = 10 x (NData + 1)
• TFrame_Maximum = 1.4 x (34 + 10 x (DLC + 1 + 1) + 1)
• TFrame_Maximum = 77 + 14 x DLC
If the Checksum is not sent (CHKDIS=1):
• TResponse_Nominal = 10 x NData
• TFrame_Maximum = 1.4 x (34 + 10 x (DLC + 1) + 1)
• TFrame_Maximum = 63 + 14 x DLC
19.6.6
LIN Errors
These error bits are cleared by writing a one to CSR.RSTSTA.
19.6.6.1
Slave Not Responding Error (CSR.LINSNRE)
This error is generated if no valid message appears within the TFrame_Maximum time frame
slot, while the USART is expecting a response from another node (NACT=SUBSCRIBE).
19.6.6.2
Checksum Error (CSR.LINCE)
This error is generated if the received checksum is wrong. This error can only be generated if the
checksum feature is enabled (CHKDIS=0).
19.6.6.3
Identifier Parity Error (CSR.LINIPE)
This error is generated if the identifier parity is wrong. This error can only be generated if parity is
enabled (PARDIS=0).
19.6.6.4
Inconsistent Sync Field Error (CSR.LINISFE)
This error is generated in slave mode if the Sync Field character received is not 0x55. Synchronization procedure is aborted.
19.6.6.5
Bit Error (CSR.LINBE)
This error is generated if the value transmitted by the USART on Tx differs from the value sampled on Rx. If a bit error is detected, the transmission is aborted at the next byte border.
19.6.7
LIN Frame Handling
19.6.7.1
Master Node Configuration
• Write a one to CR.TXEN and CR.RXEN to enable both transmitter and receiver
• Select LIN mode and master node by writing to MR.MODE
• Configure the baud rate by writing to CD and FP in BRGR
• Configure the frame transfer by writing to NACT, PARDIS, CHKDIS, CHKTYPE, DLCM,
FSDIS, and DLC in LINMR
• Check that CSR.TXRDY is one
• Send the header by writing to LINIR.IDCHR
The following procedure depends on the NACT setting:
383
32099I–01/2012
AT32UC3L016/32/64
• Case 1: NACT=PUBLISH, the USART sends a response
– Wait until TXRDY is a one
– Send a byte by writing to THR.TXCHR
– Repeat the two previous steps until there is no more data to send
– Wait until CSR.LINTC is a one
– Check for LIN errors
• Case 2: NACT=SUBSCRIBE, the USART receives a response
– Wait until RXRDY is a one
– Read RHR.RXCHR
– Repeat the two previous steps until there is no more data to read
– Wait until LINTC is a one
– Check for LIN errors
• Case 3: NACT=IGNORE, the USART is not concerned by a response
– Wait until LINTC is a one
– Check for LIN errors
Figure 19-27. Master Node Configuration, NACT=PUBLISH
Frame slot = TFrame_Maximum
Frame
Header
Break
Synch
Data3
Interframe
space
Response
space
Protected
Identifier
Response
Data 1
Data N-1
Data N
Checksum
TXRDY
FSDIS=1
FSDIS=0
RXRDY
Write
LINIR
Write
THR
Data 1
Data 2
Data 3
Data N
LINTC
384
32099I–01/2012
AT32UC3L016/32/64
Figure 19-28. Master Node Configuration, NACT=SUBSCRIBE
Frame slot = TFrame_Maximum
Frame
Header
Break
Synch
Data3
Interframe
space
Response
space
Protected
Identifier
Response
Data 1
Data N-1
Data N
Checksum
TXRDY
FSDIS=1 FSDIS=0
RXRDY
Write
LINIR
Read
RHR
Data 1
Data N-2
Data N-1
Data N
LINTC
Figure 19-29. Master Node Configuration, NACT=IGNORE
Frame slot = TFrame_Maximum
Frame
Break
Response
space
Header
Data3
Synch
Protected
Identifier
Interframe
space
Response
Data 1
Data N-1
Data N
Checksum
TXRDY
FSDIS=1
FSDIS=0
RXRDY
Write
LINIR
LINTC
19.6.7.2
Slave Node Configuration
This is identical to the master node configuration above, except for:
• LIN mode selected in MR.MODE is slave
• When the baud rate is configured, wait until CSR.LINID is a one, then;
• Check for LINISFE and LINPE errors, clear errors and LINIDby writing a one to RSTSTA
• Read IDCHR
• Configure the frame transfer by writing to NACT, PARDIS, CHKDIS, CHKTYPE, DLCM, and
DLC in LINMR
IMPORTANT: if NACT=PUBLISH, and this field is already correct, the LINMR register must still
be written with this value in order to set TXRDY, and to request the corresponding Peripheral
DMA Controller write transfer.
The different NACT settings result in the same procedure as for the master node, see page 383.
385
32099I–01/2012
AT32UC3L016/32/64
Figure 19-30. Slave Node Configuration, NACT=PUBLISH
Break
Synch
Protected
Identifier
Data 1
Data N-1
Data N
Checksum
Data N
Checksum
TXRDY
RXRDY
LINIDRX
Read
LINID
Write
THR
Data 1 Data 2
Data 3
Data N
LINTC
Figure 19-31. Slave Node Configuration, NACT=SUBSCRIBE
Break
Synch
Protected
Identifier
Data 1
Data N-1
TXRDY
RXRDY
LINIDRX
Read
LINID
Read
RHR
Data 1
Data N-2
Data N-1
Data N
LINTC
Figure 19-32. Slave Node Configuration, NACT=IGNORE
Break
Synch
Protected
Identifier
Data 1
Data N-1
Data N
Checksum
TXRDY
RXRDY
LINIDRX
Read
LINID
Read
RHR
LINTC
19.6.8
LIN Frame Handling With The Peripheral DMA Controller
The USART can be used together with the Peripheral DMA Controller in order to transfer data
without processor intervention. The DMA Controller uses the TXRDY and RXRDY bits, to trigger
one byte writes or reads. It always writes to THR, and it always reads RHR.
386
32099I–01/2012
AT32UC3L016/32/64
19.6.8.1
Master Node Configuration
The Peripheral DMA Controller Mode bit (LINMR.PDCM) allows the user to select configuration:
• PDCM=0: LIN configuration must be written to LINMR, it is not stored in the write buffer.
• PDCM=1: LIN configuration is written by the DMA Controller to THR, and is stored in the
write buffer. Since data transfer size is a byte, the transfer is split into two accesses. The first
writes the NACT, PARDIS, CHKDIS, CHKTYP, DLM and FSDIS bits, while the second writes
the DLC field. If NACT=PUBLISH, the write buffer will also contain the Identifier.
When NACT=SUBSCRIBE, the read buffer contains the data.
Figure 19-33. Master Node with Peripheral DMA Controller (PDCM=0)
WRITE BUFFER
DATA 0
Peripheral
bus
DATA 1
NODE ACTION = SUBSCRIBE
Peripheral
bus
READ BUFFER
Peripheral DMA
Controller
|
|
|
|
NODE ACTION = PUBLISH
USART LIN
CONTROLLER
DATA 0
RXRDY
Peripheral DMA
Controller
RXRDY
USART LIN
CONTROLLER
TXRDY
|
|
|
|
DATA N
DATA N
Figure 19-34. Master Node with Peripheral DMA Controller (PDCM=1)
WRITE BUFFER
WRITE BUFFER
NACT
PARDIS
CHKDIS
CHKTYP
DLM
FSDIS
NACT
PARDIS
CHKDIS
CHKTYP
DLM
FSDIS
DLC
DLC
Peripheral
bus
NODE ACTION = PUBLISH
Peripheral
bus
IDENTIFIER
IDENTIFIER
USART LIN
CONTROLLER
Peripheral DMA
Controller
DATA 0
|
|
|
|
DATA N
READ BUFFER
Peripheral DMA
Controller
RXRDY
NODE ACTION = SUBSCRIBE
USART LIN
CONTROLLER
RXRDY
DATA 0
TXRDY
|
|
|
|
DATA N
387
32099I–01/2012
AT32UC3L016/32/64
19.6.8.2
Slave Node Configuration
In this mode, the Peripheral DMA Controller transfers only data. The user reads the Identifier
from LINIR, and selects LIN mode by writing to LINMR. When NACT=PUBLISH the data is in the
write buffer, while the read buffer contains the data when NACT=SUBSCRIBE.
IMPORTANT: if in slave mode, NACT is already configured correctly as PUBLISH, the LINMR
register must still be written with this value in order to set TXRDY, and to request the corresponding Peripheral DMA Controller write transfer.
Figure 19-35. Slave Node with Peripheral DMA Controller
WRITE BUFFER
READ BUFFER
DATA 0
DATA 0
Peripheral
bus
|
|
|
|
|
|
|
|
USART LIN
CONTROLLER
Peripheral DMA
Controller
TXRDY
DATA N
19.6.9
NACT = SUBSCRIBE
Peripheral
Bus
Peripheral DMA
Controller
USART LIN
CONTROLLER
RXRDY
DATA N
Wake-up Request
Any node in a sleeping LIN cluster may request a wake-up. By writing to the Wakeup Signal
Type bit (LINMR.WKUPTYP), the user can choose to send either a LIN 1.3 (WKUPTYP=1), or a
LIN 2.0 (WKUPTYP=0) compliant wakeup request. Writing a one to the Send LIN Wakeup Signal bit (CR.LINWKUP), transmits a wakeup, and when completed sets LINTC.
According to LIN 1.3, the wakeup request should be generated with the character 0x80 in order
to impose eight successive dominant bits.
According to LIN 2.0, the wakeup request is issued by forcing the bus into the dominant state for
250µs to 5ms. Sending the character 0xF0 does this, regardless of baud rate.
• Baud rate max = 20 kbit/s -> one bit period = 50µs -> five bit periods = 250µs
• Baud rate min = 1 kbit/s -> one bit period = 1ms -> five bit periods = 5ms
19.6.10
Bus Idle Time-out
LIN bus inactivity should eventually cause slaves to time-out and enter sleep mode. LIN 1.3
specifies this to 25000 bit periods, whilst LIN 2.0 specifies 4seconds. For the time-out counter
operation see Section 19.6.3.8 ”Receiver Time-out” on page 372.
Table 19-9.
Receiver Time-out Values (RTOR.TO)
LIN Specification
2.0
1.3
Baud Rate
Time-out period
TO
1 000 bit/s
4 000
2 400 bit/s
9 600
9 600 bit/s
4s
38 400
19 200 bit/s
76 800
20 000 bit/s
80 000
-
25 000 bit periods
25 000
388
32099I–01/2012
AT32UC3L016/32/64
19.6.11
Test Modes
The internal loopback feature enables on-board diagnostics, and allows the USART to operate
in three different test modes, with reconfigured pin functionality, as shown below.
19.6.11.1
Normal Mode
During normal operation, a receivers RXD pin is connected to a transmitters TXD pin.
Figure 19-36. Normal Mode Configuration
RXD
Receiver
TXD
Transmitter
19.6.11.2
Automatic Echo Mode
Automatic echo mode allows bit-by-bit retransmission. When a bit is received on the RXD pin, it
is also sent to the TXD pin, as shown in Figure 19-37. Transmitter configuration has no effect.
Figure 19-37. Automatic Echo Mode Configuration
RXD
Receiver
TXD
Transmitter
19.6.11.3
Local Loopback Mode
Local loopback mode connects the output of the transmitter directly to the input of the receiver,
as shown in Figure 19-38. 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 19-38. Local Loopback Mode Configuration
RXD
Receiver
Transmitter
19.6.11.4
1
TXD
Remote Loopback Mode
Remote loopback mode connects the RXD pin to the TXD pin, as shown in Figure 19-39. The
transmitter and the receiver are disabled and have no effect. This mode allows bit-by-bit
retransmission.
389
32099I–01/2012
AT32UC3L016/32/64
Figure 19-39. Remote Loopback Mode Configuration
Receiver
1
RXD
TXD
Transmitter
19.6.12
Write Protection Registers
To prevent single software errors from corrupting USART behavior, certain address spaces can
be write-protected by writing the correct Write Protect KEY and a one to the Write Protect
Enable bit in the Write Protect Mode Register (WPMR.WPKEY, and WPMR.WPEN). Disabling
the write protection is done by writing the correct key, and a zero to WPEN.
Write attempts to a write protected register are detected and the Write Protect Violation Status
bit in the Write Protect Status Register (WPSR.WPVS) is set, while the Write Protect Violation
Source field (WPSR.WPVSRC) indicates the targeted register. Writing the correct key to the
Write Protect KEY bit (WPMR.WPKEY) clears WPVSRC and WPVS.
The protected registers are:
• ”Mode Register” on page 394
• ”Baud Rate Generator Register” on page 404
• ”Receiver Time-out Register” on page 405
• ”Transmitter Timeguard Register” on page 406
390
32099I–01/2012
AT32UC3L016/32/64
19.7
User Interface
Table 19-10. USART Register Memory Map
Offset
Register
Name
Access
Reset
0x0000
Control Register
CR
Write-only
0x00000000
0x0004
Mode Register
MR
Read-write
0x00000000
0x0008
Interrupt Enable Register
IER
Write-only
0x00000000
0x000C
Interrupt Disable Register
IDR
Write-only
0x00000000
0x0010
Interrupt Mask Register
IMR
Read-only
0x00000000
0x0014
Channel Status Register
CSR
Read-only
0x00000000
0x0018
Receiver Holding Register
RHR
Read-only
0x00000000
0x001C
Transmitter Holding Register
THR
Write-only
0x00000000
0x0020
Baud Rate Generator Register
BRGR
Read-write
0x00000000
0x0024
Receiver Time-out Register
RTOR
Read-write
0x00000000
0x0028
Transmitter Timeguard Register
TTGR
Read-write
0x00000000
0x0054
LIN Mode Register
LINMR
Read-write
0x00000000
0x0058
LIN Identifier Register
LINIR
Read-write
0x00000000
0x00E4
Write Protect Mode Register
WPMR
Read-write
0x00000000
0x00E8
Write Protect Status Register
WPSR
Read-only
0x00000000
0x00FC
Version Register
VERSION
Read-only
0x–(1)
Note:
1. Values in the Version Register vary with the version of the IP block implementation.
391
32099I–01/2012
AT32UC3L016/32/64
19.7.1
Name:
Control Register
CR
Access Type:
Write-only
Offset:
0x0
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
LINWKUP
20
LINABT
19
RTSDIS/RCS
18
RTSEN/FCS
17
–
16
–
15
RETTO
14
RSTNACK
13
–
12
SENDA
11
STTTO
10
STPBRK
9
STTBRK
8
RSTSTA
7
TXDIS
6
TXEN
5
RXDIS
4
RXEN
3
RSTTX
2
RSTRX
1
–
0
–
• LINWKUP: Send LIN Wakeup Signal
Writing a zero to this bit has no effect.
Writing a one to this bit will sends a wakeup signal on the LIN bus.
• LINABT: Abort LIN Transmission
Writing a zero to this bit has no effect.
Writing a one to this bit will abort the current LIN transmission.
• RTSDIS/RCS: Request to Send Disable/Release SPI Chip Select
Writing a zero to this bit has no effect.
Writing a one to this bit when USART is not in SPI master mode drives RTS pin high.
Writing a one to this bit when USART is in SPI master mode releases NSS (RTS pin).
• RTSEN/FCS: Request to Send Enable/Force SPI Chip Select
Writing a zero to this bit has no effect.
Writing a one to this bit when USART is not in SPI master mode drives RTS low.
Writing a one to this bit when USART is in SPI master mode when;
FCS=0: has no effect.
FCS=1: forces NSS (RTS pin) low, even if USART is not transmitting, in order to address SPI slave devices supporting the
CSAAT Mode (Chip Select Active After Transfer).
• RETTO: Rearm Time-out
Writing a zero to this bit has no effect.
Writing a one to this bit reloads the time-out counter and clears CSR.TIMEOUT.
• RSTNACK: Reset Non Acknowledge
Writing a zero to this bit has no effect.
Writing a one to this bit clears CSR.NACK.
• SENDA: Send Address
Writing a zero to this bit has no effect.
Writing a one to this bit will in multidrop mode send the next character written to THR as an address.
• STTTO: Start Time-out
Writing a zero to this bit has no effect.
Writing a one to this bit will abort any current time-out count down, and trigger a new count down when the next character has
been received. CSR.TIMEOUT is also cleared.
392
32099I–01/2012
AT32UC3L016/32/64
• STPBRK: Stop Break
Writing a zero to this bit has no effect.
Writing a one to this bit will stop the generation of break signal characters, and then send ones for TTGR.TG duration, or at least
12 bit periods. No effect if no break is being transmitted.
• STTBRK: Start Break
Writing a zero to this bit has no effect.
Writing a one to this bit will start transmission of break characters when current characters present in THR and the transmit shift
register have been sent. No effect if a break signal is already being generated.
• RSTSTA: Reset Status Bits
Writing a zero to this bit has no effect.
Writing a one to this bit will clear the following bits in CSR: PARE, FRAME, OVRE, LINBE, LINSFE, LINIPE, LINCE, LINSNRE,
and RXBRK.
• TXDIS: Transmitter Disable
Writing a zero to this bit has no effect.
Writing a one to this bit disables the transmitter.
• TXEN: Transmitter Enable
Writing a zero to this bit has no effect.
Writing a one to this bit enables the transmitter if TXDIS is zero.
• RXDIS: Receiver Disable
Writing a zero to this bit has no effect.
Writing a one to this bit disables the receiver.
• RXEN: Receiver Enable
Writing a zero to this bit has no effect.
Writing a one to this bit enables the receiver if RXDIS is zero.
• RSTTX: Reset Transmitter
Writing a zero to this bit has no effect.
Writing a one to this bit will reset the transmitter.
• RSTRX: Reset Receiver
Writing a zero to this bit has no effect.
Writing a one to this bit will reset the receiver.
393
32099I–01/2012
AT32UC3L016/32/64
19.7.2
Name:
Mode Register
MR
Access Type:
Read-write
Offset:
0x4
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
25
–
24
23
–
22
–
21
–
20
INACK
19
OVER
18
CLKO
17
MODE9
16
MSBF/CPOL
14
13
12
11
10
PAR
9
8
SYNC/CPHA
4
3
2
1
0
15
CHMODE
7
NBSTOP
6
CHRL
5
USCLKS
MODE
This register can only be written if the WPEN bit is cleared in the Write Protect Mode Register.
• INACK: Inhibit Non Acknowledge
0: The NACK is generated.
1: The NACK is not generated.
• OVER: Oversampling Mode
0: Oversampling at 16 times the baud rate.
1: Oversampling at 8 times the baud rate.
• CLKO: Clock Output Select
0: The USART does not drive the CLK pin.
1: The USART drives the CLK pin unless USCLKS selects the external clock.
• 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:
MSBF=0: Least Significant Bit is sent/received first.
MSBF=1: Most Significant Bit is sent/received first.
If USART operates in SPI Mode, CPOL is used with CPHA to produce the required clock/data relationship between devices.
CPOL=0: The inactive state value of CLK is logic level zero.
CPOL=1: The inactive state value of CLK is logic level one.
394
32099I–01/2012
AT32UC3L016/32/64
• CHMODE: Channel Mode
Table 19-11.
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
Table 19-12.
NBSTOP
Asynchronous (SYNC=0)
Synchronous (SYNC=1)
0
0
1 stop bit
1 stop bit
0
1
1.5 stop bits
Reserved
1
0
2 stop bits
2 stop bits
1
1
Reserved
Reserved
• PAR: Parity Type
Table 19-13.
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, CPHA determines which edge of CLK 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.
CPHA = 0: Data is changed on the leading edge of CLK and captured on the following edge of CLK.
CPHA = 1: Data is captured on the leading edge of CLK and changed on the following edge of CLK.
395
32099I–01/2012
AT32UC3L016/32/64
• CHRL: Character Length.
Table 19-14.
CHRL
Character Length
0
0
5 bits
0
1
6 bits
1
0
7 bits
1
1
8 bits
• USCLKS: Clock Selection
Table 19-15.
USCLKS
Note:
Selected Clock
0
0
CLK_USART
0
1
CLK_USART/DIV(1)
1
0
Reserved
1
1
CLK
1. The value of DIV is device dependent. Please refer to the Module Configuration section at the end of this chapter.
• MODE
Table 19-16.
MODE
Mode of the USART
0
0
0
0
Normal
0
0
1
0
Hardware Handshaking
1
0
1
0
LIN Master
1
0
1
1
LIN Slave
1
1
1
0
SPI Master
1
1
1
1
SPI Slave
Others
Reserved
396
32099I–01/2012
AT32UC3L016/32/64
19.7.3
Name:
Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x8
Reset Value:
0x00000000
31
–
30
–
29
LINSNRE
28
LINCE
27
LINIPE
26
LINISFE
25
LINBE
24
–
23
–
22
–
21
–
20
–
19
CTSIC
18
–
17
–
16
–
15
LINTC
14
LINID
13
NACK/LINBK
12
RXBUFF
11
–
10
ITER/UNRE
9
TXEMPTY
8
TIMEOUT
7
PARE
6
FRAME
5
OVRE
4
–
3
–
2
RXBRK
1
TXRDY
0
RXRDY
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
397
32099I–01/2012
AT32UC3L016/32/64
19.7.4
Name:
Interrupt Disable Register
IDR
Access Type:
Write-only
Offset:
0xC
Reset Value:
0x00000000
31
–
30
–
29
LINSNRE
28
LINCE
27
LINIPE
26
LINISFE
25
LINBE
24
–
23
–
22
–
21
–
20
–
19
CTSIC
18
–
17
–
16
–
15
LINTC
14
LINID
13
NACK/LINBK
12
RXBUFF
11
–
10
ITER/UNRE
9
TXEMPTY
8
TIMEOUT
7
PARE
6
FRAME
5
OVRE
4
–
3
–
2
RXBRK
1
TXRDY
0
RXRDY
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
398
32099I–01/2012
AT32UC3L016/32/64
19.7.5
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
Offset:
0x10
Reset Value:
0x00000000
31
–
30
–
29
LINSNRE
28
LINCE
27
LINIPE
26
LINISFE
25
LINBE
24
–
23
–
22
–
21
–
20
–
19
CTSIC
18
–
17
–
16
–
15
LINTC
14
LINID
13
NACK/LINBK
12
RXBUFF
11
–
10
ITER/UNRE
9
TXEMPTY
8
TIMEOUT
7
PARE
6
FRAME
5
OVRE
4
–
3
–
2
RXBRK
1
TXRDY
0
RXRDY
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
A bit in this register is cleared when the corresponding bit in IDR is written to one.
A bit in this register is set when the corresponding bit in IER is written to one.
399
32099I–01/2012
AT32UC3L016/32/64
19.7.6
Name:
Channel Status Register
CSR
Access Type:
Read-only
Offset:
0x14
Reset Value:
0x00000000
31
–
30
–
29
LINSNRE
28
LINCE
27
LINIPE
26
LINISFE
25
LINBE
24
–
23
CTS
22
–
21
–
20
–
19
CTSIC
18
–
17
–
16
–
15
LINTC
14
LINID
13
NACK/LINBK
12
RXBUFF
11
–
10
ITER/UNRE
9
TXEMPTY
8
TIMEOUT
7
PARE
6
FRAME
5
OVRE
4
–
3
–
2
RXBRK
1
TXRDY
0
RXRDY
• LINSNRE: LIN Slave Not Responding Error
0: No LIN Slave Not Responding Error has been detected since the last RSTSTA.
1: A LIN Slave Not Responding Error has been detected since the last RSTSTA.
• LINCE: LIN Checksum Error
0: No LIN Checksum Error has been detected since the last RSTSTA.
1: A LIN Checksum Error has been detected since the last RSTSTA.
• LINIPE: LIN Identifier Parity Error
0: No LIN Identifier Parity Error has been detected since the last RSTSTA.
1: A LIN Identifier Parity Error has been detected since the last RSTSTA.
• LINISFE: LIN Inconsistent Sync Field Error
0: No LIN Inconsistent Sync Field Error has been detected since the last RSTSTA
1: The USART is configured as a Slave node and a LIN Inconsistent Sync Field Error has been detected since the last RSTSTA.
• LINBE: LIN Bit Error
0: No Bit Error has been detected since the last RSTSTA.
1: A Bit Error has been detected since the last RSTSTA.
• CTS: Image of CTS Input
0: CTS is low.
1: CTS is high.
• CTSIC: Clear to Send Input Change Flag
0: No change has been detected on the CTS pin since the last CSR read.
1: At least one change has been detected on the CTS pin since the last CSR read.
• LINTC: LIN Transfer Completed
0: The USART is either idle or a LIN transfer is ongoing.
1: A LIN transfer has been completed since the last RSTSTA.
• LINID: LIN Identifier
0: No LIN Identifier has been sent or received.
1: A LIN Identifier has been sent (master) or received (slave), since the last RSTSTA.
• NACK: Non Acknowledge
0: No Non Acknowledge has 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 Buffer Full signal from the Peripheral DMA Controller channel is inactive.
400
32099I–01/2012
AT32UC3L016/32/64
1: The Buffer Full signal from the Peripheral DMA Controller channel is active.
• ITER/UNRE: Max number of Repetitions Reached or SPI Underrun Error
If USART does not operate in SPI Slave Mode:
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.
If USART operates in SPI Slave Mode:
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: The transmitter is either disabled or there are characters in THR, or in the transmit shift register.
1: There are no characters in neither THR, nor in the transmit shift register.
• TIMEOUT: Receiver Time-out
0: There has not been a time-out since the last Start Time-out command (CR.STTTO), or RTOR.TO is zero.
1: There has been a time-out since the last Start Time-out command.
• PARE: Parity Error
0: Either no parity error has been detected, or the parity bit is a zero in multidrop mode, since the last RSTSTA.
1: Either at least one parity error has been detected, or the parity bit is a one in multidrop mode, since the last RSTSTA.
• FRAME: Framing Error
0: No stop bit has been found as low since the last RSTSTA.
1: At least one stop bit has been found as 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.
• 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: The transmitter is either disabled, or a character in THR is waiting to be transferred to the transmit shift register, or an
STTBRK command has been requested. As soon as the transmitter is enabled, TXRDY becomes one.
1: There is no character in the THR.
• RXRDY: Receiver Ready
0: The receiver is either disabled, or no complete character has been received since the last read of RHR. If characters were
being received when the receiver was disabled, RXRDY changes to 1 when the receiver is enabled.
1: At least one complete character has been received and RHR has not yet been read.
401
32099I–01/2012
AT32UC3L016/32/64
19.7.7
Name:
Receiver Holding Register
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
–
14
–
13
–
12
–
11
–
10
–
9
–
8
RXCHR[8]
7
6
5
4
3
2
1
0
RXCHR[7:0]
• RXCHR: Received Character
Last received character.
402
32099I–01/2012
AT32UC3L016/32/64
19.7.8
Name:
Transmitter Holding Register
THR
Access Type:
Write-only
Offset:
0x1C
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
TXCHR[8]
7
6
5
4
3
2
1
0
TXCHR[7:0]
• TXCHR: Character to be Transmitted
If TXRDY is zero this field contains the next character to be transmitted.
403
32099I–01/2012
AT32UC3L016/32/64
19.7.9
Name:
Baud Rate Generator Register
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[15:8]
7
6
5
4
CD[7:0]
This register can only be written to if write protection is disabled, see ”Write Protect Mode Register” on page 410.
• FP: Fractional Part
0: Fractional divider is disabled.
1 - 7: Baud rate resolution, defined by FP x 1/8.
• CD: Clock Divider
Table 19-17.
SYNC = 1
or
MODE = SPI
(Master or Slave)
SYNC = 0
CD
OVER = 0
0
1 to 65535
OVER = 1
Baud Rate Clock Disabled
Baud Rate =
Selected Clock/16/CD
Baud Rate =
Selected Clock/8/CD
Baud Rate =
Selected Clock /CD
404
32099I–01/2012
AT32UC3L016/32/64
19.7.10
Name:
Receiver Time-out Register
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
TO[16]
15
14
13
12
11
10
9
8
3
2
1
0
TO[15:8]
7
6
5
4
TO[7:0]
This register can only be written to if write protection is disabled, see ”Write Protect Mode Register” on page 410.
• TO: Time-out Value
0: The receiver Time-out is disabled.
1 - 131071: The receiver Time-out is enabled and the time-out delay is TO x bit period.
Note that the size of the TO counter is device dependent, see the Module Configuration section.
405
32099I–01/2012
AT32UC3L016/32/64
19.7.11
Name:
Transmitter Timeguard Register
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
This register can only be written to if write protection is disabled, see ”Write Protect Mode Register” on page 410.
• 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.
406
32099I–01/2012
AT32UC3L016/32/64
19.7.12
Name:
LIN Mode Register
LINMR
Access Type:
Read-write
Offset:
0x54
Reset Value:
0x00000000
31
–
30
–
29
–
28
–
27
–
26
–
25
–
24
–
23
–
22
–
21
–
20
–
19
–
18
–
17
–
16
PDCM
15
14
13
12
11
10
9
8
3
CHKDIS
2
PARDIS
1
DLC
7
WKUPTYP
6
FSDIS
5
DLM
4
CHKTYP
0
NACT
• PDCM: Peripheral DMA Controller Mode
0: The LIN mode register is not written by the Peripheral DMA Controller.
1: The LIN mode register is, except for this bit, written by the Peripheral DMA Controller.
• DLC: Data Length Control
0 - 255: If DLM=0 this field defines the response data length to DLC+1 bytes.
• WKUPTYP: Wakeup Signal Type
0: Writing a one to CR.LINWKUP will send a LIN 2.0 wakeup signal.
1: Writing a one to CR.LINWKUP will send a LIN 1.3 wakeup signal.
• FSDIS: Frame Slot Mode Disable
0: The Frame Slot mode is enabled.
1: The Frame Slot mode is disabled.
• DLM: Data Length Mode
0: The response data length is defined by DLC.
1: The response data length is defined by bits 4 and 5 of the Identifier (LINIR.IDCHR).
• CHKTYP: Checksum Type
0: LIN 2.0 “Enhanced” checksum
1: LIN 1.3 “Classic” checksum
• CHKDIS: Checksum Disable
0: Checksum is automatically computed and sent when master, and checked when slave.
1: Checksum is not computed and sent, nor checked.
• PARDIS: Parity Disable
0: Identifier parity is automatically computed and sent when master, and checked when slave.
1: Identifier parity is not computed and sent, nor checked.
• NACT: LIN Node Action
Table 19-18.
NACT
0
Mode Description
0
PUBLISH: The USART transmits the response.
407
32099I–01/2012
AT32UC3L016/32/64
Table 19-18.
0
1
SUBSCRIBE: The USART receives the response.
1
0
IGNORE: The USART does not transmit and does not receive the response.
1
1
Reserved
408
32099I–01/2012
AT32UC3L016/32/64
19.7.13
Name:
LIN Identifier Register
LINIR
Access Type:
Read-write or Read-only
Offset:
0x58
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
IDCHR
• IDCHR: Identifier Character
If USART is in LIN master mode, the IDCHR field is read-write, and its value is the Identifier character to be transmitted.
If USART is in LIN slave mode, the IDCHR field is read-only, and its value is the last received Identifier character.
409
32099I–01/2012
AT32UC3L016/32/64
19.7.14 Write Protect Mode Register
Register Name:
WPMR
Access Type:
Read-write
Offset:
0xE4
Reset Value:
See Table 19-10
31
30
29
28
27
WPKEY[23:16]
26
25
24
23
22
21
20
19
18
17
16
11
10
9
8
3
—
2
—
1
—
0
WPEN
WPKEY[15:8]
15
14
13
12
WPKEY[7:0]
7
—
6
—
5
—
4
—
• WPKEY: Write Protect KEY
Has to be written to 0x555341 (“USA” in ASCII) in order to successfully write WPEN. Always reads as zero.
• WPEN: Write Protect Enable
0 = Write protection disabled.
1 = Write protection enabled.
Protects the registers:
• ”Mode Register” on page 394
• ”Baud Rate Generator Register” on page 404
• ”Receiver Time-out Register” on page 405
• ”Transmitter Timeguard Register” on page 406
410
32099I–01/2012
AT32UC3L016/32/64
19.7.15 Write Protect Status Register
Register Name:
WPSR
Access Type:
Read-only
Offset:
0xE8
Reset Value:
See Table 19-10
31
—
30
—
29
—
28
—
27
—
26
—
25
—
24
—
23
22
21
20
19
WPVSRC[15:8]
18
17
16
15
14
13
12
11
10
9
8
3
—
2
—
1
—
0
WPVS
WPVSRC[7:0]
7
—
6
—
5
—
4
—
• WPVSRC: Write Protect Violation Source
If WPVS=1 this field indicates which write-protected register was unsuccessfully written to, either by address offset or code.
• WPVS: Write Protect Violation Status
0= No write protect violation has occurred since the last WPSR read.
1= A write protect violation has occurred since the last WPSR read.
Note:
Reading WPSR automatically clears all fields.
411
32099I–01/2012
AT32UC3L016/32/64
19.7.16
Version Register
Name:
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
–
11
10
9
VERSION[11:8]
8
7
6
5
4
3
2
0
MFN
1
VERSION[7:0]
• MFN
Reserved. No functionality associated.
• VERSION
Version of the module. No functionality associated.
412
32099I–01/2012
AT32UC3L016/32/64
19.8
Module Configuration
The specific configuration for each USART instance is listed in the following tables.The module
bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details.
Table 19-19. USART Configuration
Feature
USART0
USART1
USART2
USART3
17 bit
17 bit
17 bit
17 bit
8
8
8
8
Receiver Time-out Counter Size
(Size of the RTOR.TO field)
DIV Value for divided CLK_USART
Table 19-20. USART Clocks
Module Name
Clock Name
Description
USART0
CLK_USART0
Clock for the USART0 bus interface
USART1
CLK_USART1
Clock for the USART1 bus interface
USART2
CLK_USART2
Clock for the USART2 bus interface
USART3
CLK_USART3
Clock for the USART3 bus interface
Table 19-21. Register Reset Values
Register
Reset Value
VERSION
0x00000440
413
32099I–01/2012
AT32UC3L016/32/64
20. Serial Peripheral Interface (SPI)
Rev: 2.1.1.3
20.1
Features
• Compatible with an embedded 32-bit microcontroller
• 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
– 4 - 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 Peripheral DMA Controller channel capabilities optimizes data transfers
– One channel for the receiver, one channel for the transmitter
– Next buffer support
– Four character FIFO in reception
20.2
Overview
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.
414
32099I–01/2012
AT32UC3L016/32/64
20.3
Block Diagram
Figure 20-1. SPI Block Diagram
Peripheral DMA
Controller
Peripheral Bus
SPCK
MISO
CLK_SPI
MOSI
Spi Interface
I/O
Controller
NPCS0/NSS
NPCS1
NPCS2
Interrupt Control
NPCS3
SPI Interrupt
20.4
Application Block Diagram
Figure 20-2. Application Block Diagram: Single Master/Multiple Slave Implementation
Spi Master
SPCK
SPCK
MISO
MISO
MOSI
MOSI
NPCS0
Slave 0
NSS
NPCS1
NPCS2
NPCS3
NC
SPCK
MISO
MOSI
Slave 1
NSS
SPCK
MISO
MOSI
Slave 2
NSS
415
32099I–01/2012
AT32UC3L016/32/64
20.5
I/O Lines Description
Table 20-1.
I/O Lines Description
Type
20.6
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
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
20.6.1
I/O Lines
The pins used for interfacing the compliant external devices may be multiplexed with I/O lines.
The user must first configure the I/O Controller to assign the SPI pins to their peripheral
functions.
20.6.2
Clocks
The clock for the SPI bus interface (CLK_SPI) is generated by the Power Manager. This clock is
enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the
SPI before disabling the clock, to avoid freezing the SPI in an undefined state.
20.6.3
Interrupts
The SPI interrupt request line is connected to the interrupt controller. Using the SPI interrupt
requires the interrupt controller to be programmed first.
20.7
20.7.1
Functional Description
Modes of Operation
The SPI operates in master mode or in slave mode.
Operation in master mode is configured by writing a one to the Master/Slave Mode bit in the
Mode Register (MR.MSTR). 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 MR.MSTR bit is written to zero, 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.
416
32099I–01/2012
AT32UC3L016/32/64
20.7.2
Data Transfer
Four combinations of polarity and phase are available for data transfers. The clock polarity is
configured with the Clock Polarity bit in the Chip Select Registers (CSRn.CPOL). The clock
phase is configured with the Clock Phase bit in the CSRn registers (CSRn.NCPHA). These two
bits determine the edges of the clock signal on which data is driven and sampled. Each of the
two bits 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 20-2 on page 417 shows the four modes and corresponding parameter settings.
Table 20-2.
SPI modes
SPI Mode
CPOL
NCPHA
0
0
1
1
0
0
2
1
1
3
1
0
Figure 20-3 on page 417 and Figure 20-4 on page 418 show examples of data transfers.
Figure 20-3. SPI Transfer Format (NCPHA = 1, 8 bits per transfer)
SPCK cycle (for reference)
2
1
3
4
5
6
7
8
SPCK
(CPOL = 0)
SPCK
(CPOL = 1)
MOSI
(from master)
MISO
(from slave)
MSB
MSB
6
5
4
3
2
1
LSB
6
5
4
3
2
1
LSB
***
NSS
(to slave)
*** Not Defined, but normaly MSB of previous character received
417
32099I–01/2012
AT32UC3L016/32/64
Figure 20-4. SPI Transfer Format (NCPHA = 0, 8 bits per transfer)
SPCK cycle (for reference)
1
2
3
4
5
6
7
8
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 normaly LSB of previous character transmitted
20.7.3
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 (TDR) and the Receive Data
Register (RDR), 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 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 to the TDR, the Peripheral Chip Select field in TDR (TDR.PCS) must be written in
order to select a slave.
If new data is written to 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 Transmit Data Register Empty bit in the Status Register (SR.TDRE). When new data is written in TDR, this bit is
cleared. The SR.TDRE bit is used to trigger the Transmit Peripheral DMA Controller channel.
The end of transfer is indicated by the Transmission Registers Empty bit in the SR register
(SR.TXEMPTY). If a transfer delay (CSRn.DLYBCT) is greater than zero for the last transfer,
SR.TXEMPTY is set after the completion of said delay. The CLK_SPI can be switched off at this
time.
During reception, received data are transferred from the Shift Register to the reception FIFO.
The FIFO can contain up to 4 characters (both Receive Data and Peripheral Chip Select fields).
While a character of the FIFO is unread, the Receive Data Register Full bit in SR remains high
(SR.RDRF). Characters are read through the RDR register. If the four characters stored in the
FIFO are not read and if a new character is stored, this sets the Overrun Error Status bit in the
SR register (SR.OVRES). The procedure to follow in such a case is described in Section
20.7.3.8.
418
32099I–01/2012
AT32UC3L016/32/64
Figure 20-5 on page 419shows a block diagram of the SPI when operating in master mode. Figure 20-6 on page 420 shows a flow chart describing how transfers are handled.
20.7.3.1
Master mode block diagram
Figure 20-5. Master Mode Block Diagram
CSR0..3
SCBR
CLK_SPI
Baud Rate Generator
SPCK
SPI
Clock
RXFIFOEN
RDR
RDRF
OVRES
RD
CSR0..3
LSB
MISO
BITS
NCPHA
0
CPOL
1
4 – Character FIFO
MSB
Shift Register
MOSI
TDR
TD
TDRE
RXFIFOEN
RDR
CSR0..3
CSNAAT
CSAAT
PS
MR
0
1
4 – Character FIFO
NPCS3
PCSDEC
PCS
0
TDR
NPCS2
Current
Peripheral
NPCS1
PCS
NPCS0
1
MSTR
MODF
NPCS0
MODFDIS
419
32099I–01/2012
AT32UC3L016/32/64
20.7.3.2
Master mode flow diagram
Figure 20-6. Master Mode Flow Diagram
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 ?
PS ?
0
1
0
Fixed
peripheral
PS ?
1
Fixed
peripheral
0
Variable
peripheral
Variable
peripheral
TDR(PCS)
= NPCS ?
no
NPCS = TDR(PCS)
NPCS = MR(PCS)
yes
MR(PCS)
= NPCS ?
no
NPCS = 0xF
NPCS = 0xF
Delay DLYBCS
Delay DLYBCS
NPCS = TDR(PCS)
NPCS = MR(PCS),
TDR(PCS)
Delay DLYBS
Serializer = TDR(TD)
TDRE = 1
Data Transfer
RDR(RD) = Serializer
RDRF = 1
Delay DLYBCT
0
TDRE ?
1
1
CSAAT ?
0
NPCS = 0xF
Delay DLYBCS
420
32099I–01/2012
AT32UC3L016/32/64
20.7.3.3
Clock generation
The SPI Baud rate clock is generated by dividing the CLK_SPI , by a value between 1 and 255.
This allows a maximum operating baud rate at up to CLK_SPI and a minimum operating baud
rate of CLK_SPI divided by 255.
Writing the Serial Clock Baud Rate field in the CSRn registers (CSRn.SCBR) to zero is forbidden. Triggering a transfer while CSRn.SCBR is zero can lead to unpredictable results.
At reset, CSRn.SCBR is zero and the user has to configure 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 configured in the
CSRn.SCBR field. This allows the SPI to automatically adapt the baud rate for each interfaced
peripheral without reprogramming.
20.7.3.4
Transfer delays
Figure 20-7 on page 421 shows a chip select transfer change and consecutive transfers on the
same chip select. Three delays can be configured to modify the transfer waveforms:
• The delay between chip selects, programmable only once for all the chip selects by writing to
the Delay Between Chip Selects field in the MR register (MR.DLYBCS). 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
Delay Before SPCK field in the CSRn registers (CSRn.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 Delay Between Consecutive Transfers field in the CSRn registers
(CSRn.DLYBCT). 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 20-7. Programmable Delays
Chip Select 1
Chip Select 2
SPCK
DLYBCS
DLYBS
DLYBCT
DLYBCT
421
32099I–01/2012
AT32UC3L016/32/64
20.7.3.5
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 a zero to the Peripheral Select bit in MR (MR.PS).
In this case, the current peripheral is defined by the MR.PCS field and the TDR.PCS field has no
effect.
Variable Peripheral Select is activated by writing a one to the MR.PS bit . The TDR.PCS field 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 Peripheral DMA Controller is an optimal means, as the size of the data transfer between the memory
and the SPI is either 4 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 MR register. Data written to TDR is 32-bits wide and defines the real data to be
transmitted and the peripheral it is destined to. Using the Peripheral DMA Controller 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 CSRn 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.
20.7.3.6
Peripheral chip select decoding
The user can configure 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 a one to
the Chip Select Decode bit in the MR register (MR.PCSDEC).
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 MR register or the TDR 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 one)
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, the CRS0
register 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.
20.7.3.7
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
422
32099I–01/2012
AT32UC3L016/32/64
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 CSRn registers can be configured with the Chip
Select Active After Transfer bit written to one (CSRn.CSAAT) . This allows the chip select lines
to remain in their current state (low = active) until transfer to another peripheral is required.
When the CSRn.CSAAT bit is written to qero, the NPCS does not rise in all cases between two
transfers on the same peripheral. During a transfer on a Chip Select, the SR.TDRE bit rises as
soon as the content of the TDR is transferred into the internal shifter. When this bit is detected
the TDR can be reloaded. If this reload occurs before the end of the current transfer and if the
next transfer is performed on the same chip select as the current transfer, the Chip Select is not
de-asserted between the two transfers. This might lead to difficulties for interfacing with some
serial peripherals requiring the chip select to be de-asserted after each transfer. To facilitate
interfacing with such devices, the CSRn registers can be configured with the Chip Select Not
Active After Transfer bit (CSRn.CSNAAT) written to one. This allows to de-assert systematically
the chip select lines during a time DLYBCS. (The value of the CSRn.CSNAAT bit is taken into
account only if the CSRn.CSAAT bit is written to zero for the same Chip Select).
Figure 20-8 on page 424 shows different peripheral deselection cases and the effect of the
CSRn.CSAAT and CSRn.CSNAAT bits.
20.7.3.8
FIFO management
A FIFO has been implemented in Reception FIFO (both in master and in slave mode), in order to
be able to store up to 4 characters without causing an overrun error. If an attempt is made to
store a fifth character, an overrun error rises. If such an event occurs, the FIFO must be flushed.
There are two ways to Flush the FIFO:
• By performing four read accesses of the RDR (the data read must be ignored)
• By writing a one to the Flush Fifo Command bit in the CR register (CR.FLUSHFIFO).
After that, the SPI is able to receive new data.
423
32099I–01/2012
AT32UC3L016/32/64
Figure 20-8. Peripheral Deselection
CSAAT = 0 and CSNAAT = 0
TDRE
NPCS[0..3]
CSAAT = 1 and CSNAAT= 0 / 1
DLYBCT
DLYBCT
A
A
A
A
DLYBCS
A
DLYBCS
PCS = A
PCS = A
Write TDR
TDRE
NPCS[0..3]
DLYBCT
DLYBCT
A
A
A
A
DLYBCS
A
DLYBCS
PCS=A
PCS = A
Write TDR
TDRE
NPCS[0..3]
DLYBCT
DLYBCT
A
B
B
A
DLYBCS
DLYBCS
PCS = B
PCS = B
Write TDR
CSAAT = 0 and CSNAAT = 0
CSAAT = 0 and CSNAAT = 1
DLYBCT
DLYBCT
TDRE
NPCS[0..3]
A
A
A
A
DLYBCS
PCS = A
PCS = A
Write TDR
Figure 20-8 on page 424 shows different peripheral deselection cases and the effect of the
CSRn.CSAAT and CSRn.CSNAAT bits.
20.7.3.9
Mode fault detection
The SPI is capable of detecting a mode fault when it is configured in master mode and NPCS0,
MOSI, MISO, and SPCK are configured as open drain through the I/O Controller with either
internal or external pullup resistors. If the I/O Controller does not have open-drain capability,
mode fault detection must be disabled by writing a one to the Mode Fault Detection bit in the MR
424
32099I–01/2012
AT32UC3L016/32/64
register (MR.MODFDIS). In systems with open-drain I/O lines, a mode fault is detected when a
low level is driven by an external master on the NPCS0/NSS signal.
When a mode fault is detected, the Mode Fault Error bit in the SR (SR.MODF) is set until the SR
is read and the SPI is automatically disabled until re-enabled by writing a one to the SPI Enable
bit in the CR register (CR.SPIEN).
By default, the mode fault detection circuitry is enabled. The user can disable mode fault detection by writing a one to the Mode Fault Detection bit in the MR register (MR.MODFDIS).
20.7.4
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 Per Transfer field of the Chip Select Register 0 (CSR0.BITS). These bits are
processed following a phase and a polarity defined respectively by the CSR0.NCPHA and
CSR0.CPOL bits. Note that the BITS, CPOL, and NCPHA bits of the other Chip Select Registers
have no effect when the SPI is configured 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 SR.RDRF bit rises. If the RDR register has not been read before new data is received,
the SR.OVRES bit is set. Data is loaded in RDR even if this flag is set. The user has to read the
SR register to clear the SR.OVRES bit.
When a transfer starts, the data shifted out is the data present in the Shift Register. If no data
has been written in the TDR register, 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 to zero.
When a first data is written in TDR, it is transferred immediately in the Shift Register and the
SR.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 SR.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 TDR. 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.
In this case the Underrun Error Status bit is set in SR (SR.UNDES).
Figure 20-9 on page 426 shows a block diagram of the SPI when operating in slave mode.
425
32099I–01/2012
AT32UC3L016/32/64
Figure 20-9. Slave Mode Functional Block Diagram
SPCK
SPI
Clock
NSS
SPIEN
RXFIFOEN
SPIENS
RDR
SPIDIS
RDRF
OVRES
RD
CSR0
BITS
NCPHA
CPOL
MOSI
LSB
0
1
4 - Character FIFO
MSB
Shift Register
MISO
UNDES
TDR
TD
TDRE
426
32099I–01/2012
AT32UC3L016/32/64
20.8
User Interface
Table 20-3.
Note:
SPI Register Memory Map
Offset
Register
Register Name
Access
Reset
0x00
Control Register
CR
Write-only
0x00000000
0x04
Mode Register
MR
Read/Write
0x00000000
0x08
Receive Data Register
RDR
Read-only
0x00000000
0x0C
Transmit Data Register
TDR
Write-only
0x00000000
0x10
Status Register
SR
Read-only
0x00000000
0x14
Interrupt Enable Register
IER
Write-only
0x00000000
0x18
Interrupt Disable Register
IDR
Write-only
0x00000000
0x1C
Interrupt Mask Register
IMR
Read-only
0x00000000
0x30
Chip Select Register 0
CSR0
Read/Write
0x00000000
0x34
Chip Select Register 1
CSR1
Read/Write
0x00000000
0x38
Chip Select Register 2
CSR2
Read/Write
0x00000000
0x3C
Chip Select Register 3
CSR3
Read/Write
0x00000000
0x E4
Write Protection Control Register
WPCR
Read/Write
0X00000000
0xE8
Write Protection Status Register
WPSR
Read-only
0x00000000
0xF8
Features Register
FEATURES
Read-only
- (1)
0xFC
Version Register
VERSION
Read-only
- (1)
1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter.
427
32099I–01/2012
AT32UC3L016/32/64
20.8.1
Name:
Control Register
CR
Access Type:
Write-only
Offset:
0x00
Reset Value:
0x00000000
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
-
-
-
-
-
-
-
FLUSHFIFO
7
6
5
4
3
2
1
0
SWRST
-
-
-
-
-
SPIDIS
SPIEN
• LASTXFER: Last Transfer
1: The current NPCS will be deasserted after the character written in TD has been transferred. When CSRn.CSAAT is one, this
allows to close the communication with the current serial peripheral by raising the corresponding NPCS line as soon as TD
transfer has completed.
0: Writing a zero to this bit has no effect.
• FLUSHFIFO: Flush Fifo Command
1: If The FIFO Mode is enabled (MR.FIFOEN written to one) and if an overrun error has been detected, this command allows to
empty the FIFO.
0: Writing a zero to this bit has no effect.
• SWRST: SPI Software Reset
1: Writing a one to this bit will reset the SPI. A software-triggered hardware reset of the SPI interface is performed. The SPI is in
slave mode after software reset. Peripheral DMA Controller channels are not affected by software reset.
0: Writing a zero to this bit has no effect.
• SPIDIS: SPI Disable
1: Writing a one to this bit will disable the SPI. As soon as SPIDIS is written to one, the 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 CR register is written, the SPI is disabled.
0: Writing a zero to this bit has no effect.
• SPIEN: SPI Enable
1: Writing a one to this bit will enable the SPI to transfer and receive data.
0: Writing a zero to this bit has no effect.
428
32099I–01/2012
AT32UC3L016/32/64
20.8.2
Name:
Mode Register
MR
Access Type:
Read/Write
Offset:
0x04
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
DLYBCS
23
22
21
20
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
LLB
RXFIFOEN
-
MODFDIS
-
PCSDEC
PS
MSTR
PCS
• DLYBCS: Delay Between Chip Selects
This field defines the delay from NPCS inactive to the activation of another NPCS. The DLYBCS time guarantees nonoverlapping 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 CLK_SPI periods will be inserted by default.
Otherwise, the following equation determines the delay:
DLYBCS
Delay Between Chip Selects = ----------------------CLKSPI
• PCS: Peripheral Chip Select
This field is only used if Fixed Peripheral Select is active (PS = 0).
If PCSDEC = 0:
PCS = xxx0NPCS[3:0] = 1110
PCS = xx01NPCS[3:0] = 1101
PCS = x011NPCS[3:0] = 1011
PCS = 0111NPCS[3:0] = 0111
PCS = 1111forbidden (no peripheral is selected)
(x = don’t care)
If PCSDEC = 1:
NPCS[3:0] output signals = PCS.
• LLB: Local Loopback Enable
1: Local loopback path enabled. LLB controls the local loopback on the data serializer for testing in master mode only (MISO is
internally connected on MOSI).
0: Local loopback path disabled.
• RXFIFOEN: FIFO in Reception Enable
1: The FIFO is used in reception (four characters can be stored in the SPI).
429
32099I–01/2012
AT32UC3L016/32/64
0: The FIFO is not used in reception (only one character can be stored in the SPI).
• MODFDIS: Mode Fault Detection
1: Mode fault detection is disabled. If the I/O controller does not have open-drain capability, mode fault detection must be
disabled for proper operation of the SPI.
0: Mode fault detection is enabled.
• 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 CSRn 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.
• PS: Peripheral Select
1: Variable Peripheral Select.
0: Fixed Peripheral Select.
• MSTR: Master/Slave Mode
1: SPI is in master mode.
0: SPI is in slave mode.
430
32099I–01/2012
AT32UC3L016/32/64
20.8.3
Name:
Receive Data Register
RDR
Access Type:
Read-only
Offset:
0x08
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
RD[15:8]
7
6
5
4
RD[7:0]
• RD: Receive Data
Data received by the SPI Interface is stored in this register right-justified. Unused bits read zero.
431
32099I–01/2012
AT32UC3L016/32/64
20.8.4
Name:
Transmit Data Register
TDR
Access Type:
Write-only
Offset:
0x0C
Reset Value:
0x00000000
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[15:8]
7
6
5
4
TD[7:0]
• LASTXFER: Last Transfer
1: The current NPCS will be deasserted after the character written in TD has been transferred. When CSRn.CSAAT is one, this
allows to close the communication with the current serial peripheral by raising the corresponding NPCS line as soon as TD
transfer has completed.
0: Writing a zero to this bit has no effect.
This field is only used if Variable Peripheral Select is active (MR.PS = 1).
• PCS: Peripheral Chip Select
If PCSDEC = 0:
PCS = xxx0NPCS[3:0] = 1110
PCS = xx01NPCS[3:0] = 1101
PCS = x011NPCS[3:0] = 1011
PCS = 0111NPCS[3:0] = 0111
PCS = 1111forbidden (no peripheral is selected)
(x = don’t care)
If PCSDEC = 1:
NPCS[3:0] output signals = PCS
This field is only used if Variable Peripheral Select is active (MR.PS = 1).
• 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 TDR
register in a right-justified format.
432
32099I–01/2012
AT32UC3L016/32/64
20.8.5
Name:
Status Register
SR
Access Type:
Read-only
Offset:
0x10
Reset Value:
0x00000000
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
-
-
-
-
-
UNDES
TXEMPTY
NSSR
7
6
5
4
3
2
1
0
-
-
-
-
OVRES
MODF
TDRE
RDRF
• SPIENS: SPI Enable Status
1: This bit is set when the SPI is enabled.
0: This bit is cleared when the SPI is disabled.
• UNDES: Underrun Error Status (Slave Mode Only)
1: This bit is set when a transfer begins whereas no data has been loaded in the TDR register.
0: This bit is cleared when the SR register is read.
• TXEMPTY: Transmission Registers Empty
1: This bit is set when TDR and internal shifter are empty. If a transfer delay has been defined, TXEMPTY is set after the
completion of such delay.
0: This bit is cleared as soon as data is written in TDR.
• NSSR: NSS Rising
1: A rising edge occurred on NSS pin since last read.
0: This bit is cleared when the SR register is read.
• OVRES: Overrun Error Status
1: This bit is set when an overrun has occurred. An overrun occurs when RDR is loaded at least twice from the serializer since
the last read of the RDR.
0: This bit is cleared when the SR register is read.
• MODF: Mode Fault Error
1: This bit is set when a Mode Fault occurred.
0: This bit is cleared when the SR register is read.
• TDRE: Transmit Data Register Empty
1: This bit is set when the last data written in the TDR register has been transferred to the serializer.
0: This bit is cleared when data has been written to TDR and not yet transferred to the serializer.
TDRE equals zero when the SPI is disabled or at reset. The SPI enable command sets this bit to one.
• RDRF: Receive Data Register Full
1: Data has been received and the received data has been transferred from the serializer to RDR since the last read of RDR.
0: No data has been received since the last read of RDR
433
32099I–01/2012
AT32UC3L016/32/64
20.8.6
Name:
Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x14
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
-
-
-
-
-
UNDES
TXEMPTY
NSSR
7
6
5
4
3
2
1
0
-
-
-
-
OVRES
MODF
TDRE
RDRF
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
434
32099I–01/2012
AT32UC3L016/32/64
20.8.7
Name:
Interrupt Disable Register
IDR
Access Type:
Write-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
-
-
-
-
-
UNDES
TXEMPTY
NSSR
7
6
5
4
3
2
1
0
-
-
-
-
OVRES
MODF
TDRE
RDRF
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
435
32099I–01/2012
AT32UC3L016/32/64
20.8.8
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
Offset:
0x1C
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
-
-
-
-
-
UNDES
TXEMPTY
NSSR
7
6
5
4
3
2
1
0
-
-
-
-
OVRES
MODF
TDRE
RDRF
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
A bit in this register is cleared when the corresponding bit in IDR is written to one.
A bit in this register is set when the corresponding bit in IER is written to one.
436
32099I–01/2012
AT32UC3L016/32/64
20.8.9
Name:
Chip Select Register 0
CSR0
Access Type:
Read/Write
Offset:
0x30
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
CSAAT
CSNAAT
NCPHA
CPOL
DLYBCT
23
22
21
20
DLYBS
15
14
13
12
SCBR
7
6
5
BITS
4
• 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:
32 × DLYBCT
Delay Between Consecutive Transfers = -----------------------------------CLKSPI
• 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.
Otherwise, the following equations determine the delay:
DLYBS
Delay Before SPCK = --------------------CLKSPI
• SCBR: Serial Clock Baud Rate
In Master Mode, the SPI Interface uses a modulus counter to derive the SPCK baud rate from the CLK_SPI. 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:
CLKSPI
SPCK Baudrate = --------------------SCBR
Writing the SCBR field to zero is forbidden. Triggering a transfer while SCBR is zero can lead to unpredictable results.
At reset, SCBR is zero and the user has to write it to a valid value before performing the first transfer.
If a clock divider (SCBRn) field is set to one and the other SCBR fields differ from one, access on CSn is correct but no correct
access will be possible on other CS.
437
32099I–01/2012
AT32UC3L016/32/64
• BITS: Bits Per Transfer
The BITS field determines the number of data bits transferred. Reserved values should not be used.
BITS
Bits Per Transfer
0000
8
0001
9
0010
10
0011
11
0100
12
0101
13
0110
14
0111
15
1000
16
1001
4
1010
5
1011
6
1100
7
1101
Reserved
1110
Reserved
1111
Reserved
• CSAAT: Chip Select Active After Transfer
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.
0: The Peripheral Chip Select Line rises as soon as the last transfer is achieved.
• CSNAAT: Chip Select Not Active After Transfer (Ignored if CSAAT = 1)
0: The Peripheral Chip Select does not rise between two transfers if the TDR is reloaded before the end of the first transfer and
if the two transfers occur on the same Chip Select.
1: The Peripheral Chip Select rises systematically between each transfer performed on the same slave for a minimal duration of:
DLYBCS
----------------------- (if DLYBCT field is different from 0)
CLKSPI
DLYBCS + 1 (if DLYBCT field equals 0)
--------------------------------CLKSPI
• NCPHA: Clock Phase
1: Data is captured after the leading (inactive-to-active) edge of SPCK and changed on the trailing (active-to-inactive) edge of
SPCK.
0: Data is changed on the leading (inactive-to-active) edge of SPCK and captured after the trailing (active-to-inactive) 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.
• CPOL: Clock Polarity
1: The inactive state value of SPCK is logic level one.
0: The inactive state value of SPCK is logic level zero.
438
32099I–01/2012
AT32UC3L016/32/64
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.
439
32099I–01/2012
AT32UC3L016/32/64
20.8.10
Name:
Chip Select Register 1
CSR1
Access Type:
Read/Write
Offset:
0x34
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
CSAAT
CSNAAT
NCPHA
CPOL
DLYBCT
23
22
21
20
DLYBS
15
14
13
12
SCBR
7
6
5
BITS
4
• 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:
32 × DLYBCT
Delay Between Consecutive Transfers = -----------------------------------CLKSPI
• 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.
Otherwise, the following equations determine the delay:
DLYBS
Delay Before SPCK = --------------------CLKSPI
• SCBR: Serial Clock Baud Rate
In Master Mode, the SPI Interface uses a modulus counter to derive the SPCK baud rate from the CLK_SPI. 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:
CLKSPI
SPCK Baudrate = --------------------SCBR
Writing the SCBR field to zero is forbidden. Triggering a transfer while SCBR is zero can lead to unpredictable results.
At reset, SCBR is zero and the user has to write it to a valid value before performing the first transfer.
If a clock divider (SCBRn) field is set to one and the other SCBR fields differ from one, access on CSn is correct but no correct
access will be possible on other CS.
440
32099I–01/2012
AT32UC3L016/32/64
• BITS: Bits Per Transfer
The BITS field determines the number of data bits transferred. Reserved values should not be used.
BITS
Bits Per Transfer
0000
8
0001
9
0010
10
0011
11
0100
12
0101
13
0110
14
0111
15
1000
16
1001
4
1010
5
1011
6
1100
7
1101
Reserved
1110
Reserved
1111
Reserved
• CSAAT: Chip Select Active After Transfer
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.
0: The Peripheral Chip Select Line rises as soon as the last transfer is achieved.
• CSNAAT: Chip Select Not Active After Transfer (Ignored if CSAAT = 1)
0: The Peripheral Chip Select does not rise between two transfers if the TDR is reloaded before the end of the first transfer and
if the two transfers occur on the same Chip Select.
1: The Peripheral Chip Select rises systematically between each transfer performed on the same slave for a minimal duration of:
DLYBCS
----------------------- (if DLYBCT field is different from 0)
CLKSPI
DLYBCS + 1 (if DLYBCT field equals 0)
--------------------------------CLKSPI
• NCPHA: Clock Phase
1: Data is captured after the leading (inactive-to-active) edge of SPCK and changed on the trailing (active-to-inactive) edge of
SPCK.
0: Data is changed on the leading (inactive-to-active) edge of SPCK and captured after the trailing (active-to-inactive) 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.
• CPOL: Clock Polarity
1: The inactive state value of SPCK is logic level one.
0: The inactive state value of SPCK is logic level zero.
441
32099I–01/2012
AT32UC3L016/32/64
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.
442
32099I–01/2012
AT32UC3L016/32/64
20.8.11
Name:
Chip Select Register 2
CSR2
Access Type:
Read/Write
Offset:
0x38
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
CSAAT
CSNAAT
NCPHA
CPOL
DLYBCT
23
22
21
20
DLYBS
15
14
13
12
SCBR
7
6
5
BITS
4
• 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:
32 × DLYBCT
Delay Between Consecutive Transfers = -----------------------------------CLKSPI
• 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.
Otherwise, the following equations determine the delay:
DLYBS
Delay Before SPCK = --------------------CLKSPI
• SCBR: Serial Clock Baud Rate
In Master Mode, the SPI Interface uses a modulus counter to derive the SPCK baud rate from the CLK_SPI. 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:
CLKSPI
SPCK Baudrate = --------------------SCBR
Writing the SCBR field to zero is forbidden. Triggering a transfer while SCBR is zero can lead to unpredictable results.
At reset, SCBR is zero and the user has to write it to a valid value before performing the first transfer.
If a clock divider (SCBRn) field is set to one and the other SCBR fields differ from one, access on CSn is correct but no correct
access will be possible on other CS.
443
32099I–01/2012
AT32UC3L016/32/64
• BITS: Bits Per Transfer
The BITS field determines the number of data bits transferred. Reserved values should not be used.
BITS
Bits Per Transfer
0000
8
0001
9
0010
10
0011
11
0100
12
0101
13
0110
14
0111
15
1000
16
1001
4
1010
5
1011
6
1100
7
1101
Reserved
1110
Reserved
1111
Reserved
• CSAAT: Chip Select Active After Transfer
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.
0: The Peripheral Chip Select Line rises as soon as the last transfer is achieved.
• CSNAAT: Chip Select Not Active After Transfer (Ignored if CSAAT = 1)
0: The Peripheral Chip Select does not rise between two transfers if the TDR is reloaded before the end of the first transfer and
if the two transfers occur on the same Chip Select.
1: The Peripheral Chip Select rises systematically between each transfer performed on the same slave for a minimal duration of:
DLYBCS
----------------------- (if DLYBCT field is different from 0)
CLKSPI
DLYBCS + 1 (if DLYBCT field equals 0)
--------------------------------CLKSPI
• NCPHA: Clock Phase
1: Data is captured after the leading (inactive-to-active) edge of SPCK and changed on the trailing (active-to-inactive) edge of
SPCK.
0: Data is changed on the leading (inactive-to-active) edge of SPCK and captured after the trailing (active-to-inactive) 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.
• CPOL: Clock Polarity
1: The inactive state value of SPCK is logic level one.
0: The inactive state value of SPCK is logic level zero.
444
32099I–01/2012
AT32UC3L016/32/64
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.
445
32099I–01/2012
AT32UC3L016/32/64
20.8.12
Name:
Chip Select Register 3
CSR3
Access Type:
Read/Write
Offset:
0x3C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
CSAAT
CSNAAT
NCPHA
CPOL
DLYBCT
23
22
21
20
DLYBS
15
14
13
12
SCBR
7
6
5
BITS
4
• 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:
32 × DLYBCT
Delay Between Consecutive Transfers = -----------------------------------CLKSPI
• 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.
Otherwise, the following equations determine the delay:
DLYBS
Delay Before SPCK = --------------------CLKSPI
• SCBR: Serial Clock Baud Rate
In Master Mode, the SPI Interface uses a modulus counter to derive the SPCK baud rate from the CLK_SPI. 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:
CLKSPI
SPCK Baudrate = --------------------SCBR
Writing the SCBR field to zero is forbidden. Triggering a transfer while SCBR is zero can lead to unpredictable results.
At reset, SCBR is zero and the user has to write it to a valid value before performing the first transfer.
If a clock divider (SCBRn) field is set to one and the other SCBR fields differ from one, access on CSn is correct but no correct
access will be possible on other CS.
446
32099I–01/2012
AT32UC3L016/32/64
• BITS: Bits Per Transfer
The BITS field determines the number of data bits transferred. Reserved values should not be used.
BITS
Bits Per Transfer
0000
8
0001
9
0010
10
0011
11
0100
12
0101
13
0110
14
0111
15
1000
16
1001
4
1010
5
1011
6
1100
7
1101
Reserved
1110
Reserved
1111
Reserved
• CSAAT: Chip Select Active After Transfer
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.
0: The Peripheral Chip Select Line rises as soon as the last transfer is achieved.
• CSNAAT: Chip Select Not Active After Transfer (Ignored if CSAAT = 1)
0: The Peripheral Chip Select does not rise between two transfers if the TDR is reloaded before the end of the first transfer and
if the two transfers occur on the same Chip Select.
1: The Peripheral Chip Select rises systematically between each transfer performed on the same slave for a minimal duration of:
DLYBCS
----------------------- (if DLYBCT field is different from 0)
CLKSPI
DLYBCS + 1 (if DLYBCT field equals 0)
--------------------------------CLKSPI
• NCPHA: Clock Phase
1: Data is captured after the leading (inactive-to-active) edge of SPCK and changed on the trailing (active-to-inactive) edge of
SPCK.
0: Data is changed on the leading (inactive-to-active) edge of SPCK and captured after the trailing (active-to-inactive) 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.
• CPOL: Clock Polarity
1: The inactive state value of SPCK is logic level one.
0: The inactive state value of SPCK is logic level zero.
447
32099I–01/2012
AT32UC3L016/32/64
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.
448
32099I–01/2012
AT32UC3L016/32/64
20.8.13 Write Protection Control Register
Register Name:
WPCR
Access Type:
Read-write
Offset:
0xE4
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
18
17
16
10
9
8
SPIWPKEY[23:16]
23
22
21
20
19
SPIWPKEY[15:8]
15
14
13
12
11
SPIWPKEY[7:0]
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
SPIWPEN
• SPIWPKEY: SPI Write Protection Key Password
If a value is written in SPIWPEN, the value is taken into account only if SPIWPKEY is written with “SPI” (SPI written in ASCII
Code, i.e. 0x535049 in hexadecimal).
• SPIWPEN: SPI Write Protection Enable
1: The Write Protection is Enabled
0: The Write Protection is Disabled
449
32099I–01/2012
AT32UC3L016/32/64
20.8.14 Write Protection Status Register
Register Name:
WPSR
Access Type:
Read-only
Offset:
0xE8
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
2
1
0
SPIWPVSRC
7
6
5
4
3
-
-
-
-
-
SPIWPVS
• SPIWPVSRC: SPI Write Protection Violation Source
This Field indicates the Peripheral Bus Offset of the register concerned by the violation (MR or CSRx)
• SPIWPVS: SPI Write Protection Violation Status
SPIWPVS value
Violation Type
1
The Write Protection has blocked a Write access to a protected register (since the last read).
2
Software Reset has been performed while Write Protection was enabled (since the last read
or since the last write access on MR, IER, IDR or CSRx).
3
Both Write Protection violation and software reset with Write Protection enabled have
occurred since the last read.
4
Write accesses have been detected on MR (while a chip select was active) or on CSRi (while
the Chip Select “i” was active) since the last read.
5
The Write Protection has blocked a Write access to a protected register and write accesses
have been detected on MR (while a chip select was active) or on CSRi (while the Chip Select
“i” was active) since the last read.
6
Software Reset has been performed while Write Protection was enabled (since the last read
or since the last write access on MR, IER, IDR or CSRx) and some write accesses have been
detected on MR (while a chip select was active) or on CSRi (while the Chip Select “i” was
active) since the last read.
7
- The Write Protection has blocked a Write access to a protected register.
and
- Software Reset has been performed while Write Protection was enabled.
and
- Write accesses have been detected on MR (while a chip select was active) or on CSRi
(while the Chip Select “i” was active) since the last read.
450
32099I–01/2012
AT32UC3L016/32/64
20.8.15 Features Register
Register Name:
FEATURES
Access Type:
Read-only
Offset:
0xF8
Reset Value:
–
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
SWIMPL
FIFORIMPL
BRPBHSB
CSNAATIMPL
EXTDEC
15
14
13
12
11
10
9
8
LENNCONF
7
6
5
4
PHZNCONF
PHCONF
PPNCONF
PCONF
LENCONF
3
2
1
0
NCS
• SWIMPL: Spurious Write Protection Implemented
•
•
•
•
•
•
•
•
0: Spurious write protection is not implemented.
1: Spurious write protection is implemented.
FIFORIMPL: FIFO in Reception Implemented
0: FIFO in reception is not implemented.
1: FIFO in reception is implemented.
BRPBHSB: Bridge Type is PB to HSB
0: Bridge type is not PB to HSB.
1: Bridge type is PB to HSB.
CSNAATIMPL: CSNAAT Features Implemented
0: CSNAAT (Chip select not active after transfer) features are not implemented.
1: CSNAAT features are implemented.
EXTDEC: External Decoder True
0: External decoder capability is not implemented.
1: External decoder capability is implemented.
LENNCONF: Character Length if not Configurable
If the character length is not configurable, this field specifies the fixed character length.
LENCONF: Character Length Configurable
0: The character length is not configurable.
1: The character length is configurable.
PHZNCONF: Phase is Zero if Phase not Configurable
0: If phase is not configurable, phase is non-zero.
1: If phase is not configurable, phase is zero.
PHCONF: Phase Configurable
0: Phase is not configurable.
1: Phase is configurable.
451
32099I–01/2012
AT32UC3L016/32/64
• PPNCONF: Polarity Positive if Polarity not Configurable
0: If polarity is not configurable, polarity is negative.
1: If polarity is not configurable, polarity is positive.
• PCONF: Polarity Configurable
0: Polarity is not configurable.
1: Polarity is configurable.
• NCS: Number of Chip Selects
This field indicates the number of chip selects implemented.
452
32099I–01/2012
AT32UC3L016/32/64
20.8.16 Version Register
Register Name:
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
MFN
11
10
VERSION[11:8]
7
6
5
4
3
2
1
0
VERSION[7:0]
• MFN
Reserved. No functionality associated.
• VERSION
Version number of the module. No functionality associated.
453
32099I–01/2012
AT32UC3L016/32/64
20.9
Module Configuration
The specific configuration for each SPI instance is listed in the following tables.The module bus
clocks listed here are connected to the system bus clocks. Please refer to the Power Manager
chapter for details.
Table 20-4.
SPI Clock Name
Module Name
Clock Name
Description
SPI
CLK_SPI
Clock for the SPI bus interface
Table 20-5.
Register
Reset Value
FEATURES
0x001F0154
VERSION
0x00000211
454
32099I–01/2012
AT32UC3L016/32/64
21. Two-wire Master Interface (TWIM)
Rev.: 1.0.1.1
21.1
Features
• Compatible with I²C standard
•
•
•
•
•
•
21.2
– Multi-master support
– Transfer speeds of 100 and 400 kbit/s
– 7- and 10-bit and General Call addressing
Compatible with SMBus standard
– Hardware Packet Error Checking (CRC) generation and verification with ACK control
– SMBus ALERT interface
– 25 ms clock low timeout delay
– 10 ms master cumulative clock low extend time
– 25 ms slave cumulative clock low extend time
Compatible with PMBus
Compatible with Atmel Two-wire Interface Serial Memories
DMA interface for reducing CPU load
Arbitrary transfer lengths, including 0 data bytes
Optional clock stretching if transmit or receive buffers not ready for data transfer
Overview
The Atmel Two-wire Master Interface (TWIM) interconnects components on a unique two-wire
bus, made up of one clock line and one data line with speeds of up to 400 kbit/s, 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 a real time clock (RTC), dot matrix/graphic LCD
controller, and temperature sensor, to name a few. The TWIM is always a bus master and can
transfer sequential or single bytes. Multiple master capability is supported. Arbitration of the bus
is performed internally and relinquishes the bus 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.Table 21-1 lists the compatibility level of the Atmel Two-wire Interface in
Master Mode and a full I²C compatible device.
Table 21-1.
Atmel TWIM Compatibility with I²C Standard
I²C Standard
Atmel TWIM
Standard-mode (100 kbit/s)
Supported
Fast-mode (400 kbit/s)
Supported
Fast-mode Plus (1 Mbit/s)
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)
Supported
Clock Stretching
Supported
Note:
1. START + b000000001 + Ack + Sr
455
32099I–01/2012
AT32UC3L016/32/64
Table 21-2 lists the compatibility level of the Atmel Two-wire Master Interface and a full SMBus
compatible master.
Table 21-2.
21.3
SMBus Standard
Atmel TWIM
Bus Timeouts
Supported
Address Resolution Protocol
Supported
Alert
Supported
Host Functionality
Supported
Packet Error Checking
Supported
List of Abbreviations
Table 21-3.
21.4
Atmel TWIM Compatibility with SMBus Standard
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 21-1. Block Diagram
Peripheral
Bus Bridge
TWCK
I/O Controller
Two-wire
Interface
Power
Manager
TWD
TWALM
CLK_TWIM
INTC
TWI Interrupt
456
32099I–01/2012
AT32UC3L016/32/64
21.5
Application Block Diagram
Figure 21-2. Application Block Diagram
VDD
Rp
Rp
Rp
TWD
TWI
Master
TWCK
TWALM
Atmel TWI
serial EEPROM
I2C RTC
I2C LCD
controller
I2C temp
sensor
Slave 1
Slave 2
Slave 3
Slave 4
Rp: pull-up value as given by the I2C Standard
21.6
I/O Lines Description
Table 21-4.
I/O Lines Description
Pin Name
Pin Description
TWD
Two-wire Serial Data
Input/Output
TWCK
Two-wire Serial Clock
Input/Output
TWALM
SMBus SMBALERT
Input/Output
21.7
Type
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
21.7.1
I/O Lines
TWD and TWCK are bidirectional lines, connected to a positive supply voltage via a current
source or pull-up resistor (see Figure 21-4 on page 459). 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.
TWALM is used to implement the optional SMBus SMBALERT signal.
The TWALM, TWD, and TWCK pins may be multiplexed with I/O Controller lines. To enable the
TWIM, the user must perform the following steps:
• Program the I/O Controller to:
– Dedicate TWD, TWCK, and optionally TWALM as peripheral lines.
– Define TWD, TWCK, and optionally TWALM as open-drain.
21.7.2
Power Management
If the CPU enters a sleep mode that disables clocks used by the TWIM, the TWIM will stop functioning and resume operation after the system wakes up from sleep mode.
457
32099I–01/2012
AT32UC3L016/32/64
21.7.3
Clocks
The clock for the TWIM bus interface (CLK_TWIM) is generated by the Power Manager. This
clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the TWIM before disabling the clock, to avoid freezing the TWIM in an undefined state.
21.7.4
DMA
The TWIM DMA handshake interface is connected to the Peripheral DMA Controller. Using the
TWIM DMA functionality requires the Peripheral DMA Controller to be programmed after setting
up the TWIM.
21.7.5
Interrupts
The TWIM interrupt request lines are connected to the interrupt controller. Using the TWIM interrupts requires the interrupt controller to be programmed first.
21.7.6
Debug Operation
When an external debugger forces the CPU into debug mode, the TWIM continues normal operation. If the TWIM is configured in a way that requires it to be periodically serviced by the CPU
through interrupts or similar, improper operation or data loss may result during debugging.
458
32099I–01/2012
AT32UC3L016/32/64
21.8
21.8.1
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
21-4).
Each transfer begins with a START condition and terminates with a STOP condition (see Figure
21-4).
• 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 21-3.
START and STOP Conditions
TWD
TWCK
Start
Stop
Figure 21-4. Transfer Format
TWD
TWCK
Start
21.8.2
Address
R/W
Ack
Data
Ack
Data
Ack
Stop
Operation
The TWIM has two modes of operation:
• Master transmitter mode
• Master receiver mode
The master is the device which starts and stops a transfer and generates the TWCK clock.
These modes are described in the following chapters.
459
32099I–01/2012
AT32UC3L016/32/64
21.8.2.1
Clock Generation
The Clock Waveform Generator Register (CWGR) is used to control the waveform of the TWCK
clock. CWGR must be written so that the desired TWI bus timings are generated. CWGR
describes bus timings as a function of cycles of a prescaled clock. The clock prescaling can be
selected through the Clock Prescaler field in CWGR (CWGR.EXP).
f CLK_TWIM
f PRESCALER = ------------------------( EXP + 1 )
2
CWGR has the following fields:
LOW: Prescaled clock cycles in clock low count. Used to time TLOW and TBUF.
HIGH: Prescaled clock cycles in clock high count. Used to time THIGH.
STASTO: Prescaled clock cycles in clock high count. Used to time THD_STA, TSU_STA, TSU_STO.
DATA: Prescaled clock cycles for data setup and hold count. Used to time THD_DAT, TSU_DAT.
EXP: Specifies the clock prescaler setting.
Note that the total clock low time generated is the sum of THD_DAT + TSU_DAT + TLOW.
Any slave or other bus master taking part in the transfer may extend the TWCK low period at any
time.
The TWIM hardware monitors the state of the TWCK line as required by the I²C specification.
The clock generation counters are started when a high/low level is detected on the TWCK line,
not when the TWIM hardware releases/drives the TWCK line. This means that the CWGR settings alone do not determine the TWCK frequency. The CWGR settings determine the clock low
time and the clock high time, but the TWCK rise and fall times are determined by the external circuitry (capacitive load, etc.).
Figure 21-5. Bus Timing Diagram
t HIGH
t LOW
S
t
HD:STA
t LOW
t
SU:DAT
t
HD:DAT
t
t
SU:DAT
t
SU:STA
SU:STO
P
Sr
460
32099I–01/2012
AT32UC3L016/32/64
21.8.2.2
Setting up and Performing a Transfer
Operation of the TWIM is mainly controlled by the Control Register (CR) and the Command Register (CMDR). TWIM status is provided in the Status Register (SR). The following list presents
the main steps in a typical communication:
1. Before any transfers can be performed, bus timings must be configured by writing to the
Clock Waveform Generator Register (CWGR). If operating in SMBus mode, the SMBus
Timing Register (SMBTR) register must also be configured.
2. If the Peripheral DMA Controller is to be used for the transfers, it must be set up.
3. CMDR or NCMDR must be written with a value describing the transfer to be performed.
The interrupt system can be set up to give interrupt requests on specific events or error conditions in the SR, for example when the transfer is complete or if arbitration is lost. The Interrupt
Enable Register (IER) and Interrupt Disable Register (IDR) can be written to specify which bits in
the SR will generate interrupt requests.
The SR.BUSFREE bit is set when activity is completed on the two-wire bus. The SR.CRDY bit is
set when CMDR and/or NCMDR is ready to receive one or more commands.
The controller will refuse to start a new transfer while ANAK, DNAK, or ARBLST in the Status
Register (SR) is one. This is necessary to avoid a race when the software issues a continuation
of the current transfer at the same time as one of these errors happen. Also, if ANAK or DNAK
occurs, a STOP condition is sent automatically. The user will have to restart the transmission by
clearing the error bits in SR after resolving the cause for the NACK.
After a data or address NACK from the slave, a STOP will be transmitted automatically. Note
that the VALID bit in CMDR is NOT cleared in this case. If this transfer is to be discarded, the
VALID bit can be cleared manually allowing any command in NCMDR to be copied into CMDR.
When a data or address NACK is returned by the slave while the master is transmitting, it is possible that new data has already been written to the THR register. This data will be transferred out
as the first data byte of the next transfer. If this behavior is to be avoided, the safest approach is
to perform a software reset of the TWIM.
21.8.3
Master Transmitter Mode
A START condition is transmitted and master transmitter mode is initiated when the bus is free
and CMDR has been written with START=1 and READ=0. START and SADR+W will then be
transmitted. During the address acknowledge clock pulse (9th pulse), the master releases the
data line (HIGH), enabling the slave to pull it down in order to acknowledge the address. The
master polls the data line during this clock pulse and sets the Address Not Acknowledged bit
(ANAK) in the Status Register if no slave acknowledges the address.
After the address phase, the following is repeated:
while (NBYTES>0)
1. Wait until THR contains a valid data byte, stretching low period of TWCK. SR.TXRDY
indicates the state of THR. Software or the Peripheral DMA Controller must write the
data byte to THR.
2. Transmit this data byte
3. Decrement NBYTES
4. If (NBYTES==0) and STOP=1, transmit STOP condition
Writing CMDR with START=STOP=1 and NBYTES=0 will generate a transmission with no data
bytes, ie START, SADR+W, STOP.
461
32099I–01/2012
AT32UC3L016/32/64
TWI transfers require the slave to acknowledge each received data 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 Data Acknowledge bit (DNACK) in the Status Register if the slave does not
acknowledge the data byte. As with the other status bits, an interrupt can be generated if
enabled in the Interrupt Enable Register (IER).
TXRDY is used as Transmit Ready for the Peripheral DMA Controller transmit channel.
The end of a command is marked when the TWIM sets the SR.CCOMP bit. See Figure 21-6 and
Figure 21-7.
Figure 21-6. Master Write with One Data Byte
TWD
S
DADR
W
A
DATA
A
P
SR.IDLE
TXRDY
Write THR (DATA)
NBYTES set to 1
STOP sent automatically
(ACK received and NBYTES=0)
Figure 21-7. Master Write with Multiple Data Bytes
TWD
S
DADR
W
A
DATAn
A
DATAn+5
A
DATAn+m
A
P
SR.IDLE
TXRDY
Write THR
(DATAn)
NBYTES set to n
21.8.4
Write THR
(DATAn+1)
Write THR
(DATAn+m)
Last data sent
STOP sent automatically
(ACK received and NBYTES=0)
Master Receiver Mode
A START condition is transmitted and master receiver mode is initiated when the bus is free and
CMDR has been written with START=1 and READ=1. START and SADR+R will then be transmitted. During the address acknowledge clock pulse (9th pulse), the master releases the data
line (HIGH), enabling the slave to pull it down in order to acknowledge the address. The master
polls the data line during this clock pulse and sets the Address Not Acknowledged bit (ANAK) in
the Status Register if no slave acknowledges the address.
After the address phase, the following is repeated:
while (NBYTES>0)
462
32099I–01/2012
AT32UC3L016/32/64
1. Wait until RHR is empty, stretching low period of TWCK. SR.RXRDY indicates the state
of RHR. Software or the Peripheral DMA Controller must read any data byte present in
RHR.
2. Release TWCK generating a clock that the slave uses to transmit a data byte.
3. Place the received data byte in RHR, set RXRDY.
4. If NBYTES=0, generate a NAK after the data byte, otherwise generate an ACK.
5. Decrement NBYTES
6. If (NBYTES==0) and STOP=1, transmit STOP condition.
Writing CMDR with START=STOP=1 and NBYTES=0 will generate a transmission with no data
bytes, ie START, DADR+R, STOP
The TWI transfers require the master to acknowledge each received data byte. During the
acknowledge clock pulse (9th pulse), the slave releases the data line (HIGH), enabling the master to pull it down in order to generate the acknowledge. All data bytes except the last are
acknowledged by the master. Not acknowledging the last byte informs the slave that the transfer
is finished.
RXRDY is used as Receive Ready for the Peripheral DMA Controller receive channel.
Figure 21-8. Master Read with One Data Byte
TWD
S
DADR
R
A
DATA
N
P
SR.IDLE
RXRDY
Write START &
STOP bit
NBYTES set to 1
Read RHR
Figure 21-9. Master Read with Multiple Data Bytes
TWD
S
DADR
R
A
DATAn
A
DATAn+1
DATAn+m-1
A
DATAn+m
N
P
SR.IDLE
RXRDY
Write START +
STOP bit
NBYTES set to m
Read RHR
DATAn
Read RHR
DATAn+m-2
Read RHR
DATAn+m-1
Read RHR
DATAn+m
Send STOP
When NBYTES=0
463
32099I–01/2012
AT32UC3L016/32/64
21.8.5
Using the Peripheral DMA Controller
The use of the Peripheral DMA Controller significantly reduces the CPU load. The user can set
up ring buffers for the Peripheral DMA Controller, containing data to transmit or free buffer space
to place received data.
To assure correct behavior, respect the following programming sequences:
21.8.5.1
Data Transmit with the Peripheral DMA Controller
1. Initialize the transmit Peripheral DMA Controller (memory pointers, size, etc.).
2. Configure the TWIM (ADR, NBYTES, etc.).
3. Start the transfer by enabling the Peripheral DMA Controller to transmit.
4. Wait for the Peripheral DMA Controller end-of-transmit flag.
5. Disable the Peripheral DMA Controller.
21.8.5.2
Data Receive with the Peripheral DMA Controller
1. Initialize the receive Peripheral DMA Controller (memory pointers, size, etc.).
2. Configure the TWIM (ADR, NBYTES, etc.).
3. Start the transfer by enabling the Peripheral DMA Controller to receive.
4. Wait for the Peripheral DMA Controller end-of-receive flag.
5. Disable the Peripheral DMA Controller.
21.8.6
Multi-master Mode
More than one master may access 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. The SR.ARBLST flag will be set. When the STOP is detected, the master who
lost arbitration may reinitiate the data transfer.
Arbitration is illustrated in Figure 21-11.
If the user starts a transfer and if the bus is busy, the TWIM automatically waits for a STOP condition on the bus before initiating the transfer (see Figure 21-10).
Note:
The state of the bus (busy or free) is not indicated in the user interface.
464
32099I–01/2012
AT32UC3L016/32/64
Figure 21-10. User 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 21-11. Arbitration Cases
TWCK
TWD
TWCK
Data from a Master
S
1
0 0 1 1
Data from TWI
S
1
0
TWD
S
1
0 0
1
P
Arbitration is lost
TWI stops sending data
1 1
Data from the master
P
Arbitration is lost
S
1
0
S
1
0 0 1
1
S
1
0
1
1
The master stops sending data
0 1
Data from the TWI
ARBLST
Bus is busy
Transfer is kept
TWI DATA transfer
A transfer is programmed
(DADR + W + START + Write THR)
21.8.7
Bus is free
Transfer is stopped
Transfer is programmed again
(DADR + W + START + Write THR)
Bus is considered as free
Transfer is initiated
Combined Transfers
CMDR and NCMDR may be used to generate longer sequences of connected transfers, since
generation of START and/or STOP conditions is programmable on a per-command basis.
Writing NCMDR with START=1 when the previous transfer was written with STOP=0 will cause
a REPEATED START on the bus. The ability to generate such connected transfers allows arbitrary transfer lengths, since it is legal to write CMDR with both START=0 and STOP=0. If this is
done in master receiver mode, the CMDR.ACKLAST bit must also be controlled.
465
32099I–01/2012
AT32UC3L016/32/64
As for single data transfers, the TXRDY and RXRDY bits in the Status Register indicates when
data to transmit can be written to THR, or when received data can be read from RHR. Transfer
of data to THR and from RHR can also be done automatically by DMA, see Section 21.8.5
21.8.7.1
Write Followed by Write
Consider the following transfer:
START, DADR+W, DATA+A, DATA+A, REPSTART, DADR+W, DATA+A, DATA+A, STOP.
To generate this transfer:
1. Write CMDR with START=1, STOP=0, DADR, NBYTES=2 and READ=0.
2. Write NCMDR with START=1, STOP=1, DADR, NBYTES=2 and READ=0.
3. Wait until SR.TXRDY==1, then write first data byte to transfer to THR.
4. Wait until SR.TXRDY==1, then write second data byte to transfer to THR.
5. Wait until SR.TXRDY==1, then write third data byte to transfer to THR.
6. Wait until SR.TXRDY==1, then write fourth data byte to transfer to THR.
21.8.7.2
Read Followed by Read
Consider the following transfer:
START, DADR+R, DATA+A, DATA+NA, REPSTART, DADR+R, DATA+A, DATA+NA, STOP.
To generate this transfer:
1. Write CMDR with START=1, STOP=0, DADR, NBYTES=2 and READ=1.
2. Write NCMDR with START=1, STOP=1, DADR, NBYTES=2 and READ=1.
3. Wait until SR.RXRDY==1, then read first data byte received from RHR.
4. Wait until SR.RXRDY==1, then read second data byte received from RHR.
5. Wait until SR.RXRDY==1, then read third data byte received from RHR.
6. Wait until SR.RXRDY==1, then read fourth data byte received from RHR.
If combining several transfers, without any STOP or REPEATED START between them, remember to write a one to the ACKLAST bit in CMDR to keep from ending each of the partial transfers
with a NACK.
21.8.7.3
Write Followed by Read
Consider the following transfer:
START, DADR+W, DATA+A, DATA+A, REPSTART, DADR+R, DATA+A, DATA+NA, STOP.
466
32099I–01/2012
AT32UC3L016/32/64
Figure 21-12. Combining a Write and Read Transfer
THR
DATA0
DATA1
RHR
TWD
DATA2
S
DADR
W
A
DATA0
A
DATA1
NA
Sr
R
DADR
A
DATA2
A
DATA3
DATA3
A
P
SR.IDLE
1
TXRDY
RXRDY
To generate this transfer:
1. Write CMDR with START=1, STOP=0, DADR, NBYTES=2 and READ=0.
2. Write NCMDR with START=1, STOP=1, DADR, NBYTES=2 and READ=1.
3. Wait until SR.TXRDY==1, then write first data byte to transfer to THR.
4. Wait until SR.TXRDY==1, then write second data byte to transfer to THR.
5. Wait until SR.RXRDY==1, then read first data byte received from RHR.
6. Wait until SR.RXRDY==1, then read second data byte received from RHR.
21.8.7.4
Read Followed by Write
Consider the following transfer:
START, DADR+R, DATA+A, DATA+NA, REPSTART, DADR+W, DATA+A, DATA+A, STOP.
Figure 21-13. Combining a Read and Write Transfer
THR
DATA2
RHR
TWD
DATA0
S
SADR
R
A
DATA0
A
DATA3
DATA1
A
DATA3
1
Sr
DADR
W
A
DATA2
A
DATA3
SR.IDLE
TXRDY
RXRDY
NA
P
2
Read
TWI_RHR
To generate this transfer:
1. Write CMDR with START=1, STOP=0, DADR, NBYTES=2 and READ=1.
2. Write NCMDR with START=1, STOP=1, DADR, NBYTES=2 and READ=0.
3. Wait until SR.RXRDY==1, then read first data byte received from RHR.
4. Wait until SR.RXRDY==1, then read second data byte received from RHR.
5. Wait until SR.TXRDY==1, then write first data byte to transfer to THR.
6. Wait until SR.TXRDY==1, then write second data byte to transfer to THR.
467
32099I–01/2012
AT32UC3L016/32/64
21.8.8
Ten Bit Addressing
Writing a one to CMDR.TENBIT enables 10-bit addressing in hardware. Performing transfers
with 10-bit addressing is similar to transfers with 7-bit addresses, except that bits 9:7 of
CMDR.SADR must be written appropriately.
In Figure 21-14 and Figure 21-15, the grey boxes represent signals driven by the master, the
white boxes are driven by the slave.
21.8.8.1
Master Transmitter
To perform a master transmitter transfer:
1. Write CMDR with TENBIT=1, REPSAME=0, READ=0, START=1, STOP=1 and the
desired address and NBYTES value.
Figure 21-14. A Write Transfer with 10-bit Addressing
1
1
S
21.8.8.2
1
1
0
X
SLAVE ADDRESS
1st 7 bits
X
0
RW A1
SLAVE ADDRESS
2nd byte
A2
DATA
A
DATA
AA P
Master Receiver
When using master receiver mode with 10-bit addressing, CMDR.REPSAME must also be controlled. CMDR.REPSAME must be written to one when the address phase of the transfer should
consist of only 1 address byte (the 11110xx byte) and not 2 address bytes. The I²C standard
specifies that such addressing is required when addressing a slave for reads using 10-bit
addressing.
To perform a master receiver transfer:
1. Write CMDR with TENBIT=1, REPSAME=0, READ=0, START=1, STOP=0,
NBYTES=0 and the desired address.
2. Write NCMDR with TENBIT=1, REPSAME=1, READ=1, START=1, STOP=1 and the
desired address and NBYTES value.
Figure 21-15. A Read Transfer with 10-bit Addressing
1
S
21.8.9
1
1
1
0
X
X
SLAVE ADDRESS
1st 7 bits
1
0
RW A1
SLAVE ADDRESS
2nd byte
A2 Sr
1
1
1
0
X
SLAVE ADDRESS
1st 7 bits
X
1
RW A3
DATA
A
DATA
A
P
SMBus Mode
SMBus mode is enabled and disabled by writing to the SMEN and SMDIS bits in CR. SMBus
mode operation is similar to I²C operation with the following exceptions:
• Only 7-bit addressing can be used.
• The SMBus standard describes a set of timeout values to ensure progress and throughput on
the bus. These timeout values must be written into SMBTR.
• Transmissions can optionally include a CRC byte, called Packet Error Check (PEC).
• A dedicated bus line, SMBALERT, allows a slave to get a master’s attention.
• A set of addresses have been reserved for protocol handling, such as Alert Response
Address (ARA) and Host Header (HH) Address.
468
32099I–01/2012
AT32UC3L016/32/64
21.8.9.1
Packet Error Checking
Each SMBus transfer can optionally end with a CRC byte, called the PEC byte. Writing a one to
CMDR.PECEN enables automatic PEC handling in the current transfer. Transfers with and without PEC can freely be intermixed in the same system, since some slaves may not support PEC.
The PEC LFSR is always updated on every bit transmitted or received, so that PEC handling on
combined transfers will be correct.
In master transmitter mode, the master calculates a PEC value and transmits it to the slave after
all data bytes have been transmitted. Upon reception of this PEC byte, the slave will compare it
to the PEC value it has computed itself. If the values match, the data was received correctly, and
the slave will return an ACK to the master. If the PEC values differ, data was corrupted, and the
slave will return a NACK value. The DNAK bit in SR reflects the state of the last received
ACK/NACK value. Some slaves may not be able to check the received PEC in time to return a
NACK if an error occurred. In this case, the slave should always return an ACK after the PEC
byte, and some other mechanism must be implemented to verify that the transmission was
received correctly.
In master receiver mode, the slave calculates a PEC value and transmits it to the master after all
data bytes have been transmitted. Upon reception of this PEC byte, the master will compare it to
the PEC value it has computed itself. If the values match, the data was received correctly. If the
PEC values differ, data was corrupted, and SR.PECERR is set. In master receiver mode, the
PEC byte is always followed by a NACK transmitted by the master, since it is the last byte in the
transfer.
The PEC byte is automatically inserted in a master transmitter transmission if PEC is enabled
when NBYTES reaches zero. The PEC byte is identified in a master receiver transmission if
PEC is enabled when NBYTES reaches zero. NBYTES must therefore be written with the total
number of data bytes in the transmission, including the PEC byte.
In combined transfers, the PECEN bit should only be written to one in the last of the combined
transfers. Consider the following transfer:
S, ADR+W, COMMAND_BYTE, ACK, SR, ADR+R, DATA_BYTE, ACK, PEC_BYTE, NACK, P
This transfer is generated by writing two commands to the command registers. The first command is a write with NBYTES=1 and PECEN=0, and the second is a read with NBYTES=2 and
PECEN=1.
Writing a one to the STOP bit in CR will place a STOP condition on the bus after the current
byte. No PEC byte will be sent in this case.
21.8.9.2
Timeouts
The TLOWS and TLOWM fields in SMBTR configure the SMBus timeout values. If a timeout
occurs, the master will transmit a STOP condition and leave the bus. The SR.TOUT bit is set.
21.8.9.3
SMBus ALERT Signal
A slave can get the master’s attention by pulling the TWALM line low. The TWIM will then set the
SR.SMBALERT bit. This can be set up to trigger an interrupt, and software can then take the
appropriate action, as defined in the SMBus standard.
469
32099I–01/2012
AT32UC3L016/32/64
21.8.10
Identifying Bus Events
This chapter lists the different bus events, and how they affect bits in the TWIM registers. This is
intended to help writing drivers for the TWIM.
Table 21-5.
Bus Events
Event
Effect
Master transmitter has sent
a data byte
SR.THR is cleared.
Master receiver has
received a data byte
SR.RHR is set.
Start+Sadr sent, no ack
received from slave
SR.ANAK is set.
SR.CCOMP not set.
CMDR.VALID remains set.
STOP automatically transmitted on bus.
Data byte sent to slave, no
ack received from slave
SR.DNAK is set.
SR.CCOMP not set.
CMDR.VALID remains set.
STOP automatically transmitted on bus.
Arbitration lost
SR.ARBLST is set.
SR.CCOMP not set.
CMDR.VALID remains set.
TWCK and TWD immediately released to a pulled-up state.
SMBus Alert received
SR.SMBALERT is set.
SMBus timeout received
SR.SMBTOUT is set.
SR.CCOMP not set.
CMDR.VALID remains set.
STOP automatically transmitted on bus.
Master transmitter receives
SMBus PEC Error
SR.DNAK is set.
SR.CCOMP not set.
CMDR.VALID remains set.
STOP automatically transmitted on bus.
Master receiver discovers
SMBus PEC Error
SR.PECERR is set.
SR.CCOMP not set.
CMDR.VALID remains set.
STOP automatically transmitted on bus.
CR.STOP is written by user
SR.STOP is set.
SR.CCOMP set.
CMDR.VALID remains set.
STOP transmitted on bus after current byte transfer has finished.
470
32099I–01/2012
AT32UC3L016/32/64
21.9
User Interface
Table 21-6.
Note:
TWIM Register Memory Map
Offset
Register
Register Name
Access
Reset
0x00
Control Register
CR
Write-only
0x00000000
0x04
Clock Waveform Generator Register
CWGR
Read/Write
0x00000000
0x08
SMBus Timing Register
SMBTR
Read/Write
0x00000000
0x0C
Command Register
CMDR
Read/Write
0x00000000
0x10
Next Command Register
NCMDR
Read/Write
0x00000000
0x14
Receive Holding Register
RHR
Read-only
0x00000000
0x18
Transmit Holding Register
THR
Write-only
0x00000000
0x1C
Status Register
SR
Read-only
0x00000002
0x20
Interrupt Enable Register
IER
Write-only
0x00000000
0x24
Interrupt Disable Register
IDR
Write-only
0x00000000
0x28
Interrupt Mask Register
IMR
Read-only
0x00000000
0x2C
Status Clear Register
SCR
Write-only
0x00000000
0x30
Parameter Register
PR
Read-only
-(1)
0x34
Version Register
VR
Read-only
-(1)
1. The reset values for these registers are device specific. Please refer to the Module Configuration section at the end of this
chapter.
471
32099I–01/2012
AT32UC3L016/32/64
21.9.1
Name:
Control Register
CR
Access Type:
Write-only
Offset:
0x00
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
-
-
-
-
-
-
-
STOP
7
6
5
4
3
2
1
0
SWRST
-
SMDIS
SMEN
-
-
MDIS
MEN
• STOP: Stop the Current Transfer
Writing a one to this bit terminates the current transfer, sending a STOP condition after the shifter has become idle. If there are
additional pending transfers, they will have to be explicitly restarted by software after the STOP condition has been successfully
sent.
Writing a zero to this bit has no effect.
• SWRST: Software Reset
If the TWIM master interface is enabled, writing a one to this bit resets the TWIM. All transfers are halted immediately, possibly
violating the bus semantics.
If the TWIM master interface is not enabled, it must first be enabled before writing a one to this bit.
Writing a zero to this bit has no effect.
• SMDIS: SMBus Disable
Writing a one to this bit disables SMBus mode.
Writing a zero to this bit has no effect.
• SMEN: SMBus Enable
Writing a one to this bit enables SMBus mode.
Writing a zero to this bit has no effect.
• MDIS: Master Disable
Writing a one to this bit disables the master interface.
Writing a zero to this bit has no effect.
• MEN: Master Enable
Writing a one to this bit enables the master interface.
Writing a zero to this bit has no effect.
472
32099I–01/2012
AT32UC3L016/32/64
21.9.2
Name:
Clock Waveform Generator Register
CWGR
Access Type:
Read/Write
Offset:
0x04
Reset Value:
0x00000000
31
30
-
23
29
28
27
26
EXP
22
21
25
24
DATA
20
19
18
17
16
11
10
9
8
3
2
1
0
STASTO
15
14
13
12
HIGH
7
6
5
4
LOW
• EXP: Clock Prescaler
Used to specify how to prescale the TWCK clock. Counters are prescaled according to the following formula
f CLK_TWIM
f PRESCALER = ------------------------( EXP + 1 )
2
• DATA: Data Setup and Hold Cycles
Clock cycles for data setup and hold count. Prescaled by CWGR.EXP. Used to time THD_DAT, TSU_DAT.
• STASTO: START and STOP Cycles
Clock cycles in clock high count. Prescaled by CWGR.EXP. Used to time THD_STA, TSU_STA, TSU_STO
• HIGH: Clock High Cycles
Clock cycles in clock high count. Prescaled by CWGR.EXP. Used to time THIGH.
• LOW: Clock Low Cycles
Clock cycles in clock low count. Prescaled by CWGR.EXP. Used to time TLOW, TBUF.
473
32099I–01/2012
AT32UC3L016/32/64
21.9.3
Name:
SMBus Timing Register
SMBTR
Access Type:
Read/Write
Offset:
0x08
Reset Value:
0x00000000
31
30
29
28
EXP
23
22
21
20
27
26
25
24
-
-
-
-
19
18
17
16
11
10
9
8
3
2
1
0
THMAX
15
14
13
12
TLOWM
7
6
5
4
TLOWS
• EXP: SMBus Timeout Clock Prescaler
Used to specify how to prescale the TIM and TLOWM counters in SMBTR. Counters are prescaled according to the following
formula
f CLKTWIM
f prescaled, SMBus = -----------------------( EXP + 1 )
2
• THMAX: Clock High Maximum Cycles
Clock cycles in clock high maximum count. Prescaled by SMBTR.EXP. Used for bus free detection. Used to time THIGH:MAX.
NOTE: Uses the prescaler specified by CWGR, NOT the prescaler specified by SMBTR.
• TLOWM: Master Clock Stretch Maximum Cycles
Clock cycles in master maximum clock stretch count. Prescaled by SMBTR.EXP. Used to time TLOW:MEXT
• TLOWS: Slave Clock Stretch Maximum Cycles
Clock cycles in slave maximum clock stretch count. Prescaled by SMBTR.EXP. Used to time TLOW:SEXT.
474
32099I–01/2012
AT32UC3L016/32/64
21.9.4
Name:
Command Register
CMDR
Access Type:
Read/Write
Offset:
0x0C
Reset Value:
0x00000000
31
30
-
23
29
28
-
22
21
20
27
26
25
24
-
-
ACKLAST
PECEN
19
18
17
16
10
9
8
NBYTES
15
14
13
12
11
VALID
STOP
START
REPSAME
TENBIT
7
6
5
4
3
SADR[6:0]
SADR[9:7]
2
1
0
READ
• ACKLAST: ACK Last Master RX Byte
0: Causes the last byte in master receive mode (when NBYTES has reached 0) to be NACKed. This is the standard way of
ending a master receiver transfer.
1: Causes the last byte in master receive mode (when NBYTES has reached 0) to be ACKed. Used for performing linked
transfers in master receiver mode with no STOP or REPEATED START between the subtransfers. This is needed when more
than 255 bytes are to be received in one single transmission.
• PECEN: Packet Error Checking Enable
0: Causes the transfer not to use PEC byte verification. The PEC LFSR is still updated for every bit transmitted or received. Must
be used if SMBus mode is disabled.
1: Causes the transfer to use PEC. PEC byte generation (if master transmitter) or PEC byte verification (if master receiver) will
be performed.
• NBYTES: Number of Data Bytes in Transfer
The number of data bytes in the transfer. After the specified number of bytes have been transferred, a STOP condition is
transmitted if CMDR.STOP is one. In SMBus mode, if PEC is used, NBYTES includes the PEC byte, i.e. there are NBYTES-1
data bytes and a PEC byte.
• VALID: CMDR Valid
0: Indicates that CMDR does not contain a valid command.
1: Indicates that CMDR contains a valid command. This bit is cleared when the command is finished.
• STOP: Send STOP Condition
0: Do not transmit a STOP condition after the data bytes have been transmitted.
1: Transmit a STOP condition after the data bytes have been transmitted.
• START: Send START Condition
0: The transfer in CMDR should not commence with a START or REPEATED START condition.
1: The transfer in CMDR should commence with a START or REPEATED START condition. If the bus is free when the command
is executed, a START condition is used. If the bus is busy, a REPEATED START is used.
• REPSAME: Transfer is to Same Address as Previous Address
Only used in 10-bit addressing mode, always write to 0 in 7-bit addressing mode.
475
32099I–01/2012
AT32UC3L016/32/64
Write this bit to one if the command in CMDR performs a repeated start to the same slave address as addressed in the previous
transfer in order to enter master receiver mode.
Write this bit to zero otherwise.
• TENBIT: Ten Bit Addressing Mode
0: Use 7-bit addressing mode.
1: Use 10-bit addressing mode. Must not be used when the TWIM is in SMBus mode.
• SADR: Slave Address
Address of the slave involved in the transfer. Bits 9-7 are don’t care if 7-bit addressing is used.
• READ: Transfer Direction
0: Allow the master to transmit data.
1: Allow the master to receive data.
476
32099I–01/2012
AT32UC3L016/32/64
21.9.5
Name:
Next Command Register
NCMDR
Access Type:
Read/Write
Offset:
0x10
Reset Value:
0x00000000
31
30
-
29
28
-
23
22
21
20
27
26
25
24
-
-
ACKLAST
PECEN
19
18
17
16
10
9
8
NBYTES
15
14
13
12
11
VALID
STOP
START
REPSAME
TENBIT
7
6
5
4
3
SADR[6:0]
SADR[9:7]
2
1
0
READ
This register is identical to CMDR. When the VALID bit in CMDR becomes 0, the content of NCMDR is copied into CMDR,
clearing the VALID bit in NCMDR. If the VALID bit in CMDR is cleared when NCMDR is written, the content is copied
immediately.
477
32099I–01/2012
AT32UC3L016/32/64
21.9.6
Name:
Receive Holding Register
RHR
Access Type:
Read-only
Offset:
0x14
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: Received Data
When the RXRDY bit in the Status Register (SR) is one, this field contains a byte received from the TWI bus.
478
32099I–01/2012
AT32UC3L016/32/64
21.9.7
Name:
Transmit Holding Register
THR
Access Type:
Write-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
TXDATA
• TXDATA: Data to Transmit
Write data to be transferred on the TWI bus here.
479
32099I–01/2012
AT32UC3L016/32/64
21.9.8
Name:
Status Register
SR
Access Type:
Read-only
Offset:
0x1C
Reset Value:
0x00000002
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
MENB
15
14
13
12
11
10
9
8
-
STOP
PECERR
TOUT
SMBALERT
ARBLST
DNAK
ANAK
7
6
5
4
3
2
1
0
-
-
BUSFREE
IDLE
CCOMP
CRDY
TXRDY
RXRDY
• MENB: Master Interface Enable
0: Master interface is disabled.
1: Master interface is enabled.
• STOP: Stop Request Accepted
This bit is one when a STOP request caused by writing a one to CR.STOP has been accepted, and transfer has stopped.
This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR).
• PECERR: PEC Error
This bit is one when a SMBus PEC error occurred.
This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR).
• TOUT: Timeout
This bit is one when a SMBus timeout occurred.
This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR).
• SMBALERT: SMBus Alert
This bit is one when an SMBus Alert was received.
This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR).
• ARBLST: Arbitration Lost
This bit is one when the actual state of the SDA line did not correspond to the data driven onto it, indicating a higher-priority
transmission in progress by a different master.
This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR).
• DNAK: NAK in Data Phase Received
This bit is one when no ACK was received form slave during data transmission.
This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR).
• ANAK: NAK in Address Phase Received
This bit is one when no ACK was received from slave during address phase
This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR).
• BUSFREE: Two-wire Bus is Free
This bit is one when activity has completed on the two-wire bus.
Otherwise, this bit is cleared.
480
32099I–01/2012
AT32UC3L016/32/64
• IDLE: Master Interface is Idle
This bit is one when no command is in progress, and no command waiting to be issued.
Otherwise, this bit is cleared.
• CCOMP: Command Complete
This bit is one when the current command has completed successfully.
This bit is zero if the command failed due to conditions such as a NAK receved from slave.
This bit is cleared by writing 1 to the corresponding bit in the Status Clear Register (SCR).
• CRDY: Ready for More Commands
This bit is one when CMDR and/or NCMDR is ready to receive one or more commands.
This bit is cleared when this is no longer true.
• TXRDY: THR Data Ready
This bit is one when THR is ready for one or more data bytes.
This bit is cleared when this is no longer true (i.e. THR is full or transmission has stopped).
• RXRDY: RHR Data Ready
This bit is one when RX data are ready to be read from RHR.
This bit is cleared when this is no longer true.
481
32099I–01/2012
AT32UC3L016/32/64
21.9.9
Name:
Interrupt Enable Register
IER
Access Type:
Write-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
8
-
STOP
PECERR
TOUT
SMBALERT
ARBLST
DNAK
ANAK
7
6
5
4
3
2
1
0
-
-
BUSFREE
IDLE
CCOMP
CRDY
TXRDY
RXRDY
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR
482
32099I–01/2012
AT32UC3L016/32/64
21.9.10
Name:
Interrupt Disable Register
IDR
Access Type:
Write-only
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
-
STOP
PECERR
TOUT
SMBALERT
ARBLST
DNAK
ANAK
7
6
5
4
3
2
1
0
-
-
BUSFREE
IDLE
CCOMP
CRDY
TXRDY
RXRDY
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR
483
32099I–01/2012
AT32UC3L016/32/64
21.9.11
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
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
-
STOP
PECERR
TOUT
SMBALERT
ARBLST
DNAK
ANAK
7
6
5
4
3
2
1
0
-
-
BUSFREE
IDLE
CCOMP
CRDY
TXRDY
RXRDY
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
This bit is cleared when the corresponding bit in IDR is written to one.
This bit is set when the corresponding bit in IER is written to one.
484
32099I–01/2012
AT32UC3L016/32/64
21.9.12
Name:
Status Clear Register
SCR
Access Type :
Write-only
Offset:
0x2C
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
-
STOP
PECERR
TOUT
SMBALERT
ARBLST
DNAK
ANAK
7
6
5
4
3
2
1
0
-
-
-
-
CCOMP
-
-
-
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in SR and the corresponding interrupt request.
485
32099I–01/2012
AT32UC3L016/32/64
21.9.13
Name:
Parameter Register (PR)
PR
Access Type:
Read-only
Offset:
0x30
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
-
-
-
-
-
-
-
-
486
32099I–01/2012
AT32UC3L016/32/64
21.9.14
Name:
Version Register (VR)
VR
Access Type:
Read-only
Offset:
0x34
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
VARIANT
11
10
VERSION [11:8]
3
2
1
0
VERSION [7:0]
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the module. No functionality associated.
487
32099I–01/2012
AT32UC3L016/32/64
21.10 Module Configuration
The specific configuration for each TWIM instance is listed in the following tables.The module
bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details.
Table 21-7.
Module Clock Name
Module Name
Clock Name
Description
TWIM0
CLK_TWIM0
Clock for the TWIM0 bus interface
TWIM1
CLK_TWIM1
Clock for the TWIM1 bus interface
Table 21-8.
Register Reset Values
Register
Reset Value
VERSION
0x0000 0101
PARAMETER
0x0000 0000
488
32099I–01/2012
AT32UC3L016/32/64
22. Two-wire Slave Interface (TWIS)
Rev.: 1.1.2.1
22.1
Features
• Compatible with I²C standard
•
•
•
•
•
•
22.2
– Transfer speeds of 100 and 400 kbit/s
– 7 and 10-bit and General Call addressing
Compatible with SMBus standard
– Hardware Packet Error Checking (CRC) generation and verification with ACK response
– SMBALERT interface
– 25 ms clock low timeout delay
– 25 ms slave cumulative clock low extend time
Compatible with PMBus
DMA interface for reducing CPU load
Arbitrary transfer lengths, including 0 data bytes
Optional clock stretching if transmit or receive buffers not ready for data transfer
32-bit Peripheral Bus interface for configuration of the interface
Overview
The Atmel Two-wire Slave Interface (TWIS) interconnects components on a unique two-wire
bus, made up of one clock line and one data line with speeds of up to 400 kbit/s, based on a
byte-oriented transfer format. It can be used with any Atmel Two-wire Interface bus, I²C, or
SMBus-compatible master. The TWIS is always a bus slave and can transfer sequential or single bytes.
Below, Table 22-1 lists the compatibility level of the Atmel Two-wire Slave Interface and a full I²C
compatible device.
Table 22-1.
Atmel TWIS Compatibility with I²C Standard
I²C Standard
Atmel TWIS
Standard-mode (100 kbit/s)
Supported
Fast-mode (400 kbit/s)
Supported
7 or 10 bits Slave Addressing
Supported
START BYTE(1)
Not Supported
Repeated Start (Sr) Condition
Supported
ACK and NAK Management
Supported
Slope control and input filtering (Fast mode)
Supported
Clock stretching
Supported
Note:
1. START + b000000001 + Ack + Sr
489
32099I–01/2012
AT32UC3L016/32/64
Below, Table 22-2 lists the compatibility level of the Atmel Two-wire Slave Interface and a full
SMBus compatible device.
Table 22-2.
22.3
SMBus Standard
Atmel TWIS
Bus Timeouts
Supported
Address Resolution Protocol
Supported
Alert
Supported
Packet Error Checking
Supported
List of Abbreviations
Table 22-3.
22.4
Atmel TWIS Compatibility with SMBus Standard
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 22-1. Block Diagram
Peripheral
Bus Bridge
TWCK
I/O Controller
Two-wire
Interface
Power
Manager
TWD
TWALM
Interrupt
Controller
CLK_TWIS
TWI Interrupt
490
32099I–01/2012
AT32UC3L016/32/64
22.5
Application Block Diagram
Figure 22-2. 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
22.6
I/O Lines Description
Table 22-4.
I/O Lines Description
Pin Name
Pin Description
TWD
Two-wire Serial Data
Input/Output
TWCK
Two-wire Serial Clock
Input/Output
TWALM
SMBus SMBALERT
Input/Output
22.7
Type
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
22.7.1
I/O Lines
TWDand TWCK are bidirectional lines, connected to a positive supply voltage via a current
source or pull-up resistor (see Figure 22-5 on page 493). 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.
TWALM is used to implement the optional SMBus SMBALERT signal.
TWALM, TWD, and TWCK pins may be multiplexed with I/O Controller lines. To enable the
TWIS, the user must perform the following steps:
• Program the I/O Controller to:
– Dedicate TWD, TWCK, and optionally TWALM as peripheral lines.
– Define TWD, TWCK, and optionally TWALM as open-drain.
491
32099I–01/2012
AT32UC3L016/32/64
22.7.2
Power Management
If the CPU enters a sleep mode that disables clocks used by the TWIS, the TWIS will stop functioning and resume operation after the system wakes up from sleep mode. The TWIS is able to
wake the system from sleep mode upon address match, see Section 22.8.7 on page 500.
22.7.3
Clocks
The clock for the TWIS bus interface (CLK_TWIS) is generated by the Power Manager. This
clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the TWIS before disabling the clock, to avoid freezing the TWIS in an undefined state.
22.7.4
DMA
The TWIS DMA handshake interface is connected to the Peripheral DMA Controller. Using the
TWIS DMA functionality requires the Peripheral DMA Controller to be programmed after setting
up the TWIS.
22.7.5
Interrupts
The TWIS interrupt request lines are connected to the interrupt controller. Using the TWIS interrupts requires the interrupt controller to be programmed first.
22.7.6
22.8
22.8.1
Debug Operation
When an external debugger forces the CPU into debug mode, the TWIS continues normal operation. If the TWIS is configured in a way that requires it to be periodically serviced by the CPU
through interrupts or similar, improper operation or data loss may result during debugging.
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
22-4 on page 493).
Each transfer begins with a START condition and terminates with a STOP condition (see Figure
22-3).
• 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 22-3.
START and STOP Conditions
TWD
TWCK
Start
Stop
492
32099I–01/2012
AT32UC3L016/32/64
Figure 22-4. Transfer Format
TWD
TWCK
Start
22.8.2
Address
R/W
Ack
Data
Ack
Data
Ack
Stop
Operation
The TWIS has two modes of operation:
• Slave transmitter mode
• Slave receiver mode
A master is a device which starts and stops a transfer and generates the TWCK clock. A slave is
assigned an address and responds to requests from the master. These modes are described in
the following chapters.
Figure 22-5. 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
22.8.2.1
Bus Timing
The Timing Register (TR) is used to control the timing of bus signals driven by the TWIS. TR
describes bus timings as a function of cycles of the prescaled CLK_TWIS. The clock prescaling
can be selected through TR.EXP.
f CLK_TWIS
f PRESCALED = -----------------------( EXP + 1 )
2
TR has the following fields:
TLOWS: Prescaled clock cycles used to time SMBUS timeout TLOW:SEXT.
493
32099I–01/2012
AT32UC3L016/32/64
TTOUT: Prescaled clock cycles used to time SMBUS timeout TTIMEOUT.
SUDAT: Non-prescaled clock cycles for data setup and hold count. Used to time TSU_DAT.
EXP: Specifies the clock prescaler setting used for the SMBUS timeouts.
Figure 22-6. Bus Timing Diagram
t HIGH
t LOW
S
t
HD:STA
t LOW
t
SU:DAT
t
HD:DAT
t
t
22.8.2.2
t
SU:DAT
SU:STA
SU:STO
P
Sr
Setting Up and Performing a Transfer
Operation of the TWIS is mainly controlled by the Control Register (CR). The following list presents the main steps in a typical communication:
3. Before any transfers can be performed, bus timings must be configured by writing to the
Timing Register (TR).If the Peripheral DMA Controller is to be used for the transfers, it
must be set up.
4. The Control Register (CR) must be configured with information such as the slave
address, SMBus mode, Packet Error Checking (PEC), number of bytes to transfer, and
which addresses to match.
The interrupt system can be set up to generate interrupt request on specific events or error conditions, for example when a byte has been received.
The NBYTES register is only used in SMBus mode, when PEC is enabled. In I²C mode or in
SMBus mode when PEC is disabled, the NBYTES register is not used, and should be written to
zero. NBYTES is updated by hardware, so in order to avoid hazards, software updates of
NBYTES can only be done through writes to the NBYTES register.
22.8.2.3
Address Matching
The TWIS can be set up to match several different addresses. More than one address match
may be enabled simultaneously, allowing the TWIS to be assigned to several addresses. The
address matching phase is initiated after a START or REPEATED START condition. When the
TWIS receives an address that generates an address match, an ACK is automatically returned
to the master.
494
32099I–01/2012
AT32UC3L016/32/64
In I²C mode:
• The address in CR.ADR is checked for address match if CR.SMATCH is one.
• The General Call address is checked for address match if CR.GCMATCH is one.
In SMBus mode:
• The address in CR.ADR is checked for address match if CR.SMATCH is one.
• The Alert Response Address is checked for address match if CR.SMAL is one.
• The Default Address is checked for address match if CR.SMDA is one.
• The Host Header Address is checked for address match if CR.SMHH is one.
22.8.2.4
Clock Stretching
Any slave or bus master taking part in a transfer may extend the TWCK low period at any time.
The TWIS may extend the TWCK low period after each byte transfer if CR.STREN is one and:
• Module is in slave transmitter mode, data should be transmitted, but THR is empty, or
• Module is in slave receiver mode, a byte has been received and placed into the internal
shifter, but the Receive Holding Register (RHR) is full, or
• Stretch-on-address-match bit CR.SOAM=1 and slave was addressed. Bus clock remains
stretched until all address match bits in the Status Register (SR) have been cleared.
If CR.STREN is zero and:
• Module is in slave transmitter mode, data should be transmitted but THR is empty: Transmit
the value present in THR (the last transmitted byte or reset value), and set SR.URUN.
• Module is in slave receiver mode, a byte has been received and placed into the internal
shifter, but RHR is full: Discard the received byte and set SR.ORUN.
22.8.2.5
Bus Errors
If a bus error (misplaced START or STOP) condition is detected, the SR.BUSERR bit is set and
the TWIS waits for a new START condition.
22.8.3
Slave Transmitter Mode
If the TWIS matches an address in which the R/W bit in the TWI address phase transfer is set, it
will enter slave transmitter mode and set the SR.TRA bit (note that SR.TRA is set one
CLK_TWIS cycle after the relevant address match bit in the same register is set).
After the address phase, the following actions are performed:
1. If SMBus mode and PEC is used, NBYTES must be set up with the number of bytes to
transmit. This is necessary in order to know when to transmit the PEC byte. NBYTES
can also be used to count the number of bytes received if using DMA.
2. Byte to transmit depends on I²C/SMBus mode and CR.PEC:
– If in I²C mode or CR.PEC is zero or NBYTES is non-zero: The TWIS waits until THR
contains a valid data byte, possibly stretching the low period of TWCK. After THR
contains a valid data byte, the data byte is transferred to a shifter, and then
SR.TXRDY is changed to one because the THR is empty again.
– SMBus mode and CR.PEC is one: If NBYTES is zero, the generated PEC byte is
automatically transmitted instead of a data byte from THR. TWCK will not be
stretched by the TWIS.
3. The data byte in the shifter is transmitted.
495
32099I–01/2012
AT32UC3L016/32/64
4. NBYTES is updated. If CR.CUP is one, NBYTES is incremented, otherwise NBYTES is
decremented.
5. After each data byte has been transmitted, the master transmits an ACK (Acknowledge)
or NAK (Not Acknowledge) bit. If a NAK bit is received by the TWIS, the SR.NAK bit is
set. Note that this is done two CLK_TWIS cycles after TWCK has been sampled by the
TWIS to be HIGH (see Figure 22-9). The NAK indicates that the transfer is finished, and
the TWIS will wait for a STOP or REPEATED START. If an ACK bit is received, the
SR.NAK bit remains LOW. The ACK indicates that more data should be transmitted,
jump to step 2. At the end of the ACK/NAK clock cycle, the Byte Transfer Finished
(SR.BTF) bit is set. Note that this is done two CLK_TWIS cycles after TWCK has been
sampled by the TWIS to be LOW (see Figure 22-9). Also note that in the event that
SR.NAK bit is set, it must not be cleared before the SR.BTF bit is set to ensure correct
TWIS behavior.
6. If STOP is received, SR.TCOMP and SR.STO will be set.
7. If REPEATED START is received, SR.REP will be set.
The TWI transfers require the receiver to acknowledge each received data byte. During the
acknowledge clock pulse (9th pulse), the slave releases the data line (HIGH), enabling the master to pull it down in order to generate the acknowledge. The slave polls the data line during this
clock pulse and sets the NAK bit in SR if the master does not acknowledge the data byte. A NAK
means that the master does not wish to receive additional data bytes. As with the other status
bits, an interrupt can be generated if enabled in the Interrupt Enable Register (IER).
SR.TXRDY is used as Transmit Ready for the Peripheral DMA Controller transmit channel.
The end of the complete transfer is marked by the SR.TCOMP bit changing from zero to one.
See Figure 22-7 and Figure 22-8.
Figure 22-7. Slave Transmitter with One Data Byte
TWD
S
DADR
R
A
DATA
N
P
TCOMP
TXRDY
Write THR (DATA)
NBYTES set to 1
STOP sent by master
496
32099I–01/2012
AT32UC3L016/32/64
Figure 22-8. Slave Transmitter with Multiple Data Bytes
TWD
S
DADR
R
A
DATA n
A
DATA n+5
A
DATA n+m
N
P
TCOMP
TXRDY
Write THR (Data n)
NBYTES set to m
Write THR (Data n+1)
Write THR (Data n+m)
Last data sent
STOP sent by master
Figure 22-9. Timing Relationship between TWCK, SR.NAK, and SR.BTF
TWD
DATA (LSB)
N
P
TWCK
SR.NAK
SR.BTF
t1
t1
t1: (CLK_TWIS period) x 2
22.8.4
Slave Receiver Mode
If the TWIS matches an address in which the R/W bit in the TWI address phase transfer is
cleared, it will enter slave receiver mode and clear SR.TRA (note that SR.TRA is cleared one
CLK_TWIS cycle after the relevant address match bit in the same register is set).
After the address phase, the following is repeated:
1. If SMBus mode and PEC is used, NBYTES must be set up with the number of bytes to
receive. This is necessary in order to know which of the received bytes is the PEC byte.
NBYTES can also be used to count the number of bytes received if using DMA.
2. Receive a byte. Set SR.BTF when done.
3. Update NBYTES. If CR.CUP is written to one, NBYTES is incremented, otherwise
NBYTES is decremented. NBYTES is usually configured to count downwards if PEC is
used.
4. After a data byte has been received, the slave transmits an ACK or NAK bit. For ordinary data bytes, the CR.ACK field controls if an ACK or NAK should be returned. If PEC
is enabled and the last byte received was a PEC byte (indicated by NBYTES equal to
zero), The TWIS will automatically return an ACK if the PEC value was correct, otherwise a NAK will be returned.
5. If STOP is received, SR.TCOMP will be set.
6. If REPEATED START is received, SR.REP will be set.
The TWI transfers require the receiver to acknowledge each received data byte. During the
acknowledge clock pulse (9th pulse), the master releases the data line (HIGH), enabling the
497
32099I–01/2012
AT32UC3L016/32/64
slave to pull it down in order to generate the acknowledge. The master polls the data line during
this clock pulse.
The SR.RXRDY bit indicates that a data byte is available in the RHR. The RXRDY bit is also
used as Receive Ready for the Peripheral DMA Controller receive channel.
Figure 22-10. Slave Receiver with One Data Byte
TWD
S
DADR
W
A
DATA
A
P
TCOMP
RXRDY
Read RHR
Figure 22-11. Slave Receiver with Multiple Data Bytes
TWD
S
DADR
W
A
DATA n
A
DATA (n+1)
A
DAT A (n+m)-1
A
DATA (n+m)
P
A
TCOMP
RXRDY
Read RHR
DATA n
22.8.5
Read RHR
DATA (n+1)
Read RHR
DAT A (n+m)-1
Read RHR
DATA (n+m)
Using the Peripheral DMA Controller
The use of the Peripheral DMA Controller significantly reduces the CPU load. The user can set
up ring buffers for the Peripheral DMA Controller, containing data to transmit or free buffer space
to place received data. By initializing NBYTES to zero before a transfer, and writing a one to
CR.CUP, NBYTES is incremented by one each time a data has been transmitted or received.
This allows the user to detect how much data was actually transferred by the DMA system.
To assure correct behavior, respect the following programming sequences:
22.8.5.1
Data Transmit with the Peripheral DMA Controller
1. Initialize the transmit Peripheral DMA Controller (memory pointers, size, etc.).
2. Configure the TWIS (ADR, NBYTES, etc.).
3. Start the transfer by enabling the Peripheral DMA Controller to transmit.
4. Wait for the Peripheral DMA Controller end-of-transmit flag.
5. Disable the Peripheral DMA Controller.
22.8.5.2
Data Receive with the Peripheral DMA Controller
1. Initialize the receive Peripheral DMA Controller (memory pointers, size - 1, etc.).
2. Configure the TWIS (ADR, NBYTES, etc.).
498
32099I–01/2012
AT32UC3L016/32/64
3. Start the transfer by enabling the Peripheral DMA Controller to receive.
4. Wait for the Peripheral DMA Controller end-of-receive flag.
5. Disable the Peripheral DMA Controller.
22.8.6
SMBus Mode
SMBus mode is enabled by writing a one to the SMBus Mode Enable (SMEN) bit in CR. SMBus
mode operation is similar to I²C operation with the following exceptions:
• Only 7-bit addressing can be used.
• The SMBus standard describes a set of timeout values to ensure progress and throughput on
the bus. These timeout values must be written to TR.
• Transmissions can optionally include a CRC byte, called Packet Error Check (PEC).
• A dedicated bus line, SMBALERT, allows a slave to get a master’s attention.
• A set of addresses have been reserved for protocol handling, such as Alert Response
Address (ARA) and Host Header (HH) Address. Address matching on these addresses can
be enabled by configuring CR appropriately.
22.8.6.1
Packet Error Checking (PEC)
Each SMBus transfer can optionally end with a CRC byte, called the PEC byte. Writing a one to
the Packet Error Checking Enable (PECEN) bit in CR enables automatic PEC handling in the
current transfer. The PEC generator is always updated on every bit transmitted or received, so
that PEC handling on following linked transfers will be correct.
In slave receiver mode, the master calculates a PEC value and transmits it to the slave after all
data bytes have been transmitted. Upon reception of this PEC byte, the slave will compare it to
the PEC value it has computed itself. If the values match, the data was received correctly, and
the slave will return an ACK to the master. If the PEC values differ, data was corrupted, and the
slave will return a NAK value. The SR.SMBPECERR bit is set automatically if a PEC error
occurred.
In slave transmitter mode, the slave calculates a PEC value and transmits it to the master after
all data bytes have been transmitted. Upon reception of this PEC byte, the master will compare
it to the PEC value it has computed itself. If the values match, the data was received correctly. If
the PEC values differ, data was corrupted, and the master must take appropriate action.
The PEC byte is automatically inserted in a slave transmitter transmission if PEC enabled when
NBYTES reaches zero. The PEC byte is identified in a slave receiver transmission if PEC
enabled when NBYTES reaches zero. NBYTES must therefore be set to the total number of
data bytes in the transmission, including the PEC byte.
22.8.6.2
Timeouts
The Timing Register (TR) configures the SMBus timeout values. If a timeout occurs, the slave
will leave the bus. The SR.SMBTOUT bit is also set.
22.8.6.3
SMBALERT
A slave can get the master’s attention by pulling the SMBALERT line low. This is done by writing
a one to the SMBus Alert (SMBALERT) bit in CR. This will also enable address match on the
Alert Response Address (ARA).
499
32099I–01/2012
AT32UC3L016/32/64
22.8.7
Wakeup from Sleep Modes by TWI Address Match
The TWIS is able to wake the device up from a sleep mode upon an address match, including
sleep modes where CLK_TWIS is stopped. After detecting the START condition on the bus, The
TWIS will stretch TWCK until CLK_TWIS has started. The time required for starting CLK_TWIS
depends on which sleep mode the device is in. After CLK_TWIS has started, the TWIS releases
its TWCK stretching and receives one byte of data on the bus. At this time, only a limited part of
the device, including the TWIS, receives a clock, thus saving power. The TWIS goes on to
receive the slave address. If the address phase causes a TWIS address match, the entire
device is wakened and normal TWIS address matching actions are performed. Normal TWI
transfer then follows. If the TWIS is not addressed, CLK_TWIS is automatically stopped and the
device returns to its original sleep mode.
22.8.8
Identifying Bus Events
This chapter lists the different bus events, and how these affects the bits in the TWIS registers.
This is intended to help writing drivers for the TWIS.
Table 22-5.
Bus Events
Event
Effect
Slave transmitter has sent a
data byte
SR.THR is cleared.
SR.BTF is set.
The value of the ACK bit sent immediately after the data byte is given
by CR.ACK.
Slave receiver has received
a data byte
SR.RHR is set.
SR.BTF is set.
SR.NAK updated according to value of ACK bit received from master.
Start+Sadr on bus, but
address is to another slave
None.
Start+Sadr on bus, current
slave is addressed, but
address match enable bit in
CR is not set
None.
Start+Sadr on bus, current
slave is addressed,
corresponding address
match enable bit in CR set
Correct address match bit in SR is set.
SR.TRA updated according to transfer direction (updating is done one
CLK_TWIS cycle after address match bit is set)
Slave enters appropriate transfer direction mode and data transfer
can commence.
Start+Sadr on bus, current
slave is addressed,
corresponding address
match enable bit in CR set,
SR.STREN and SR.SOAM
are set.
Correct address match bit in SR is set.
SR.TRA updated according to transfer direction (updating is done one
CLK_TWIS cycle after address match bit is set).
Slave stretches TWCK immediately after transmitting the address
ACK bit. TWCK remains stretched until all address match bits in SR
have been cleared.
Slave enters appropriate transfer direction mode and data transfer
can commence.
Repeated Start received
after being addressed
SR.REP set.
SR.TCOMP unchanged.
Stop received after being
addressed
SR.STO set.
SR.TCOMP set.
500
32099I–01/2012
AT32UC3L016/32/64
Table 22-5.
Bus Events
Event
Effect
Start, Repeated Start, or
Stop received in illegal
position on bus
SR.BUSERR set.
SR.STO and SR.TCOMP may or may not be set depending on the
exact position of an illegal stop.
Data is to be received in
slave receiver mode,
SR.STREN is set, and RHR
is full
TWCK is stretched until RHR has been read.
Data is to be transmitted in
slave receiver mode,
SR.STREN is set, and THR
is empty
TWCK is stretched until THR has been written.
Data is to be received in
slave receiver mode,
SR.STREN is cleared, and
RHR is full
TWCK is not stretched, read data is discarded.
SR.ORUN is set.
Data is to be transmitted in
slave receiver mode,
SR.STREN is cleared, and
THR is empty
TWCK is not stretched, previous contents of THR is written to bus.
SR.URUN is set.
SMBus timeout received
SR.SMBTOUT is set.
TWCK and TWD are immediately released.
Slave transmitter in SMBus
PEC mode has transmitted
a PEC byte, that was not
identical to the PEC
calculated by the master
receiver.
Slave receiver discovers
SMBus PEC Error
Master receiver will transmit a NAK as usual after the last byte of a
master receiver transfer.
Master receiver will retry the transfer at a later time.
SR.SMBPECERR is set.
NAK returned after the data byte.
501
32099I–01/2012
AT32UC3L016/32/64
22.9
User Interface
Table 22-6.
TWIS Register Memory Map
Offset
Register
Register Name
Access
Reset
0x00
Control Register
CR
Read/Write
0x00000000
0x04
NBYTES Register
NBYTES
Read/Write
0x00000000
0x08
Timing Register
TR
Read/Write
0x00000000
0x0C
Receive Holding Register
RHR
Read-only
0x00000000
0x10
Transmit Holding Register
THR
Write-only
0x00000000
0x14
Packet Error Check Register
PECR
Read-only
0x00000000
0x18
Status Register
SR
Read-only
0x00000002
0x1C
Interrupt Enable Register
IER
Write-only
0x00000000
0x20
Interrupt Disable Register
IDR
Write-only
0x00000000
0x24
Interrupt Mask Register
IMR
Read-only
0x00000000
0x28
Status Clear Register
SCR
Write-only
0x00000000
0x2C
Parameter Register
PR
Read-only
-(1)
0x30
Version Register
VR
Read-only
-(1)
Note:
1. The reset values for these registers are device specific. Please refer to the Module Configuration section at the end of this
chapter.
502
32099I–01/2012
AT32UC3L016/32/64
22.9.1
Name:
Control Register
CR
Access Type:
Read/Write
Offset:
0x00
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
TENBIT
23
22
21
20
19
18
17
16
ADR[9:8]
ADR[7:0]
15
14
13
12
11
10
9
8
-
SOAM
CUP
ACK
PECEN
SMHH
SMDA
SMBALERT
7
6
5
4
3
2
1
0
SWRST
-
-
STREN
GCMATCH
SMATCH
SMEN
SEN
• TENBIT: Ten Bit Address Match
0: Disables Ten Bit Address Match.
1: Enables Ten Bit Address Match.
• ADR: Slave Address
Slave address used in slave address match. Bits 9:0 are used if in 10-bit mode, bits 6:0 otherwise.
• SOAM: Stretch Clock on Address Match
0: Does not stretch bus clock after address match.
1: Stretches bus clock after address match.
• CUP: NBYTES Count Up
0: Causes NBYTES to count down (decrement) per byte transferred.
1: Causes NBYTES to count up (increment) per byte transferred.
• ACK: Slave Receiver Data Phase ACK Value
0: Causes a low value to be returned in the ACK cycle of the data phase in slave receiver mode.
1: Causes a high value to be returned in the ACK cycle of the data phase in slave receiver mode.
• PECEN: Packet Error Checking Enable
0: Disables SMBus PEC (CRC) generation and check.
1: Enables SMBus PEC (CRC) generation and check.
• SMHH: SMBus Host Header
0: Causes the TWIS not to acknowledge the SMBus Host Header.
1: Causes the TWIS to acknowledge the SMBus Host Header.
• SMDA: SMBus Default Address
0: Causes the TWIS not to acknowledge the SMBus Default Address.
1: Causes the TWIS to acknowledge the SMBus Default Address.
• SMBALERT: SMBus Alert
0: Causes the TWIS to release the SMBALERT line and not to acknowledge the SMBus Alert Response Address (ARA).
1: Causes the TWIS to pull down the SMBALERT line and to acknowledge the SMBus Alert Response Address (ARA).
• SWRST: Software Reset
This bit will always read as 0.
Writing a zero to this bit has no effect.
503
32099I–01/2012
AT32UC3L016/32/64
Writing a one to this bit resets the TWIS.
• STREN: Clock Stretch Enable
0: Disables clock stretching if RHR/THR buffer full/empty. May cause over/underrun.
1: Enables clock stretching if RHR/THR buffer full/empty.
• GCMATCH: General Call Address Match
0: Causes the TWIS not to acknowledge the General Call Address.
1: Causes the TWIS to acknowledge the General Call Address.
• SMATCH: Slave Address Match
0: Causes the TWIS not to acknowledge the Slave Address.
1: Causes the TWIS to acknowledge the Slave Address.
• SMEN: SMBus Mode Enable
0: Disables SMBus mode.
1: Enables SMBus mode.
• SEN: Slave Enable
0: Disables the slave interface.
1: Enables the slave interface.
504
32099I–01/2012
AT32UC3L016/32/64
22.9.2
Name:
NBYTES Register
NBYTES
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
2
1
0
NBYTES
• NBYTES: Number of Bytes to Transfer
Writing to this field updates the NBYTES counter. The field can also be read to learn the progress of the transfer. NBYTES can
be incremented or decremented automatically by hardware.
505
32099I–01/2012
AT32UC3L016/32/64
22.9.3
Name:
Timing Register
TR
Access Type:
Read/Write
Offset:
0x08
Reset Value:
0x00000000
31
30
29
28
EXP
23
22
21
20
27
26
25
24
-
-
-
-
19
18
17
16
11
10
9
8
3
2
1
0
SUDAT
15
14
13
12
TTOUT
7
6
5
4
TLOWS
• EXP: Clock Prescaler
Used to specify how to prescale the SMBus TLOWS counter. The counter is prescaled according to the following formula:
f CLK_TWIS
f PRESCALED = -----------------------( EXP + 1 )
2
• SUDAT: Data Setup Cycles
Non-prescaled clock cycles for data setup count. Used to time TSU_DAT. Data is driven SUDAT cycles after TWCK low detected.
This timing is used for timing the ACK/NAK bits, and any data bits driven in slave transmitter mode.
• TTOUT: SMBus TTIMEOUT Cycles
Prescaled clock cycles used to time SMBus TTIMEOUT.
• TLOWS: SMBus TLOW:SEXT Cycles
Prescaled clock cycles used to time SMBus TLOW:SEXT.
506
32099I–01/2012
AT32UC3L016/32/64
22.9.4
Name:
Receive Holding Register
RHR
Access Type:
Read-only
Offset:
0x0C
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: Received Data Byte
When the RXRDY bit in the Status Register (SR) is one, this field contains a byte received from the TWI bus.
507
32099I–01/2012
AT32UC3L016/32/64
22.9.5
Name:
Transmit Holding Register
THR
Access Type:
Write-only
Offset:
0x10
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: Data Byte to Transmit
Write data to be transferred on the TWI bus here.
508
32099I–01/2012
AT32UC3L016/32/64
22.9.6
Name:
Packet Error Check Register
PECR
Access Type:
Read-only
Offset:
0x14
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
PEC
• PEC: Calculated PEC Value
The calculated PEC value. Updated automatically by hardware after each byte has been transferred. Reset by hardware after a
STOP condition. Provided if the user manually wishes to control when the PEC byte is transmitted, or wishes to access the PEC
value for other reasons. In ordinary operation, the PEC handling is done automatically by hardware.
509
32099I–01/2012
AT32UC3L016/32/64
22.9.7
Name:
Status Register
SR
Access Type:
Read-only
Offset:
0x18
Reset Value:
0x000000002
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
BTF
REP
STO
SMBDAM
SMBHHM
SMBALERTM
GCM
SAM
15
14
13
12
11
10
9
8
-
BUSERR
SMBPECERR
SMBTOUT
-
-
-
NAK
7
6
5
4
3
2
1
0
ORUN
URUN
TRA
-
TCOMP
SEN
TXRDY
RXRDY
• BTF: Byte Transfer Finished
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when byte transfer has completed.
• REP: Repeated Start Received
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when a REPEATED START condition is received.
• STO: Stop Received
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when the STOP condition is received.
• SMBDAM: SMBus Default Address Match
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when the received address matched the SMBus Default Address.
• SMBHHM: SMBus Host Header Address Match
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when the received address matched the SMBus Host Header Address.
• SMBALERTM: SMBus Alert Response Address Match
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when the received address matched the SMBus Alert Response Address.
• GCM: General Call Match
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when the received address matched the General Call Address.
• SAM: Slave Address Match
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when the received address matched the Slave Address.
• BUSERR: Bus Error
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when a misplaced START or STOP condition has occurred.
510
32099I–01/2012
AT32UC3L016/32/64
• SMBPECERR: SMBus PEC Error
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when a SMBus PEC error has occurred.
• SMBTOUT: SMBus Timeout
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when a SMBus timeout has occurred.
• NAK: NAK Received
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when a NAK was received from the master during slave transmitter operation.
• ORUN: Overrun
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when an overrun has occurred in slave receiver mode. Can only occur if CR.STREN is zero.
• URUN: Underrun
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when an underrun has occurred in slave transmitter mode. Can only occur if CR.STREN is zero.
• TRA: Transmitter Mode
0: The slave is in slave receiver mode.
1: The slave is in slave transmitter mode.
• TCOMP: Transmission Complete
This bit is cleared when the corresponding bit in SCR is written to one.
This bit is set when transmission is complete. Set after receiving a STOP after being addressed.
• SEN: Slave Enabled
0: The slave interface is disabled.
1: The slave interface is enabled.
• TXRDY: TX Buffer Ready
0: The TX buffer is full and should not be written to.
1: The TX buffer is empty, and can accept new data.
• RXRDY: RX Buffer Ready
0: No RX data ready in RHR.
1: RX data is ready to be read from RHR.
511
32099I–01/2012
AT32UC3L016/32/64
22.9.8
Name:
Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x1C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
BTF
REP
STO
SMBDAM
SMBHHM
SMBALERTM
GCM
SAM
15
14
13
12
11
10
9
8
-
BUSERR
SMBPECERR
SMBTOUT
-
-
-
NAK
7
6
5
4
3
2
1
0
ORUN
URUN
-
-
TCOMP
-
TXRDY
RXRDY
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will write a one to the corresponding bit in IMR.
512
32099I–01/2012
AT32UC3L016/32/64
22.9.9
Name:
Interrupt Disable Register
IDR
Access Type:
Write-only
Offset:
0x20
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
BTF
REP
STO
SMBDAM
SMBHHM
SMBALERTM
GCM
SAM
15
14
13
12
11
10
9
8
-
BUSERR
SMBPECERR
SMBTOUT
-
-
-
NAK
7
6
5
4
3
2
1
0
ORUN
URUN
-
-
TCOMP
-
TXRDY
RXRDY
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
513
32099I–01/2012
AT32UC3L016/32/64
22.9.10
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
Offset:
0x24
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
BTF
REP
STO
SMBDAM
SMBHHM
SMBALERTM
GCM
SAM
15
14
13
12
11
10
9
8
-
BUSERR
SMBPECERR
SMBTOUT
-
-
-
NAK
7
6
5
4
3
2
1
0
ORUN
URUN
-
-
TCOMP
-
TXRDY
RXRDY
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
This bit is cleared when the corresponding bit in IDR is written to one.
This bit is set when the corresponding bit in IER is written to one.
514
32099I–01/2012
AT32UC3L016/32/64
22.9.11
Name:
Status Clear Register
SCR
Access Type:
Write-only
Offset:
0x28
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
BTF
REP
STO
SMBDAM
SMBHHM
SMBALERTM
GCM
SAM
15
14
13
12
11
10
9
8
-
BUSERR
SMBPECERR
SMBTOUT
-
-
-
NAK
7
6
5
4
3
2
1
0
ORUN
URUN
-
-
TCOMP
-
-
-
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in SR and the corresponding interrupt request.
515
32099I–01/2012
AT32UC3L016/32/64
22.9.12
Name:
Parameter Register
PR
Access Type:
Read-only
Offset:
0x2C
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
-
-
-
-
-
-
-
-
516
32099I–01/2012
AT32UC3L016/32/64
22.9.13
Name:
Version Register (VR)
VR
Access Type:
Read-only
Offset:
0x30
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
VARIANT
11
10
VERSION [11:8]
3
2
1
0
VERSION [7:0]
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the module. No functionality associated.
517
32099I–01/2012
AT32UC3L016/32/64
22.10 Module Configuration
The specific configuration for each TWIS instance is listed in the following tables.The module
bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details.
Table 22-7.
Module Clock Name
Module Name
Clock Name
Description
TWIS0
CLK_TWIS0
Clock for the TWIS0 bus interface
TWIS1
CLK_TWIS1
Clock for the TWIS1 bus interface
Table 22-8.
Register Reset Values
Register
Reset Value
VERSION
0x00000112
PARAMETER
0x00000000
518
32099I–01/2012
AT32UC3L016/32/64
23. Pulse Width Modulation Controller (PWMA)
Rev 1.0.1.0
23.1
Features
• Left-aligned non-inverted 8-bit PWM
• Common 8-bit timebase counter
•
•
•
•
•
•
•
•
23.2
– Asynchronous clock source supported
– Spread-spectrum counter to allow a constantly varying duty cycle
Separate 8-bit duty cycle register per channel
Synchronized channel updates
– No glitches when changing the duty cycles
Interlinked operation supported
– Multiple channels can be updated with the same duty cycle value at a time
– Up to four channels can be updated with different duty cycle values at a time
Interrupt on PWM timebase overflow
Incoming peripheral events supported
– Pre-defined channels support incoming (increase/decrease) peripheral events from the
Peripheral Event System
– Increase event will increase the duty cycle by one
– Decrease event will decrease the duty cycle by one
One output peripheral event supported
– Connected to channel 0 and asserted when the common timebase counter is equal to the
programmed duty cycle for channel 0
Output PWM waveform for each channel
Open drain driving on selected pins for 5V PWM operation
Overview
The Pulse Width Modulation Controller (PWMA) controls several pulse width modulation (PWM)
channels. The number of channels is specific to the device. Each channel controls one square
output PWM waveform. Characteristics of the output PWM waveforms such as period and duty
cycle are configured through the user interface. All user interface registers are mapped on the
peripheral bus.
The duty cycle value for each channel can be set independently, while the period is determined
by a common timebase counter (TC). The timebase for the counter is selected by using the allocated asynchronous Generic Clock (GCLK). The user interface for the PWMA contains
handshake and synchronizing logic to ensure that no glitches occur on the output PWM waveforms while changing the duty cycle values.
PWMA duty cycle values can be changed using two approaches, either an interlinked singlevalue mode or an interlinked multi-value mode. In the interlinked single-value mode, any set of
channels, up to 32 channels, can be updated simultaneously with the same value while the other
channels remain unchanged. In the interlinked multi-value mode, up to 4 selected channels can
be updated with 4 different values while the other channels remain unchanged.
Some pins can be driven in open drain mode, allowing the PWMA to generate a 5V waveform
using an external pullup resistor.
519
32099I–01/2012
AT32UC3L016/32/64
23.3
Block Diagram
Figure 23-1. PWMA Block Diagram
PB
IRQ
PB Clock Domain
Control
CLK_PWMA
TOP TCLR SPREAD
BUSY
Duty Cycle
Interrupt
Handling
Channel
Select
Adjust
Channel_0
Sync
Spread
Spectrum
Counter
Timebase
Counter
GCLK
PWM Blocks
Duty Cycle
Register
TOFL
ETV
COMP
Channel 0
Channel 1
Channel m
GCLK Domain
PWMA[m:0]
23.4
I/O Lines Description
Each channel outputs one PWM waveform on one external I/O line.
Table 23-1.
23.5
I/O Line Description
Pin Name
Pin Description
Type
PWMA[n]
Output PWM waveform for one channel n
Output
PWMMOD[n]
Output PWM waveform for one channel n, open drain mode
Output
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
520
32099I–01/2012
AT32UC3L016/32/64
23.5.1
I/O Lines
The pins used for interfacing the PWMA may be multiplexed with I/O Controller lines. The programmer must first program the I/O Controller to assign the desired PWMA pins to their
peripheral function.
It is only required to enable the PWMA outputs actually in use.
23.5.2
Power Management
If the CPU enters a sleep mode that disables clocks used by the PWMA, the PWMA will stop
functioning and resume operation after the system wakes up from sleep mode.
23.5.3
Clocks
The clock for the PWMA bus interface (CLK_PWMA) is generated by the Power Manager. This
clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the PWMA before disabling the clock, to avoid freezing the PWMA in an undefined state.
Additionally, the PWMA depends on a dedicated Generic Clock (GCLK). The GCLK can be set
to a wide range of frequencies and clock sources and must be enabled in the System Control
Interface (SCIF) before the PWMA can be used.
23.5.4
Interrupts
The PWMA interrupt request lines are connected to the interrupt controller. Using the PWMA
interrupts requires the interrupt controller to be programmed first.
23.5.5
Peripheral Events
The PWMA peripheral events are connected via the Peripheral Event System. Refer to the
Peripheral Event System chapter for details.
23.5.6
Debug Operation
When an external debugger forces the CPU into debug mode, the PWMA continues normal
operation. If the PWMA is configured in a way that requires it to be periodically serviced by the
CPU through interrupts, improper operation or data loss may result during debugging.
23.6
Functional Description
The PWMA embeds a number of PWM channel submodules, each providing an output PWM
waveform. Each PWM channel contains a duty cycle register and a comparator. A common
timebase counter for all channels determines the frequency and the period for all the PWM
waveforms.
23.6.1
Enabling the PWMA
Once the GCLK has been enabled, the PWMA is enabled by writing a one to the EN bit in the
Control Register (CR).
23.6.2
Timebase Counter
The top value of the timebase counter defines the period of the PWMA output waveform. The
timebase counter starts at zero when the PWMA is enabled and counts upwards until it reaches
its effective top value (ETV). The effective top value is defined by specifying the desired number
of GCLK clock cycles in the TOP field of CR (CR.TOP) in normal operation (CR.SPREAD is
521
32099I–01/2012
AT32UC3L016/32/64
zero). When the timebase counter reaches its effective top value, it restarts counting from zero.
The period of the PWMA output waveform is then:
T PWMA = ( ETV + 1 ) ⋅ T GCLK
The timebase counter can be reset by writing a one to the Timebase Clear bit in CR (CR.TCLR).
Note that this can cause a glitch to the output PWM waveforms in use.
23.6.3
Spread Spectrum Counter
The spread spectrum counter allows the generation of constantly varying duty cycles on the output PWM waveforms. This is achieved by varying the effective top value of the timebase counter
in a range defined by the spread spectrum counter value.
When CR.SPREAD is not zero, the spread spectrum counter is enabled. Its range is defined by
CR.SPREAD. It starts to count from -CR.SPREAD when the PWMA is enabled or after reset
and counts upwards. When it reaches CR.SPREAD, it restarts to count from -CR.SPREAD
again. The spread spectrum counter will cause the effective top value (ETV) to vary from TOPSPREAD to TOP+SPREAD. Figure 23-2 on page 522 illustrates this. This leads to a constantly
varying duty cycle on the PWM output waveforms though the duty cycle values stored are
unchanged.
Figure 23-2. PWMA Adjusting Top Value for Timebase Counter
0xFF
SPREAD
TOP
Adjusting top value range
for the timerbase counter
-SPREAD
Duty Cycle
0x0
23.6.3.1
Special considerations
The maximum value of the timebase counter is 255. If SPREAD is written to a value that will
cause the ETV to exceed this value, the spread spectrum counter’s range will be limited to prevent the timebase counter to exceed its maximum value.
If SPREAD is written to a value causing (TOP-SPREAD) to be below zero, the spread specturm
counter’s range will be limited to prevent the timebase counter to count below zero.
In both cases, the SPREAD value read from the Control Register will be the same value as written to the SPREAD field.
522
32099I–01/2012
AT32UC3L016/32/64
When writing a one to CR.TCLR, the timebase counter and the spread spectrum counter are
reset at their lower limit values and the effective top value of the timebase counter will also be
reset.
23.6.4
Duty Cycle and Waveform Properties
Each PWM channel has its own duty cycle value (DCV) which is write-only and cannot be read
out. The duty cycle value can be changed in two approaches as described in Section23.6.5.
When the duty cycle value is zero, the PWM output is zero. Otherwise, the PWM output is set
when the timebase counter is zero, and cleared when the timebase counter reaches the duty
cycle value. This is summarized as:
⎧ low when DCV = 0 or TC > DCV
PWM Waveform = ⎨
⎩ high when TC ≤ DCV and DCV ≠ 0
Note that when increasing the duty cycle value for one channel from 0 to 1, the number of GCLK
cycles when the PWM waveform is high will jump from 0 to 2. When incrementing the duty cycle
value by one for any other values, the number of GCLK cycle when the waveform is high will
increase by one. This is summarized in Table 23-2.
Table 23-2.
PMW Waveform Duty Cycles
Duty Cycle Value
#Clock Cycles
When Waveform is High
#Clock Cycles
When Waveform is Low
0
0
ETV+1
1
2
ETV-1
2
3
ETV-2
...
...
...
ETV-1
ETV
1
ETV
ETV+1
0
Every other output PWM waveform toggles on the negative edge of the GCLK instead of the
positive edge. This is to avoid too many I/O toggling simultaneously on the output I/O lines.
23.6.5
Updating Duty Cycle Values
23.6.5.1
Interlinked Single Value PWM Operation
The PWM channels can be interlinked to allow multiple channels to be updated simultaneously
with the same duty cycle value. This value must be written to the Interlinked Single Value Duty
(ISDUTY) register. Each channel has a corresponding enabling bit in the Interlinked Single
Value Channel Set (ISCHSETm) register. When a bit is written to one in the ISCHSETm register,
the duty cycle register for the corresponding channel will be updated with the value stored in the
ISDUTY register. It can only be updated when the READY bit in the Status Register
(SR.READY) is one, indicating that the PWMA is ready for writing. Figure 23-3 on page 524
shows the writing procedure. It is thus possible to update the duty cycle values for up to 32 PWM
channels within one ISCHSETm register at a time.
523
32099I–01/2012
AT32UC3L016/32/64
Figure 23-3. Interlinked Single Value PWM Operation Flow
ISDUTY
DUTYm
23.6.5.2
ISCHSET
...
Write
Enable
DUTY1
DUTY0
Interlinked Multiple Value PWM Operation
The interlinked multiple value PWM operation allows up to four channels to be updated simultaneously with different duty cycle values. These duty cycle values must be written to the IMDUTY
register. The index number of the four channels to be updated is written to the four SEL fields in
the Interlinked Multiple Value Channel Select (IMCHSEL) register (IMCHSEL.SEL). When the
IMCHSEL register is written, the values stored in the IMDUTY register are synchronized to the
duty cycle registers for the channels selected by the SEL fields. Figure 23-4 on page 524 shows
the writing procedure.
Note that only writes to the implemented channels will be effective. If one of the IMCHSEL.SEL
fields points to a non-existing channel, the corresponding value in the IMDUTY register will not
be written. If the same channel is specified in multiple IMCHSEL.SEL fields, the channel will be
updated with the value stored in the corresponding upper field of the IMDUTY register.
Figure 23-4. Interlinked Multiple Value PWM Operation Flow
IMDUTY
IMCHSEL
MUX
DUTYm
23.6.6
...
DUTY1
DUTY0
Open Drain Mode
Some pins can be used in open drain mode, allowing the PWMA waveform to toggle between
0V and up to 5V on these pins. In this mode the PWMA will drive the pin to zero or leave the output open. An external pullup can be used to pull the pin up to the desired voltage.
To enable open drain mode on a pin the PWMAOD function must be selected instead of the
PWMA function in the I/O Controller. Please refer to the Module Configuration chapter for information about which pins are available in open drain mode.
23.6.7
Synchronization
Both the timebase counter and the spread spectrum counter can be reset and the duty cycle
registers can be written through the user interface of the module. This requires a synchronization between the PB and GCLK clock domains, which takes a few clock cycles of each clock
domain. The BUSY bit in SR indicates when the synchronization is ongoing. Writing to the module while the BUSY bit is set will result in discarding the new value.
524
32099I–01/2012
AT32UC3L016/32/64
Note that the duty cycle registers will not be updated with the new values until the timebase
counter reaches its top value, in order to avoid glitches. The BUSY bit in SR will always be set
during this updating and synchronization period.
23.6.8
Interrupts
When the timebase counter overflows, the Timebase Overflow bit in the Status Register
(SR.TOFL) is set. If the corresponding bit in the Interrupt Mask Register (IMR) is set, an interrupt
request will be generated.
Since the user needs to wait until the user interface is available between each write due to synchronization, a READY bit is provided in SR, which can be used to generate an interrupt
request.
The interrupt request will be generated if the corresponding bit in IMR is set. Bits in IMR are set
by writing a one to the corresponding bit in the Interrupt Enable Register (IER), and cleared by
writing a one to the corresponding bit in the Interrupt Disable Register (IDR). The interrupt
request remains active until the corresponding bit in SR is cleared by writing a one to the corresponding bit in the Status Clear Register (SCR).
23.6.9
Peripheral Events
23.6.9.1
Input Peripheral Events
The pre-defined channels support input peripheral events from the Peripheral Event System. An
increase event (event_incr) will increase the duty cycle value by one, and a decrease event
(event_decr) will decrease the duty cycle value by one. If an increase event and a decrease
event occur at the same time, the duty cycle value will not be changed.
The number of channels supporting input peripheral events is device specific. Please refer to the
Module Configuration section at the end of this chapter for details.
Input peripheral events must be enabled by writing a one to the corresponding bit in the Chanel
Event Enable Register (CHEERm) before peripheral events can be used to control the duty
cycle value. Each bit in the register corresponds to one channel, where bit 0 corresponds to
channel 0 and so on. Both the increase and decrease events are enabled for the corresponding
channel when a bit in the CHEERm register is written to one.
23.6.9.2
Output Peripheral Event
The PWMA also supports one output peripheral event (event_ch0) to the Peripheral Event System. This output peripheral event is connected to channel 0 and will be asserted when the
timebase counter reaches the duty cycle value for channel 0. This output event is always
enabled.
525
32099I–01/2012
AT32UC3L016/32/64
23.7
User Interface
Table 23-3.
Note:
PWMA Register Memory Map
Offset
Register
Register Name
Access
Reset
0x00
Control Register
CR
Read/Write
0x00000000
0x04
Interlinked Single Value Duty Register
ISDUTY
Write-only
0x00000000
0x08
Interlinked Multiple Value Duty Register
IMDUTY
Write-only
0x00000000
0x0C
Interlinked Multiple Value Channel Select
IMCHSEL
Write-only
0x00000000
0x10
Interrupt Enable Register
IER
Write-only
0x00000000
0x14
Interrupt Disable Register
IDR
Write-only
0x00000000
0x18
Interrupt Mask Register
IMR
Read-only
0x00000000
0x1C
Status Register
SR
Read-only
0x00000000
0x20
Status Clear Register
SCR
Write-only
0x00000000
0x24
Parameter Register
PARAMETER
Read-only
- (1)
0x28
Version Register
VERSION
Read-only
- (1)
0x30
Interlinked Single Value Channel Set 0
ISCHSET0
Write-only
0x00000000
0x38
Channel Event Enable Register 0
CHEER0
Write-only
0x00000000
0x40
Interlinked Single Value Channel Set 1
ISCHSET1
Write-only
0x00000000
0x48
Channel Event Enable Register 1
CHEER1
Write-only
0x00000000
1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter.
526
32099I–01/2012
AT32UC3L016/32/64
23.7.1
Name:
Control Register
CR
Access Type:
Read/Write
Offset:
0x00
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
15
14
13
SPREAD
12
11
10
9
8
TOP
7
6
5
4
3
2
1
0
-
-
-
-
-
-
TCLR
EN
• SPREAD: Spread Spectrum Limit Value
The spread spectrum limit value, together with the TOP field, defines the range for the spread spectrum counter. It is introduced
in order to achieve constant varying duty cycles on the output PWM waveforms. Refer to Section23.6.3 for more information.
• TOP: Timebase Counter Top Value
The top value for the timebase counter. The effective top value of the timebase counter is defined by both the TOP and the
SPREAD fields. Refer to Section23.6.2 for more information.
• TCLR: Timebase Clear
Writing a zero to this bit has no effect.
Writing a one to this bit will clear the timebase counter.
This bit is always read as zero.
• EN: Module Enable
0: The PWMA is disabled
1: The PWMA is enabled
527
32099I–01/2012
AT32UC3L016/32/64
23.7.2
Name:
Interlinked Single Value Duty Register
ISDUTY
Access Type:
Write-only
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
2
1
0
DUTY
• DUTY: Duty Cycle Value
The duty cycle value written to this field is written simultaneously to all channels selected in the ISCHSET registers.
If the value zero is written to DUTY all affected channels will be disabled. In this state the output waveform will be zero all the
time.
528
32099I–01/2012
AT32UC3L016/32/64
23.7.3
Name:
Interlinked Multiple Value Duty Register
IMDUTY
Access Type:
Write-only
Offset:
0x08
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
DUTY3
23
22
21
20
DUTY2
15
14
13
12
DUTY1
7
6
5
4
DUTY0
• DUTYn: Duty Cycle
The value written to DUTY field n will be written to the PWMA channel selected by the corresponding SEL field in the IMCHSEL
register.
If the value zero is written to DUTY all affected channels will be disabled. In this state the output waveform will be zero all the
time.
529
32099I–01/2012
AT32UC3L016/32/64
23.7.4
Name:
Interlinked Multiple Value Channel Select
IMCHSEL
Access Type:
Write-only
Offset:
0x0C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
SEL3
23
22
21
20
SEL2
15
14
13
12
SEL1
7
6
5
4
SEL0
• SELn: Channel Select
The duty cycle of the PWMA channel SELn will be updated with the value in the DUTYn field of the IMDUTY register when
IMCHSEL is written. If SELn points to a non-implemented channel, the write will be discarded.
Note:
The duty registers will be updated with the value stored in the IMDUTY register when the IMCHSEL register is written. Synchronization takes place immeidately when an IMCHSEL register is written. The duty cycle registers will, however, not be updated
until the synchronization is completed and the timebase counter reaches its top value in order to avoid glitches.
530
32099I–01/2012
AT32UC3L016/32/64
23.7.5
Name:
Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x10
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
-
-
-
-
-
READY
-
TOFL
Writing a zero to a bit in this register has no effect
Writing a one to a bit in this register will set the corresponding bit in IMR.
531
32099I–01/2012
AT32UC3L016/32/64
23.7.6
Interrupt Disable Register
Name:
IDR
Access Type:
Write-only
Offset:
0x14
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
-
-
-
-
-
READY
-
TOFL
Writing a zero to a bit in this register has no effect
Writing a one to a bit in this register will clear the corresponding bit in IMR.
532
32099I–01/2012
AT32UC3L016/32/64
23.7.7
Interrupt Mask Register
Name:
IMR
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
-
-
-
-
-
READY
-
TOFL
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
A bit in this register is cleared when the corresponding bit in IDR is written to one.
A bit in this register is set when the corresponding bit in IER is written to one.
533
32099I–01/2012
AT32UC3L016/32/64
23.7.8
Name:
Status Register
SR
Access Type:
Read-only
Offset:
0x1C
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
-
-
-
-
BUSY
READY
-
TOFL
• BUSY: Interface Busy
This bit is automatically cleared when the interface is no longer busy.
This bit is set when the user interface is busy and will not respond to new write operations.
• READY: Interface Ready
This bit is cleared by writing a one to the corresponding bit in the SCR register.
This bit is set when the BUSY bit has a 1-to-0 transition.
• TOFL: Timebase Overflow
This bit is cleared by writing a one to corresponding bit in the SCR register.
This bit is set when the timebase counter has wrapped at its top value.
534
32099I–01/2012
AT32UC3L016/32/64
23.7.9
Name:
Status Clear Register
SCR
Access Type:
Write-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
1
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
READY
-
TOFL
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in SR and the corresponding interrupt request.
This register always reads as zero.
535
32099I–01/2012
AT32UC3L016/32/64
23.7.10
Name:
Parameter Register
PARAMETER
Access Type:
Read-only
Offset:
0x24
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
CHANNELS
• CHANNELS: Channels Implemented
This field contains the number of channels implemented on the device.
536
32099I–01/2012
AT32UC3L016/32/64
23.7.11
Name:
Version Register
VERSION
Access Type:
Read-only
Offset:
0x28
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the module. No functionality associated.
537
32099I–01/2012
AT32UC3L016/32/64
23.7.12
Name:
Interlinked Single Value Channel Set
ISCHSETm
Access Type:
Write-only
Offset:
0x30+m*0x10
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
SET
23
22
21
20
SET
15
14
13
12
SET
7
6
5
4
SET
• SET: Single Value Channel Set
If the bit n in SET is one, the duty cycle of PWMA channel n will be updated with the value written to ISDUTY.
If more than one ISCHSET register is present, ISCHSET0 controls channels 31 to 0 and ISCHSET1 controls channels 63 to 32.
Note:
The duty registers will be updated with the value stored in the ISDUTY register when any ISCHSETm register is written. Synchronization takes place immeidately when an ISCHSET register is written. The duty cycle registers will, however, not be
updated until the synchronization is completed and the timebase counter reaches its top value in order to avoid glitches.
538
32099I–01/2012
AT32UC3L016/32/64
23.7.13
Name:
Channel Event Enable Register
CHEERm
Access Type:
Write-only
Offset:
0x38+m*0x10
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
CHEE
23
22
21
20
CHEE
15
14
13
12
CHEE
7
6
5
4
CHEE
• CHEE: Channel Event Enable
0: The input peripheral event for the corresponding channel is disabled.
1: The input peripheral event for the corresponding channel is enabled.
Both increase and decrease events for channel n are enabled if bit n is one.
If more than one CHEER register is present, CHEER0 controls channels 31-0 and CHEER1 controls channels 64-32 and so on.
539
32099I–01/2012
AT32UC3L016/32/64
23.8
Module Configuration
The specific configuration for each PWMA instance is listed in the following tables. The module
bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details.
Table 23-4.
PWMA Configuration
Feature
PWMA
Number of PWM channels
36
Channels supporting incoming peripheral events
0, 6, 8, 9, 11, 14, 19, and 20
PWMA channels with Open Drain mode
21, 27, and 28
Table 23-5.
PWMA Clocks
Clock Name
Descripton
CLK_PWMA
Clock for the PWMA bus interface
GCLK
The generic clock used for the PWMA is GCLK3
Table 23-6.
Register Reset Values
Register
Reset Value
VERSION
0x00000101
PARAMETER
0x00000024
540
32099I–01/2012
AT32UC3L016/32/64
24. Timer/Counter (TC)
Rev: 2.2.3.1.3
24.1
Features
• Three 16-bit Timer Counter channels
• A wide range of functions including:
•
•
•
•
24.2
– 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
Peripheral event input on all A lines in capture mode
Overview
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 TC block has two global registers which act upon all three TC channels.
The Block Control Register (BCR) allows the three channels to be started simultaneously with
the same instruction.
The Block Mode Register (BMR) defines the external clock inputs for each channel, allowing
them to be chained.
541
32099I–01/2012
AT32UC3L016/32/64
24.3
Block Diagram
Figure 24-1. TC Block Diagram
I/O
Contr oller
TIMER_CLOCK1
TCLK0
TIMER_CLOCK2
TIOA1
TIMER_CLOCK4
XC0
TIOA2
TIMER_CLOCK3
TCLK1
XC1
TCLK2
XC2
TIMER_CLOCK5
Timer/Counter
Channel 0
TIOA
TIOB
A0
TIOA0
B0
TIOB0
TC0XC0S
SYNC
CLK0
CLK1
CLK2
INT0
TCLK0
XC0
TCLK1
TIOA0
XC1
TIOA2
XC2
TCLK2
Timer/Counter
Channel 1
XC0
TCLK1
XC1
TCLK2
XC2
TIOA0
TIOA1
TC2XC2S
TIOB
A1
TIOA1
B1
TIOB1
SYNC
TC1XC1S
TCLK0
TIOA
Timer/Counter
Channel 2
INT1
TIOA
TIOB
SYNC
A2
TIOA2
B2
TIOB2
INT2
Timer Count er
Interrupt
Controller
24.4
I/O Lines Description
Table 24-1.
24.5
I/O Lines Description
Pin Name
Description
Type
CLK0-CLK2
External Clock Input
Input
A0-A2
I/O Line A
Input/Output
B0-B2
I/O Line B
Input/Output
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
24.5.1
I/O Lines
The pins used for interfacing the compliant external devices may be multiplexed with I/O lines.
The user must first program the I/O Controller to assign the TC pins to their peripheral functions.
542
32099I–01/2012
AT32UC3L016/32/64
When using the TIOA lines as inputs the user must make sure that no peripheral events are generated on the line. Refer to the Peripheral Event System chapter for details.
24.5.2
Power Management
If the CPU enters a sleep mode that disables clocks used by the TC, the TC will stop functioning
and resume operation after the system wakes up from sleep mode.
24.5.3
Clocks
The clock for the TC bus interface (CLK_TC) is generated by the Power Manager. This clock is
enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the
TC before disabling the clock, to avoid freezing the TC in an undefined state.
24.5.4
Interrupts
The TC interrupt request line is connected to the interrupt controller. Using the TC interrupt
requires the interrupt controller to be programmed first.
24.5.5
Peripheral Events
The TC peripheral events are connected via the Peripheral Event System. Refer to the Peripheral Event System chapter for details.
24.5.6
Debug Operation
The Timer Counter clocks are frozen during debug operation, unless the OCD system keeps
peripherals running in debug operation.
24.6
Functional Description
24.6.1
TC Description
The three channels of the Timer Counter are independent and identical in operation. The registers for channel programming are listed in Figure 24-3 on page 558.
24.6.1.1
Channel I/O Signals
As described in Figure 24-1 on page 542, each Channel has the following I/O signals.
Table 24-2.
Channel I/O Signals Description
Block/Channel
Signal Name
XC0, XC1, XC2
Channel Signal
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
24.6.1.2
Description
Interrupt Signal Output
Synchronization Input Signal
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 Counter Overflow Status bit in the Channel n Status Register (SRn.COVFS) is set.
543
32099I–01/2012
AT32UC3L016/32/64
The current value of the counter is accessible in real time by reading the Channel n Counter
Value Register (CVn). 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.
24.6.1.3
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 A0, A1 or A2 for
chaining by writing to the BMR register. See Figure 24-2 on page 544.
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. See the Module Configuration Chapter for details about
the connection of these clock sources.
• External clock signals: XC0, XC1 or XC2. See the Module Configuration Chapter for details
about the connection of these clock sources.
This selection is made by the Clock Selection field in the Channel n Mode Register
(CMRn.TCCLKS).
The selected clock can be inverted with the Clock Invert bit in CMRn (CMRn.CLKI). 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
Signal Selection field in the CMRn register (CMRn.BURST) defines this signal.
Note:
In all cases, if an external clock is used, the duration of each of its levels must be longer than the
CLK_TC period. The external clock frequency must be at least 2.5 times lower than the CLK_TC.
Figure 24-2. Clock Selection
TCCLKS
TIMER_CLOCK1
TIMER_CLOCK2
CLKI
TIMER_CLOCK3
TIMER_CLOCK4
TIMER_CLOCK5
Selected
Clock
XC0
XC1
XC2
BURST
1
544
32099I–01/2012
AT32UC3L016/32/64
24.6.1.4
Clock control
The clock of each counter can be controlled in two different ways: it can be enabled/disabled
and started/stopped. See Figure 24-3 on page 545.
• The clock can be enabled or disabled by the user by writing to the Counter Clock
Enable/Disable Command bits in the Channel n Clock Control Register (CCRn.CLKEN and
CCRn.CLKDIS). In Capture mode it can be disabled by an RB load event if the Counter Clock
Disable with RB Loading bit in CMRn is written to one (CMRn.LDBDIS). In Waveform mode,
it can be disabled by an RC Compare event if the Counter Clock Disable with RC Compare
bit in CMRn is written to one (CMRn.CPCDIS). When disabled, the start or the stop actions
have no effect: only a CLKEN command in CCRn can re-enable the clock. When the clock is
enabled, the Clock Enabling Status bit is set in SRn (SRn.CLKSTA).
• The clock can also be started or stopped: a trigger (software, synchro, external or compare)
always starts the clock. In Capture mode the clock can be stopped by an RB load event if the
Counter Clock Stopped with RB Loading bit in CMRn is written to one (CMRn.LDBSTOP). In
Waveform mode it can be stopped by an RC compare event if the Counter Clock Stopped
with RC Compare bit in CMRn is written to one (CMRn.CPCSTOP). The start and the stop
commands have effect only if the clock is enabled.
Figure 24-3. Clock Control
Selected
Clock
Trigger
CLKSTA
Q
Q
S
CLKEN
CLKDIS
S
R
R
Counter
Clock
24.6.1.5
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 selection is done by writing to the Wave bit in the CCRn register
(CCRn.WAVE).
In Capture mode, TIOA and TIOB are configured as inputs.
545
32099I–01/2012
AT32UC3L016/32/64
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.
24.6.1.6
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 writing a one to the
Software Trigger Command bit in CCRn (CCRn.SWTRG).
• 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 a one to the Synchro Command bit in the BCR register
(BCR.SYNC).
• Compare RC Trigger: RC is implemented in each channel and can provide a trigger when the
counter value matches the RC value if the RC Compare Trigger Enable bit in CMRn
(CMRn.CPCTRG) is written to one.
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 to be one of the following signals: TIOB, XC0, XC1, or XC2. This external
event can then be programmed to perform a trigger by writing a one to the External Event Trigger Enable bit in CMRn (CMRn.ENETRG).
If an external trigger is used, the duration of the pulses must be longer than the CLK_TC 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.
24.6.1.7
24.6.2
Peripheral events on TIOA inputs
The TIOA input lines are ored internally with peripheral events from the Peripheral Event System. To capture using events the user must ensure that the corresponding pin functions for the
TIOA line are disabled. When capturing on the external TIOA pin the user must ensure that no
peripheral events are generated on this pin.
Capture Operating Mode
This mode is entered by writing a zero to the CMRn.WAVE bit.
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 24-4 on page 548 shows the configuration of the TC channel when programmed in Capture mode.
24.6.2.1
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.
546
32099I–01/2012
AT32UC3L016/32/64
The RA Loading Selection field in CMRn (CMRn.LDRA) defines the TIOA edge for the loading of
the RA register, and the RB Loading Selection field in CMRn (CMRn.LDRB) defines the TIOA
edge for the loading of the RB register.
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 Load Overrun Status bit in
SRn (SRn.LOVRS). In this case, the old value is overwritten.
24.6.2.2
Trigger conditions
In addition to the SYNC signal, the software trigger and the RC compare trigger, an external trigger can be defined.
The TIOA or TIOB External Trigger Selection bit in CMRn (CMRn.ABETRG) selects TIOA or
TIOB input signal as an external trigger. The External Trigger Edge Selection bit in CMRn
(CMRn.ETREDG) defines the edge (rising, falling or both) detected to generate an external trigger. If CMRn.ETRGEDG is zero (none), the external trigger is disabled.
547
32099I–01/2012
32099I–01/2012
TIOA
TIOB
SYNC
MTIOA
MTIOB
TIMER_CLOCK2
TIMER_CLOCK3
TIMER_CLOCK4
TIMER_CLOCK5
XC0
XC1
XC2
TIMER_CLOCK1
1
Edge
Detector
ETRGEDG
SWTRG
CLKI
Edge
Detector
LDRA
CLK
Trig
S
R
OVF
If RA is Loaded
CPCTRG
16-bit
Counter
RESET
Q
LDBSTOP
R
S
CLKEN
Edge
Detector
LDRB
Capture
Register A
Q
CLKSTA
LDBDIS
Capture
Register B
CLKDIS
SR
Timer/Counter Channel
If RA is not Loaded
or RB is Loaded
ABETRG
BURST
TCCLKS
Compare RC =
Register C
COVFS
LDRBS
INT
AT32UC3L016/32/64
Figure 24-4. Capture Mode
LOVRS
CPCS
LDRAS
ETRGS
IMR
548
AT32UC3L016/32/64
24.6.3
Waveform Operating Mode
Waveform operating mode is entered by writing a one to the CMRn.WAVE bit.
In Waveform operating mode the TC channel generates one or two PWM signals with the same
frequency and independently programmable duty cycles, or generates different types of oneshot 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.
Figure 24-5 on page 550 shows the configuration of the TC channel when programmed in
Waveform operating mode.
24.6.3.1
Waveform selection
Depending on the Waveform Selection field in CMRn (CMRn.WAVSEL), the behavior of CVn
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.
549
32099I–01/2012
32099I–01/2012
TIOB
SYNC
XC2
XC1
TIMER_CLOCK5
XC0
TIMER_CLOCK4
TIMER_CLOCK3
TIMER_CLOCK2
TIMER_CLOCK1
1
Edge
Detector
EEVTEDG
SWTRG
ENETRG
Trig
CLK
R
S
Register A
Q
CLKSTA
Compare RA =
OVF
WAVSEL
RESET
16-bit
Counter
WAVSEL
Q
SR
Timer/Counter Channel
EEVT
BURST
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
O utput Contr oller
O utput Cont r oller
TCCLKS
TIOB
MTIOB
TIOA
MTIOA
AT32UC3L016/32/64
Figure 24-5. Waveform Mode
CPCS
CPBS
COVFS
ETRGS
IMR
550
AT32UC3L016/32/64
24.6.3.2
WAVSEL = 0
When CMRn.WAVSEL is zero, the value of CVn is incremented from 0 to 0xFFFF. Once
0xFFFF has been reached, the value of CVn is reset. Incrementation of CVn starts again and
the cycle continues. See Figure 24-6 on page 551.
An external event trigger or a software trigger can reset the value of CVn. It is important to note
that the trigger may occur at any time. See Figure 24-7 on page 552.
RC Compare cannot be programmed to generate a trigger in this configuration. At the same
time, RC Compare can stop the counter clock (CMRn.CPCSTOP = 1) and/or disable the counter
clock (CMRn.CPCDIS = 1).
Figure 24-6. WAVSEL= 0 Without Trigger
Counter Value
Counter cleared by compare match with
0xFFFF
0xFFFF
RC
RB
RA
Waveform Examples
Time
TIOB
TIOA
551
32099I–01/2012
AT32UC3L016/32/64
Figure 24-7. WAVSEL= 0 With Trigger
Counter Value
Counter cleared by compare match with 0xFFFF
0xFFFF
RC
Counter cleared by trigger
RB
RA
Waveform Examples
Time
TIOB
TIOA
24.6.3.3
WAVSEL = 2
When CMRn.WAVSEL is two, the value of CVn is incremented from zero to the value of RC,
then automatically reset on a RC Compare. Once the value of CVn has been reset, it is then
incremented and so on. See Figure 24-8 on page 553.
It is important to note that CVn can be reset at any time by an external event or a software trigger if both are programmed correctly. See Figure 24-9 on page 553.
In addition, RC Compare can stop the counter clock (CMRn.CPCSTOP) and/or disable the
counter clock (CMRn.CPCDIS = 1).
552
32099I–01/2012
AT32UC3L016/32/64
Figure 24-8. WAVSEL = 2 Without Trigger
Counter Value
0xFFFF
Counter cleared by compare match
with RC
RC
RB
RA
Waveform Examples
Time
TIOB
TIOA
Figure 24-9. WAVSEL = 2 With Trigger
Counter Value
0xFFFF
Counter cleared by compare match with RC
Counter cleared by trigger
RC
RB
RA
Waveform Examples
Time
TIOB
TIOA
24.6.3.4
WAVSEL = 1
When CMRn.WAVSEL is one, the value of CVn is incremented from 0 to 0xFFFF. Once 0xFFFF
is reached, the value of CVn is decremented to 0, then re-incremented to 0xFFFF and so on.
See Figure 24-10 on page 554.
553
32099I–01/2012
AT32UC3L016/32/64
A trigger such as an external event or a software trigger can modify CVn at any time. If a trigger
occurs while CVn is incrementing, CVn then decrements. If a trigger is received while CVn is
decrementing, CVn then increments. See Figure 24-11 on page 554.
RC Compare cannot be programmed to generate a trigger in this configuration.
At the same time, RC Compare can stop the counter clock (CMRn.CPCSTOP = 1) and/or disable the counter clock (CMRn.CPCDIS = 1).
Figure 24-10. WAVSEL = 1 Without Trigger
Counter Value
Counter decremented by compare match
with 0xFFFF
0xFFFF
RC
RB
RA
Time
Waveform Examples
TIOB
TIOA
Figure 24-11. WAVSEL = 1 With Trigger
Counter Value
Counter decremented by compare match with 0xFFFF
0xFFFF
Counter decremented by trigger
RC
RB
Counter incremented by trigger
RA
Waveform Examples
Time
TIOB
TIOA
554
32099I–01/2012
AT32UC3L016/32/64
24.6.3.5
WAVSEL = 3
When CMRn.WAVSEL is three, the value of CVn is incremented from zero to RC. Once RC is
reached, the value of CVn is decremented to zero, then re-incremented to RC and so on. See
Figure 24-12 on page 555.
A trigger such as an external event or a software trigger can modify CVn at any time. If a trigger
occurs while CVn is incrementing, CVn then decrements. If a trigger is received while CVn is
decrementing, CVn then increments. See Figure 24-13 on page 556.
RC Compare can stop the counter clock (CMRn.CPCSTOP = 1) and/or disable the counter clock
(CMRn.CPCDIS = 1).
Figure 24-12. WAVSEL = 3 Without Trigger
Counter Value
0xFFFF
Counter cleared by compare match with RC
RC
RB
RA
Waveform Examples
Time
TIOB
TIOA
555
32099I–01/2012
AT32UC3L016/32/64
Figure 24-13. WAVSEL = 3 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
TIOB
Time
TIOA
24.6.3.6
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 External Event Selection field in CMRn (CMRn.EEVT) selects the external trigger. The
External Event Edge Selection field in CMRn (CMRn.EEVTEDG) defines the trigger edge for
each of the possible external triggers (rising, falling or both). If CMRn.EEVTEDG is written to
zero, no external event is defined.
If TIOB is defined as an external event signal (CMRn.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 writing a one to the
CMRn.ENETRG bit.
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 CMRn.WAVSEL field.
24.6.3.7
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
• 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 following fields in CMRn:
• RC Compare Effect on TIOB (CMRn.BCPC)
556
32099I–01/2012
AT32UC3L016/32/64
• RB Compare Effect on TIOB (CMRn.BCPB)
• RC Compare Effect on TIOA (CMRn.ACPC)
• RA Compare Effect on TIOA (CMRn.ACPA)
557
32099I–01/2012
AT32UC3L016/32/64
24.7
User Interface
Table 24-3.
TC Register Memory Map
Offset
Register
Register Name
Access
Reset
0x00
Channel 0 Control Register
CCR0
Write-only
0x00000000
0x04
Channel 0 Mode Register
CMR0
Read/Write
0x00000000
0x10
Channel 0 Counter Value
CV0
Read-only
0x00000000
0x14
Channel 0 Register A
RA0
Read/Write(1)
0x00000000
0x18
Channel 0 Register B
RB0
Read/Write(1)
0x00000000
0x1C
Channel 0 Register C
RC0
Read/Write
0x00000000
0x20
Channel 0 Status Register
SR0
Read-only
0x00000000
0x24
Interrupt Enable Register
IER0
Write-only
0x00000000
0x28
Channel 0 Interrupt Disable Register
IDR0
Write-only
0x00000000
0x2C
Channel 0 Interrupt Mask Register
IMR0
Read-only
0x00000000
0x40
Channel 1 Control Register
CCR1
Write-only
0x00000000
0x44
Channel 1 Mode Register
CMR1
Read/Write
0x00000000
0x50
Channel 1 Counter Value
CV1
Read-only
0x00000000
0x54
Channel 1 Register A
RA1
(1)
0x00000000
(1)
0x00000000
Read/Write
0x58
Channel 1 Register B
RB1
Read/Write
0x5C
Channel 1 Register C
RC1
Read/Write
0x00000000
0x60
Channel 1 Status Register
SR1
Read-only
0x00000000
0x64
Channel 1 Interrupt Enable Register
IER1
Write-only
0x00000000
0x68
Channel 1 Interrupt Disable Register
IDR1
Write-only
0x00000000
0x6C
Channel 1 Interrupt Mask Register
IMR1
Read-only
0x00000000
0x80
Channel 2 Control Register
CCR2
Write-only
0x00000000
0x84
Channel 2 Mode Register
CMR2
Read/Write
0x00000000
0x90
Channel 2 Counter Value
CV2
Read-only
0x00000000
0x94
Channel 2 Register A
RA2
Read/Write(1)
0x00000000
0x98
Channel 2 Register B
RB2
Read/Write(1)
0x00000000
0x9C
Channel 2 Register C
RC2
Read/Write
0x00000000
0xA0
Channel 2 Status Register
SR2
Read-only
0x00000000
0xA4
Channel 2 Interrupt Enable Register
IER2
Write-only
0x00000000
0xA8
Channel 2 Interrupt Disable Register
IDR2
Write-only
0x00000000
0xAC
Channel 2 Interrupt Mask Register
IMR2
Read-only
0x00000000
0xC0
Block Control Register
BCR
Write-only
0x00000000
0xC4
Block Mode Register
BMR
Read/Write
0x00000000
0xF8
Features Register
FEATURES
Read-only
-(2)
0xFC
Version Register
VERSION
Read-only
-(2)
558
32099I–01/2012
AT32UC3L016/32/64
Notes:
1. Read-only if CMRn.WAVE is zero.
2. The reset values are device specific. Please refer to the Module Configuration section at the
end of this chapter.
559
32099I–01/2012
AT32UC3L016/32/64
24.7.1
Name:
Channel Control Register
CCR
Access Type:
Write-only
Offset:
0x00 + n * 0x40
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
-
-
-
-
-
SWTRG
CLKDIS
CLKEN
• SWTRG: Software Trigger Command
1: Writing a one to this bit will perform a software trigger: the counter is reset and the clock is started.
0: Writing a zero to this bit has no effect.
• CLKDIS: Counter Clock Disable Command
1: Writing a one to this bit will disable the clock.
0: Writing a zero to this bit has no effect.
• CLKEN: Counter Clock Enable Command
1: Writing a one to this bit will enable the clock if CLKDIS is not one.
0: Writing a zero to this bit has no effect.
560
32099I–01/2012
AT32UC3L016/32/64
24.7.2
Name:
Channel Mode Register: Capture Mode
CMR
Access Type:
Read/Write
Offset:
0x04 + n * 0x40
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
WAVE
CPCTRG
-
-
-
ABETRG
7
6
5
4
3
2
LDBDIS
LDBSTOP
BURST
LDRB
CLKI
LDRA
9
8
ETRGEDG
1
0
TCCLKS
• LDRB: RB Loading Selection
LDRB
Edge
0
none
1
rising edge of TIOA
2
falling edge of TIOA
3
each edge of TIOA
• LDRA: RA Loading Selection
LDRA
Edge
0
none
1
rising edge of TIOA
2
falling edge of TIOA
3
each edge of TIOA
• WAVE
1: Capture mode is disabled (Waveform mode is enabled).
0: Capture mode is enabled.
• CPCTRG: RC Compare Trigger Enable
1: RC Compare resets the counter and starts the counter clock.
0: RC Compare has no effect on the counter and its clock.
• ABETRG: TIOA or TIOB External Trigger Selection
1: TIOA is used as an external trigger.
561
32099I–01/2012
AT32UC3L016/32/64
0: TIOB is used as an external trigger.
• ETRGEDG: External Trigger Edge Selection
ETRGEDG
Edge
0
none
1
rising edge
2
falling edge
3
each edge
• LDBDIS: Counter Clock Disable with RB Loading
1: Counter clock is disabled when RB loading occurs.
0: Counter clock is not disabled when RB loading occurs.
• LDBSTOP: Counter Clock Stopped with RB Loading
1: Counter clock is stopped when RB loading occurs.
0: Counter clock is not stopped when RB loading occurs.
• BURST: Burst Signal Selection
BURST
Burst Signal Selection
0
The clock is not gated by an external signal
1
XC0 is ANDed with the selected clock
2
XC1 is ANDed with the selected clock
3
XC2 is ANDed with the selected clock
• CLKI: Clock Invert
1: The counter is incremented on falling edge of the clock.
0: The counter is incremented on rising edge of the clock.
• TCCLKS: Clock Selection
TCCLKS
Clock Selected
0
TIMER_CLOCK1
1
TIMER_CLOCK2
2
TIMER_CLOCK3
3
TIMER_CLOCK4
4
TIMER_CLOCK5
5
XC0
6
XC1
7
XC2
562
32099I–01/2012
AT32UC3L016/32/64
24.7.3
Name:
Channel Mode Register: Waveform Mode
CMR
Access Type:
Read/Write
Offset:
0x04 + n * 0x40
Reset Value:
0x00000000
31
30
29
BSWTRG
23
27
BEEVT
22
21
ASWTRG
15
28
20
WAVE
13
7
6
19
CPCDIS
CPCSTOP
4
BURST
BCPB
18
11
ENETRG
5
24
17
16
ACPC
12
WAVSEL
25
BCPC
AEEVT
14
26
ACPA
10
9
EEVT
3
CLKI
8
EEVTEDG
2
1
0
TCCLKS
• BSWTRG: Software Trigger Effect on TIOB
BSWTRG
Effect
0
none
1
set
2
clear
3
toggle
• BEEVT: External Event Effect on TIOB
BEEVT
Effect
0
none
1
set
2
clear
3
toggle
563
32099I–01/2012
AT32UC3L016/32/64
• BCPC: RC Compare Effect on TIOB
BCPC
Effect
0
none
1
set
2
clear
3
toggle
• BCPB: RB Compare Effect on TIOB
BCPB
Effect
0
none
1
set
2
clear
3
toggle
• ASWTRG: Software Trigger Effect on TIOA
ASWTRG
Effect
0
none
1
set
2
clear
3
toggle
• AEEVT: External Event Effect on TIOA
AEEVT
Effect
0
none
1
set
2
clear
3
toggle
• ACPC: RC Compare Effect on TIOA
ACPC
Effect
0
none
1
set
2
clear
3
toggle
564
32099I–01/2012
AT32UC3L016/32/64
• ACPA: RA Compare Effect on TIOA
ACPA
Effect
0
none
1
set
2
clear
3
toggle
• WAVE
1: Waveform mode is enabled.
0: Waveform mode is disabled (Capture mode is enabled).
• WAVSEL: Waveform Selection
WAVSEL
Effect
0
UP mode without automatic trigger on RC Compare
1
UPDOWN mode without automatic trigger on RC Compare
2
UP mode with automatic trigger on RC Compare
3
UPDOWN mode with automatic trigger on RC Compare
• ENETRG: External Event Trigger Enable
1: The external event resets the counter and starts the counter clock.
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.
• EEVT: External Event Selection
EEVT
Note:
Signal selected as external event
TIOB Direction
0
TIOB
input(1)
1
XC0
output
2
XC1
output
3
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.
• EEVTEDG: External Event Edge Selection
EEVTEDG
Edge
0
none
1
rising edge
2
falling edge
3
each edge
• CPCDIS: Counter Clock Disable with RC Compare
1: Counter clock is disabled when counter reaches RC.
0: Counter clock is not disabled when counter reaches RC.
565
32099I–01/2012
AT32UC3L016/32/64
• CPCSTOP: Counter Clock Stopped with RC Compare
1: Counter clock is stopped when counter reaches RC.
0: Counter clock is not stopped when counter reaches RC.
• BURST: Burst Signal Selection
BURST
Burst Signal Selection
0
The clock is not gated by an external signal.
1
XC0 is ANDed with the selected clock.
2
XC1 is ANDed with the selected clock.
3
XC2 is ANDed with the selected clock.
• CLKI: Clock Invert
1: Counter is incremented on falling edge of the clock.
0: Counter is incremented on rising edge of the clock.
• TCCLKS: Clock Selection
TCCLKS
Clock Selected
0
TIMER_CLOCK1
1
TIMER_CLOCK2
2
TIMER_CLOCK3
3
TIMER_CLOCK4
4
TIMER_CLOCK5
5
XC0
6
XC1
7
XC2
566
32099I–01/2012
AT32UC3L016/32/64
24.7.4
Name:
Channel Counter Value Register
CV
Access Type:
Read-only
Offset:
0x10 + n * 0x40
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
CV[15:8]
7
6
5
4
CV[7:0]
• CV: Counter Value
CV contains the counter value in real time.
567
32099I–01/2012
AT32UC3L016/32/64
24.7.5
Name:
Channel Register A
RA
Access Type:
Read-only if CMRn.WAVE = 0, Read/Write if CMRn.WAVE = 1
Offset:
0x14 + n * 0X40
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
RA[15:8]
7
6
5
4
RA[7:0]
• RA: Register A
RA contains the Register A value in real time.
568
32099I–01/2012
AT32UC3L016/32/64
24.7.6
Name:
Channel Register B
RB
Access Type:
Read-only if CMRn.WAVE = 0, Read/Write if CMRn.WAVE = 1
Offset:
0x18 + n * 0x40
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
RB[15:8]
7
6
5
4
RB[7:0]
• RB: Register B
RB contains the Register B value in real time.
569
32099I–01/2012
AT32UC3L016/32/64
24.7.7
Name:
Channel Register C
RC
Access Type:
Read/Write
Offset:
0x1C + n * 0x40
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
RC[15:8]
7
6
5
4
RC[7:0]
• RC: Register C
RC contains the Register C value in real time.
570
32099I–01/2012
AT32UC3L016/32/64
24.7.8
Name:
Channel Status Register
SR
Access Type:
Read-only
Offset:
0x20 + n * 0x40
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
MTIOB
MTIOA
CLKSTA
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
ETRGS
LDRBS
LDRAS
CPCS
CPBS
CPAS
LOVRS
COVFS
Note: Reading the Status Register will also clear the interrupt bit for the corresponding interrupts.
• MTIOB: TIOB Mirror
1: TIOB is high. If CMRn.WAVE is zero, this means that TIOB pin is high. If CMRn.WAVE is one, this means that TIOB is driven
high.
0: TIOB is low. If CMRn.WAVE is zero, this means that TIOB pin is low. If CMRn.WAVE is one, this means that TIOB is driven
low.
• MTIOA: TIOA Mirror
1: TIOA is high. If CMRn.WAVE is zero, this means that TIOA pin is high. If CMRn.WAVE is one, this means that TIOA is driven
high.
0: TIOA is low. If CMRn.WAVE is zero, this means that TIOA pin is low. If CMRn.WAVE is one, this means that TIOA is driven
low.
• CLKSTA: Clock Enabling Status
1: This bit is set when the clock is enabled.
0: This bit is cleared when the clock is disabled.
• ETRGS: External Trigger Status
1: This bit is set when an external trigger has occurred.
0: This bit is cleared when the SR register is read.
• LDRBS: RB Loading Status
1: This bit is set when an RB Load has occurred and CMRn.WAVE is zero.
0: This bit is cleared when the SR register is read.
• LDRAS: RA Loading Status
1: This bit is set when an RA Load has occurred and CMRn.WAVE is zero.
0: This bit is cleared when the SR register is read.
• CPCS: RC Compare Status
1: This bit is set when an RC Compare has occurred.
0: This bit is cleared when the SR register is read.
571
32099I–01/2012
AT32UC3L016/32/64
• CPBS: RB Compare Status
1: This bit is set when an RB Compare has occurred and CMRn.WAVE is one.
0: This bit is cleared when the SR register is read.
• CPAS: RA Compare Status
1: This bit is set when an RA Compare has occurred and CMRn.WAVE is one.
0: This bit is cleared when the SR register is read.
• LOVRS: Load Overrun Status
1: This bit is set when RA or RB have been loaded at least twice without any read of the corresponding register and
CMRn.WAVE is zero.
0: This bit is cleared when the SR register is read.
• COVFS: Counter Overflow Status
1: This bit is set when a counter overflow has occurred.
0: This bit is cleared when the SR register is read.
572
32099I–01/2012
AT32UC3L016/32/64
24.7.9
Name:
Channel Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x24 + n * 0x40
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
ETRGS
LDRBS
LDRAS
CPCS
CPBS
CPAS
LOVRS
COVFS
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
573
32099I–01/2012
AT32UC3L016/32/64
24.7.10
Name:
Channel Interrupt Disable Register
IDR
Access Type:
Write-only
Offset:
0x28 + n * 0x40
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
ETRGS
LDRBS
LDRAS
CPCS
CPBS
CPAS
LOVRS
COVFS
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
574
32099I–01/2012
AT32UC3L016/32/64
24.7.11
Name:
Channel Interrupt Mask Register
IMR
Access Type:
Read-only
Offset:
0x2C + n * 0x40
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
ETRGS
LDRBS
LDRAS
CPCS
CPBS
CPAS
LOVRS
COVFS
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
A bit in this register is cleared when the corresponding bit in IDR is written to one.
A bit in this register is set when the corresponding bit in IER is written to one.
575
32099I–01/2012
AT32UC3L016/32/64
24.7.12
Name:
Block Control Register
BCR
Access Type:
Write-only
Offset:
0xC0
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
-
-
-
-
-
-
-
SYNC
• SYNC: Synchro Command
1: Writing a one to this bit asserts the SYNC signal which generates a software trigger simultaneously for each of the channels.
0: Writing a zero to this bit has no effect.
576
32099I–01/2012
AT32UC3L016/32/64
24.7.13
Name:
Block Mode Register
BMR
Access Type:
Read/Write
Offset:
0xC4
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
-
-
TC2XC2S
TC1XC1S
TC0XC0S
• TC2XC2S: External Clock Signal 2 Selection
TC2XC2S
Signal Connected to XC2
0
TCLK2
1
none
2
TIOA0
3
TIOA1
• TC1XC1S: External Clock Signal 1 Selection
TC1XC1S
Signal Connected to XC1
0
TCLK1
1
none
2
TIOA0
3
TIOA2
577
32099I–01/2012
AT32UC3L016/32/64
• TC0XC0S: External Clock Signal 0 Selection
TC0XC0S
Signal Connected to XC0
0
TCLK0
1
none
2
TIOA1
3
TIOA2
578
32099I–01/2012
AT32UC3L016/32/64
24.7.14
Name:
Features Register
FEATURES
Access Type:
Read-only
Offset:
0xF8
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
-
-
-
-
-
-
BRPBHSB
UPDNIMPL
7
6
5
4
3
2
1
0
CTRSIZE
• BRPBHSB: Bridge type is PB to HSB
1: Bridge type is PB to HSB.
0: Bridge type is not PB to HSB.
• UPDNIMPL: Up/down is implemented
1: Up/down counter capability is implemented.
0: Up/down counter capability is not implemented.
• CTRSIZE: Counter size
This field indicates the size of the counter in bits.
579
32099I–01/2012
AT32UC3L016/32/64
24.7.15
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
580
32099I–01/2012
AT32UC3L016/32/64
24.8
Module Configuration
The specific configuration for each Timer/Counter instance is listed in the following tables.The
module bus clocks listed here are connected to the system bus clocks. Please refer to the Power
Manager chapter for details.
Table 24-4.
TC Bus Interface Clocks
Clock Name
Description
CLK_TC0
Clock for the TC0 bus interface
CLK_TC1
Clock for the TC1 bus interface
Table 24-5.
Register Reset Values
Register
Reset Value
FEATURES
0x00000310
VERSION
0x00000223
Table 24-6.
Features
Feature
Implementation
Bridge type
PB to HSB
Up/down counter capability implemented
Yes
Counter size
16
Table 24-7.
Event Sources
Instance
Source(s)
TC0
TIOA
TC1
-
581
32099I–01/2012
AT32UC3L016/32/64
24.8.1
Clock Connections
There are two Timer/Counter modules, TC0 and TC1, with three channels each, giving a total of
six Timer/Counter channels. Each Timer/Counter channel can independently select an internal
or external clock source for its counter:
Table 24-8.
Timer/Counter Clock Connections
Module
Source
Name
Connection
TC0
Internal
TIMER_CLOCK1
32 KHz oscillator clock (OSC32K)
TIMER_CLOCK2
PBA Clock / 2
TIMER_CLOCK3
PBA Clock / 8
TIMER_CLOCK4
PBA Clock / 32
TIMER_CLOCK5
PBA Clock / 128
XC0
See Section 3.2 on page 9
External
XC1
XC2
TC1
Internal
External
TIMER_CLOCK1
32 KHz oscillator clock (OSC32K)
TIMER_CLOCK2
PBA Clock / 2
TIMER_CLOCK3
PBA Clock / 8
TIMER_CLOCK4
PBA Clock / 32
TIMER_CLOCK5
PBA Clock / 128
XC0
See Section 3.2 on page 9
XC1
XC2
582
32099I–01/2012
AT32UC3L016/32/64
25. Peripheral Event System
Rev: 1.0.0.1
25.1
Features
•
•
•
•
25.2
Direct peripheral to peripheral communication system
Allows peripherals to receive, react to, and send peripheral events without CPU intervention
Cycle deterministic event communication
Asynchronous interrupts allow advanced peripheral operation in low power sleep modes
Overview
Several peripheral modules can be configured to emit or respond to signals known as peripheral
events. The exact condition to trigger a peripheral event, or the action taken upon receiving a
peripheral event, is specific to each module. Peripherals that respond to peripheral events are
called peripheral event users and peripherals that emit peripheral events are called peripheral
event generators. A single module can be both a peripheral event generator and user.
The peripheral event generators and users are interconnected by a network known as the
Peripheral Event System. This allows low latency peripheral-to-peripheral signaling without CPU
intervention, and without consuming system resources such as bus or RAM bandwidth. This
offloads the CPU and system resources compared to a traditional interrupt-based software
driven system.
25.3
Peripheral Event System Block Diagram
Figure 25-1. Peripheral Event System Block Diagram
Generator
Generator
25.4
25.4.1
Peripheral
Event
System
User
Generator/
User
Functional Description
Configuration
The Peripheral Event System in the AT32UC3L016/32/64 has a fixed mapping of peripheral
events between generators and users, as described in Table 25-1 to Table 25-4. Thus, the user
does not need to configure the interconnection between the modules, although each peripheral
event can be enabled or disabled at the generator or user side as described in the peripheral
chapter for each module.
583
32099I–01/2012
AT32UC3L016/32/64
Table 25-1.
Peripheral Event Mapping from ACIFB to PWMA
Generator
Generated Event
ACIFB channel 0
AC0 VINP > AC0 VINN
User
Effect
PWMA duty cycle value increased by one
PWMA channel 0
AC0 VINN > AC0 VINP
ACIFB channel 1
AC1 VINP > AC1 VINN
PWMA duty cycle value decreased by one
PWMA duty cycle value increased by one
PWMA channel 6
AC1 VINN > AC1 VINP
ACIFB channel 2
AC2 VINP > AC2 VINN
PWMA duty cycle value decreased by one
PWMA duty cycle value increased by one
PWMA channel 8
AC2 VINN > AC2 VINP
ACIFB channel 3
ACIFB channel 4
AC3 VINP > AC3 VINN
PWMA duty cycle value decreased by one
PWMA duty cycle value increased by one
PWMA channel 9
AC3 VINN > AC3 VINP
PWMA duty cycle value decreased by one
AC4 VINP > AC4 VINN
PWMA duty cycle value increased by one
No
PWMA channel 11
AC4 VINN > AC4 VINP
ACIFB channel 5
AC5 VINP > AC5 VINN
PWMA duty cycle value decreased by one
PWMA duty cycle value increased by one
PWMA channel 14
AC5 VINN > AC5 VINP
ACIFB channel 6
AC6 VINP > AC6 VINN
PWMA duty cycle value decreased by one
PWMA duty cycle value increased by one
PWMA channel 19
AC6 VINN > AC6 VINP
ACIFB channel 7
AC7 VINP > AC7 VINN
PWMA duty cycle value decreased by one
PWMA duty cycle value increased by one
PWMA channel 20
AC7 VINN > AC7 VINP
ACIFB channel n
Table 25-2.
Generator
ACn VINN > ACn VINP
PWMA duty cycle value decreased by one
Automatically used by the CAT when
performing QMatrix acquisition.
CAT
No
Peripheral Event Mapping from GPIO to TC
Generated Event
User
Pin change on PA00-PA07
Pin change on PA08-PA15
GPIO
Asynchronous
Effect
Asynchronous
A0 capture
TC0
Pin change on PA16-PA23
A1 capture
A2 capture
Pin change on PB00-PB07
No
A1 capture
TC1
Pin change on PB08-PB15
A2 capture
584
32099I–01/2012
AT32UC3L016/32/64
Table 25-3.
Peripheral Event Mapping from AST
Generator
Generated Event
User
Effect
ACIFB
Comparison is triggered if the ACIFB.CONFn
register is written to 11 (Event Triggered Single
Measurement Mode) and the EVENTEN bit in
the ACIFB.CTRL register is written to 1.
ADCIFB
Conversion is triggered if the TRGMOD bit in
the ADCIFB.TRGR register is written to 111
(Peripheral Event Trigger).
CAT
Trigger one iteration of autonomous touch
detection.
Overflow event
Periodic event
Alarm event
Overflow event
AST
Periodic event
Alarm event
Asynchronous
Yes
Overflow event
Periodic event
Alarm event
Table 25-4.
Peripheral Event Mapping from PWMA
Generator
PWMA channel 0
Generated Event
User
Effect
Asynchronous
ACIFB
Comparison is triggered if the ACIFB.CONFn
register is written to 11 (Event Triggered Single
Measurement Mode) and the EVENTEN bit in
the ACIFB.CTRL register is written to 1.
No
Timebase counter
reaches the duty cycle
value.
ADCIFB
25.4.2
Conversion is triggered if the TRGMOD bit in
the ADCIFB.TRGR register is written to 111
(Peripheral Event Trigger).
Peripheral Event Connections
Each generated peripheral event is connected to one or more users. If a peripheral event is connected to multiple users, the peripheral event can trigger actions in multiple modules.
A peripheral event user can likewise be connected to one or more peripheral event generators. If
a peripheral event user is connected to multiple generators, the peripheral events are OR’ed
together to a single peripheral event. This means that peripheral events from either one of the
generators will result in a peripheral event to the user.
To configure a peripheral event, the peripheral event must be enabled at both the generator and
user side. Even if a generator is connected to multiple users, only the users with the peripheral
event enabled will trigger on the peripheral event.
25.4.3
Low Power Operation
As the peripheral events do not require CPU intervention, they are available in Idle mode. They
are also available in deeper sleep modes if both the generator and user remain clocked in that
mode.
Certain events are known as asynchronous peripheral events, as identified in Table 25-1 to
Table 25-4. These can be issued even when the system clock is stopped, and revive unclocked
user peripherals. The clock will be restarted for this module only, without waking the system from
sleep mode. The clock remains active only as long as required by the triggered function, before
being switched off again, and the system remains in the original sleep mode. The CPU and sys-
585
32099I–01/2012
AT32UC3L016/32/64
tem will only be woken up if the user peripheral generates an interrupt as a result of the
operation. This concept is known as SleepWalking and is described in further detail in the Power
Manager chapter. Note that asynchronous peripheral events may be associated with a delay
due to the need to restart the system clock source if this has been stopped in the sleep mode.
25.5
Application Example
This application example shows how the Peripheral Event System can be used to program the
ADC Interface to perform ADC conversions at selected intervals.
Conversions of the active analog channels are started with a software or a hardware trigger.
One of the possible hardware triggers is a peripheral event trigger, allowing the Peripheral Event
System to synchronize conversion with some configured peripheral event source. From Table
25-3 and Table 25-4, it can be read that this peripheral event source can be either an AST
peripheral event, or an event from the PWM Controller. The AST can generate periodic peripheral events at selected intervals, among other types of peripheral events. The Peripheral Event
System can then be used to set up the ADC Interface to sample an analog signal at regular
intervals.
The user must enable peripheral events in the AST and in the ADC Interface to accomplish this.
The periodic peripheral event in the AST is enabled by writing a one to the corresponding bit in
the AST Event Enable Register (EVE). To select the peripheral event trigger for the ADC Interface, the user must write the value 0x7 to the Trigger Mode (TRGMOD) field in the ADC
Interface Trigger Register (TRGR). When the peripheral events are enabled, the AST will generate peripheral events at the selected intervals, and the Peripheral Event System will route the
peripheral events to the ADC Interface, which will perform ADC conversions at the selected
intervals.
Figure 25-2. Application Example
Trigger
Periodic peripheral
conversion
Peripheral
event
AST
Event
System
ADC
Interface
Since the AST peripheral event is asynchronous, the description above will also work in sleep
modes where the ADC clock is stopped. In this case, the ADC clock (and clock source, if
needed) will be restarted during the ADC conversion. After the conversion, the ADC clock and
clock source will return to the sleep state, unless the ADC generates an interrupt, which in turn
will wake up the system. Using asynchronous interrupts thus allows ADC operation in much
lower power states than would otherwise be possible.
586
32099I–01/2012
AT32UC3L016/32/64
26. ADC Interface (ADCIFB)
Rev:1.0.1.1
26.1
Features
• Multi-channel Analog-to-Digital Converter with up to 12-bit resolution
• Enhanced Resolution Mode
•
•
•
•
•
26.2
– 11-bit resolution obtained by interpolating 4 samples
– 12-bit resolution obtained by interpolating 16 samples
Glueless interface with resistive touch screen panel, allowing
– Resistive Touch Screen position measurement
– Pen detection and pen loss detection
Integrated enhanced sequencer
– ADC Mode
– Resistive Touch Screen Mode
Numerous trigger sources
– Software
– Embedded 16-bit timer for periodic trigger
– Pen detect trigger
– Continuous trigger
– External trigger, rising, falling, or any-edge trigger
– Peripheral event trigger
ADC Sleep Mode for low power ADC applications
Programmable ADC timings
– Programmable ADC clock
– Programmable startup time
Overview
The ADC Interface (ADCIFB) converts analog input voltages to digital values. The ADCIFB is
based on a Successive Approximation Register (SAR) 10-bit Analog-to-Digital Converter (ADC).
The conversions extend from 0V to ADVREFP.
The ADCIFB supports 8-bit and 10-bit resolution mode, in addition to enhanced resolution mode
with 11-bit and 12-bit resolution. Conversion results are reported in a common register for all
channels.
The 11-bit and 12-bit resolution modes are obtained by interpolating multiple samples to acquire
better accuracy. For 11-bit mode 4 samples are used, which gives an effective sample rate of
1/4 of the actual sample frequency. For 12-bit mode 16 samples are used, giving a effective
sample rate of 1/16 of actual. This arrangement allows conversion speed to be traded for better
accuracy.
Conversions can be started for all enabled channels, either by a software trigger, by detection of
a level change on the external trigger pin (TRIGGER), or by an integrated programmable timer.
When the Resistive Touch Screen Mode is enabled, an integrated sequencer automatically configures the pad control signals and performs resistive touch screen conversions.
The ADCIFB also integrates an ADC Sleep Mode, a Pen-Detect Mode, and an Analog Compare
Mode, and connects with one Peripheral DMA Controller channel. These features reduce both
power consumption and processor intervention.
587
32099I–01/2012
AT32UC3L016/32/64
26.3
Block Diagram
Figure 26-1. ADCIFB Block Diagram
ADCIFB
I/O Controller
TRIGGER
ADP0
ADP1
Trigger
Timer
Resisitve Touch
Screen
Sequencer
ADC Control
Logic
CLK_ADC
CLK_ADCIFB
User
Interface
Peripheral
Bus
AD1
AD2
....
AD3
ADn
Analog Multiplexer
AD0
Successive
Approximation
Register
Analog-to-Digital
Converter
DMA
Request
Interrupt
Request
ADVREFP
588
32099I–01/2012
AT32UC3L016/32/64
26.4
I/O Lines Description
Table 26-1.
I/O Lines Description
Pin Name
Description
Type
ADVREFP
Reference voltage
Analog
TRIGGER
External trigger
Digital
ADP0
Drive Pin 0 for Resistive Touch Screen top channel (Xp)
Digital
ADP1
Drive Pin 1 for Resistive Touch Screen right channel (Yp)
Digital
AD0-ADn
Analog input channels 0 to n
Analog
26.5
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
26.5.1
I/O Lines
The analog input pins can be multiplexed with I/O Controller lines. The user must make sure the
I/O Controller is configured correctly to allow the ADCIFB access to the AD pins before the
ADCIFB is instructed to start converting data. If the user fails to do this the converted data may
be wrong.
The number of analog inputs is device dependent, please refer to the ADCIFB Module Configuration chapter for the number of available AD inputs on the current device.
The ADVREFP pin must be connected correctly prior to using the ADCIFB. Failing to do so will
result in invalid ADC operation. See the Electrical Characteristics chapter for details.
If the TRIGGER, ADP0, and ADP1 pins are to be used in the application, the user must configure the I/O Controller to assign the needed pins to the ADCIFB function.
26.5.2
Power Management
If the CPU enters a sleep mode that disables clocks used by the ADCIFB, the ADCIFB will stop
functioning and resume operation after the system wakes up from sleep mode.
If the Peripheral Event System is configured to send asynchronous peripheral events to the
ADCIFB and the clock used by the ADCIFB is stopped, a local and temporary clock will automatically be requested so the event can be processed. Refer to Section 26.6.13, Section 26.6.12,
and the Peripheral Event System chapter for details.
Before entering a sleep mode where the clock to the ADCIFB is stopped, make sure the Analogto-Digital Converter cell is put in an inactive state. Refer to Section 26.6.13 for more information.
26.5.3
Clocks
The clock for the ADCIFB bus interface (CLK_ADCIFB) is generated by the Power Manager.
This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to
disable the ADCIFB before disabling the clock, to avoid freezing the ADCIFB in an undefined
state.
589
32099I–01/2012
AT32UC3L016/32/64
26.5.4
DMA
The ADCIFB DMA handshake interface is connected to the Peripheral DMA Controller. Using
the ADCIFB DMA functionality requires the Peripheral DMA Controller to be programmed first.
When using the ADCIFB DMA functionality CLK_ADCIFB can not be divided.
26.5.5
Interrupts
The ADCIFB interrupt request line is connected to the interrupt controller. Using the ADCIFB
interrupt request functionality requires the interrupt controller to be programmed first.
26.5.6
Peripheral Events
The ADCIFB peripheral events are connected via the Peripheral Event System. Refer to the
Peripheral Event System chapter for details
26.5.7
26.6
Debug Operation
When an external debugger forces the CPU into debug mode, this module continues normal
operation. If this module is configured in a way that requires it to be periodically serviced by the
CPU through interrupt requests or similar, improper operation or data loss may result during
debugging.
Functional Description
The ADCIFB embeds a Successive Approximation Register (SAR) Analog-to-Digital Converter
(ADC). The ADC supports 8-bit or 10-bit resolution, which can be extended to 11 or 12 bits by
the Enhanced Resolution Mode.
The conversion is performed on a full range between 0 V and the reference voltage pin
ADVREFP. Analog inputs between these voltages converts to digital values (codes) based on a
linear conversion. This linear conversion is described in the expression below where M is the
number of bits used to represent the analog value, Vin is the voltage of the analog value to convert, Vref is the maximum voltage, and Code is the converted digital value.
M
2 ⋅ V in
Code = -------------------V ref
26.6.1
Initializing the ADCIFB
The ADC Interface is enabled by writing a one to the Enable bit in the Control Register (CR.EN).
After the ADC Interface is enabled, the ADC timings needs to be configured by writing the correct values to the RES, PRESCAL, and STARTUP fields in the ADC Configuration Register
(ACR). See Section 26.6.5, and Section 26.6.7 for details. Before the ADCIFB can be used, the
I/O Controller must be configured correctly and the Reference Voltage (ADVREFP) signal must
be connected. Refer to Section 26.5.1 for details.
26.6.2
Basic Operation
To convert analog values to digital values the user must first initialize the ADCIFB as described
in Section 26.6.1. When the ADCIFB is initialized the channels to convert must be enabled by
writing a one the corresponding bits in the Channel Enable Register (CHER). Enabling channel
N instructs the ADCIFB to convert the analog voltage applied to AD pin N at each conversion
590
32099I–01/2012
AT32UC3L016/32/64
sequence. Multiple channels can be enabled resulting in multiple AD pins being converted at
each conversion sequence.
To start converting data the user can either manually start a conversion sequence by writing a
one to the START bit in the Control Register (CR.START) or configure an automatic trigger to
initiate the conversions. The automatic trigger can be configured to trig on many different conditions. Refer to Section 26.8.1 for details.
The result of the conversion is stored in the Last Converted Data Register (LCDR) as they
become available, overwriting the result from the previous conversion. To avoid data loss if more
than one channel is enabled, the user must read the conversion results as they become available either by using an interrupt handler or by using a Peripheral DMA channel to copy the
results to memory. Failing to do so will result in an Overrun Error condition, indicated by the
OVRE bit in the Status Register (SR).
To use an interrupt handler the user must enable the Data Ready (DRDY) interrupt request by
writing a one to the corresponding bit in the Interrupt Enable Register (IER). To clear the interrupt after the conversion result is read, the user must write a one to the corresponding bit in the
Interrupt Clear Register (ICR). See Section 26.6.11 for details.
To use a Peripheral DMA Controller channel the user must configure the Peripheral DMA Controller appropriately. The Peripheral DMA Controller will, when configured, automatically read
converted data as they become available. There is no need to manually clear any bits in the
Interrupt Status Register as this is performed by the hardware. If an Overrun Error condition happens during DMA operation, the OVRE bit in the SR will be set.
26.6.3
ADC Resolution
The Analog-to-Digital Converter cell supports 8-bit or 10-bit resolution, which can be extended to
11-bit and 12-bit with the Enhanced Resolution Mode. The resolution is selected by writing the
selected resolution value to the RES field in the ADC Configuration Register (ACR). See Section
26.9.3.
By writing a zero to the RES field, the ADC switches to the lowest resolution and the conversion
results can be read in the eight lowest significant bits of the Last Converted Data Register
(LCDR). The four highest bits of the Last Converted Data (LDATA) field in the LCDR register
reads as zero. Writing a one to the RES field enables 10-bit resolution, the optimal resolution for
both sampling speed and accuracy. Writing two or three automatically enables Enhanced Resolution Mode with 11-bit or 12-bit resolution, see Section 26.6.4 for details.
When a Peripheral DMA Controller channel is connected to the ADCIFB in 10-bit, 11-bit, or 12bit resolution mode, a transfer size of 16 bits must be used. By writing a zero to the RES field,
the destination buffers can be optimized for 8-bit transfers.
26.6.4
Enhanced Resolution Mode
The Enhanced Resolution Mode is automatically enabled when 11-bit or 12-bit mode is selected
in the ADC Configuration Register (ACR). In this mode the ADCIFB will trade conversion performance for accuracy by averaging multiple samples.
To be able to increase the accuracy by averaging multiple samples it is important that some
noise is present in the input signal. The noise level should be between one and two LSB peakto-peak to get good averaging performance.
591
32099I–01/2012
AT32UC3L016/32/64
The performance cost of enabling 11-bit mode is 4 ADC samples, which reduces the effective
ADC performance by a factor 4. For 12-bit mode this factor is 16. For 12-bit mode the effective
sample rate is maximum ADC sample rate divided by 16.
26.6.5
ADC Clock
The ADCIFB generates an internal clock named CLK_ADC that is used by the Analog-to-Digital
Converter cell to perform conversions. The CLK_ADC frequency is selected by writing to the
PRESCAL field in the ADC Configuration Register (ACR). The CLK_ADC range is between
CLK_ADCIFB/2, if PRESCAL is 0, and CLK_ADCIFB/128, if PRESCAL is 63 (0x3F).
A sensible PRESCAL value must be used in order to provide an ADC clock frequency according
to the maximum sampling rate parameter given in the Electrical Characteristics section. Failing
to do so may result in incorrect Analog-to-Digital Converter operation.
26.6.6
ADC Sleep Mode
The ADC Sleep Mode maximizes power saving by automatically deactivating the Analog-to-Digital Converter cell when it is not being used for conversions. The ADC Sleep Mode is enabled by
writing a one to the SLEEP bit in the ADC Configuration Register (ACR).
When a trigger occurs while the ADC Sleep Mode is enabled, the Analog-to-Digital Converter
cell is automatically activated. As the analog cell requires a startup time, the logic waits during
this time and then starts the conversion of the enabled channels. When conversions of all
enabled channels are complete, the ADC is deactivated until the next trigger.
26.6.7
Startup Time
The Analog-to-Digital Converter cell has a minimal startup time when the cell is activated. This
startup time is given in the Electrical Characteristics chapter and must be written to the
STARTUP field in the ADC Configuration Register (ACR) to get correct conversion results.
The STARTUP field expects the startup time to be represented as the number of CLK_ADC
cycles between 8 and 1024 and in steps of 8 that is needed to cover the ADC startup time as
specified in the Electrical Characteristics chapter.
The Analog-to-Digital Converter cell is activated at the first conversion after reset and remains
active if ACR.SLEEP is zero. If ACR.SLEEP is one, the Analog-to-Digital Converter cell is automatically deactivated when idle and thus each conversion sequence will have a initial startup
time delay.
26.6.8
Sample and Hold Time
A minimal Sample and Hold Time is necessary for the ADCIFB to guarantee the best converted
final value when switching between ADC channels. This time depends on the input impedance
of the analog input, but also on the output impedance of the driver providing the signal to the
analog input, as there is no input buffer amplifier.
The Sample and Hold time has to be programmed through the SHTIM field in the ADC Configuration Register (ACR). This field can define a Sample and Hold time between 1 and 16
CLK_ADC cycles.
26.6.9
ADC Conversion
ADC conversions are performed on all enabled channels when a trigger condition is detected.
For details regarding trigger conditions see Section 26.8.1. The term channel is used to identify
a specific analog input pin so it can be included or excluded in an Analog-to-Digital conversion
592
32099I–01/2012
AT32UC3L016/32/64
sequence and to identify which AD pin was used to convert the current value in the Last Converted Data Register (LCDR). Channel number N corresponding to AD pin number N.
Channels are enabled by writing a one to the corresponding bit in the Channel Enable Register
(CHER), and disabled by writing a one to the corresponding bit in the Channel Disable Register
(CHDR). Active channels are listed in the Channel Status Register (CHSR).
When a conversion sequence is started, all enabled channels will be converted in one sequence
and the result will be placed in the Last Converted Data Register (LCDR) with the channel number used to produce the result. It is important to read out the results while the conversion
sequence is ongoing, as new values will automatically overwrite any old value and the old value
will be lost if not previously read by the user.
If the Analog-to-Digital Converter cell is inactive when starting a conversion sequence, the conversion logic will wait a configurable number of CLK_ADC cycles as defined in the startup time
field in the ADC Configuration Register (ACR). After the cell is activated all enabled channels is
converted one by one until no more enabled channels exist. The conversion sequence converts
each enabled channel in order starting with the channel with the lowest channel number. If the
ACR.SLEEP bit is one, the Analog-to-Digital Converter cell is deactivated after the conversion
sequence has finished.
For each channel converted, the ADCIFB waits a Sample and Hold number of CLK_ADC cycles
as defined in the SHTIM field in ACR, and then instructs the Analog-to-Digital Converter cell to
start converting the analog voltage. The ADC cell requires 10 CLK_ADC cycles to actually convert the value, so the total time to convert a channel is Sample and Hold + 10 CLK_ADC cycles.
26.6.10
Analog Compare Mode
The ADCIFB can test if the converted values, as they become available, are below, above, or
inside a specified range and generate interrupt requests based on this information. This is useful
for applications where the user wants to monitor some external analog signal and only initiate
actions if the value is above, below, or inside some specified range.
The Analog Compare mode is enabled by writing a one to the Analog Compare Enable (ACE) bit
in the Mode Register (MR). The values to compare must be written to the Low Value (LV) field
and the High Value (HV) field in the Compare Value Register (CVR). The Analog Compare
mode will, when enabled, check all enabled channels against the pre-programmed high and low
values and set status bits.
To generate an interrupt request if a converted value is below a limit, write the limit to the
CVR.LV field and enable interrupt request on the Compare Lesser Than (CLT) bit by writing a
one to the corresponding bit in the Interrupt Enable Register (IER). To generate an interrupt
request if a converted value is above a limit, write the limit to the CVR.HV field and enable interrupt for Compare Greater Than (CGT) bit. To generate an interrupt request if a converted value
is inside a range, write the low and high limit to the LV and HV fields and enable the Compare
Else (CELSE) interrupt. To generate an interrupt request if a value is outside a range, write the
LV and HV fields to the low and high limits of the range and enable CGT and CLT interrupts.
Note that the values written to LV and HV must match the resolution selected in the ADC Configuration Register (ACR).
26.6.11
Interrupt Operation
Interrupt requests are enabled by writing a one to the corresponding bit in the Interrupt Enable
Register (IER) and disabled by writing a one to the corresponding bit in the Interrupt Disable
593
32099I–01/2012
AT32UC3L016/32/64
Register (IDR). Enabled interrupts can be read from the Interrupt Mask Register (IMR). Active
interrupt requests, but potentially masked, are visible in the Interrupt Status Register (ISR). To
clear an active interrupt request, write a one to the corresponding bit in the Interrupt Clear Register (ICR).
The source for the interrupt requests are the status bits in the Status Register (SR). The SR
shows the ADCIFB status at the time the register is read. The Interrupt Status Register (ISR)
shows the status since the last write to the Interrupt Clear Register. The combination of ISR and
SR allows the user to react to status change conditions but also allows the user to read the current status at any time.
26.6.12
Peripheral Events
The Peripheral Event System can be used together with the ADCIFB to allow any peripheral
event generator to be used as a trigger source. To enable peripheral events to trigger a conversion sequence the user must write the Peripheral Event Trigger value (0x7) to the Trigger Mode
(TRGMOD) field in the Trigger Register (TRGR). Refer to Table 26-4 on page 606. The user
must also configure a peripheral event generator to emit peripheral events for the ADCIFB to
trigger on. Refer to the Peripheral Event System chapter for details.
26.6.13
Sleep Mode
Before entering sleep modes the user must make sure the ADCIFB is idle and that the Analogto-Digital Converter cell is inactive. To deactivate the Analog-to-Digital Converter cell the SLEEP
bit in the ADC Configuration Register (ACR) must be written to one and the ADCIFB must be
idle. To make sure the ADCIFB is idle, write a zero the Trigger Mode (TRGMOD) field in the
Trigger Register (TRGR) and wait for the READY bit in the Status Register (SR) to be set.
Note that by deactivating the Analog-to-Digital Converter cell, a startup time penalty as defined
in the STARTUP field in the ADC Configuration Register (ACR) will apply on the next
conversion.
26.6.14
26.7
Conversion Performances
For performance and electrical characteristics of the ADCIFB, refer to the Electrical Characteristics chapter.
Resistive Touch Screen
The ADCIFB embeds an integrated Resistive Touch Screen Sequencer that can be used to calculate contact coordinates on a resistive touch screen film. When instructed to start, the
integrated Resistive Touch Screen Sequencer automatically applies a sequence of voltage patterns to the resistive touch screen films and the Analog-to-Digital Conversion cell is used to
measure the effects. The resulting measurements can be used to calculate the horizontal and
vertical contact coordinates. It is recommended to use a high resistance touch screen for optimal
resolution.
The resistive touch screen film is connected to the ADCIFB using the AD and ADP pins. See
Section 26.7.3 for details.
Resistive Touch Screen Mode is enabled by writing a one to the Touch Screen ADC Mode field
in the Mode Register (MR.TSAMOD). In this mode, channels TSPO+0 though TSPO+3 are
automatically enabled where TSPO refers to the Touch Screen Pin Offset field in the Mode Register (MR.TSPO). For each conversion sequence, all enabled channels before TSPO+0 and
after TSPO+3 are converted as ordinary ADC channels, producing 1 conversion result each.
594
32099I–01/2012
AT32UC3L016/32/64
When the sequencer enters the TSPO+0 channel the Resistive Touch Screen Sequencer will
take over control and convert the next 4 channels as described in Section 26.7.4.
26.7.1
Resistive Touch Screen Principles
A resistive touch screen is based on two resistive films, each one fitted with a pair of electrodes,
placed at the top and bottom on one film, and on the right and left on the other. Between the two,
there is a layer that acts as an insulator, but makes a connection when pressure is applied to the
screen. This is illustrated in Figure 26-2 on page 595.
Figure 26-2. Resistive Touch Screen Position Measurement
Pen
Contact
XP
YM
YP
XM
VDD
XP
YP
XP
Volt
XM
Vertical Position Detection
26.7.2
VDD
YP
GND
Volt
YM
GND
Horizontal Position Detection
Position Measurement Method
As shown in Figure 26-2 on page 595, to detect the position of a contact, voltage is first applied
to XP (top) and Xm (bottom) leaving Yp and Ym tristated. Due to the linear resistance of the film,
there is a voltage gradient from top to bottom on the first film. When a contact is performed on
the screen, the voltage at the contact point propagates to the second film. If the input impedance
on the YP (right) and Ym (left) electrodes are high enough, no current will flow, allowing the voltage at the contact point to be measured at Yp. The value measured represents the vertical
position component of the contact point.
For the horizontal direction, the same method is used, but by applying voltage from YP (right) to
Ym (left) and measuring at XP.
In an ideal world (linear, with no loss), the vertical position is equal to:
VYP / VDD
To compensate for some of the real world imperfections, VXP and VXm can be measured and
used to improve accuracy at the cost of two more conversions per axes. The new expression for
the vertical position then becomes:
(VYP - VXM) / (VXP - VXM)
595
32099I–01/2012
AT32UC3L016/32/64
26.7.3
Resistive Touch Screen Pin Connections
Table 26-2.
Resistive Touch Screen Pin Connections
ADCIFB Pin
TS Signal, APOE == 0
TS Signal, APOE == 1
ADP0
Xp through a resistor
No Connect
ADP1
Yp through a resistor
No Connect
ADtspo+0
Xp
Xp
ADtspo+1
Xm
Xm
ADtspo+2
Yp
Yp
ADtspo+3
Ym
Ym
The resistive touch screen film signals connects to the ADCIFB using the AD and ADP pins. The
XP (top) and XM (bottom) film signals are connected to ADtspo+0 and ADtspo+1 pins, and the YP
(right) and YM (left) signals are connected to ADtspo+2 and ADtspo+3 pins. The tspo index is
configurable through the Touch Screen Pin Offset (TSPO) field in the Mode Register (MR) and
allows the user to configure which AD pins to use for resistive touch screen applications. Writing
a zero to the TSPO field instructs the ADCIFB to use AD0 through AD3, where AD0 is connected to XP, AD1 is connected to XM and so on. Writing a one to the TSPO field instructs the
ADCIFB to use AD1 through AD4 for resistive touch screen sequencing, where AD1 is connected to XP and AD0 is free to be used as an ordinary ADC channel.
When the Analog Pin Output Enable (APOE) bit in the Mode Register (MR) is zero, the AD pins
are used to measure input voltage and drive the GND sequences, while the ADP pins are used
to drive the VDD sequences. This arrangement allows the user to reduce the voltage seen at the
AD input pins by inserting external resistors between ADP0 and XP and ADP1 and YP signals
which are again directly connected to the AD pins. It is important that the voltages observed at
the AD pins are not higher than the maximum allowed ADC input voltage. See Figure 26-3 on
page 597 for details regarding how to connect the resistive touch screen films to the AD and
ADP pins.
By adding a resistor between ADP0 and XP, and ADP1 and YP, the maximum voltage observed
at the AD pins can be controlled by the following voltage divider expressions:
R filmx
V ( AD tspo + 0 ) = --------------------------------------------- ⋅ V ( DP 0 )
R filmx + R resistorx
The Rfilmx parameter is the film resistance observed when measuring between XP and XM. The
Rresistorx parameter is the resistor size inserted between ADP0 and XP. The definition of Rfilmy
and Rresistory is the same but for ADP1, YP, and YM instead.
596
32099I–01/2012
AT32UC3L016/32/64
R filmy
V ( AD tspo + 2 ) = --------------------------------------------- ⋅ V ( DP 1 )
R filmy + R resistory
The ADP pins are used by default, as the APOE bit is zero after reset. Writing a one to the
APOE bit instructs the ADCIFB Resistive Touch Screen Sequencer to use the already connected ADtspo+0 and ADtspo+2 pins to drive VDD to XP and YP signals directly. In this mode the
ADP pins can be used as general purpose I/O pins.
Before writing a one to the APOE bit the user must make sure that the I/O voltage is compatible
with the ADC input voltage. If the I/O voltage is higher than the maximum input voltage of the
ADC, permanent damage may occur. Refer to the Electrical Characteristics chapter for details.
Figure 26-3. Resistive Touch Screen Pin Connections
DP0
Analog Pin Output Enable (MR.APOE) == 0
DP1
ADtspo+0
XP
YM
ADtspo+1
YP
ADtspo+2
XM
ADtspo+3
Analog Pin Output Enable (MR.APOE) == 1
NC
DP0
NC
DP1
ADtspo+0
XP
YM
ADtspo+1
YP
ADtspo+2
XM
ADtspo+3
597
32099I–01/2012
AT32UC3L016/32/64
26.7.4
Resistive Touch Screen Sequencer
The Resistive Touch Screen Sequencer is responsible for applying voltage to the resistive touch
screen films as described in Section 26.7.2. This is done by controlling the output enable and the
output value of the ADP and AD pins. This allows the Resistive Touch Screen Sequencer to add
a voltage gradient on one film while keeping the other film floating so a touch can be measured.
The Resistive Touch Screen Sequencer will when measuring the vertical position, apply VDD
and GND to the pins connected to XP and XM. The YP and YM pins are put in tristate mode so the
measurement of YP can proceed without interference. To compensate for ADC offset errors and
non ideal pad drivers, the actual voltage of XP and XM is measured as well, so the real values for
VDD and GND can be used in the contact point calculation to increase accuracy. See second
formula in Section 26.7.2.
When the vertical values are converted the same setup is applies for the second axes, by setting
XP and XM in tristate mode and applying VDD and GND to YP and YM. Refer to Section 26.8.3 for
details.
26.7.5
Pen Detect
If no contact is applied to the resistive touch screen films, any resistive touch screen conversion
result will be undefined as the film being measured is floating. This can be avoided by enabling
Pen Detect and only trigger resistive touch screen conversions when the Pen Contact
(PENCNT) status bit in the Status Register (SR) is one. Pen Detect is enabled by writing a one
to the Pen Detect (PENDET) bit in the Mode Register (MR).
When Pen Detect is enabled, the ADCIFB grounds the vertical panel by applying GND to XP and
XM and polarizes the horizontal panel by enabling pull-up on the pin connected to YP. The YM pin
will in this mode be tristated. Since there is no contact, no current is flowing and there is no
related power consumption. As soon as a contact occurs, GND will propagate to YM by pulling
down YP, allowing the contact to be registered by the ADCIFB.
A programmable debouncing filter can be used to filter out false pen detects because of noise.
The debouncing filter is programmable from one CLK_ADC period and up to 215 CLK_ADC periods. The debouncer length is set by writing to the PENDBC field in MR.
598
32099I–01/2012
AT32UC3L016/32/64
Figure 26-4. Resistive Touch Screen Pen Detect
Resistive
Touch Screen
Sequencer
XP
GND
XM
GND
YP
Pullup
YM
Tristate
T o the ADC
PENDBC
D ebouncer
Pen Int er r upt
The Resistive Touch Screen Pen Detect can be used to generate an ADCIFB interrupt request
or it can be used to trig a conversion, so that a position can be measured as soon as a contact is
detected.
The Pen Detect Mode generates two types of status signals, reported in the Status Register
(SR):
• The bit PENCNT is set when current flows and remains set until current stops.
• The bit NOCNT is set when no current flows and remains set until current flows.
Before a current change is reflected in the SR, the new status must be stable for the duration of
the debouncing time.
Both status conditions can generate an interrupt request if the corresponding bit in the Interrupt
Mask Register (IMR) is one. Refer to Section 26.6.11 on page 593.
599
32099I–01/2012
AT32UC3L016/32/64
26.8
Operating Modes
The ADCIFB features two operating modes, each defining a separate conversion sequence:
• ADC Mode: At each trigger, all the enabled channels are converted.
• Resistive Touch Screen Mode: At each trigger, all enabled channels plus the resistive touch
screen channels are converted as described in Section 26.8.3. If channels except the
dedicated resistive touch screen channels are enabled, they are converted normally before
and after the resistive touch screen channels are converted.
The operating mode is selected by the TSAMOD field in the Mode Register (MR).
26.8.1
Conversion Triggers
A conversion sequence is started either by a software or by a hardware trigger. When a conversion sequence is started, all enabled channels will be converted and made available in the
shared Last Converted Register (LCDR).
The software trigger is asserted by writing a one to the START field in the Control Register (CR).
The hardware trigger can be selected by the TRGMOD field in the Trigger Register (TRGR). Different hardware triggers exist:
• External trigger, either rising or falling or any, detected on the external trigger pin TRIGGER
• Pen detect trigger, depending the PENDET bit in the Mode Register (MR)
• Continuous trigger, meaning the ADCIFB restarts the next sequence as soon as it finishes
the current one
• Periodic trigger, which is defined by the TRGR.TRGPER field
• Peripheral event trigger, allowing the Peripheral Event System to synchronize conversion with
some configured peripheral event source.
Enabling a hardware trigger does not disable the software trigger functionality. Thus, if a hardware trigger is selected, the start of a conversion can still be initiated by the software trigger.
26.8.2
ADC Mode
In the ADC Mode, the active channels are defined by the Channel Status Register (CHSR). A
channel is enabled by writing a one to the corresponding bit in the Channel Enable Register
(CHER), and disabled by writing a one to the corresponding bit in the Channel Disable Register
(CHDR). The conversion results are stored in the Last Converted Data Register (LCDR) as they
become available, overwriting old conversions.
At each trigger, the following sequence is performed:
1. If ACR.SLEEP is one, wake up the ADC and wait for the startup time.
2. If Channel 0 is enabled, convert Channel 0 and store result in LCDR.
3. If Channel 1 is enabled, convert Channel 1 and store result in LCDR.
4. If Channel N is enabled, convert Channel N and store result in LCDR.
5. If ACR.SLEEP is one, place the ADC cell in a low-power state.
If the Peripheral DMA Controller is enabled, all converted values are transferred continuously
into the memory buffer.
26.8.3
Resistive Touch Screen Mode
Writing a one to the TSAMOD field in the Mode Register (MR) enables Resistive Touch Screen
Mode. In this mode the channels TSPO+0 to TSPO+3, corresponding to the resistive touch
600
32099I–01/2012
AT32UC3L016/32/64
screen inputs, are automatically activated. In addition, if any other channels are enabled, they
will be converted before and after the resistive touch screen conversion.
At each trigger, the following sequence is performed:
1. If ACR.SLEEP is one, wake up the ADC cell and wait for the startup time.
2. Convert all enabled channels before TSPO and store the results in the LCDR.
3. Apply supply on the inputs XP and XM during the Sample and Hold Time.
4. Convert Channel XM and store the result in TMP.
5. Apply supply on the inputs XP and XM during the Sample and Hold Time.
6. Convert Channel XP, subtract TMP from the result and store the subtracted result in
LCDR.
7. Apply supply on the inputs XP and XM during the Sample and Hold Time.
8. Convert Channel YP, subtract TMP from the result and store the subtracted result in
LCDR.
9. Apply supply on the inputs YP and YM during the Sample and Hold Time.
10. Convert Channel YM and store the result in TMP.
11. Apply supply on the inputs YP and YM during the Sample and Hold Time.
12. Convert Channel YP, subtract TMP from the result and store the subtracted result in
LCDR.
13. Apply supply on the inputs YP and YM during the Sample and Hold Time.
14. Convert Channel XP, subtract TMP from the result and store the subtracted result in
LCDR.
15. Convert all enabled channels after TSPO + 3 and store results in the LCDR.
16. If ACR.SLEEP is one, place the ADC cell in a low-power state.
The resulting buffer structure stored in memory is:
1. XP - XM
2. YP - XM
3. YP - YM
4. XP - YM.
The vertical position can be easily calculated by dividing the data at offset 1(XP - XM) by the data
at offset 2(YP - XM).
The horizontal position can be easily calculated by dividing the data at offset 3(YP - YM) by the
data at offset 4(XP - YM).
601
32099I–01/2012
AT32UC3L016/32/64
26.9
User Interface
Table 26-3.
Note:
ADCIFB Register Memory Map
Offset
Register
Name
Access
Reset
0x00
Control Register
CR
Write-only
-
0x04
Mode Register
MR
Read/Write
0x00000000
0x08
ADC Configuration Register
ACR
Read/Write
0x00000000
0x0C
Trigger Register
TRGR
Read/Write
0x00000000
0x10
Compare Value Register
CVR
Read/Write
0x00000000
0x14
Status Register
SR
Read-only
0x00000000
0x18
Interrupt Status Register
ISR
Read-only
0x00000000
0x1C
Interrupt Clear Register
ICR
Write-only
-
0x20
Interrupt Enable Register
IER
Write-only
-
0x24
Interrupt Disable Register
IDR
Write-only
-
0x28
Interrupt Mask Register
IMR
Read-only
0x00000000
0x2C
Last Converted Data Register
LCDR
Read-only
0x00000000
0x30
Parameter Register
PARAMETER
Read-only
-(1)
0x34
Version Register
VERSION
Read-only
-(1)
0x40
Channel Enable Register
CHER
Write-only
-
0x44
Channel Disable Register
CHDR
Write-only
-
0x48
Channel Status Register
CHSR
Read-only
0x00000000
1. The reset values for these registers are device specific. Please refer to the Module Configuration section at the end of this
chapter.
602
32099I–01/2012
AT32UC3L016/32/64
26.9.1
Control Register
Register Name:
CR
Access Type:
Write-only
Offset:
0x00
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
-
-
-
-
-
-
DIS
EN
7
6
5
4
3
2
1
0
-
-
-
-
-
-
START
SWRST
• DIS: ADCDIFB Disable
Writing a zero to this bit has no effect.
Writing a one to this bit disables the ADCIFB.
Note: Disabling the ADCIFB effectively stops all clocks in the module so the user must make sure the ADCIFB is idle before
disabling the ADCIFB.
• EN: ADCIFB Enable
Writing a zero to this bit has no effect.
Writing a one to this bit enables the ADCIFB.
Note: The ADCIFB must be enabled before use.
• START: Start Conversion
Writing a zero to this bit has no effect.
Writing a one to this bit starts an Analog-to-Digital conversion.
• SWRST: Software Reset
Writing a zero to this bit has no effect.
Writing a one to this bit resets the ADCIFB, simulating a hardware reset.
603
32099I–01/2012
AT32UC3L016/32/64
26.9.2
Name:
Mode Register
MR
Access Type:
Read/Write
Offset:
0x04
Reset Value:
0x00000000
31
30
29
28
PENDBC
23
22
21
20
27
26
25
24
-
-
-
-
19
18
17
16
TSPO
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
APOE
ACE
PENDET
-
-
-
TSAMOD
• PENDBC: Pen Detect Debouncing Period
Period = 2PENDBC*TCLK_ADC
• TSPO: Touch Screen Pin Offset
The Touch Screen Pin Offset field is used to indicate which AD pins are connected to the resistive touch screen film edges. Only
an offset is specified and it is assumed that the resistive touch screen films are connected sequentially from the specified offset
pin and up to and including offset + 3 (4 pins).
• APOE: Analog Pin Output Enable
0: AD pins are not used to drive VDD in resistive touch screen sequence.
1: AD pins are used to drive VDD in resistive touch screen sequence.
Note: If the selected I/O voltage configuration is incompatible with the Analog-to-Digital converter cell voltage specification, this
bit must stay cleared to avoid damaging the ADC. In this case the ADP pins must be used to drive VDD instead, as described in
Section 26.7.3. If the I/O and ADC voltages are compatible, the AD pins can be used directly by writing a one to this bit. In this
case the ADP pins can be ignored.
• ACE: Analog Compare Enable
0: The analog compare functionality is disabled.
1: The analog compare functionality is enabled.
• PENDET: Pen Detect
0: The pen detect functionality is disabled.
1: The pen detect functionality is enabled.
Note: Touch detection logic can only be enabled when the ADC sequencer is idle. For successful pen detection the user must
make sure there is enough idle time between consecutive scans for the touch detection logic to settle.
• TSAMOD: Touch Screen ADC Mode
0: Touch Screen Mode is disabled.
1: Touch Screen Mode is enabled.
604
32099I–01/2012
AT32UC3L016/32/64
26.9.3
Name:
ADC Configuration Register
ACR
Access Type:
Read/Write
Offset:
0x08
Reset Value:
0x00000000
31
30
29
28
-
-
-
-
23
22
21
20
-
27
26
25
24
18
17
16
10
9
8
3
2
1
0
-
-
-
SLEEP
SHTIM
19
STARTUP
15
14
-
-
7
6
-
-
13
12
11
PRESCAL
5
4
RES
• SHTIM: Sample & Hold Time for ADC Channels
T SAMPLE&HOLD = ( SHTIM + 3 ) ⋅ T CLK_ADC
• STARTUP: Startup Time
TARTUP
= ( STARTUP + 1 ) ⋅ 8 ⋅ T CLK_AD
• PRESCAL: Prescaler Rate Selection
T CLK_ADC = ( PRESCAL + 1 ) ⋅ 2 ⋅ T CLK_ADCIFB
• RES: Resolution Selection
0: 8-bit resolution.
1: 10-bit resolution.
2: 11-bit resolution, interpolated.
3: 12-bit resolution, interpolated.
• SLEEP: ADC Sleep Mode
0: ADC Sleep Mode is disabled.
1: ADC Sleep Mode is enabled.
605
32099I–01/2012
AT32UC3L016/32/64
26.9.4
Name:
Trigger Register
TRGR
Access Type:
Read/Write
Offset:
0x0C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
TRGPER[15:8]
23
22
21
20
TRGPER[7:0]
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
-
TRGMOD
• TRGPER: Trigger Period
Effective only if TRGMOD defines a Periodic Trigger.
Defines the periodic trigger period, with the following equations:
Trigger Period = TRGPER *TCLK_ADC
• TRGMOD: Trigger Mode
Table 26-4.
Trigger Modes
TRGMOD
Selected Trigger Mode
0
0
0
No trigger, only software trigger can start conversions
0
0
1
External Trigger Rising Edge
0
1
0
External Trigger Falling Edge
0
1
1
External Trigger Any Edge
1
0
0
Pen Detect Trigger (shall be selected only if PENDET is set and TSAMOD = Touch
Screen mode)
1
0
1
Periodic Trigger (TRGPER shall be initiated appropriately)
1
1
0
Continuous Mode
1
1
1
Peripheral Event Trigger
606
32099I–01/2012
AT32UC3L016/32/64
26.9.5
Name:
Compare Value Register
CVR
Access Type:
Read/Write
Offset:
0x10
Reset Value:
0x00000000
31
30
29
28
-
-
-
-
23
22
21
20
27
26
25
24
HV[11:8]
19
18
17
16
11
10
9
8
1
0
HV[7:0]
15
14
13
12
-
-
-
-
7
6
5
4
LV[11:8]
3
2
LV[7:0]
• HV: High Value
Defines the high value used when comparing analog input.
• LV: Low Value
Defines the low value used when comparing analog input.
607
32099I–01/2012
AT32UC3L016/32/64
26.9.6
Name:
Status Register
SR
Access Type:
Read-only
Offset:
0x14
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
EN
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
CELSE
CGT
CLT
-
-
BUSY
READY
7
6
5
4
3
2
1
0
-
-
NOCNT
PENCNT
-
-
OVRE
DRDY
• EN: Enable Status
0: The ADCIFB is disabled.
1: The ADCIFB is enabled.
This bit is cleared when CR.DIS is written to one.
This bit is set when CR.EN is written to one.
• CELSE: Compare Else Status
This bit is cleared when either CLT or CGT are detected or when analog compare is disabled.
This bit is set when no CLT or CGT are detected on the last converted data and analog compare is enabled.
• CGT: Compare Greater Than Status
This bit is cleared when no compare greater than CVR.HV is detected on the last converted data or when analog compare is
disabled.
This bit is set when compare greater than CVR.HV is detected on the last converted data and analog compare is enabled.
• CLT: Compare Lesser Than Status
This bit is cleared when no compare lesser than CVR.LV is detected on the last converted data or when analog compare is
disabled.
This bit is set when compare lesser than CVR.LV is detected on the last converted data and analog compare is enabled.
• BUSY: Busy Status
This bit is cleared when the ADCIFB is ready to perform a conversion sequence.
This bit is set when the ADCIFB is busy performing a convention sequence.
• READY: Ready Status
This bit is cleared when the ADCIFB is busy performing a conversion sequence
This bit is set when the ADCIFB is ready to perform a conversion sequence.
• NOCNT: No Contact Status
This bit is cleared when no contact loss is detected or pen detect is disabled
This bit is set when contact loss is detected and pen detect is enabled.
• PENCNT: Pen Contact Status
This bit is cleared when no contact is detected or pen detect is disabled.
608
32099I–01/2012
AT32UC3L016/32/64
This bit is set when pen contact is detected and pen detect is enabled.
• OVRE: Overrun Error Status
This bit is cleared when no Overrun Error has occurred since the start of a conversion sequence.
This bit is set when one or more Overrun Error has occurred since the start of a conversion sequence.
• DRDY: Data Ready Status
0: No data has been converted since the last reset.
1: One or more conversions have completed since the last reset and data is available in LCDR.
This bit is cleared when CR.SWRST is written to one.
This bit is set when one or more conversions have completed and data is available in LCDR.
609
32099I–01/2012
AT32UC3L016/32/64
26.9.7
Name:
Interrupt Status Register
ISR
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
-
CELSE
CGT
CLT
-
-
BUSY
READY
7
6
5
4
3
2
1
0
-
-
NOCNT
PENCNT
-
-
OVRE
DRDY
• CELSE: Compare Else Status
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set when the corresponding bit in SR has a zero-to-one transition.
• CGT: Compare Greater Than Status
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set when the corresponding bit in SR has a zero-to-one transition.
• CLT: Compare Lesser Than Status
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set when the corresponding bit in SR has a zero-to-one transition.
• BUSY: Busy Status
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set when the corresponding bit in SR has a zero-to-one transition.
• READY: Ready Status
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set when the corresponding bit in SR has a zero-to-one transition.
• NOCNT: No Contact Status
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set when the corresponding bit in SR has a zero-to-one transition.
• PENCNT: Pen Contact Status
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set when the corresponding bit in SR has a zero-to-one transition.
• OVRE: Overrun Error Status
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set when the corresponding bit in SR has a zero-to-one transition.
• DRDY: Data Ready Status
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set when a conversion has completed and new data is available in LCDR.
610
32099I–01/2012
AT32UC3L016/32/64
26.9.8
Name:
Interrupt Clear Register
ICR
Access Type:
Write-only
Offset:
0x1C
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
-
CELSE
CGT
CLT
-
-
BUSY
READY
7
6
5
4
3
2
1
0
-
-
NOCNT
PENCNT
-
-
OVRE
DRDY
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in ISR and the corresponding interrupt request.
611
32099I–01/2012
AT32UC3L016/32/64
26.9.9
Name:
Interrupt Enable Register
IER
Access Type:
Write-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
8
-
CELSE
CGT
CLT
-
-
BUSY
READY
7
6
5
4
3
2
1
0
-
-
NOCNT
PENCNT
-
-
OVRE
DRDY
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
612
32099I–01/2012
AT32UC3L016/32/64
26.9.10
Name:
Interrupt Disable Register
IDR
Access Type:
Write-only
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
-
CELSE
CGT
CLT
-
-
BUSY
READY
7
6
5
4
3
2
1
0
-
-
NOCNT
PENCNT
-
-
OVRE
DRDY
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
613
32099I–01/2012
AT32UC3L016/32/64
26.9.11
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
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
-
CELSE
CGT
CLT
-
-
BUSY
READY
7
6
5
4
3
2
1
0
-
-
NOCNT
PENCNT
-
-
OVRE
DRDY
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
A bit in this register is cleared by writing a one to the corresponding bit in Interrupt Disable Register (IDR).
A bit in this register is set by writing a one to the corresponding bit in Interrupt Enable Register (IER).
614
32099I–01/2012
AT32UC3L016/32/64
26.9.12
Name:
Last Converted Data Register
LCDR
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
11
10
9
8
1
0
LCCH
15
14
13
12
-
-
-
-
7
6
5
4
LDATA[11:8]
3
2
LDATA[7:0]
• LCCH: Last Converted Channel
This field indicates what channel was last converted, i.e. what channel the LDATA represents.
• LDATA: Last Data Converted
The analog-to-digital conversion data is placed in this register at the end of a conversion on any analog channel and remains
until a new conversion on any analog channel is completed.
615
32099I–01/2012
AT32UC3L016/32/64
26.9.13
Name:
Parameter Register
PARAMETER
Access Type:
Read-only
Offset:
0x30
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
CH31
CH30
CH29
CH28
CH27
CH26
CH25
CH24
23
22
21
20
19
18
17
16
CH23
CH22
CH21
CH20
CH19
CH18
CH17
CH16
15
14
13
12
11
10
9
8
CH15
CH14
CH13
CH12
CH11
CH10
CH9
CH8
7
6
5
4
3
2
1
0
CH7
CH6
CH5
CH4
CH3
CH2
CH1
CH0
• CHn: Channel n Implemented
0: The corresponding channel is not implemented.
1: The corresponding channel is implemented.
616
32099I–01/2012
AT32UC3L016/32/64
26.9.14
Name:
Version Register
VERSION
Access Type:
Read-only
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
9
8
-
-
-
-
7
6
5
4
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the Module. No functionality associated.
617
32099I–01/2012
AT32UC3L016/32/64
26.9.15
Name:
Channel Enable Register
CHER
Access Type:
Write-only
Offset:
0x40
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
CH31
CH30
CH29
CH28
CH27
CH26
CH25
CH24
23
22
21
20
19
18
17
16
CH23
CH22
CH21
CH20
CH19
CH18
CH17
CH16
15
14
13
12
11
10
9
8
CH15
CH14
CH13
CH12
CH11
CH10
CH9
CH8
7
6
5
4
3
2
1
0
CH7
CH6
CH5
CH4
CH3
CH2
CH1
CH0
• CHn: Channel n Enable
Writing a zero to a bit in this register has no effect
Writing a one to a bit in this register enables the corresponding channel
The number of available channels is device dependent. Please refer to the Module Configuration section at the end of this
chapter for information regarding which channels are implemented.
618
32099I–01/2012
AT32UC3L016/32/64
26.9.16
Name:
Channel Disable Register
CHDR
Access Type:
Write-only
Offset:
0x44
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
CH31
CH30
CH29
CH28
CH27
CH26
CH25
CH24
23
22
21
20
19
18
17
16
CH23
CH22
CH21
CH20
CH19
CH18
CH17
CH16
15
14
13
12
11
10
9
8
CH15
CH14
CH13
CH12
CH11
CH10
CH9
CH8
7
6
5
4
3
2
1
0
CH7
CH6
CH5
CH4
CH3
CH2
CH1
CH0
• CHn: Channel N Disable
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register disables the corresponding channel.
Warning: If the corresponding channel is disabled during a conversion, or if it is disabled and then re-enabled during a
conversion, its associated data and its corresponding DRDY and OVRE bits in SR are unpredictable.
The number of available channels is device dependent. Please refer to the Module Configuration section at the end of this
chapter for information regarding how many channels are implemented.
619
32099I–01/2012
AT32UC3L016/32/64
26.9.17
Name:
Channel Status Register
CHSR
Access Type:
Read-only
Offset:
0x48
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
CH31
CH30
CH29
CH28
CH27
CH26
CH25
CH24
23
22
21
20
19
18
17
16
CH23
CH22
CH21
CH20
CH19
CH18
CH17
CH16
15
14
13
12
11
10
9
8
CH15
CH14
CH13
CH12
CH11
CH10
CH9
CH8
7
6
5
4
3
2
1
0
CH7
CH6
CH5
CH4
CH3
CH2
CH1
CH0
• CHn: Channel N Status
0: The corresponding channel is disabled.
1: The corresponding channel is enabled.
A bit in this register is cleared by writing a one to the corresponding bit in Channel Disable Register (CHDR).
A bit in this register is set by writing a one to the corresponding bit in Channel Enable Register (CHER).
The number of available channels is device dependent. Please refer to the Module Configuration section at the end of this
chapter for information regarding how many channels are implemented.
620
32099I–01/2012
AT32UC3L016/32/64
26.10 Module Configuration
The specific configuration for each ADCIFB instance is listed in the following tables.The module
bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details.
Table 26-5.
Module Configuration
Feature
ADCIFB
Number of ADC channels
9 (8 + 1 internal temperature sensor channel)
Table 26-6.
ADCIFB Clocks
Clock Name
Description
CLK_ADCIFB
Clock for the ADCIFB bus interface
Table 26-7.
Register Reset Values
Register
Reset Value
VERSION
0x00000101
PARAMETER
0x000003FF
Table 26-8.
ADC Input Channels
Channel
Input
CH0
AD0
CH1
AD1
CH2
AD2
CH4
AD4
CH5
AD5
CH6
AD6
CH7
AD7
CH8
AD8
CH9
Temperature sensor
621
32099I–01/2012
AT32UC3L016/32/64
27. Analog Comparator Interface (ACIFB)
Rev: 2.0.2.2
27.1
Features
• Controls an array of Analog Comparators
• Low power option
•
•
•
•
27.2
– Single shot mode support
Selectable settings for filter option
– Filter length and hysteresis
Window Mode
– Detect inside/outside window
– Detect above/below window
Interrupt
– On comparator result rising edge, falling edge, toggle
– Inside window, outside window, toggle
– When startup time has passed
Can generate events to the peripheral event system
Overview
The Analog Comparator Interface (ACIFB) is able to control a number of Analog Comparators
(AC) with identical behavior. An Analog Comparator compares two voltages and gives a compare output depending on this comparison.
The ACIFB can be configured in normal mode using each comparator independently or in window mode using defined comparator pairs to observe a window.
The number of channels implemented is device specific. Refer to the Module Configuration section at the end of this chapter for details.
622
32099I–01/2012
AT32UC3L016/32/64
27.3
Block Diagram
Figure 27-1. ACIFB Block Diagram
Analog
Comparators
ACP0
ACIFB
TR.ACTEST0
CLK_ACIFB
INP
+
INN
ACOUT0
FILTER
AC
ACN0
GCLK
INTERRUPT
GENERATION
CONF0.INSELN
……………...
ACREFN
ACPn
Peripheral Bus
INP
IRQ
PERIPHERAL
EVENT
GENERATION
CTRL.ACTEST
+
ACOUTn
INN
EVENTS
FILTER
AC
ACNn
TRIGGER
CONFn.INSELN
TR.ACTESTn
27.4
I/O Lines Description
There are two groups of analog comparators, A and B, as shown in Table 27-1. In normal mode,
this grouping does not have any meaning. In window mode, two analog comparators, one from
group A and the corresponding comparator from group B, are paired.
Table 27-1.
Analog Comparator Groups for Window Mode
Group A
Group B
Pair Number
AC0
AC1
0
AC2
AC3
1
AC4
AC5
2
AC6
AC7
3
Table 27-2.
I/O Line Description
Pin Name
Pin Description
Type
ACAPn
Positive reference pin for Analog Comparator A n
Analog
ACANn
Negative reference pin for Analog Comparator A n
Analog
623
32099I–01/2012
AT32UC3L016/32/64
Table 27-2.
I/O Line Description
Pin Name
Pin Description
Type
ACBPn
Positive reference pin for Analog Comparator B n
Analog
ACBNn
Negative reference pin for Analog Comparator B n
Analog
ACREFN
Reference Voltage for all comparators selectable for INN
Analog
The signal names corresponds to the groups A and B of analog comparators. For normal mode,
the mapping from input signal names in the block diagram to the signal names is given in Table
27-3.
Table 27-3.
27.5
Signal Name Mapping
Pin Name
Channel Number
Normal Mode
ACAP0/ACAN0
0
ACP0/ACN0
ACBP0/ACBN0
1
ACP1/ACN1
ACAP1/ACAN1
2
ACP2/ACN2
ACBP1/ACBN1
3
ACP3/ACN3
ACAP2/ACAN2
4
ACP4/ACN4
ACBP2/ACBN2
5
ACP5/ACN5
ACAP3/ACAN3
6
ACP6/ACN6
ACBP3/ACBN3
7
ACP7/ACN7
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
27.5.1
I/O Lines
The ACIFB pins are multiplexed with other peripherals. The user must first program the I/O Controller to give control of the pins to the ACIFB.
27.5.2
Power Management
If the CPU enters a sleep mode that disables clocks used by the ACIFB, the ACIFB will stop
functioning and resume operation after the system wakes up from sleep mode.
27.5.3
Clocks
The clock for the ACIFB bus interface (CLK_ACIFB) is generated by the Power Manager. This
clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the ACIFB before disabling the clock, to avoid freezing the ACIFB in an undefined state.
The ACIFB uses a GCLK as clock source for the Analog Comparators. The user must set up this
GCLK at the right frequency. The CLK_ACIFB clock of the interface must be at least 4x the
GCLK frequency used in the comparators. The GCLK is used both for measuring the startup
time of a comparator, and to give a frequency for the comparisons done in Continuous Measurement Mode, see Section 27.6.
Refer to the Electrical Characteristics chapter for GCLK frequency limitations.
624
32099I–01/2012
AT32UC3L016/32/64
27.5.4
Interrupts
The ACIFB interrupt request line is connected to the interrupt controller. Using the ACIFB interrupt requires the interrupt controller to be programmed first.
27.5.5
Peripheral Events
The ACIFB peripheral events are connected via the Peripheral Event System. Refer to the
Peripheral Event System chapter for details.
27.5.6
Debug Operation
When an external debugger forces the CPU into debug mode, the ACIFB continues normal
operation. If the ACIFB is configured in a way that requires it to be periodically serviced by the
CPU through interrupts or similar, improper operation or data loss may result during debugging.
27.6
Functional Description
The ACIFB is enabled by writing a one to the Control Register Enable bit (CTRL.EN). Additionally, the comparators must be individually enabled by programming the MODE field in the AC
Configuration Register (CONFn.MODE).
The results from the individual comparators can either be used directly (normal mode), or the
results from two comparators can be grouped to generate a comparison window (window mode).
All comparators need not be in the same mode, some comparators may be in normal mode,
while others are in window mode. There are restrictions on which AC channels that can be
grouped together in a window pair, see Section 27.6.5.
27.6.1
Analog Comparator Operation
Each AC channel can be in one of four different modes, determined by CONFn.MODE:
• Off
• Continuous Measurement Mode (CM)
• User Triggered Single Measurement Mode (UT)
• Event Triggered Single Measurement Mode (ET)
After being enabled, a startup time defined in CTRL.SUT is required before the result of the
comparison is ready. The GCLK is used for measuring the startup time of a comparator,
During the startup time the AC output is not available. When the ACn Ready bit in the Status
Register (SR.ACRDYn) is one, the output of ACn is ready. In window mode the result is available when both the comparator outputs are ready (SR.ACRDYn=1 and SR.ACRDYn+1=1).
27.6.1.1
Continuous Measurement Mode
In CM, the Analog Comparator is continuously enabled and performing comparisons. This
ensures that the result of the latest comparison is always available in the ACn Current Comparison Status bit in the Status Register (SR.ACCSn). Comparisons are done on every positive
edge of GCLK.
CM is enabled by writing CONFn.MODE to 1. After the startup time has passed, a comparison is
done and SR is updated. Appropriate peripheral events and interrupts are also generated. New
comparisons are performed continuously until the CONFn.MODE field is written to 0.
625
32099I–01/2012
AT32UC3L016/32/64
27.6.1.2
User Triggered Single Measurement Mode
In the UT mode, the user starts a single comparison by writing a one to the User Start Single
Comparison bit (CTRL.USTART). This mode is enabled by writing CONFn.MODE to 2. After the
startup time has passed, a single comparison is done and SR is updated. Appropriate peripheral
events and interrupts are also generated. No new comparisons will be performed.
CTRL.USTART is cleared automatically by hardware when the single comparison has been
done.
27.6.1.3
Event Triggered Single Measurement Mode
This mode is enabled by writing CONFn.MODE to 3 and Peripheral Event Trigger Enable
(CTRL.EVENTEN) to one. The ET mode is similar to the UT mode, the difference is that a
peripheral event from another hardware module causes the hardware to automatically set the
Peripheral Event Start Single Comparison bit (CTRL.ESTART). After the startup time has
passed, a single comparison is done and SR is updated. Appropriate peripheral events and
interrupts are also generated. No new comparisons will be performed. CTRL.ESTART is cleared
automatically by hardware when the single comparison has been done.
27.6.1.4
Selecting Comparator Inputs
Each Analog Comparator has one positive (INP) and one negative (INN) input. The positive
input is fed from an external input pin (ACPn). The negative input can either be fed from an
external input pin (ACNn) or from a reference voltage common to all ACs (ACREFN).
The user selects the input source as follows:
• In normal mode with the Negative Input Select and Positive Input Select fields
(CONFn.INSELN and CONFn.INSELP).
• In window mode with CONFn.INSELN, CONFn.INSELP and CONFn+1.INSELN,
CONFn+1,INSELP. The user must configure CONFn.INSELN and CONFn+1.INSELP to the
same source.
27.6.2
Interrupt Generation
The interrupt request will be generated if the corresponding bit in the Interrupt Mask Register
(IMR) is set. Bits in IMR are set by writing a one to the corresponding bit in the Interrupt Enable
Register (IER), and cleared by writing a one to the corresponding bit in the Interrupt Disable
Register (IDR). The interrupt request remains active until the corresponding bit in ISR is cleared
by writing a one to the corresponding bit in the Interrupt Status Clear Register (ICR).
27.6.3
Peripheral Event Generation
The ACIFB can be set up so that certain comparison results notify other parts of the device via
the Peripheral Event system. Refer to Section 27.6.4.3 and Section 27.6.5.3 for information on
which comparison results can generate events, and how to configure the ACIFB to achieve this.
Zero or one event will be generated per comparison.
27.6.4
Normal Mode
In normal mode all Analog Comparators are operating independently.
27.6.4.1
Normal Mode Output
Each Analog Comparator generates one output ACOUT according to the input voltages on INP
(AC positive input) and INN (AC negative input):
626
32099I–01/2012
AT32UC3L016/32/64
• ACOUT = 1 if VINP > VINN
• ACOUT = 0 if VINP < VINN
• ACOUT = 0 if the AC output is not available (SR.ACRDY = 0)
The output can optionally be filtered, as described in Section 27.6.6.
27.6.4.2
Normal Mode Interrupt
The AC channels can generate interrupts. The Interrupt Settings field in the Configuration Register (CONFn.IS) can be configured to select when the AC will generate an interrupt:
• When VINP > VINN
• When VINP < VINN
• On toggle of the AC output (ACOUT)
• When comparison has been done
27.6.4.3
Normal Mode Peripheral Events
The ACIFB can generate peripheral events according to the configuration of CONFn.EVENN
and CONFn.EVENP.
• When VINP > VINN or
• When VINP < VINN or
• On toggle of the AC output (ACOUT)
27.6.5
Window Mode
In window mode, two ACs (an even and the following odd build up a pair) are grouped.
The negative input of ACn (even) and the positive input of ACn+1 (odd) has to be connected
together externally to the device and are controlled by the Input Select fields in the AC Configuration Registers (CONFn.INSELN and CONFn+1.INSELP). The positive input of ACn (even) and
the negative input of ACn+1 (odd) can still be configured independently by CONFn.INSELP and
CONFn+1.INSELN, respectively.
627
32099I–01/2012
AT32UC3L016/32/64
Figure 27-2. Analog Comparator Interface in Window Mode
ACAP0
+
ACAN0
AC0
Filter
-
Interrupt
Generator
ACOUT0
ACWOUT
COMMON
Comparator pair 0
Window
Module
window event
SR.WFCS0
-
Filter
SR.ACCS0
27.6.5.1
AC1
Peripheral Event
Generator
Window
ACOUT1
+
ACBP0
ACBN0
IRQ
Window Mode Output
When operating in window mode, each channel generates the same ACOUT outputs as in normal mode, see Section 27.6.4.1.
Additionally, the ACIFB generates a window mode signal (acwout) according to the common
input voltage to be compared:
• ACWOUT = 1 if the common input voltage is inside the window, VACN(N+1) < Vcommon < VACP(N)
• ACWOUT = 0 if the common input voltage is outside the window, Vcommon < VACN(N+1) or
Vcommon > VACP(N)
• ACWOUT = 0 if the window mode output is not available (SR.ACRDYn=0 or
SR.ACRDYn+1=0)
27.6.5.2
Window Mode Interrupts
When operating in window mode, each channel can generate the same interrupts as in normal
mode, see Section 27.6.4.2.
Additionally, when channels operate in window mode, programming Window Mode Interrupt Settings in the Window Mode Configuration Register (CONFWn.WIS) can cause interrupts to be
generated when:
• As soon as the common input voltage is inside the window.
• As soon as the common input voltage is outside the window.
• On toggle of the window compare output (ACWOUT).
• When the comparison in both channels in the window pair is ready.
628
32099I–01/2012
AT32UC3L016/32/64
27.6.5.3
Window Mode Peripheral Events
When operating in window mode, each channel can generate the same peripheral events as in
normal mode, see Section 27.6.4.3.
Additionally, when channels operate in window mode, programming Window Mode Event Selection Source (CONFWn.WEVSRC) can cause peripheral events to be generated when:
• As soon as the common input voltage is inside the window.
• As soon as the common input voltage is outside the window.
• On toggle of the window compare output (ACWOUT)
• Whenever a comparison is ready and the common input voltage is inside the window.
• Whenever a comparison is ready and the common input voltage is outside the window.
• When the comparison in both channels in the window pair is ready.
27.6.6
Filtering
The output of the comparator can be filtered to reduce noise. The filter length is determined by
the Filter Length field in the CONFn register (CONFn.FLEN). The filter samples the Analog
Comparator output at the GCLK frequency for 2CONFn.FLEN samples. A separate counter (CNT)
counts the number of cycles the AC output was one. This filter is deactivated if CONFn.FLEN
equals 0.
If the filter is enabled, the Hysteresis Value field HYS in the CONFn register (CONFn.HYS) can
be used to define a hysteresis value. The hysteresis value should be chosen so that:
2 FLEN
---------------- > HYS
2
The filter function is defined by:
2 FLEN
CNT ≥ ⎛ ---------------- + HYS⎞ ⇒ comp = 1
⎝ 2
⎠
FLEN
FLEN
2
2
⎛ --------------- + HYS⎞ > CNT ≥ ⎛ ---------------- – H YS⎞ ⇒ comp unchanged
⎝ 2
⎠
⎝ 2
⎠
2 FLEN
CNT < ⎛ ---------------- – H YS⎞ ⇒ comp = 0
⎝ 2
⎠
The filtering algorithm is explained in Figure 27-3. 2FLEN measurements are sampled. If the number of measurements that are zero is less than (2FLEN/2 - HYS), the filtered result is zero. If the
number of measurements that are one is more than (2FLEN/2 + HYS), the filtered result is one.
Otherwise, the result is unchanged.
629
32099I–01/2012
AT32UC3L016/32/64
Figure 27-3. The Filtering Algorithm
2
FLEN
2
0
”Result=0"
Result =
UNCHANGED
HYS
27.7
”Result=1"
2
FLEN
HYS
Peripheral Event Triggers
Peripheral events from other modules can trigger comparisons in the ACIFB. All channels that
are set up in Event Triggered Single Measurement Mode will be started simultaneously when a
peripheral event is received. Channels that are operating in Continuous Measurement Mode or
User Triggered Single Measurement Mode will be unaffected by the received event. The software can still operate these channels independently of channels in Event Triggered Single
Measurement Mode.
A peripheral event will trigger one or more comparisons, in normal or window mode.
27.8
AC Test mode
By writing the Analog Comparator Test Mode (CR.ACTEST) bit to one, the outputs from the ACs
are overridden by the value in the Test Register (TR), see Figure 27-1. This is useful for software
development.
630
32099I–01/2012
AT32UC3L016/32/64
27.9
User Interface
Table 27-4.
ACIFB Register Memory Map
Offset
Register
Register Name
Access
Reset
0x00
Control Register
CTRL
Read/Write
0x00000000
0x04
Status Register
SR
Read-only
0x00000000
0x10
Interrupt Enable Register
IER
Write-only
0x00000000
0x14
Interrupt Disable Register
IDR
Write-only
0x00000000
0x18
Interrupt Mask Register
IMR
Read-only
0x00000000
0x1C
Interrupt Status Register
ISR
Read-only
0x00000000
0x20
Interrupt Status Clear Register
ICR
Write-only
0x00000000
0x24
Test Register
TR
Read/Write
0x00000000
0x30
Parameter Register
PARAMETER
Read-only
-(1)
0x34
Version Register
VERSION
Read-only
-(1)
0x80
Window0 Configuration Register
CONFW0
Read/Write
0x00000000
0x84
Window1 Configuration Register
CONFW1
Read/Write
0x00000000
0x88
Window2 Configuration Register
CONFW2
Read/Write
0x00000000
0x8C
Window3 Configuration Register
CONFW3
Read/Write
0x00000000
0xD0
AC0 Configuration Register
CONF0
Read/Write
0x00000000
0xD4
AC1 Configuration Register
CONF1
Read/Write
0x00000000
0xD8
AC2 Configuration Register
CONF2
Read/Write
0x00000000
0xDC
AC3 Configuration Register
CONF3
Read/Write
0x00000000
0xE0
AC4 Configuration Register
CONF4
Read/Write
0x00000000
0xE4
AC5 Configuration Register
CONF5
Read/Write
0x00000000
0xE8
AC6 Configuration Register
CONF6
Read/Write
0x00000000
0xEC
AC7 Configuration Register
CONF7
Read/Write
0x00000000
Note:
1. The reset values for these registers are device specific. Please refer to the Module Configuration section at the end of this
chapter.
631
32099I–01/2012
AT32UC3L016/32/64
27.9.1
Name:
Control Register
CTRL
Access Type:
Read/Write
Offset:
0x00
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
SUT[9:8]
SUT[7:0]
7
6
5
4
3
2
1
0
ACTEST
-
ESTART
USTART
-
-
-EVENTEN
EN
• SUT: Startup Time
SUT
Analog Comparator startup time = ----------------- .
F GCLK
Each time an AC is enabled, the AC comparison will be enabled after the startup time of the AC.
• ACTEST: Analog Comparator Test Mode
0: The Analog Comparator outputs feeds the channel logic in ACIFB.
1: The Analog Comparator outputs are bypassed with the AC Test Register.
• ESTART: Peripheral Event Start Single Comparison
Writing a zero to this bit has no effect.
Writing a one to this bit starts a comparison and can be used for test purposes.
This bit is cleared when comparison is done.
This bit is set when an enabled peripheral event is received.
• USTART: User Start Single Comparison
Writing a zero to this bit has no effect.
Writing a one to this bit starts a Single Measurement Mode comparison.
This bit is cleared when comparison is done.
• EVENTEN: Peripheral Event Trigger Enable
0: A peripheral event will not trigger a comparison.
1: Enable comparison triggered by a peripheral event.
• EN: ACIFB Enable
0: The ACIFB is disabled.
1: The ACIFB is enabled.
632
32099I–01/2012
AT32UC3L016/32/64
27.9.2
Name:
Status Register
SR
Access Type:
Read-only
Offset:
0x04
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
WFCS3
WFCS2
WFCS1
WFCS0
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
ACRDY7
ACCS7
ACRDY6
ACCS6
ACRDY5
ACCS5
ACRDY4
ACCS4
7
6
5
4
3
2
1
0
ACRDY3
ACCS3
ACRDY2
ACCS2
ACRDY1
ACCS1
ACRDY0
ACCS0
• WFCSn: Window Mode Current Status
This bit is cleared when the common input voltage is outside the window.
This bit is set when the common input voltage is inside the window.
• ACRDYn: ACn Ready
This bit is cleared when the AC output (ACOUT) is not ready.
This bit is set when the AC output (ACOUT) is ready, AC is enabled and its startup time is over.
• ACCSn: ACn Current Comparison Status
This bit is cleared when VINP is currently lower than VINN
This bit is set when VINP is currently greater than VINN.
633
32099I–01/2012
AT32UC3L016/32/64
27.9.3
Name:
Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x10
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
WFINT3
WFINT2
WFINT1
WFINT0
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
SUTINT7
ACINT7
SUTINT6
ACINT6
SUTINT5
ACINT5
SUTINT4
ACINT4
7
6
5
4
3
2
1
0
SUTINT3
ACINT3
SUTINT2
ACINT2
SUTINT1
ACINT1
SUTINT0
ACINT0
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
634
32099I–01/2012
AT32UC3L016/32/64
27.9.4
Name:
Interrupt Disable Register
IDR
Access Type:
Write-only
Offset:
0x14
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
WFINT3
WFINT2
WFINT1
WFINT0
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
SUTINT7
ACINT7
SUTINT6
ACINT6
SUTINT5
ACINT5
SUTINT4
ACINT4
7
6
5
4
3
2
1
0
SUTINT3
ACINT3
SUTINT2
ACINT2
SUTINT1
ACINT1
SUTINT0
ACINT0
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
635
32099I–01/2012
AT32UC3L016/32/64
27.9.5
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
Offset:
0x18
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
WFINT3
WFINT2
WFINT1
WFINT0
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
SUTINT7
ACINT7
SUTINT6
ACINT6
SUTINT5
ACINT5
SUTINT4
ACINT4
7
6
5
4
3
2
1
0
SUTINT3
ACINT3
SUTINT2
ACINT2
SUTINT1
ACINT1
SUTINT0
ACINT0
• WFINTn: Window Mode Interrupt Mask
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
This bit is cleared when the corresponding bit in IDR is written to one.
This bit is set when the corresponding bit in IER is written to one.
• SUTINTn: ACn Startup Time Interrupt Mask
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
This bit is cleared when the corresponding bit in IDR is written to one.
This bit is set when the corresponding bit in IER is written to one.
• ACINTn: ACn Interrupt Mask
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
This bit is cleared when the corresponding bit in IDR is written to one.
This bit is set when the corresponding bit in IER is written to one.
636
32099I–01/2012
AT32UC3L016/32/64
27.9.6
Name:
Interrupt Status Register
ISR
Access Type:
Read-only
Offset:
0x1C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
WFINT3
WFINT2
WFINT1
WFINT0
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
SUTINT7
ACINT7
SUTINT6
ACINT6
SUTINT5
ACINT5
SUTINT4
ACINT4
7
6
5
4
3
2
1
0
SUTINT3
ACINT3
SUTINT2
ACINT2
SUTINT1
ACINT1
SUTINT0
ACINT0
• WFINTn: Window Mode Interrupt Status
0: No Window Mode Interrupt is pending.
1: Window Mode Interrupt is pending.
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set when the corresponding channel pair operating in window mode generated an interrupt.
• SUTINTn: ACn Startup Time Interrupt Status
0: No Startup Time Interrupt is pending.
1: Startup Time Interrupt is pending.
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set when the startup time of the corresponding AC has passed.
• ACINTn: ACn Interrupt Status
0: No Normal Mode Interrupt is pending.
1: Normal Mode Interrupt is pending.
This bit is cleared when the corresponding bit in ICR is written to one.
This bit is set when the corresponding channel generated an interrupt.
637
32099I–01/2012
AT32UC3L016/32/64
27.9.7
Name:
Interrupt Status Clear Register
ICR
Access Type:
Write-only
Offset:
0x20
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
WFINT3
WFINT2
WFINT1
WFINT0
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
SUTINT7
ACINT7
SUTINT6
ACINT6
SUTINT5
ACINT5
SUTINT4
ACINT4
7
6
5
4
3
2
1
0
SUTINT3
ACINT3
SUTINT2
ACINT2
SUTINT1
ACINT1
SUTINT0
ACINT0
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in ISR and the corresponding interrupt request.
638
32099I–01/2012
AT32UC3L016/32/64
27.9.8
Name:
Test Register
TR
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
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
ACTEST7
ACTEST6
ACTEST5
ACTEST4
ACTEST3
ACTEST2
ACTEST1
ACTEST0
• ACTESTn: AC Output Override Value
If CTRL.ACTEST is set, the ACn output is overridden with the value of ACTESTn.
639
32099I–01/2012
AT32UC3L016/32/64
27.9.9
Name:
Parameter Register
PARAMETER
Access Type:
Read-only
Offset:
0x30
Reset Value:
-
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
WIMPL3
WIMPL2
WIMPL1
WIMPL0
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
ACIMPL7
ACIMPL6
ACIMPL5
ACIMPL4
ACIMPL3
ACIMPL2
ACIMPL1
ACIMPL0
• WIMPLn: Window Pair n Implemented
0: Window Pair not implemented.
1: Window Pair implemented.
• ACIMPLn: Analog Comparator n Implemented
0: Analog Comparator not implemented.
1: Analog Comparator implemented.
640
32099I–01/2012
AT32UC3L016/32/64
27.9.10
Name:
Version Register
VERSION
Access Type:
Read-only
Offset:
0x34
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the module. No functionality associated.
641
32099I–01/2012
AT32UC3L016/32/64
27.9.11
Name:
Window Configuration Register
CONFWn
Access Type:
Read/Write
Offset:
0x80,0x84,0x88,0x8C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
-
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
WFEN
15
14
13
12
11
10
9
8
-
-
-
-
WEVEN
7
6
5
4
3
2
-
-
-
-
-
-
WEVSRC
0
1
WIS
• WFEN: Window Mode Enable
0: The window mode is disabled.
1: The window mode is enabled.
• WEVEN: Window Event Enable
0: Event from awout is disabled.
1: Event from awout is enabled.
• WEVSRC: Event Source Selection for Window Mode
000: Event on acwout rising edge.
001: Event on acwout falling edge.
010: Event on awout rising or falling edge.
011: Inside window.
100: Outside window.
101: Measure done.
110-111: Reserved.
• WIS: Window Mode Interrupt Settings
00: Window interrupt as soon as the input voltage is inside the window.
01: Window interrupt as soon as the input voltage is outside the window.
10: Window interrupt on toggle of window compare output.
11: Window interrupt when evaluation of input voltage is done.
642
32099I–01/2012
AT32UC3L016/32/64
27.9.12
Name:
AC Configuration Register
CONFn
Access Type:
Read/Write
Offset:
0xD0,0xD4,0xD8,0xDC,0xE0,0xE4,0xE8,0xEC
Reset Value:
0x00000000
31
30
-
29
28
27
26
FLEN
25
24
HYS
23
22
21
20
19
18
17
16
-
-
-
-
-
-
EVENP
EVENN
15
14
13
12
11
10
9
8
-
-
-
-
7
6
5
4
-
-
MODE
INSELP
INSELN
3
2
-
-
0
1
IS
• FLEN: Filter Length
000: Filter off.
n: Number of samples to be averaged =2n.
• HYS: Hysteresis Value
0000: No hysteresis.
1111: Max hysteresis.
• EVENN: Event Enable Negative
0: Do not output event when ACOUT is zero.
1: Output event when ACOUT is zero.
• EVENP: Event Enable Positive
0: Do not output event when ACOUT is one.
1: Output event when ACOUT is one.
• INSELP: Positive Input Select
00: ACPn pin selected.
01: Reserved.
10: Reserved.
11: Reserved.
• INSELN: Negative Input Select
00: ACNn pin selected.
01: ACREFN pin selected.
10: Reserved.
11: Reserved.
• MODE: Mode
00: Off.
01: Continuous Measurement Mode.
10: User Triggered Single Measurement Mode.
11: Event Triggered Single Measurement Mode.
643
32099I–01/2012
AT32UC3L016/32/64
• IS: Interrupt Settings
00: Comparator interrupt when as VINP > VINN.
01: Comparator interrupt when as VINP < VINN.
10: Comparator interrupt on toggle of Analog Comparator output.
11: Comparator interrupt when comparison of VINP and VINN is done.
644
32099I–01/2012
AT32UC3L016/32/64
27.10 Module Configuration
The specific configuration for each ACIFB instance is listed in the following tables.The module
bus clocks listed here are connected to the system bus clocks. Refer to the Power Manager
chapter for details.
Table 27-5.
ACIFB Configuration
Feature
ACIFB
Number of channels
8
Table 27-6.
ACIFB Clocks
Clock Name
Description
CLK_ACIFB
Clock for the ACIFB bus interface
GCLK
The generic clock used for the ACIFB is GCLK4
Table 27-7.
Register Reset Values
Register
Reset Value
VERSION
0x00000202
PARAMETER
0x000F00FF
645
32099I–01/2012
AT32UC3L016/32/64
28. Capacitive Touch Module (CAT)
Rev: 2.0.0.1
28.1
Features
•
•
•
•
•
•
28.2
QTouch® method allows N touch sensors to be implemented using 2N physical pins
QMatrix method allows X by Y matrix of sensors to be implemented using (X+2Y) physical pins
One autonomous QTouch sensor operates without CPU intervention
Interfaced with peripheral DMA controller to reduce processor overhead
External synchronization to reduce 50 or 60 Hz mains interference
Spread spectrum sensor drive capability
Overview
The Capacitive Touch Module (CAT) senses touch on external capacitive touch sensors. Capacitive touch sensors use no external mechanical components, and therefore demand less
maintenance in the user application.
The module implements the QTouch method of capturing signals from capacitive touch sensors.
The QTouch method is generally suitable for small numbers of sensors since it requires 2 physical pins per sensor. The module also implements the QMatrix method, which is more
appropriate for large numbers of sensors since it allows an X by Y matrix of sensors to be implemented using only (X+2Y) physical pins. The module allows methods to function together, so N
touch sensors and an X by Y matrix of sensors can be implemented using (2N+X+2Y) physical
pins.
In addition, the module allows sensors using the QTouch method to be divided into two groups.
Each QTouch group can be configured with different properties. This eases the implementation
of multiple kinds of controls such as push buttons, wheels, and sliders.
The module also implements one autonomous QTouch sensor that is capable of detecting touch
or proximity without CPU intervention. This allows proximity or activation detection in low-power
sleep modes.
646
32099I–01/2012
AT32UC3L016/32/64
28.3
Block Diagram
Figure 28-1. CAT Block Diagram
Capacitive Touch Module (CAT)
CLK_CAT
Peripheral Bus
Interface
Registers
Counters
Finite State
Machine
Capacitor
Charge and
Discharge
Sequence
Generator
GCLK_CAT
CSAn
CSBn
I/O
Controller
Pins
SMP
VDIVEN
NOTE:
Italicized
signals and
blocks are
used only for
QMatrix
operation
DIS
Discharge
Current
Sources
Yn
SYNC
Peripheral
Event System
Analog
Comparator
Interface
Analog
Comparators
647
32099I–01/2012
AT32UC3L016/32/64
28.4
I/O Lines Description
Table 28-1.
28.5
I/O Lines Description
Name
Description
Type
CSAn
Capacitive sense A line n
I/O
CSBn
Capacitive sense B line n
I/O
DIS
Discharge current control (only used for QMatrix)
Analog
SMP
SMP line (only used for QMatrix)
Output
SYNC
Synchronize signal
VDIVEN
Voltage divider enable (only used for QMatrix)
Input
Output
Product Dependencies
In order to use the CAT module, other parts of the system must be configured correctly, as
described below.
28.5.1
I/O Lines
The CAT pins may be multiplexed with other peripherals. The user must first program the I/O
Controller to give control of the pins to the CAT module. In QMatrix mode, the Y lines must be
driven by the CAT and analog comparators sense the voltage on the Y lines. Thus, the CAT (not
the Analog Comparator Interface) must be the selected function for the Y lines in the I/O
Controller.
By writing ones and zeros to bits in the Pin Mode Registers (PINMODEx), most of the CAT pins
can be individually selected to implement the QTouch method or the QMatrix method. Each pin
has a different name and function depending on whether it is implementing the QTouch method
or the QMatrix method. The following table shows the pin names for each method and the bits in
the PINMODEx registers which control the selection of the QTouch or QMatrix method.
Table 28-2.
Pin Selection Guide
CAT Module Pin
Name
QTouch Method
Pin Name
QMatrix Method Pin
Name
Selection Bit in
PINMODEx Register
CSA0
SNS0
X0
SP0
CSB0
SNSK0
X1
SP0
CSA1
SNS1
Y0
SP1
CSB1
SNSK1
YK0
SP1
CSA2
SNS2
X2
SP2
CSB2
SNSK2
X3
SP2
CSA3
SNS3
Y1
SP3
CSB3
SNSK3
YK1
SP3
CSA4
SNS4
X4
SP4
CSB4
SNSK4
X5
SP4
CSA5
SNS5
Y2
SP5
648
32099I–01/2012
AT32UC3L016/32/64
Table 28-2.
28.5.2
Pin Selection Guide
CAT Module Pin
Name
QTouch Method
Pin Name
QMatrix Method Pin
Name
Selection Bit in
PINMODEx Register
CSB5
SNSK5
YK2
SP5
CSA6
SNS6
X6
SP6
CSB6
SNSK6
X7
SP6
CSA7
SNS7
Y3
SP7
CSB7
SNSK7
YK3
SP7
CSA8
SNS8
X8
SP8
CSB8
SNSK8
X9
SP8
CSA9
SNS9
Y4
SP9
CSB9
SNSK9
YK4
SP9
CSA10
SNS10
X10
SP10
CSB10
SNSK10
X11
SP10
CSA11
SNS11
Y5
SP11
CSB11
SNSK11
YK5
SP11
CSA12
SNS12
X12
SP12
CSB12
SNSK12
X13
SP12
CSA13
SNS13
Y6
SP13
CSB13
SNSK13
YK6
SP13
CSA14
SNS14
X14
SP14
CSB14
SNSK14
X15
SP14
CSA15
SNS15
Y7
SP15
CSB15
SNSK15
YK7
SP15
CSA16
SNS16
X16
SP16
CSB16
SNSK16
X17
SP16
Clocks
The clock for the CAT module, CLK_CAT, is generated by the Power Manager (PM). This clock
is turned on by default, and can be enabled and disabled in the PM. The user must ensure that
CLK_CAT is enabled before using the CAT module.
QMatrix operations also require the CAT generic clock, GCLK_CAT. This generic clock is generated by the System Control Interface (SCIF), and is shared between the CAT and the Analog
Comparator Interface. The user must ensure that the GCLK_CAT is enabled in the SCIF before
using QMatrix functionality in the CAT module. For proper QMatrix operation, the frequency of
GCLK_CAT must be less than half the frequency of CLK_CAT. If only QTouch functionality is
used, then GCLK_CAT is unnecessary.
28.5.3
Interrupts
The CAT interrupt request line is connected to the interrupt controller. Using CAT interrupts
requires the interrupt controller to be programmed first.
649
32099I–01/2012
AT32UC3L016/32/64
28.5.4
Peripheral Events
The CAT module receives nine peripheral events, one from the AST and the remaining eight
from on-chip analog comparators. These peripheral events are connected via the Peripheral
Event System. Refer to the Peripheral Event System chapter for additional details.
28.5.5
Peripheral Direct Memory Access
The CAT module provides handshake capability for a Peripheral DMA Controller. One handshake controls transfers from the Acquired Count Register (ACOUNT) to memory. A second
handshake requests burst lengths for each (X,Y) pair to the Matrix Burst Length Register
(MBLEN) when using the QMatrix acquisition method. The Peripheral DMA Controller must be
configured properly and enabled in order to perform direct memory access transfers to/from the
CAT module.
28.5.6
Analog Comparators
When the CAT module is performing QMatrix acquisition, it requires that on-chip analog comparators be used as part of the process. These analog comparators are not controlled directly by
the CAT module, but by a separate Analog Comparator (AC) Interface. This interface must be
configured properly and enabled before the CAT module is used. This includes configuring the
generic clock input for the analog comparators to the proper sampling frequency.
28.5.7
28.6
28.6.1
Debug Operation
When an external debugger forces the CPU into debug mode, the CAT continues normal operation. If the CAT is configured in a way that requires it to be periodically serviced by the CPU
through interrupts or similar, improper operation or data loss may result during debugging.
Functional Description
Acquisition Types
The CAT module can perform three types of QTouch acquisition from capacitive touch sensors:
autonomous QTouch (one sensor only), QTouch group A, and QTouch group B. The CAT module can also perform QMatrix acquisition. Each type of acquisition has an associated set of pin
selection and configuration registers that allow a large degree of flexibility.
The following schematic diagrams show typical hardware connections for QTouch and QMatrix
sensors, respectively:
Figure 28-2. CAT Touch Connections
SNSKn
QTouch
Sensor
32-bit AVR
Microcontroller
Cs (Sense Capacitor)
SNSn
650
32099I–01/2012
AT32UC3L016/32/64
Figure 28-3. CAT Matrix Connections
X2
X3
X6
QMatrix Sensor Array
X7
YK0
Y0
Cs0 (Sense Capacitor)
YK1
Y1
Cs1 (Sense Capacitor)
32-bit AVR
Microcontroller
Rsmp1
Rsmp0
SMP
DIS
Rdis
VDIVEN
Ra
ACREFN
Rb
NOTE: If the CAT internal
current sources will be enabled,
the SMP signal and Rsmp
resistors should NOT be included
in the design. If the CAT internal
current sources will NOT be
enabled, the DIS signal and Rdis
resistor should NOT be included
in the design.
In order to use the autonomous QTouch detection capability, the user must first set up the
Autonomous Touch Pin Select Register (ATPINS) and Autonomous Touch Configuration Registers (ATCFG0 through 3) with appropriate values. The module can then be enabled using the
Control Register (CTRL). After the module is enabled, the module will acquire data from the
autonomous QTouch sensor and use it to determine whether the sensor is activated. The
active/inactive status of the autonomous QTouch sensor is reported in the Status Register (SR),
and it is also possible to configure the CAT to generate an interrupt whenever the status
changes. The module will continue acquiring autonomous QTouch sensor data and updating
autonomous QTouch status until the module is disabled or reset.
In order to use the QMatrix, QTouch group A, or QTouch group B acquisition capabilities, it is
first necessary to set up the appropriate pin mode registers (PINMODE0 and PINMODE1) and
configuration registers (MGCFG0, MGCFG1, TGACFG0, TGACFG1, TGBCFG0, and
TGBCFG1). The module must then be enabled using the CTRL register. In order to initiate
acquisition, it is necessary to perform a write to the Acquisition Initiation and Selection Register
(AISR). The specific value written to AISR determines which type of acquisition will be performed: QMatrix, QTouch group A, or QTouch group B. The CPU can initiate acquisition by
writing to the AISR.
While QMatrix, QTouch group A, or QTouch group B acquisition is in progress, the module collects count values from the sensors and buffers them. Availability of acquired count data is
indicated by the Acquisition Ready (ACREADY) bit in the Status Register (SR). The CPU or the
Peripheral DMA Controller can then read the acquired counts from the ACOUNT register.
651
32099I–01/2012
AT32UC3L016/32/64
Because the CAT module is configured with Peripheral DMA Controller capability that can transfer data from memory to MBLEN and from ACOUNT to memory, the Peripheral DMA Controller
can perform long acquisition sequences and store results in memory without CPU intervention.
28.6.2
Prescaler and Charge Length
Each QTouch acquisition type (autonomous QTouch, QTouch group A, and QTouch group B)
has its own prescaler. Each QTouch prescaler divides down the CLK_CAT clock to an appropriate sampling frequency for its particular acquisition type. Typical frequencies are 1 MHz for
QTouch acquisition and 4MHz for QMatrix burst timing control.
Each QTouch prescaler is controlled by the DIV field in the appropriate Configuration Register 0
(ATCFG0, TGACFG0, or TGBCFG0). The QMatrix burst timing prescaler is controlled by the
DIV field in MGCFG0. Each prescaler uses the following formula to generate the sampling clock:
Sampling clock = CLK_CAT / (2(DIV+1))
The capacitive sensor charge length, discharge length, and settle length can be determined for
each acquisition type using the CHLEN, DILEN, and SELEN fields in Configuration Registers 0
and 1. The lengths are specified in terms of prescaler clocks. In addition, the QMatrix Cx discharge length can be determined using the CXDILEN field in MGCFG2.
For QMatrix acquisition, the duration of CHLEN should not be set to the same value as the
period of any periodic signal on any other pin. If the duration of CHLEN is the same as the
period of a signal on another pin, it is likely that the other signal will significantly affect measurements due to stray capacitive coupling. For example, if a 1 MHz signal is generated on another
pin of the device, then CHLEN should not be 1 microsecond.
For the QMatrix method, burst and capture lengths are set for each (X,Y) pair by writing the
desired length values to the MBLEN register. The write must be done before each X line can
start its acquisition and is indicated by the status bit MBLREQ in the Status Register (SR). A
DMA handshake interface is also connected to this status bit to reduce CPU overhead during
QMatrix acquisitions.
Four burst lengths (BURST0..3) can be written at one time into the MBLEN register. If the current configuration uses Y lines larger than Y3 the register has to be written a second time. The
first write to MBLEN specifies the burst length for Y lines 0 to 3 in the BURST0 to BURST3 fields,
respectively. The second write specifies the burst length for Y lines 4 to 7 in fields BURST0 to
BURST3, respectively, and so on.
The Y and YK pins remain clamped to ground apart from the specified number of burst pulses,
when charge is transferred and captured into the sampling capacitor.
28.6.3
Capacitive Count Acquisition
For the QMatrix, QTouch group A, and QTouch group B types of acquisition, the module
acquires count values from the sensors, buffers them, and makes them available for reading in
the ACOUNT register. Further processing of the count values must be performed by the CPU.
When the module performs QMatrix acquisition using multiple Y lines, it starts the capture for
each Y line at the appropriate time in the burst sequence so that all captures finish simultaneously. For example, suppose that an acquisition is performed on Y0 and Y1 with BURST0=53
and BURST1=60. The module will first toggle the X line 7 times while capturing on Y1 while Y0
and YK0 are clamped to ground. The module will then toggle the X line 53 times while capturing
on both Y1 and Y0.
652
32099I–01/2012
AT32UC3L016/32/64
28.6.4
Autonomous QTouch
For autonomous QTouch, a complete detection algorithm is implemented within the CAT module. The additional parameters needed to control the autonomous QTouch detection algorithm
must be specified by the user in the ATCFG2 and ATCFG3 registers.
Autonomous QTouch sensitivity and out-of-touch sensitivity can be adjusted with the SENSE
and OUTSENS fields, respectively, in ATCFG2. Each field accepts values from one to 255
where 255 is the least sensitive setting. The value in the OUTSENS field should be smaller than
the value in the SENSE field.
To avoid false positives a detect integration filtering technique can be used. The number of successive detects required is specified in the FILTER field of the ATCFG2 register.
To compensate for changes in capacitance the CAT can recalibrate the autonomous QTouch
sensor periodically. The timing of this calibration is done with the NDRIFT and PDRIFT fields in
the Configuration register, ATCFG3. It is recommended that the PDRIFT value is smaller than
the NDRIFT value.
The autonomous QTouch sensor will also recalibrate if the count value goes too far positive
beyond a threshold. This positive recalibration threshold is specified by the PTHR field in the
ATCFG3 register.
The following block diagram shows the sequence of acquisition and processing operations used
by the CAT module. The AISR written bit is internal and not visible in the user interface.
28.6.5
Peripheral Events
The peripheral event from the AST is used to trigger one iteration of autonomous touch detection when the device is in a sleep mode that has disabled CLK_CAT. When CLK_CAT is
disabled and the peripheral event from the AST becomes active, a request will automatically be
made to enable CLK_CAT. When CLK_CAT is enabled, the CAT will perform one iteration of the
autonomous touch detection algorithm. After this, the CLK_CAT will be disabled, and the CAT
module will remain in a frozen state until the next AST peripheral event.
The eight peripheral events from the analog comparators are automatically used by the CAT
when performing QMatrix acquisition. The CAT will automatically use the negative peripheral
events from the AC Interface on every Y pin in QMatrix mode. When QMatrix acquisition is used
the analog comparator corresponding to the selected Y pins must be enabled and converting
continuously, using the Y pin as the positive reference and the ACREFN as negative reference.
653
32099I–01/2012
AT32UC3L016/32/64
Figure 28-4. CAT Acquisition and Processing Sequence
Idle
AISR written flag set?
No
Yes
Autonomous touch
enabled (ATEN)?
No
28.6.6
Acquire counts
Yes
Acquire
autonomous
touch count
Wait for all
acquired counts
to be transferred
Update
autonomous
touch detection
algorithm
Clear AISR
written flag
Spread Spectrum Sensor Drive
To reduce electromagnetic compatibility issues, the capacitive sensors can be driven with a
spread spectrum signal. To enable spread spectrum drive for a specific acquisition type, the
user must write a one to the SPREAD bit in the appropriate Configuration Register 1 (MGCFG1,
ATCFG1, TGACFG1, or TGBCFG1).
During spread spectrum operation, the length of each pulse within a burst is varied in a deterministic pattern, so that the exact same burst pattern is used for a specific burst length. The
maximum spread is determined by the MAXDEV field in the Spread Spectrum Configuration
Register (SSCFG) register. The prescaler divisor is varied in a sawtooth pattern from
(2(DIV+1))-MAXDEV to (2(DIV+1))+MAXDEV and then back to (2(DIV+1))-MAXDEV. For example, if DIV is 2 and MAXDEV is 3, the prescaler divisor will have the following sequence: 6, 7, 8,
654
32099I–01/2012
AT32UC3L016/32/64
9, 3, 4, 5, 6, 7, 8, 9, 3, 4, etc. MAXDEV must not exceed the value of (2(DIV+1)), or undefined
behavior will occur.
28.6.7
Synchronization
To prevent interference from the 50 or 60Hz mains line the CAT can trigger acquisition on the
SYNC signal. The SYNC signal should be derived from the mains line. The acquisition will trigger on a falling edge of this signal. To enable synchronization for a specific acquisition type, the
user must write a one to the SYNC bit in the appropriate Configuration Register 1 (MGCFG1,
ATCFG1, TGACFG1, or TGBCFG1).
For QMatrix acquisition, all X lines must be sampled at a specific phase of the noise signal for
the synchronization to be effective. This can be accomplished by the synchronization timer,
which is enabled by writing a non-zero value to the SYNCTIM field in the MGCFG2 register. This
ensures that the start of the acquisition of each X line is spaced at regular intervals, defined by
the SYNCTIM field.
28.6.8
Resistive Drive
By default, the CAT pins are driven with normal I/O drive properties. Some of the CSA and CSB
pins can optionally drive with a 1k output resistance for improved EMC. The pins that have this
capability are listed in the Module Configuration section.
28.6.9
Discharge Current Sources
The device integrates discharge current sources, which can be used to discharge the sampling
capacitors during the QMatrix measurement phase. The discharge current sources are enabled
by writing the GLEN bit in the Discharge Current Source (DICS) register to one. This enables an
internal reference voltage, which can be either the internal 1.1V band gap voltage or VDDIO/3,
as selected by the INTVREFSEL bit in the DICS register. If the DICS.INTREFSEL bit is one, the
reference voltage is applied across an internal reference resistor, Rint. Otherwise, the voltage is
applied to the DIS pin, and an external reference resistor must be connected between DIS and
ground. The nominal discharge current is given by the following formula, where Vref is the reference voltage, R ref is the value of the reference resistor, trim is the value written to the
DICS.TRIM field, and k is a constant of proportionality:
I = (Vref/Rref)*(1+(k*trim))
The values for the internal reference resistor, Rint, and the constant, k, may be found in the Electrical Characteristics section. The nominal discharge current may be programmed between 2
and 20µA. The reference current can be fine-tuned by adjusting the trim value in the DICS.TRIM
field.
The reference current is mirrored to each Y-pin if the corresponding bit is written to one in the
DICS.SOURCES field.
28.6.10
Voltage Divider Enable (VDIVEN) Capability
In many QMatrix applications, the sense capacitors will be charged to 50 mV or more and the
negative reference pin (ACREFN) of the analog comparators can be tied directly to ground. In
that case, the relatively small input offset voltage of the comparators will not cause acquisition
problems. However, in certain specialized QMatrix applications such as interpolated touch
screens, it may be desirable for the sense capacitors to be charged to less than 25 mV. When
such small voltages are used on the sense capacitors, the input offset voltage of the comparators becomes an issue and can cause QMatrix acquisition problems.
655
32099I–01/2012
AT32UC3L016/32/64
Problems with QMatrix acquisition of small sense capapcitor voltages can be solved by connecting the negative reference pin (ACREFN) to a voltage divider that produces a small positive
voltage (20 mV, typically) to cancel any negative input offset voltage. With a 3.3V supply, recommended values for the voltage divider are Ra (resistor from positive supply to ACREFN) of 8200
ohm and Rb (resistor from ACREFN to ground) of 50 ohm. These recommended values will produce 20 mV on the ACREFN pin, which should generally be enough to compensate for the
worst-case negative input offset of the analog comparators.
Unfortunately, such a voltage divider constantly draws a small current from the power supply,
reducing battery life in portable applications. In order to prevent this constant power drain, the
CAT module provides a voltage divider enable pin (VDIVEN) that can be used for driving the
voltage divider. The VDIVEN pin provides power to the voltage divider only when the comparators are actually performing QMatrix comparisions. When the comparators are inactive, the
VDIVEN output is zero. This minimizes the power consumed by the voltage divider.
656
32099I–01/2012
AT32UC3L016/32/64
28.7
User Interface
Table 28-3.
CAT Register Memory Map
Offset
Register
Register Name
Access
Reset
0x00
Control Register
CTRL
Read/Write
0x00000000
0x04
Autonomous Touch Pin Selection Register
ATPINS
Read/Write
0x00000000
0x08
Pin Mode Register 0
PINMODE0
Read/Write
0x00000000
0x0C
Pin Mode Register 1
PINMODE1
Read/Write
0x00000000
0x10
Autonomous Touch Configuration Register 0
ATCFG0
Read/Write
0x00000000
0x14
Autonomous Touch Configuration Register 1
ATCFG1
Read/Write
0x00000000
0x18
Autonomous Touch Configuration Register 2
ATCFG2
Read/Write
0x00000000
0x1C
Autonomous Touch Configuration Register 3
ATCFG3
Read/Write
0x00000000
0x20
Touch Group A Configuration Register 0
TGACFG0
Read/Write
0x00000000
0x24
Touch Group A Configuration Register 1
TGACFG1
Read/Write
0x00000000
0x28
Touch Group B Configuration Register 0
TGBCFG0
Read/Write
0x00000000
0x2C
Touch Group B Configuration Register 1
TGBCFG1
Read/Write
0x00000000
0x30
Matrix Group Configuration Register 0
MGCFG0
Read/Write
0x00000000
0x34
Matrix Group Configuration Register 1
MGCFG1
Read/Write
0x00000000
0x38
Matrix Group Configuration Register 2
MGCFG2
Read/Write
0x00000000
0x3C
Status Register
SR
Read-only
0x00000000
0x40
Status Clear Register
SCR
Write-only
-
0x44
Interrupt Enable Register
IER
Write-only
-
0x48
Interrupt Disable Register
IDR
Write-only
-
0x4C
Interrupt Mask Register
IMR
Read-only
0x00000000
0x50
Acquisition Initiation and Selection Register
AISR
Read/Write
0x00000000
0x54
Acquired Count Register
ACOUNT
Read-only
0x00000000
0x58
Matrix Burst Length Register
MBLEN
Write-only
-
0x5C
Discharge Current Source Register
DICS
Read/Write
0x00000000
0x60
Spread Spectrum Configuration Register
SSCFG
Read/Write
0x00000000
0x64
CSA Resistor Control Register
CSARES
Read/Write
0x00000000
0x68
CSB Resistor Control Register
CSBRES
Read/Write
0x00000000
0x6C
Autonomous Touch Base Count Register
ATBASE
Read-only
0x00000000
0x70
Autonomous Touch Current Count Register
ATCURR
Read-only
0x00000000
0x80
Analog Comparator Shift Offset Register 0
ACSHI0
Read/Write
0x00000000
0x84
Analog Comparator Shift Offset Register 1
ACSHI1
Read/Write
0x00000000
0x88
Analog Comparator Shift Offset Register 2
ACSHI2
Read/Write
0x00000000
0x8C
Analog Comparator Shift Offset Register 3
ACSHI3
Read/Write
0x00000000
0x90
Analog Comparator Shift Offset Register 4
ACSHI4
Read/Write
0x00000000
0x94
Analog Comparator Shift Offset Register 5
ACSHI5
Read/Write
0x00000000
657
32099I–01/2012
AT32UC3L016/32/64
Table 28-3.
CAT Register Memory Map
Offset
Register
Register Name
Access
Reset
0x98
Analog Comparator Shift Offset Register 6
ACSHI6
Read/Write
0x00000000
0x9C
Analog Comparator Shift Offset Register 7
ACSHI7
Read/Write
0x00000000
0xF8
Parameter Register
PARAMETER
Read-only
0xFC
Version Register
VERSION
Read-only
658
32099I–01/2012
AT32UC3L016/32/64
28.7.1
Name:
Control Register
CTRL
Access Type:
Read/Write
Offset:
0x00
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
SWRST
-
-
-
-
-
-
EN
• SWRST: Software reset
Writing a zero to this bit has no effect.
Writing a one to this bit resets the module. The module will be disabled after the reset.
This bit always reads as zero.
• EN: Module enable
0: Module is disabled.
1: Module is enabled.
659
32099I–01/2012
AT32UC3L016/32/64
28.7.2
Name:
Autonomous Touch Pin Selection Register
ATPINS
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
-
-
-
-
-
-
-
ATEN
7
6
5
4
3
2
1
0
-
-
-
ATSP
• ATEN: Autonomous Touch Enable
0: Autonomous QTouch acquisition and detection is disabled.
1: Autonomous QTouch acquisition and detection is enabled using the sense pair specified in ATSP.
• ATSP: Autonomous Touch Sense Pair
Selects the sense pair that will be used by the autonomous QTouch sensor. A value of n will select sense pair n (CSAn and
CSBn pins).
660
32099I–01/2012
AT32UC3L016/32/64
28.7.3
Name:
Pin Mode Registers 0 and 1
PINMODE0 and PINMODE1
Access Type:
Read/Write
Offset:
0x08, 0x0C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
23
22
21
20
-
15
14
13
SP[16]
12
11
10
9
8
3
2
1
0
SP[15:8]
7
6
5
4
SP[7:0]
• SP: Sense Pair Mode Selection
Each SP[n] bit determines the operation mode of sense pair n (CSAn and CSBn pins). The (PINMODE1.SP[n]
PINMODE0.SP[n]) bits have the following definitions:
00: Sense pair n disabled.
01: Sense pair n is assigned to QTouch Group A.
10: Sense pair n is assigned to QTouch Group B.
11: Sense pair n is assigned to the QMatrix Group.
661
32099I–01/2012
AT32UC3L016/32/64
28.7.4
Name:
Autonomous Touch Configuration Register 0
ATCFG0
Access Type:
Read/Write
Offset:
0x10
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
DIV[15:8]
23
22
21
20
DIV[7:0]
15
14
13
12
CHLEN
7
6
5
4
SELEN
• DIV: Clock Divider
The prescaler is used to ensure that the CLK_CAT clock is divided to around 1 MHz to produce the sampling clock.The
prescaler uses the following formula to generate the sampling clock:
Sampling clock = CLK_CAT / (2(DIV+1))
• CHLEN: Charge Length
For the autonomous QTouch sensor, specifies how many sample clock cycles should be used for transferring charge to the
sense capacitor.
• SELEN: Settle Length
For the autonomous QTouch sensor, specifies how many sample clock cycles should be used for settling after charge transfer.
662
32099I–01/2012
AT32UC3L016/32/64
28.7.5
Name:
Autonomous Touch Configuration Register 1
ATCFG1
Access Type:
Read/Write
Offset:
0x14
Reset Value:
0x00000000
31
30
29
23
28
27
DISHIFT
22
21
26
-
20
25
24
SYNC
SPREAD
19
18
17
16
11
10
9
8
3
2
1
0
DILEN
15
14
13
12
MAX[15:8]
7
6
5
4
MAX[7:0]
• DISHIFT: Discharge Shift
For the autonomous QTouch sensor, specifies how many bits the DILEN field should be shifted before using it to determine the
discharge time.
• SYNC: Sync Pin
For the autonomous QTouch sensor, specifies that acquisition shall begin when a falling edge is received on the SYNC line.
• SPREAD: Spread Spectrum Sensor Drive
For the autonomous QTouch sensor, specifies that spread spectrum sensor drive shall be used.
• DILEN: Discharge Length
For the autonomous QTouch sensor, specifies how many sample clock cycles the CAT should use to discharge the capacitors
before charging them.
• MAX: Maximum Count
For the autonomous QTouch sensor, specifies how many counts the maximum acquisition should be.
663
32099I–01/2012
AT32UC3L016/32/64
28.7.6
Name:
Autonomous Touch Configuration Register 2
ATCFG2
Access Type:
Read/Write
Offset:
0x18
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
23
22
21
20
15
14
FILTER
13
12
11
10
9
8
3
2
1
0
OUTSENS
7
6
5
4
SENSE
• FILTER: Autonomous Touch Filter Setting
For the autonomous QTouch sensor, specifies how many positive detects in a row the CAT needs to have on the autonomous
QTouch sensor before reporting it as a touch. A FILTER value of 0 is not allowed and will result in undefined behavior.
• OUTSENS: Out-of-Touch Sensitivity
For the autonomous QTouch sensor, specifies how sensitive the out-of-touch detector should be.
• SENSE: Sensitivity
For the autonomous QTouch sensor, specifies how sensitive the touch detector should be.
664
32099I–01/2012
AT32UC3L016/32/64
28.7.7
Name:
Autonomous Touch Configuration Register 3
ATCFG3
Access Type:
Read/Write
Offset:
0x1C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
23
22
21
20
PTHR
15
14
13
12
PDRIFT
7
6
5
4
NDRIFT
• PTHR: Positive Recalibration Threshold
For the autonomous QTouch sensor, specifies how far a sensor’s signal must move in a positive direction from the reference in
order to cause a recalibration.
• PDRIFT: Positive Drift Compensation
For the autonomous QTouch sensor, specifies how often a positive drift compensation should be performed. When this field is
zero, positive drift compensation will never be performed. When this field is non-zero, the positive drift compensation time
interval is given by the following formula:
Tpdrift = PDRIFT * 65536 * (sample clock period)
• NDRIFT: Negative Drift Compensation
For the autonomous QTouch sensor, specifies how often a negative drift compensation should be performed. When this field is
zero, negative drift compensation will never be performed. When this field is non-zero, the negative drift compensation time
interval is given by the following formula:
Tndrift = NDRIFT * 65536 * (sample clock period)
With the typical sample clock frequency of 1 MHz, PDRIFT and NDRIFT can be set from 0.066 seconds to 16.7 seconds
with 0.066 second resolution.
665
32099I–01/2012
AT32UC3L016/32/64
28.7.8
Name:
Touch Group x Configuration Register 0
TGxCFG0
Access Type:
Read/Write
Offset:
0x20, 0x28
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
DIV[15:8]
23
22
21
20
DIV[7:0]
15
14
13
12
CHLEN
7
6
5
4
SELEN
• DIV: Clock Divider
The prescaler is used to ensure that the CLK_CAT clock is divided to around 1 MHz to produce the sampling clock.The
prescaler uses the following formula to generate the sampling clock:
Sampling clock = CLK_CAT / (2(DIV+1))
• CHLEN: Charge Length
For the QTouch method, specifies how many sample clock cycles should be used for transferring charge to the sense capacitor.
• SELEN: Settle Length
For the QTouch method, specifies how many sample clock cycles should be used for settling after charge transfer.
666
32099I–01/2012
AT32UC3L016/32/64
28.7.9
Name:
Touch Group x Configuration Register 1
TGxCFG1
Access Type:
Read/Write
Offset:
0x24, 0x2C
Reset Value:
0x00000000
31
30
-
-
23
22
29
28
DISHIFT
21
20
27
26
25
24
-
-
SYNC
SPREAD
19
18
17
16
11
10
9
8
3
2
1
0
DILEN
15
14
13
12
MAX[15:8]
7
6
5
4
MAX[7:0]
• DISHIFT: Discharge Shift
For the sensors in QTouch group x, specifies how many bits the DILEN field should be shifted before using it to determine the
discharge time.
• SYNC: Sync Pin
For sensors in QTouch group x, specifies that acquisition shall begin when a falling edge is received on the SYNC line.
• SPREAD: Spread Spectrum Sensor Drive
For sensors in QTouch group x, specifies that spread spectrum sensor drive shall be used.
• DILEN: Discharge Length
For sensors in QTouch group x, specifies how many clock cycles the CAT should use to discharge the capacitors before
charging them.
• MAX: Touch Maximum Count
For sensors in QTouch group x, specifies how many counts the maximum acquisition should be.
667
32099I–01/2012
AT32UC3L016/32/64
28.7.10
Name:
Matrix Group Configuration Register 0
MGCFG0
Access Type:
Read/Write
Offset:
0x30
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
DIV[15:8]
23
22
21
20
DIV[7:0]
15
14
13
12
CHLEN
7
6
5
4
SELEN
• DIV: Clock Divider
The prescaler is used to ensure that the CLK_CAT clock is divided to around 4 MHz to produce the burst timing clock.The
prescaler uses the following formula to generate the burst timing clock:
Burst timing clock = CLK_CAT / (2(DIV+1))
• CHLEN: Charge Length
For QMatrix sensors, specifies how many burst prescaler clock cycles should be used for transferring charge to the sense
capacitor.
• SELEN: Settle Length
For QMatrix sensors, specifies how many burst prescaler clock cycles should be used for settling after charge transfer.
668
32099I–01/2012
AT32UC3L016/32/64
28.7.11
Name:
Matrix Group Configuration Register 1
MGCFG1
Access Type:
Read/Write
Offset:
0x34
Reset Value:
0x00000000
31
30
29
23
28
27
DISHIFT
22
21
26
-
20
25
24
SYNC
SPREAD
19
18
17
16
11
10
9
8
3
2
1
0
DILEN
15
14
13
12
MAX[15:8]
7
6
5
4
MAX[7:0]
• DISHIFT: Discharge Shift
For QMatrix sensors, specifies how many bits the DILEN field should be shifted before using it to determine the discharge time.
• SYNC: Sync Pin
For QMatrix sensors, specifies that acquisition shall begin when a falling edge is received on the SYNC line.
• SPREAD: Spread Spectrum Sensor Drive
For QMatrix sensors, specifies that spread spectrum sensor drive shall be used.
• DILEN: Discharge Length
For QMatrix sensors, specifies how many burst prescaler clock cycles the CAT should use to discharge the capacitors at the
beginning of a burst sequence.
• MAX: Maximum Count
For QMatrix sensors, specifies how many counts the maximum acquisition should be.
669
32099I–01/2012
AT32UC3L016/32/64
28.7.12
Name:
Matrix Group Configuration Register 2
MGCFG2
Access Type:
Read/Write
Offset:
0x38
Reset Value:
0x00000000
31
30
29
ACCTRL
23
28
27
26
CONSEN
22
21
25
24
20
19
18
17
16
11
10
9
8
CXDILEN
15
14
13
12
7
6
SYNCTIM[11:8]
5
4
3
2
1
0
SYNCTIM[7:0]
• ACCTRL: Analog Comparator Control
When written to one, allows the CAT to disable the analog comparators when they are not needed. When written to zero, the
analog comparators are always enabled.
• CONSEN: Consensus Filter Length
For QMatrix sensors, specifies that discharge will be terminated when CONSEN out of the most recent 5 comparator samples
are positive. For example, a value of 3 in the CONSEN field will terminate discharge when 3 out of the most recent 5 comparator
samples are positive. When CONSEN has the default value of 0, discharge will be terminated immediately when the comparator
output goes positive.
• CXDILEN: Cx Capacitor Discharge Length
For QMatrix sensors, specifies how many burst prescaler clock cycles the CAT should use to discharge the Cx capacitor at the
end of each burst cycle.
• SYNCTIM: Sync Time Interval
When non-zero, determines the number of prescaled clock cycles between the start of the acquisition on each X line for QMatrix
acquisition.
670
32099I–01/2012
AT32UC3L016/32/64
28.7.13
Name:
Status Register
SR
Access Type:
Read-only
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
-
-
-
-
-
-
ACQDONE
ACREADY
7
6
5
4
3
2
1
0
-
-
-
MBLREQ
ATSTATE
ATSC
ATCAL
ENABLED
• ACQDONE: Acquisition Done
0: Acquisition is not done (still in progress).
1: Acquisition is complete.
• ACREADY: Acquired Count Data is Ready
0: Acquired count data is not available in the ACOUNT register.
1: Acquired count data is available in the ACOUNT register.
• MBLREQ: Matrix Burst Length Required
0: The QMatrix acquisition does not require any burst lengths.
1: The QMatrix acquisition requires burst lengths for the current X line.
• ATSTATE: Autonomous Touch Sensor State
0: The autonomous QTouch sensor is not active.
1: The autonomous QTouch sensor is active.
• ATSC: Autonomous Touch Sensor Status Interrupt
0: No status change in the autonomous QTouch sensor.
1: Status change in the autonomous QTouch sensor.
• ATCAL: Autonomous Touch Calibration Ongoing
0: The autonomous QTouch sensor is not calibrating.
1: The autonomous QTouch sensor is calibrating.
• ENABLED: Module Enabled
0: The module is disabled.
1: The module is enabled.
671
32099I–01/2012
AT32UC3L016/32/64
28.7.14
Name:
Status Clear Register
SCR
Access Type:
Write-only
Offset:
0x40
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
-
-
-
-
-
-
ACQDONE
ACREADY
7
6
5
4
3
2
1
0
-
-
-
-
-
ATSC
ATCAL
-
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in SR and the corresponding interrupt request.
672
32099I–01/2012
AT32UC3L016/32/64
28.7.15
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
10
9
8
-
-
-
-
-
-
ACQDONE
ACREADY
7
6
5
4
3
2
1
0
-
-
-
-
-
ATSC
ATCAL
-
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
673
32099I–01/2012
AT32UC3L016/32/64
28.7.16
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
10
9
8
-
-
-
-
-
-
ACQDONE
ACREADY
7
6
5
4
3
2
1
0
-
-
-
-
-
ATSC
ATCAL
-
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
674
32099I–01/2012
AT32UC3L016/32/64
28.7.17
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
10
9
8
-
-
-
-
-
-
ACQDONE
ACREADY
7
6
5
4
3
2
1
0
-
-
-
-
-
ATSC
ATCAL
-
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
A bit in this register is cleared when the corresponding bit in IDR is written to one.
A bit in this register is set when the corresponding bit in IER is written to one.
675
32099I–01/2012
AT32UC3L016/32/64
28.7.18
Name:
Acquisition Initiation and Selection Register
AISR
Access Type:
Read/Write
Offset:
0x50
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
23
22
21
20
-
15
14
13
12
-
7
6
5
4
-
ACQSEL
• ACQSEL: Acquisition Type Selection
A write to this register initiates an acquisition of the following type:
00: QTouch Group A.
01: QTouch Group B.
10: QMatrix Group.
11: Undefined behavior.
A read of this register will return the value that was previously written.
676
32099I–01/2012
AT32UC3L016/32/64
28.7.19
Name:
Acquired Count Register
ACOUNT
Access Type:
Read-only
Offset:
0x54
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
Y
23
22
21
20
SPORX
15
14
13
12
COUNT[15:8]
7
6
5
4
COUNT[7:0]
• Y: Y index
The Y index (for QMatrix method) associated with this count value.
• SPORX: Sensor pair or X index
The sensor pair index (for QTouch method) or X index (for QMatrix method) associated with this count value.
• COUNT: Count value
The signal (number of counts) acquired on the channel specified in the SPORX and Y fields.
When multiple acquired count values are read from a QTouch acquisition, the Y field will always be 0 and the SPORX value will
increase monotonically. For example, suppose a QTouch acquisition is performed using sensor pairs SP1, SP4, and SP9. The
first count read will have SPORX=1, the second read will have SPORX=4, and the third read will have SPORX=9.
When multiple acquired count values are read from a QMatrix acquisition, the SPORX value will stay the same while Y
increases monotonically through all Y values in the group. Then SPORX will increase to the next X value in the group. For
example, a QMatrix acquisition with X=2,3 and Y=4,7 would provide count values in the following order: X=2 and Y=4, then X=2
and Y=7, then X=3 and Y=4, and finally X=3 and Y=7.
677
32099I–01/2012
AT32UC3L016/32/64
28.7.20
Name:
Matrix Burst Length Register
MBLEN
Access Type:
Write-only
Offset:
0x58
Reset Value:
-
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
BURST0
23
22
21
20
BURST1
15
14
13
12
BURST2
7
6
5
4
BURST3
• BURSTx: Burst Length x
For QMatrix sensors, specifies how many times the switching sequence should be repeated before acquisition begins for each
channel. Each count in the BURSTx field specifies 1 repeat of the switching sequence, so the actual burst length will be BURST.
Before doing a QMatrix acquisition on one X line this register has to be written with the burst values for the current XY pairs. For
each X line this register needs to be programmed with all the Y values. If Y values larger than 3 are used the register has to be
written several times in order to specify all burst lengths.
The Status Register bit MBLREQ is set to 1 when the CAT is waiting for values to be written into this register.
678
32099I–01/2012
AT32UC3L016/32/64
28.7.21
Name:
Discharge Current Source Register
DICS
Access Type:
Read/Write
Offset:
0x5C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
FSOURCES[7:0]
23
22
21
20
19
18
17
16
GLEN
-
-
-
-
-
INTVREFSEL
INTREFSEL
15
14
13
12
11
10
9
8
-
-
-
7
6
5
1
0
TRIM
4
3
2
SOURCES[7:0]
• FSOURCES: Force Discharge Current Sources
When FSOURCES[n] is 0, the corresponding discharge current source behavior depends on SOURCES[n].
When FSOURCES[n] is 1, the corresponding discharge current source is forced to be enabled continuously. This is useful for
testing or debugging but should not be done during normal acquisition.
• GLEN: Global Enable
0: The current source module is globally disabled.
1: The current source module is globally enabled.
• INTVREFSEL: Internal Voltage Reference Select
0: The voltage for the reference resistor is generated from the internal band gap circuit.
1: The voltage for the reference resistor is VDDIO/3.
• INTREFSEL: Internal Reference Select
0: The reference current flows through an external resistor on the DIS pin.
1: The reference current flows through the internal reference resistor.
• TRIM: Reference Current Trimming
This field is used to trim the discharge current. 0x00 corresponds to the minimum current value, and 0x1F corresponds to the
maximum current value.
• SOURCES: Enable Discharge Current Sources
When SOURCES[n] is 0, the corresponding discharge current source is disabled.
When SOURCES[n] is 1, the corresponding discharge current source is enabled at appropriate times during acquisition.
679
32099I–01/2012
AT32UC3L016/32/64
28.7.22
Name:
Spread Spectrum Configuration Register
SSCFG
Access Type:
Read/Write
Offset:
0x60
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
MAXDEV
• MAXDEV: Maximum Deviation
When spread spectrum burst is enabled, MAXDEV indicates the maximum number of prescaled clock cycles the burst pulse will
be extended or shortened.
680
32099I–01/2012
AT32UC3L016/32/64
28.7.23
Name:
CSA Resistor Control Register
CSARES
Access Type:
Read/Write
Offset:
0x64
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
23
22
21
20
-
15
14
13
RES[16]
12
11
10
9
8
3
2
1
0
RES[15:8]
7
6
5
4
RES[7:0]
• RES: Resistive Drive Enable
When RES[n] is 0, CSA[n] has the same drive properties as normal I/O pads.
When RES[n] is 1, CSA[n] has a nominal output resistance of 1kOhm during the burst phase.
681
32099I–01/2012
AT32UC3L016/32/64
28.7.24
Name:
CSB Resistor Control Register
CSBRES
Access Type:
Read/Write
Offset:
0x68
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
23
22
21
20
-
15
14
13
RES[16]
12
11
10
9
8
3
2
1
0
RES[15:8]
7
6
5
4
RES[7:0]
• RES: Resistive Drive Enable
When RES[n] is 0, CSB[n] has the same drive properties as normal I/O pads.
When RES[n] is 1, CSB[n] has a nominal output resistance of 1kOhm during the burst phase.
682
32099I–01/2012
AT32UC3L016/32/64
28.7.25
Name:
Autonomous Touch Base Count Register
ATBASE
Access Type:
Read-only
Offset:
0x6C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
23
22
21
20
-
15
14
13
12
COUNT[15:8]
7
6
5
4
COUNT[7:0]
• COUNT: Count value
The base count currently stored by the autonomous touch sensor. This is useful for autonomous touch debugging purposes.
683
32099I–01/2012
AT32UC3L016/32/64
28.7.26
Name:
Autonomous Touch Current Count Register
ATCURR
Access Type:
Read-only
Offset:
0x70
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
23
22
21
20
-
15
14
13
12
COUNT[15:8]
7
6
5
4
COUNT[7:0]
• COUNT: Count value
The current count acquired by the autonomous touch sensor. This is useful for autonomous touch debugging purposes.
684
32099I–01/2012
AT32UC3L016/32/64
28.7.27
Name:
Analog Comparator Shift Offset Register x
ACSHIx
Access Type:
Read/Write
Offset:
0x80, 0x84, 0x88, 0x8C, 0x90, 0x94, 0x98, and 0x9C
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
1
0
23
22
21
20
-
15
14
13
12
7
6
SHIVAL[11:8]
5
4
3
2
SHIVAL[7:0]
• SHIVAL: Shift Offset Value
Specifies the amount to shift the count value from each comparator. This allows the offset of each comparator to be
compensated.
685
32099I–01/2012
AT32UC3L016/32/64
28.7.28
Name:
Parameter Register
PARAMETER
Access Type:
Read-only
Offset:
0xF8
Reset Value:
-
31
30
29
28
27
26
25
24
19
18
17
16
11
10
9
8
3
2
1
0
23
22
21
20
SP[23:16]
15
14
13
12
SP[15:8]
7
6
5
4
SP[7:0]
• SP[n]: Sensor pair implemented
0: The corresponding sensor pair is not implemented
1: The corresponding sensor pair is implemented.
686
32099I–01/2012
AT32UC3L016/32/64
28.7.29
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant number
Reserved. No functionality associated.
• VERSION: Version number
Version number of the module. No functionality associated.
687
32099I–01/2012
AT32UC3L016/32/64
28.8
Module Configuration
The specific configuration the CAT module is listed in the following tables.The module bus
clocks listed here are connected to the system bus clocks. Please refer to the Power Manager
chapter for details.
Table 28-4.
Feature
CAT
Number of touch sensors/Size of matrix
Allows up to 17 touch sensors, or up to 16 by 8
matrix sensors to be interfaced.
Table 28-5.
CAT Clocks
Clock Name
Description
CLK_CAT
Clock for the CAT bus interface
GCLK
The generic clock used for the CAT is GCLK4
Table 28-6.
28.8.1
CAT Configuration
Register Reset Values
Register
Reset Value
VERSION
0x00000200
PARAMETER
0x0001FFFF
Resistive Drive
By default, the CAT pins are driven with normal I/O drive properties. Some of the CSA and CSB
pins can optionally drive with a 1k output resistance for improved EMC.
To enable resistive drive on a pin, the user must write a one to the corresponding bit in the CSA
Resistor Control Register (CSARES) or CSB Resistor Control Register (CSBRES) register.
688
32099I–01/2012
AT32UC3L016/32/64
29. Glue Logic Controller (GLOC)
Rev: 1.0.0.0
29.1
Features
•
•
•
•
29.2
Glue logic for general purpose PCB design
Programmable lookup table
Up to four inputs supported per lookup table
Optional filtering of output
Overview
The Glue Logic Controller (GLOC) contains programmable logic which can be connected to the
device pins. This allows the user to eliminate logic gates for simple glue logic functions on the
PCB.
The GLOC consists of a number of lookup table (LUT) units. Each LUT can generate an output
as a user programmable logic expression with four inputs. Inputs can be individually masked.
The output can be combinatorially generated from the inputs, or filtered to remove spikes.
29.3
Block Diagram
TRUTH
PERIPHERAL BUS
Figure 29-1. GLOC Block Diagram
OUT[0]
...
OUT[n]
FILTER
FILTEN
GCLK
AEN
CLK_GLOC
IN[3:0]
…
IN[(4n+3):4n]
689
32099I–01/2012
AT32UC3L016/32/64
29.4
I/O Lines Description
Table 29-1.
I/O Lines Description
Pin Name
Pin Description
Type
IN0-INm
Inputs to lookup tables
Input
OUT0-OUTn
Output from lookup tables
Output
Each LUT have 4 inputs and one output. The inputs and outputs for the LUTs are mapped
sequentially to the inputs and outputs. This means that LUT0 is connected to IN0 to IN3 and
OUT0. LUT1 is connected to IN4 to IN7 and OUT1. In general, LUTn is connected to IN[4n] to
IN[4n+3] and OUTn.
29.5
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
29.5.1
I/O Lines
The pins used for interfacing the GLOC may be multiplexed with I/O Controller lines. The programmer must first program the I/O Controller to assign the desired GLOC pins to their
peripheral function. If I/O lines of the GLOC are not used by the application, they can be used for
other purposes by the I/O Controller.
It is only required to enable the GLOC inputs and outputs actually in use. Pullups for pins configured to be used by the GLOC will be disabled.
29.5.2
Clocks
The clock for the GLOC bus interface (CLK_GLOC) is generated by the Power Manager. This
clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to disable the GLOC before disabling the clock, to avoid freezing the module in an undefined state.
Additionally, the GLOC depends on a dedicated Generic Clock (GCLK). The GCLK can be set to
a wide range of frequencies and clock sources, and must be enabled by the System Control
Interface (SCIF) before the GLOC filter can be used.
29.5.3
Debug Operation
When an external debugger forces the CPU into debug mode, the GLOC continues normal
operation.
29.6
29.6.1
Functional Description
Enabling the Lookup Table Inputs
Since the inputs to each lookup table (LUT) unit can be multiplexed with other peripherals, each
input must be explicitly enabled by writing a one to the corresponding enable bit (AEN) in the
corresponding Control Register (CR).
If no inputs are enabled, the output OUTn will be the least significant bit in the TRUTHn register.
690
32099I–01/2012
AT32UC3L016/32/64
29.6.2
Configuring the Lookup Table
The lookup table in each LUT unit can generate any logic expression OUT as a function of up to
four inputs, IN[3:0]. The truth table for the expression is written to the TRUTH register for the
LUT. Table 29-2 shows the truth table for LUT0. The truth table for LUTn is written to TRUTHn,
and the corresponding input and outputs will be IN[4n] to IN[4n+3] and OUTn.
Table 29-2.
29.6.3
Truth Table for the Lookup Table in LUT0
IN[3]
IN[2]
IN[1]
IN[0]
OUT[0]
0
0
0
0
TRUTH0[0]
0
0
0
1
TRUTH0[1]
0
0
1
0
TRUTH0[2]
0
0
1
1
TRUTH0[3]
0
1
0
0
TRUTH0[4]
0
1
0
1
TRUTH0[5]
0
1
1
0
TRUTH0[6]
0
1
1
1
TRUTH0[7]
1
0
0
0
TRUTH0[8]
1
0
0
1
TRUTH0[9]
1
0
1
0
TRUTH0[10]
1
0
1
1
TRUTH0[11]
1
1
0
0
TRUTH0[12]
1
1
0
1
TRUTH0[13]
1
1
1
0
TRUTH0[14]
1
1
1
1
TRUTH0[15]
Output Filter
By default, the output OUTn is a combinatorial function of the inputs IN[4n] to IN[4n+3]. This may
cause some short glitches to occur when the inputs change value.
It is also possible to clock the output through a filter to remove glitches. This requires that the
corresponding generic clock (GCLK) has been enabled before use. The filter can then be
enabled by writing a one to the Filter Enable (FILTEN) bit in CRn. The OUTn output will be
delayed by three to four GCLK cycles when the filter is enabled.
691
32099I–01/2012
AT32UC3L016/32/64
29.7
User Interface
Table 29-3.
GLOC Register Memory Map
Offset
Register
Register Name
Access
Reset
0x00+n*0x08
Control Register n
CRn
Read/Write
0x00000000
0x04+n*0x08
Truth Table Register n
TRUTHn
Read/Write
0x00000000
0x38
Parameter Register
PARAMETER
Read-only
- (1)
0x3C
Version Register
VERSION
Read-only
- (1)
Note:
1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter.
692
32099I–01/2012
AT32UC3L016/32/64
29.7.1
Name:
Control Register n
CRn
Access Type:
Read/Write
Offset:
0x00+n*0x08
Reset Value:
0x00000000
31
30
29
28
27
26
25
24
FILTEN
-
-
-
-
-
-
-
23
22
21
20
19
18
17
16
-
-
-
-
-
-
-
-
15
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
-
-
-
-
AEN
• FILTEN: Filter Enable
1: The output is glitch filtered
0: The output is not glitch filtered
• AEN: Enable IN Inputs
Input IN[n] is enabled when AEN[n] is one.
Input IN[n] is disabled when AEN[n] is zero, and will not affect the OUT value.
693
32099I–01/2012
AT32UC3L016/32/64
29.7.2
Name:
Truth Table Register n
TRUTHn
Access Type:
Read/Write
Offset:
0x04+n*0x08
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
TRUTH[15:8]
7
6
5
4
TRUTH[7:0]
• TRUTH: Truth Table Value
This value defines the output OUT as a function of inputs IN:
OUT = TRUTH[IN]
694
32099I–01/2012
AT32UC3L016/32/64
29.7.3
Name:
Parameter Register
PARAMETER
Access Type:
Read-only
Offset:
0x38
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
LUTS
• LUTS: Lookup Table Units Implemented
This field contains the number of lookup table units implemented in this device.
695
32099I–01/2012
AT32UC3L016/32/64
29.7.4
Name:
VERSION Register
VERSION
Access Type:
Read-only
Offset:
0x3C
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
VARIANT
11
10
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VARIANT: Variant Number
Reserved. No functionality associated.
• VERSION: Version Number
Version number of the module. No functionality associated.
696
32099I–01/2012
AT32UC3L016/32/64
29.8
Module Configuration
The specific configuration for each GLOC instance is listed in the following tables.The GLOC
bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details.
Table 29-4.
GLOC Configuration
Feature
GLOC
Number of LUT units
2
Table 29-5.
GLOC Clocks
Clock Name
Description
CLK_GLOC
Clock for the GLOC bus interface
GCLK
The generic clock used for the GLOC is GCLK5
Table 29-6.
Register Reset Values
Register
Reset Value
VERSION
0x00000100
PARAMETER
0x00000002
697
32099I–01/2012
AT32UC3L016/32/64
30. aWire UART (AW)
Rev: 2.1.0.0
30.1
Features
• Asynchronous receiver or transmitter when the aWire system is not used for debugging.
• One- or two-pin operation supported.
30.2
Overview
If the AW is not used for debugging, the aWire UART can be used by the user to send or receive
data with one start bit, eight data bits, no parity bits, and one stop bit. This can be controlled
through the aWire UART user interface.
This chapter only describes the aWire UART user interface. For a description of the aWire
Debug Interface, please see the Programming and Debugging chapter.
30.3
Block Diagram
Figure 30-1. aWire Debug Interface Block Diagram
PB
aWire Debug Interface
Flash
Controller
CHIP_ERASE command
AW User Interface
CPU
HALT command
RESET command
Power
Manager
External reset
AW_ENABLE
AW CONTROL
Reset
filter
RESET_N
Baudrate Detector
SAB interface
UART
RW
SZ
ADDR
DATA
CRC
SAB
698
32099I–01/2012
AT32UC3L016/32/64
30.4
I/O Lines Description
Table 30-1.
I/O Lines Description
Name
Description
Type
DATA
aWire data multiplexed with the RESET_N pin.
Input/Output
30.5
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
30.5.1
I/O Lines
The pin used by AW is multiplexed with the RESET_N pin. The reset functionality is the default
function of this pin. To enable the aWire functionality on the RESET_N pin the user must enable
the aWire UART user interface.
30.5.2
Power Management
If the CPU enters a sleep mode that disables clocks used by the aWire UART user interface, the
aWire UART user interface will stop functioning and resume operation after the system wakes
up from sleep mode.
30.5.3
Clocks
The aWire UART uses the internal 120 MHz RC oscillator (RC120M) as clock source for its
operation. When using the aWire UART user interface RC120M must enabled using the Clock
Request Register (see Section 30.6.1).
The clock for the aWire UART user interface (CLK_AW) is generated by the Power Manager.
This clock is enabled at reset, and can be disabled in the Power Manager. It is recommended to
disable the aWire UART user interface before disabling the clock, to avoid freezing the aWire
UART user interface in an undefined state.
30.5.4
Interrupts
The aWire UART user interface interrupt request line is connected to the interrupt controller.
Using the aWire UART user interface interrupt requires the interrupt controller to be programmed first.
30.5.5
Debug Operation
If the AW is used for debugging the aWire UART user interface will not be usable.
When an external debugger forces the CPU into debug mode, the aWire UART user interface
continues normal operation. If the aWire UART user interface is configured in a way that
requires it to be periodically serviced by the CPU through interrupts or similar, improper operation or data loss may result during debugging.
30.5.6
30.6
External Components
The AW needs an external pullup on the RESET_N pin to ensure that the pin is pulled up when
the bus is not driven.
Functional Description
The aWire UART user interface can be used as a spare Asynchronous Receiver or Transmitter
when AW is not used for debugging.
699
32099I–01/2012
AT32UC3L016/32/64
30.6.1
How to Initialize The Module
To initialize the aWire UART user interface the user must first enable the clock by writing a one
to the Clock Enable bit in the Clock Request Register (CLKR.CLKEN) and wait for the Clock
Enable bit in the Status Register (SR.CENABLED) to be set. After doing this either receive,
transmit or receive with resync must be selected by writing the corresponding value into the
Mode field of the Control (CTRL.MODE) Register. Due to the RC120M being asynchronous with
the system clock values must be allowed to propagate in the system. During this time the aWire
master will set the Busy bit in the Status Register (SR.BUSY).
After the SR.BUSY bit is cleared the Baud Rate field in the Baud Rate Register (BRR.BR) can
be written with the wanted baudrate ( f br ) according to the following formula ( f aw is the RC120M
clock frequency):
8f aw
f br = ----------BR
After this operation the user must wait until the SR.BUSY is cleared. The interface is now ready
to be used.
30.6.2
Basic Asynchronous Receiver Operation
The aWire UART user interface must be initialized according to the sequence above, but the
CTRL.MODE field must be written to one (Receive mode).
When a data byte arrives the aWire UART user interface will indicate this by setting the Data
Ready Interrupt bit in the Status Register (SR.DREADYINT). The user must read the Data in the
Receive Holding Register (RHR.RXDATA) and clear the Interrupt bit by writing a one to the Data
Ready Interrupt Clear bit in the Status Clear Register (SCR.DREADYINT). The interface is now
ready to receive another byte.
30.6.3
Basic Asynchronous Transmitter Operation
The aWire UART user interface must be initialized according to the sequence above, but the
CTRL.MODE field must be written to two (Transmit mode).
To transmit a data byte the user must write the data to the Transmit Holding Register
(THE.TXDATA). Before the next byte can be written the SR.BUSY must be cleared.
30.6.4
Basic Asynchronous Receiver with Resynchronization
By writing three into CTRL.MODE the aWire UART user interface will assume that the first byte
it receives is a sync byte (0x55) and set BRR.BR according to this. All subsequent transfers will
assume this baudrate, unless BRR.BR is rewritten by the user.
To make the aWire UART user interface accept a new sync resynchronization the aWire UART
user interface must be disabled by writing zero to CTRL.MODE and then reenable the interface.
30.6.5
Overrun
In Receive mode an overrun can occur if the user has not read the previous received data from
the RHR.RXDATA when the newest data should be placed there. Such a condition is flagged by
setting the Overrun bit in the Status Register (SR.OVERRUN). If SR.OVERRUN is set the newest data received is placed in RHR.RXDATA and the data that was there before is overwritten.
700
32099I–01/2012
AT32UC3L016/32/64
30.6.6
Interrupts
To make the CPU able to do other things while waiting for the aWire UART user interface to finish its operations the aWire UART user interface supports generating interrupts. All status bits in
the Status Register can be used as interrupt sources, except the SR.BUSY and SR.CENABLED
bits.
To enable an interrupt the user must write a one to the corresponding bit in the Interrupt Enable
Register (IER). Upon the next zero to one transition of this SR bit the aWire UART user interface
will flag this interrupt to the CPU. To clear the interrupt the user must write a one to the corresponding bit in the Status Clear Register (SCR).
Interrupts can be disabled by writing a one to the corresponding bit in the Interrupt Disable Register (IDR). The interrupt Mask Register (IMR) can be read to check if an interrupt is enabled or
disabled.
30.6.7
Using the Peripheral DMA Controller
To relieve the CPU of data transfers the aWire UART user interface support using the Peripheral
DMA controller.
To transmit using the Peripheral DMA Controller do the following:
1. Setup the aWire UART user interface in transmit mode.
2. Setup the Peripheral DMA Controller with buffer address and length, use byte as transfer size.
3. Enable the Peripheral DMA Controller.
4. Wait until the Peripheral DMA Controller is done.
To receive using the Peripheral DMA Controller do the following:
1. Setup the aWire UART user interface in receive mode
2. Setup the Peripheral DMA Controller with buffer address and length, use byte as transfer size.
3. Enable the Peripheral DMA Controller.
4. Wait until the Peripheral DMA Controller is ready.
701
32099I–01/2012
AT32UC3L016/32/64
30.7
User Interface
Table 30-2.
Note:
aWire UART user interface Register Memory Map
Offset
Register
Register Name
Access
Reset
0x00
Control Register
CTRL
Read/Write
0x00000000
0x04
Status Register
SR
Read-only
0x00000000
0x08
Status Clear Register
SCR
Write-only
-
0x0C
Interrupt Enable Register
IER
Write-only
-
0x10
Interrupt Disable Register
IDR
Write-only
-
0x14
Interrupt Mask Register
IMR
Read-only
0x00000000
0x18
Receive Holding Register
RHR
Read-only
0x00000000
0x1C
Transmit Holding Register
THR
Read/Write
0x00000000
0x20
Baud Rate Register
BRR
Read/Write
0x00000000
0x24
Version Register
VERSION
Read-only
-(1)
0x28
Clock Request Register
CLKR
Read/Write
0x00000000
1. The reset values are device specific. Please refer to the Module Configuration section at the end of this chapter.
702
32099I–01/2012
AT32UC3L016/32/64
30.7.1
Name:
Control Register
CTRL
Access Type:
Read/Write
Offset:
0x00
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: aWire UART user interface mode
Table 30-3.
aWire UART user interface Modes
MODE
Mode Description
0
Disabled
1
Receive
2
Transmit
3
Receive with resync.
703
32099I–01/2012
AT32UC3L016/32/64
30.7.2
Name:
Status Register
SR
Access Type:
Read-only
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
-
-
TRMIS
-
-
OVERRUN
DREADYINT
READYINT
7
6
5
4
3
2
1
0
-
-
-
-
-
CENABLED
-
BUSY
• TRMIS: Transmit Mismatch
0: No transfers mismatches.
1: The transceiver was active when receiving.
This bit is set when the transceiver is active when receiving.
This bit is cleared when corresponding bit in SCR is written to one.
• OVERRUN: Data Overrun
0: No data overwritten in RHR.
1: Data in RHR has been overwritten before it has been read.
This bit is set when data in RHR is overwritten before it has been read.
This bit is cleared when corresponding bit in SCR is written to one.
• DREADYINT: Data Ready Interrupt
0: No new data in the RHR.
1: New data received and placed in the RHR.
This bit is set when new data is received and placed in the RHR.
This bit is cleared when corresponding bit in SCR is written to one.
• READYINT: Ready Interrupt
0: The interface has not generated an ready interrupt.
1: The interface has had a transition from busy to not busy.
This bit is set when the interface has transition from busy to not busy.
This bit is cleared when corresponding bit in SCR is written to one.
• CENABLED: Clock Enabled
0: The aWire clock is not enabled.
1: The aWire clock is enabled.
704
32099I–01/2012
AT32UC3L016/32/64
This bit is set when the clock is disabled.
This bit is cleared when the clock is enabled.
• BUSY: Synchronizer Busy
0: The asynchronous interface is ready to accept more data.
1: The asynchronous interface is busy and will block writes to CTRL, BRR, and THR.
This bit is set when the asynchronous interface becomes busy.
This bit is cleared when the asynchronous interface becomes ready.
705
32099I–01/2012
AT32UC3L016/32/64
30.7.3
Name:
Status Clear Register
SCR
Access Type:
Write-only
Offset:
0x08
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
-
-
TRMIS
-
-
OVERRUN
DREADYINT
READYINT
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
-
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in SR and the corresponding interrupt request.
706
32099I–01/2012
AT32UC3L016/32/64
30.7.4
Name:
Interrupt Enable Register
IER
Access Type:
Write-only
Offset:
0x0C
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
-
-
TRMIS
-
-
OVERRUN
DREADYINT
READYINT
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
-
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will set the corresponding bit in IMR.
707
32099I–01/2012
AT32UC3L016/32/64
30.7.5
Name:
Interrupt Disable Register
IDR
Access Type:
Write-only
Offset:
0x10
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
-
-
TRMIS
-
-
OVERRUN
DREADYINT
READYINT
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
-
Writing a zero to a bit in this register has no effect.
Writing a one to a bit in this register will clear the corresponding bit in IMR.
708
32099I–01/2012
AT32UC3L016/32/64
30.7.6
Name:
Interrupt Mask Register
IMR
Access Type:
Read-only
Offset:
0x14
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
-
-
TRMIS
-
-
OVERRUN
DREADYINT
READYINT
7
6
5
4
3
2
1
0
-
-
-
-
-
-
-
-
0: The corresponding interrupt is disabled.
1: The corresponding interrupt is enabled.
A bit in this register is cleared when the corresponding bit in IDR is written to one.
A bit in this register is set when the corresponding bit in IER is written to one.
709
32099I–01/2012
AT32UC3L016/32/64
30.7.7
Name:
Receive Holding Register
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
14
13
12
11
10
9
8
-
-
-
-
-
-
-
-
7
6
5
4
3
2
1
0
RXDATA
• RXDATA: Received Data
The last byte received.
710
32099I–01/2012
AT32UC3L016/32/64
30.7.8
Name:
Transmit Holding Register
THR
Access Type:
Read/Write
Offset:
0x1C
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: Transmit Data
The data to send.
711
32099I–01/2012
AT32UC3L016/32/64
30.7.9
Name:
Baud Rate Register
BRR
Access Type:
Read/Write
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
8
3
2
1
0
BR[15:8]
7
6
5
4
BR[7:0]
• BR: Baud Rate
The baud rate ( f br ) of the transmission, calculated using the following formula ( f aw is the RC120M frequency):
8f aw
f br = ----------BR
BR should not be set to a value smaller than 32.
Writing a value to this field will update the baud rate of the transmission.
Reading this field will give the current baud rate of the transmission.
712
32099I–01/2012
AT32UC3L016/32/64
30.7.10
Name:
Version Register
VERSION
Access Type:
Read-only
Offset:
0x24
Reset Value:
0x00000200
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
VERSION[11:8]
3
2
1
0
VERSION[7:0]
• VERSION: Version Number
Version number of the module. No functionality associated.
713
32099I–01/2012
AT32UC3L016/32/64
30.7.11
Name:
Clock Request Register
CLKR
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
-
-
-
-
-
-
-
CLKEN
• CLKEN: Clock Enable
0: The aWire clock is disabled.
1: The aWire clock is enabled.
Writing a zero to this bit will disable the aWire clock.
Writing a one to this bit will enable the aWire clock.
714
32099I–01/2012
AT32UC3L016/32/64
30.8
Module Configuration
The specific configuration for each aWire instance is listed in the following tables.The module
bus clocks listed here are connected to the system bus clocks. Please refer to the Power Manager chapter for details.
Table 30-4.
AW Clocks
Clock Name
Description
CLK_AW
Clock for the AW bus interface
Table 30-5.
Register Reset Values
Register
Reset Value
VERSION
0x00000210
715
32099I–01/2012
AT32UC3L016/32/64
31. Programming and Debugging
31.1
Overview
The AT32UC3L016/32/64 supports programming and debugging through two interfaces, JTAG
and aWire. JTAG is an industry standard interface and allows boundary scan for PCB testing, as
well as daisy-chaining of multiple devices on the PCB. aWire is an Atmel proprietary protocol
which offers higher throughput and robust communication, and does not require application pins
to be reserved. Either interface provides access to the internal Service Access Bus (SAB), which
offers a bridge to the High Speed Bus, giving access to memories and peripherals in the device.
By using this bridge to the bus system, the flash and fuses can thus be programmed by accessing the Flash Controller in the same manner as the CPU.
The SAB also provides access to the Nexus-compliant On-Chip Debug (OCD) system in the
device, which gives the user non-intrusive run-time control of the program execution. Additionally, trace information can be output on the Auxiliary (AUX) debug port or buffered in internal
RAM for later retrieval by JTAG or aWire.
31.2
Service Access Bus
The AVR32 architecture offers a common interface for access to On-Chip Debug, programming,
and test functions. These are mapped on a common bus called the Service Access Bus (SAB),
which is linked to the JTAG and aWire port through a bus master module, which also handles
synchronization between the debugger and SAB clocks.
When accessing the SAB through the debugger there are no limitations on debugger frequency
compared to device frequency, although there must be an active system clock in order for the
SAB accesses to complete. If the system clock is switched off in sleep mode, activity on the
debugger will restart the system clock automatically, without waking the device from sleep.
Debuggers may optimize the transfer rate by adjusting the frequency in relation to the system
clock. This ratio can be measured with debug protocol specific instructions.
The Service Access Bus uses 36 address bits to address memory or registers in any of the
slaves on the bus. The bus supports sized accesses of bytes (8 bits), halfwords (16 bits), or
words (32 bits). All accesses must be aligned to the size of the access, i.e. halfword accesses
must have the lowest address bit cleared, and word accesses must have the two lowest address
bits cleared.
31.2.1
SAB Address Map
The 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 31-1.
Table 31-1.
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
716
32099I–01/2012
AT32UC3L016/32/64
Table 31-1.
31.2.2
SAB Slaves, Addresses and Descriptions
Slave
Address [35:32]
Description
HSB
0x5
Alternative mapping for HSB space, for compatibility with
other 32-bit AVR devices.
Memory Service
Unit
0x6
Memory Service Unit registers
Reserved
Other
Unused
SAB Security Restrictions
The Service Access bus can be restricted by internal security measures. A short description of
the security measures are found in the table below.
31.2.2.1
Security measure and control location
A security measure is a mechanism to either block or allow SAB access to a certain address or
address range. A security measure is enabled or disabled by one or several control signals. This
is called the control location for the security measure.
These security measures can be used to prevent an end user from reading out the code programmed in the flash, for instance.
Table 31-2.
SAB Security Measures
Security Measure
Control Location
Description
Secure mode
FLASHCDW
SECURE bits set
Allocates a portion of the flash for secure code. This code
cannot be read or debugged. The User page is also locked.
Security bit
FLASHCDW
security bit set
Programming and debugging not possible, very restricted
access.
User code
programming
FLASHCDW
UPROT + security
bit set
Restricts all access except parts of the flash and the flash
controller for programming user code. Debugging is not
possible unless an OS running from the secure part of the
flash supports it.
Below follows a more in depth description of what locations are accessible when the security
measures are active.
Table 31-3.
Secure Mode SAB Restrictions
Name
Address Start
Address End
Access
Secure flash area
0x580000000
0x580000000 +
(USERPAGE[15:0] << 10)
Blocked
Secure RAM area
0x500000000
0x500000000 +
(USERPAGE[31:16] << 10)
Blocked
User page
0x580800000
0x581000000
Other accesses
-
-
Note:
Read
As normal
1. Second Word of the User Page, refer to the Fuses Settings section for details.
717
32099I–01/2012
AT32UC3L016/32/64
Table 31-4.
Security Bit SAB Restrictions
Name
Address start
Address end
Access
OCD DCCPU,
OCD DCEMU,
OCD DCSR
0x100000110
0x100000118
Read/Write
User page
0x580800000
0x581000000
Read
Other accesses
-
-
Table 31-5.
Blocked
User Code Programming SAB Restrictions
Name
Address start
Address end
Access
OCD DCCPU,
OCD DCEMU,
OCD DCSR
0x100000110
0x100000118
Read/Write
User page
0x580800000
0x581000000
Read
FLASHCDW PB
interface
0x5FFFE0000
0x5FFFE0400
Read/Write
FLASH pages
outside
BOOTPROT
0x580000000 +
BOOTPROT size
0x580000000 + Flash size
Read/Write
Other accesses
-
-
Blocked
718
32099I–01/2012
AT32UC3L016/32/64
31.3
On-Chip Debug
Rev: 2.1.0.0
31.3.1
Features
•
•
•
•
•
•
•
•
31.3.2
Debug interface in compliance with IEEE-ISTO 5001-2003 (Nexus 2.0) Class 2+
JTAG or aWire access to all on-chip debug functions
Advanced Program, Data, Ownership, and Watchpoint trace supported
NanoTrace aWire- or 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 AT32UC3L016/32/64 is facilitated by a powerful On-Chip Debug (OCD) system. The user accesses this through an external debug tool which connects to the JTAG or
aWire port and the Auxiliary (AUX) port if implemented. The AUX port is primarily used for trace
functions, and an aWire- or 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 AT32UC3L016/32/64 implements several useful OCD features, such as:
• Debug Communication Channel between CPU and debugger
• 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 the debugger, for
instance 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.
31.3.3
I/O Lines Description
The OCD AUX trace port contains a number of pins, as shown in Table 31-6 on page 720.
These are multiplexed with I/O Controller lines, and must explicitly be enabled by writing OCD
registers before the debug session starts. The AUX port is mapped to two different locations,
719
32099I–01/2012
AT32UC3L016/32/64
selectable by OCD Registers, minimizing the chance that the AUX port will need to be shared
with an application.
Table 31-6.
31.3.4
Auxiliary Port Signals
Pin Name
Pin Description
Direction
Active Level
Type
MCKO
Trace data output clock
Output
Digital
MDO[5:0]
Trace data output
Output
Digital
MSEO[1:0]
Trace frame control
Output
Digital
EVTI_N
Event In
EVTO_N
Event Out
Input
Low
Digital
Output
Low
Digital
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
31.3.4.1
Power Management
The OCD clock operates independently of the CPU clock. If enabled in the Power Manager, the
OCD clock (CLK_OCD) will continue running even if the CPU enters a sleep mode that disables
the CPU clock.
31.3.4.2
Clocks
The OCD has a clock (CLK_OCD) running synchronously with the CPU clock. This clock is generated by the Power Manager. The clock is enabled at reset, and can be disabled by writing to
the Power Manager.
31.3.4.3
Interrupt
The OCD system interrupt request lines are connected to the interrupt controller. Using the OCD
interrupts requires the interrupt controller to be programmed first.
720
32099I–01/2012
AT32UC3L016/32/64
31.3.5
Block Diagram
Figure 31-1. On-Chip Debug Block Diagram
aWire
JTAG
aWire
JTAG
AUX
On-Chip Debug
Memory
Service
Unit
Service Access Bus
Transmit Queue
Watchpoints
Debug PC
Debug
Instruction
Breakpoints
CPU
31.3.6
Program
Trace
Internal
SRAM
HSB Bus Matrix
Data Trace
Ownership
Trace
Memories and
peripherals
SAB-based Debug Features
A debugger can control all OCD features by writing OCD registers over the SAB interface. Many
of these do not depend on output on the AUX port, allowing an aWire- or 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.
An aWire-based debugger should connect to the device through the RESET_N pin.
721
32099I–01/2012
AT32UC3L016/32/64
Figure 31-2. JTAG-based Debugger
PC
JTAG -based
debug tool
10-pin IDC
JTAG
32-bit AVR
Figure 31-3. aWire-based Debugger
PC
aWire-based
debug tool
aWire
32-bit AVR
31.3.6.1
Debug Communication Channel
The Debug Communication Channel (DCC) consists of a pair OCD registers with associated
handshake logic, accessible to both CPU and debugger. The registers can be used to exchange
data between the CPU and the debugmaster, both runtime as well as in debug mode.
722
32099I–01/2012
AT32UC3L016/32/64
The OCD system can generate an interrupt to the CPU when DCCPU is read and when DCEMU
is written. This enables the user to build a custum debug protocol using only these registers. The
DCCPU and DCEMU registers are available even when the security bit in the flash is active.
For more information refer to the AVR32UC Technical Reference Manual.
31.3.6.2
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 the debugger, 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 the debugger, or Monitor Mode, running instructions from program memory.
31.3.6.3
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 the debugger, the instruction is executed, allowing the debugger to execute CPU instructions directly. The debug 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.
31.3.6.4
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 the debug master.
31.3.6.5
Program Counter Monitoring
Normally, the CPU would need to be halted for a debugger to examine the current PC value.
However, the AT32UC3L016/32/64 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.
31.3.7
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 Service Access Bus.
723
32099I–01/2012
AT32UC3L016/32/64
31.3.7.1
Cyclic Redundancy Check (CRC)
The MSU can be used to automatically calculate the CRC of a block of data in memory. The
MSU will then read out each word in the specified memory block and report the CRC32-value in
an MSU register.
31.3.7.2
NanoTrace
The MSU additionally supports NanoTrace. This is a 32-bit AVR-specific feature, in which trace
data is output to memory instead of the AUX port. This allows the trace data to be extracted by
the debugger through the SAB, enabling trace features for aWire- or 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.
31.3.8
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.
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.
724
32099I–01/2012
AT32UC3L016/32/64
Figure 31-4. AUX+JTAG Based Debugger
PC
T ra c e b u ffe r
A U X +JTA G
d e b u g to o l
M ic t o r 3 8
AUX
h ig h s p e e d
JTA G
AVR 32
31.3.8.1
Trace Operation
Trace features are enabled by writing OCD registers by the debugger. 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.
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.
31.3.8.2
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.
31.3.8.3
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 AT32UC3L016/32/64 contains two data trace channels,
725
32099I–01/2012
AT32UC3L016/32/64
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.
31.3.8.4
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.
31.3.8.5
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.
31.3.8.6
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.
31.3.8.7
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 AT32UC3L016/32/64 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.
726
32099I–01/2012
AT32UC3L016/32/64
31.3.9
Module Configuration
The bit mapping of the Peripheral Debug Register (PDBG) is described in Table 31-7. Please
refer to the On-chip Debug chapter in the AVR32UC Technical Reference Manual for details.
Table 31-7.
Bit mapping of the Peripheral Debug Register (PDBG)
Bit
Peripheral
0
AST
1
WDT
727
32099I–01/2012
AT32UC3L016/32/64
31.4
JTAG and Boundary-scan (JTAG)
Rev: 2.2.1.4
31.4.1
Features
• IEEE1149.1 compliant JTAG Interface
• Boundary-scan Chain for board-level testing
• Direct memory access and programming capabilities through JTAG Interface
31.4.2
Overview
The JTAG Interface offers a four pin programming and debug solution, including boundary-scan
support for board-level testing.
Figure 31-5 on page 729 shows how the JTAG is connected in an 32-bit AVR 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
Device Identification 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 described
in Section 31.4.11.
Section 31.5 lists the supported JTAG instructions, with references to the description in this
document.
728
32099I–01/2012
AT32UC3L016/32/64
31.4.3
Block Diagram
Figure 31-5. JTAG and Boundary-scan Access
32-bit AVR device
JTAG
JTAG master
Boundary scan enable
TAP
Controller
TDO TDI
JTAG Pins
TMS TCK
TCK
TMS
TDI
TDO
Instruction register
scan enable
Data register
scan enable
Instruction Register
TMS TCK
TDO TDI
JTAG data registers
2nd JTAG
device
Device Identification
Register
Boundary Scan Chain
Pins and analog blocks
By-pass Register
Reset Register
Part specific registers
...
Service Access Bus
interface
SAB
31.4.4
Internal I/O
lines
I/O Lines Description
Table 31-8.
I/O Line Description
Pin Name
Pin Description
Type
Active Level
RESET_N
External reset pin. Used when enabling and disabling the JTAG.
Input
Low
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
31.4.5
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
729
32099I–01/2012
AT32UC3L016/32/64
31.4.5.1
I/O Lines
The TMS, TDI, TDO, and TCK pins are multiplexed with I/O lines. When the JTAG is used the
associated pins must be enabled. To enable the JTAG pins, refer to Section 31.4.7.
While using the multiplexed JTAG lines all normal peripheral activity on these lines is disabled.
The user must make sure that no external peripheral is blocking the JTAG lines while
debugging.
31.4.5.2
Power Management
When an instruction that accesses the SAB is loaded in the instruction register, before entering
a sleep mode, the system clocks are not switched off to allow debugging in sleep modes. This
can lead to a program behaving differently when debugging.
31.4.5.3
Clocks
The JTAG Interface uses the external TCK pin as clock source. This clock must be provided by
the JTAG master.
Instructions that use the SAB bus requires the internal main clock to be running.
31.4.6
JTAG Interface
The JTAG Interface is accessed through the dedicated JTAG pins shown in Table 31-8 on page
729. The TMS control line navigates the TAP controller, as shown in Figure 31-6 on page 731.
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.
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 31.5.2, as well as a number of 32-bit AVR-specific private JTAG instructions described in Section 31.5.3. Each
instruction selects a specific data register for the Shift-DR path, as described for each
instruction.
730
32099I–01/2012
AT32UC3L016/32/64
Figure 31-6. TAP Controller State Diagram
1
Test-LogicReset
0
0
Run-Test/
Idle
1
Select-DR
Scan
1
Select-IR
Scan
1
0
1
0
Capture-DR
1
0
Shift-DR
0
0
Shift-IR
1
1
Exit1-DR
Exit1-IR
0
0
Pause-DR
1
0
Exit2-DR
0
Pause-IR
1
1
0
1
1
Capture-IR
Update-DR
0
0
0
1
Exit2-IR
1
1
Update-IR
0
731
32099I–01/2012
AT32UC3L016/32/64
31.4.7
How to Initialize the Module
To enable the JTAG pins the TCK pin must be held low while the RESET_N pin is released.
After enabling the JTAG interface the halt bit is set automatically to prevent the system from running code after the interface is enabled. To make the CPU run again set halt to zero using the
HALT command..
JTAG operation when RESET_N is pulled low is not possible.
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 and after enabling the JTAG pins 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 Run-Test/Idle state, which is the starting point for JTAG operations.
31.4.8
How to disable the module
To disable the JTAG pins the TCK pin must be held high while RESET_N pin is released.
31.4.9
Typical Sequence
Assuming Run-Test/Idle is the present state, a typical scenario for using the JTAG Interface
follows.
31.4.9.1
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. During shifting,
the JTAG outputs status bits on TDO, refer to Section 31.5 for a description of these. 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.
Figure 31-7. Scanning in JTAG Instruction
TCK
TAP State
TLR
RTI
SelDR SelIR CapIR ShIR
Ex1IR UpdIR RTI
TMS
TDI
TDO
31.4.9.2
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
732
32099I–01/2012
AT32UC3L016/32/64
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.
31.4.10
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 32-bit AVR 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. 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
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. The internal chip clock is not required to run during boundary-scan
operations.
NOTE: For pins connected to 5V lines care should be taken to not drive the pins to a logic one
using boundary-scan, as this will create a current flowing from the 3,3V driver to the 5V pull-up
on the line. Optionally a series resistor can be added between the line and the pin to reduce the
current.
Details about the boundary-scan chain can be found in the BSDL file for the device. This can be
found on the Atmel website.
31.4.11
Service Access Bus
The AVR32 architecture offers a common interface for access to On-Chip Debug, programming,
and test functions. These are mapped on a common bus called the Service Access Bus (SAB),
733
32099I–01/2012
AT32UC3L016/32/64
which is linked to the JTAG through a bus master module, which also handles synchronization
between the TCK and SAB clocks.
For more information about the SAB and a list of SAB slaves see the Service Access Bus
chapter.
31.4.11.1
SAB Address Mode
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.
31.4.11.2
Block Transfer
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.
31.4.11.3
Canceling a SAB Access
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.
31.4.11.4
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.
• 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
734
32099I–01/2012
AT32UC3L016/32/64
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 is invalid. The SAB stays in data mode. Repeat
scanning until the busy bit clears.
• During Shift-DR of write data: The write data is ignored. The SAB stays in data mode. Repeat
scanning until the busy bit clears.
31.4.11.5
Error Reporting
The Service Access Bus 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 is invalid.
• During Shift-DR of write data: The previous write operation failed. The new data is 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.
• After power-up: The error bit is set after power up, but there has been no previous SAB
instruction so this error can be discarded.
31.4.11.6
Protected Reporting
A protected status may be reported during Shift-IR or Shift-DR. This indicates that the security
bit in the Flash Controller is set and that the chip is locked for access, according to Section
31.5.1.
The protected state is reported when:
• The Flash Controller is under reset. This can be due to the AVR_RESET command or the
RESET_N line.
• The Flash Controller has not read the security bit from the flash yet (This will take a a few
ms). Happens after the Flash Controller reset has been released.
• The security bit in the Flash Controller is set.
What to do if the protected bit is set:
• Release all active AVR_RESET domains, if any.
• Release the RESET_N line.
• Wait a few ms for the security bit to clear. It can be set temporarily due to a reset.
735
32099I–01/2012
AT32UC3L016/32/64
• Perform a CHIP_ERASE to clear the security bit. NOTE: This will erase all the contents of the
non-volatile memory.
31.5
JTAG Instruction Summary
The implemented JTAG instructions in the 32-bit AVR are shown in the table below.
Table 31-9.
Instruction
OPCODE
JTAG Instruction Summary
Instruction
Description
0x01
IDCODE
Select the 32-bit Device Identification register as data register.
0x02
SAMPLE_PRELOAD
Take a snapshot of external pin values without affecting system operation.
0x03
EXTEST
Select boundary-scan chain as data register for testing circuitry external to
the device.
0x04
INTEST
Select boundary-scan chain for internal testing of the device.
0x06
CLAMP
Bypass device through Bypass register, while driving outputs from boundaryscan register.
0x0C
AVR_RESET
Apply or remove a static reset to the device
0x0F
CHIP_ERASE
Erase the device
0x10
NEXUS_ACCESS
Select the SAB Address and Data registers as data register for the TAP. The
registers are accessed in Nexus mode.
0x11
MEMORY_WORD_ACCESS
Select the SAB Address and Data registers as data register for the TAP.
0x12
MEMORY_BLOCK_ACCESS
Select the SAB Data register as data register for the TAP. The address is
auto-incremented.
0x13
CANCEL_ACCESS
Cancel an ongoing Nexus or Memory access.
0x14
MEMORY_SERVICE
Select the SAB Address and Data registers as data register for the TAP. The
registers are accessed in Memory Service mode.
0x15
MEMORY_SIZED_ACCESS
Select the SAB Address and Data registers as data register for the TAP.
0x17
SYNC
Synchronization counter
0x1C
HALT
Halt the CPU for safe programming.
0x1F
BYPASS
Bypass this device through the bypass register.
N/A
Acts as BYPASS
Others
31.5.1
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
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.
736
32099I–01/2012
AT32UC3L016/32/64
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.
31.5.1.1
Notation
Table 31-11 on page 737 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 least
significantbit is always shifted first, and the most significant bit shifted last. The symbols used
are shown in Table 31-10.
Table 31-10. Symbol Description
Symbol
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 31.4.11.4 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 31.4.11.5 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
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 31-11. 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
737
32099I–01/2012
AT32UC3L016/32/64
Table 31-11. Instruction Description (Continued)
31.5.2
Instruction
Description
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
Public JTAG Instructions
The JTAG standard defines a number of public JTAG instructions. These instructions are
described in the sections below.
31.5.2.1
IDCODE
This instruction selects the 32 bit Device Identification register (DID) as Data Register. The DID
register consists of a version number, a device number, and the manufacturer code chosen by
JEDEC. This is the default instruction after a JTAG reset. Details about the DID register can be
found in the module configuration section at the end of this chapter.
Starting in Run-Test/Idle, the Device Identification register is accessed in the following way:
1. Select the IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. Return to Run-Test/Idle.
5. Select the DR Scan path.
6. In Capture-DR: The IDCODE value is latched into the shift register.
7. In Shift-DR: The IDCODE scan chain is shifted by the TCK input.
8. Return to Run-Test/Idle.
Table 31-12. IDCODE Details
31.5.2.2
Instructions
Details
IR input value
00001 (0x01)
IR output value
p0001
DR Size
32
DR input value
xxxxxxxx xxxxxxxx xxxxxxxx xxxxxxxx
DR output value
Device Identification Register
SAMPLE_PRELOAD
This instruction takes 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.
Starting in Run-Test/Idle, the Device Identification register is accessed in the following way:
738
32099I–01/2012
AT32UC3L016/32/64
1. Select the IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. Return to Run-Test/Idle.
5. Select the DR Scan path.
6. In Capture-DR: The Data on the external pins are sampled into the boundary-scan
chain.
7. In Shift-DR: The boundary-scan chain is shifted by the TCK input.
8. Return to Run-Test/Idle.
Table 31-13. SAMPLE_PRELOAD Details
31.5.2.3
Instructions
Details
IR input value
00010 (0x02)
IR output value
p0001
DR Size
Depending on boundary-scan chain, see BSDL-file.
DR input value
Depending on boundary-scan chain, see BSDL-file.
DR output value
Depending on boundary-scan chain, see BSDL-file.
EXTEST
This instruction selects the boundary-scan chain as Data Register for testing circuitry external to
the 32-bit AVR 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.
Starting in Run-Test/Idle, the EXTEST instruction is accessed the following way:
1. Select the IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. In Update-IR: The data from the boundary-scan chain is applied to the output pins.
5. Return to Run-Test/Idle.
6. Select the DR Scan path.
7. In Capture-DR: The data on the external pins is sampled into the boundary-scan chain.
8. In Shift-DR: The boundary-scan chain is shifted by the TCK input.
9. In Update-DR: The data from the scan chain is applied to the output pins.
10. Return to Run-Test/Idle.
Table 31-14. EXTEST Details
Instructions
Details
IR input value
00011 (0x03)
IR output value
p0001
DR Size
Depending on boundary-scan chain, see BSDL-file.
DR input value
Depending on boundary-scan chain, see BSDL-file.
DR output value
Depending on boundary-scan chain, see BSDL-file.
739
32099I–01/2012
AT32UC3L016/32/64
31.5.2.4
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 boundary-scan
chain.
Starting in Run-Test/Idle, the INTEST instruction is accessed the following way:
1. Select the IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. In Update-IR: The data from the boundary-scan chain is applied to the internal logic
inputs.
5. Return to Run-Test/Idle.
6. Select the DR Scan path.
7. In Capture-DR: The data on the internal logic is sampled into the boundary-scan chain.
8. In Shift-DR: The boundary-scan chain is shifted by the TCK input.
9. In Update-DR: The data from the boundary-scan chain is applied to internal logic
inputs.
10. Return to Run-Test/Idle.
Table 31-15. INTEST Details
31.5.2.5
Instructions
Details
IR input value
00100 (0x04)
IR output value
p0001
DR Size
Depending on boundary-scan chain, see BSDL-file.
DR input value
Depending on boundary-scan chain, see BSDL-file.
DR output value
Depending on boundary-scan chain, see BSDL-file.
CLAMP
This instruction selects the Bypass register as Data Register. The device output pins are driven
from the boundary-scan chain.
Starting in Run-Test/Idle, the CLAMP instruction is accessed the following way:
1. Select the IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. In Update-IR: The data from the boundary-scan chain is applied to the output pins.
5. Return to Run-Test/Idle.
6. Select the DR Scan path.
7. In Capture-DR: A logic ‘0’ is loaded into the Bypass Register.
8. In Shift-DR: Data is scanned from TDI to TDO through the Bypass register.
740
32099I–01/2012
AT32UC3L016/32/64
9. Return to Run-Test/Idle.
Table 31-16. CLAMP Details
31.5.2.6
Instructions
Details
IR input value
00110 (0x06)
IR output value
p0001
DR Size
1
DR input value
x
DR output value
x
BYPASS
This instruction selects the 1-bit Bypass Register as Data Register.
Starting in Run-Test/Idle, the CLAMP instruction is accessed the following way:
1. Select the IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. Return to Run-Test/Idle.
5. Select the DR Scan path.
6. In Capture-DR: A logic ‘0’ is loaded into the Bypass Register.
7. In Shift-DR: Data is scanned from TDI to TDO through the Bypass register.
8. Return to Run-Test/Idle.
Table 31-17. BYPASS Details
31.5.3
Instructions
Details
IR input value
11111 (0x1F)
IR output value
p0001
DR Size
1
DR input value
x
DR output value
x
Private JTAG Instructions
The 32-bit AVR defines a number of private JTAG instructions, not defined by the JTAG standard. Each instruction is briefly described in text, with details following in table form.
31.5.3.1
NEXUS_ACCESS
This instruction allows Nexus-compliant access to the On-Chip Debug registers through the
SAB. The 7-bit register index, 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.
NOTE: The polarity of the direction bit is inverse of the Nexus standard.
741
32099I–01/2012
AT32UC3L016/32/64
Starting in Run-Test/Idle, OCD registers are accessed in the following way:
1. Select the IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. Return to Run-Test/Idle.
5. Select the DR Scan path.
6. In Shift-DR: Scan in the direction bit (1=read, 0=write) and the 7-bit address for the
OCD register.
7. Go to Update-DR and re-enter Select-DR Scan.
8. In Shift-DR: For a read operation, scan out the contents of the addressed register. For a
write operation, scan in the new contents of the register.
9. Return to Run-Test/Idle.
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 31-18. NEXUS_ACCESS Details
31.5.3.2
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. The 7-bit register
index, 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 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 IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. Return to Run-Test/Idle.
5. Select the DR Scan path.
6. In Shift-DR: Scan in the direction bit (1=read, 0=write) and the 7-bit address for the
Memory Service register.
742
32099I–01/2012
AT32UC3L016/32/64
7. Go to Update-DR and re-enter Select-DR Scan.
8. In Shift-DR: For a read operation, scan out the contents of the addressed register. For a
write operation, scan in the new contents of the register.
9. Return to Run-Test/Idle.
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 31-19. MEMORY_SERVICE Details
31.5.3.3
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 is 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.
743
32099I–01/2012
AT32UC3L016/32/64
The size field is encoded as i Table 31-20.
Table 31-20. Size Field Semantics
Size field value
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
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
00
01
Starting in Run-Test/Idle, SAB data is accessed in the following way:
1. Select the IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. Return to Run-Test/Idle.
5. Select the DR Scan path.
6. In Shift-DR: Scan in the direction bit (1=read, 0=write), 2-bit access size, and the 36-bit
address of the data to access.
7. Go to Update-DR and re-enter Select-DR Scan.
8. In Shift-DR: For a read operation, scan out the contents of the addressed area. For a
write operation, scan in the new contents of the area.
9. 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 31-21. 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
744
32099I–01/2012
AT32UC3L016/32/64
Table 31-21. MEMORY_SIZED_ACCESS Details (Continued)
31.5.3.4
Instructions
Details
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 is accessed
through the 34 MSB of the SAB address, 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 and not scanned in.
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 is accessed in the following way:
1. Select the IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. Return to Run-Test/Idle.
5. Select the DR Scan path.
6. In Shift-DR: Scan in the direction bit (1=read, 0=write) and the 34-bit address of the
data to access.
7. Go to Update-DR and re-enter Select-DR Scan.
8. In Shift-DR: For a read operation, scan out the contents of the addressed area. For a
write operation, scan in the new contents of the area.
9. 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 31-22. 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
745
32099I–01/2012
AT32UC3L016/32/64
Table 31-22. MEMORY_WORD_ACCESS Details (Continued)
31.5.3.5
Instructions
Details
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 is 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:
1. Use the MEMORY_SIZE_ACCESS or MEMORY_WORD_ACCESS to read or write the
first location.
2. Return to Run-Test/Idle.
3. Select the IR Scan path.
4. In Capture-IR: The IR output value is latched into the shift register.
5. In Shift-IR: The instruction register is shifted by the TCK input.
6. Return to Run-Test/Idle.
7. 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).
8. In Shift-DR: 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.
9. Go to Update-DR.
10. If the block access is not complete, return to Select-DR Scan and repeat the access.
11. 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 31-23. 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
746
32099I–01/2012
AT32UC3L016/32/64
Table 31-23. MEMORY_BLOCK_ACCESS Details (Continued)
Instructions
Details
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.
31.5.3.6
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 JTAG master.
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.
Starting in Run-Test/Idle, CANCEL_ACCESS is accessed in the following way:
1. Select the IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. Return to Run-Test/Idle.
Table 31-24. CANCEL_ACCESS Details
31.5.3.7
Instructions
Details
IR input value
10011 (0x13)
IR output value
peb01
DR Size
1
DR input value
x
DR output value
0
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 IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. Return to Run-Test/Idle.
5. Select the DR Scan path.
747
32099I–01/2012
AT32UC3L016/32/64
6. Scan in an 16-bit counter value.
7. Go to Update-DR and re-enter Select-DR Scan.
8. In Shift-DR: Scan out the busy bit, and until the busy bit clears goto 7.
9. Calculate an approximation to the internal clock speed using the elapsed time and the
counter value.
10. 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 31-25. SYNC_ACCESS Details
31.5.3.8
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.
The AVR_RESET instruction can be used in the following way:
1. Select the IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. Return to Run-Test/Idle.
5. Select the DR Scan path.
6. In Shift-DR: Scan in the value corresponding to the reset domains the JTAG master
wants to reset into the data register.
7. Return to Run-Test/Idle.
8. Stay in run test idle for at least 10 TCK clock cycles to let the reset propagate to the
system.
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 31-26. AVR_RESET Details
Instructions
Details
IR input value
01100 (0x0C)
IR output value
p0001
748
32099I–01/2012
AT32UC3L016/32/64
Table 31-26. AVR_RESET Details (Continued)
31.5.3.9
Instructions
Details
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
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. Reset the system and stop the CPU from executing.
2. Select the IR Scan path.
3. In Capture-IR: The IR output value is latched into the shift register.
4. In Shift-IR: The instruction register is shifted by the TCK input.
5. Check the busy bit that was scanned out during Shift-IR. If the busy bit was set goto 2.
6. Return to Run-Test/Idle.
Table 31-27. CHIP_ERASE Details
31.5.3.10
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. Before releasing the halt command the CPU needs to be reset to ensure that
it will start at the reset startup address.
The HALT instruction can be used in the following way:
1. Select the IR Scan path.
2. In Capture-IR: The IR output value is latched into the shift register.
3. In Shift-IR: The instruction register is shifted by the TCK input.
4. Return to Run-Test/Idle.
5. Select the DR Scan path.
749
32099I–01/2012
AT32UC3L016/32/64
6. In Shift-DR: Scan in the value 1 to halt the CPU, 0 to start CPU execution.
7. Return to Run-Test/Idle.
Table 31-28. 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
750
32099I–01/2012
AT32UC3L016/32/64
31.5.4
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.
31.5.4.1
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.
MSB
Bit
LSB
31
Device ID
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 31-29.
Note that if the flash controller is statically reset, the ID code will be undefined.
Table 31-29. Device and JTAG ID
31.5.4.2
Device Name
JTAG ID Code (R is the revision number)
AT32UC3L064
0xR203003F
AT32UC3L032
0xR203403F
AT32UC3L016
0xR203803F
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.
Bit
Reset
domain
System
0
System
Resets the whole chip, except the JTAG itself.
751
32099I–01/2012
AT32UC3L016/32/64
31.5.4.3
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.
752
32099I–01/2012
AT32UC3L016/32/64
31.6
aWire Debug Interface (AW)
Rev.: 2.1.0.1
31.6.1
Features
•
•
•
•
•
•
•
•
31.6.2
Single pin debug system.
Half Duplex asynchronous communication (UART compatible).
Full duplex mode for direct UART connection.
Compatible with JTAG functionality, except boundary scan.
Failsafe packet-oriented protocol.
Read and write on-chip memory and program on-chip flash and fuses through SAB interface.
On-Chip Debug access through SAB interface.
Asynchronous receiver or transmitter when the aWire system is not used for debugging.
Overview
The aWire Debug Interface (AW) offers a single pin debug solution that is fully compatible with
the functionality offered by the JTAG interface, except boundary scan. This functionality includes
memory access, programming capabilities, and On-Chip Debug access.
Figure 31-8 on page 754 shows how the AW is connected in a 32-bit AVR device. The
RESET_N pin is used both as reset and debug pin. A special sequence on RESET_N is needed
to block the normal reset functionality and enable the AW.
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 31.6.6.8.
Section 31.6.7 lists the supported aWire commands and responses, with references to the
description in this document.
If the AW is not used for debugging, the aWire UART can be used by the user to send or receive
data with one stop bit, eight data bits, no parity bits, and one stop bit. This can be controlled
through the aWire user interface.
753
32099I–01/2012
AT32UC3L016/32/64
31.6.3
Block Diagram
Figure 31-8. aWire Debug Interface Block Diagram
PB
aWire Debug Interface
Flash
Controller
CHIP_ERASE command
AW User Interface
CPU
HALT command
RESET command
Power
Manager
External reset
AW_ENABLE
AW CONTROL
Reset
filter
RESET_N
Baudrate Detector
SAB interface
UART
RW
SZ
ADDR
DATA
CRC
SAB
31.6.4
I/O Lines Description
Table 31-30. I/O Lines Description
Name
Description
Type
DATA
aWire data multiplexed with the RESET_N pin.
Input/Output
DATAOUT
aWire data output in 2-pin mode.
Output
31.6.5
Product Dependencies
In order to use this module, other parts of the system must be configured correctly, as described
below.
754
32099I–01/2012
AT32UC3L016/32/64
31.6.5.1
I/O Lines
The pin used by AW is multiplexed with the RESET_N pin. The reset functionality is the default
function of this pin. To enable the aWire functionality on the RESET_N pin the user must enable
the AW either by sending the enable sequence over the RESET_N pin from an external aWire
master or by enabling the aWire user interface.
In 2-pin mode data is received on the RESET_N line, but transmitted on the DATAOUT line.
After sending the 2_PIN_MODE command the DATAOUT line is automatically enabled. All other
peripheral functions on this pin is disabled.
31.6.5.2
Power Management
When debugging through AW the system clocks are automatically turned on to allow debugging
in sleep modes.
31.6.5.3
Clocks
The aWire UART uses the internal 120 MHz RC oscillator (RC120M) as clock source for its
operation. When enabling the AW the RC120M is automatically started.
31.6.5.4
31.6.6
External Components
The AW needs an external pullup on the RESET_N pin to ensure that the pin is pulled up when
the bus is not driven.
Functional Description
31.6.6.1
aWire Communication Protocol
The AW is accessed through the RESET_N pin shown in Table 31-30 on page 754. The AW
communicates through a UART operating at variable baud rate (depending on a sync pattern)
with one start bit, 8 data bits (LSB first), one stop bit, and no parity bits. The aWire protocol is
based upon command packets from an externalmaster and response packets from the slave
(AW). The master always initiates communication and decides the baud rate.
The packet contains a sync byte (0x55), a command/response byte, two length bytes (optional),
a number of data bytes as defined in the length field (optional), and two CRC bytes. If the command/response has the most significant bit set, the command/response also carries the optional
length and data fields. The CRC field is not checked if the CRC value transmitted is 0x0000.
Table 31-31. aWire Packet Format
Field
SYNC
COMMAND/
RESPONSE
Number of bytes
Description
Comment
Optional
1
Sync pattern (0x55).
Used by the receiver to set the baud rate
clock.
No
1
Command from the master or
response from the slave.
When the most significant bit is set the
command/response has a length field. A
response has the next most significant bit
set. A command does not have this bit set.
No
755
32099I–01/2012
AT32UC3L016/32/64
Table 31-31. aWire Packet Format
Field
Number of bytes
Description
2
The number of bytes in the DATA
field.
Yes
DATA
LENGTH
Data according to command/
response.
Yes
CRC
2
CRC calculated with the FCS16
polynomial.
LENGTH
Comment
Optional
CRC value of 0x0000 makes the aWire
disregard the CRC if the master does not
support it.
No
CRC calculation
The CRC is calculated from the command/response, length, and data fields. The polynomial
used is the FCS16 (or CRC-16-CCIT) in reverse mode (0x8408) and the starting value is
0x0000.
Example command
Below is an example command from the master with additional data.
Figure 31-9. Example Command
baud_rate_clk
data_pin
field
...
sync(0x55)
command(0x81)
length(MSB)
length(lsb)
data(MSB)
data(LSB)
CRC(MSB)
CRC(lsb)
...
Example response
Below is an example response from the slave with additional data.
Figure 31-10. Example Response
baud_rate_clk
data_pin
field
...
sync(0x55)
response(0xC1)
length(MSB)
length(lsb)
data(MSB)
data(LSB)
CRC(MSB)
CRC(lsb)
...
756
32099I–01/2012
AT32UC3L016/32/64
Avoiding drive contention when changing direction
The aWire debug protocol uses one dataline in both directions. To avoid both the master and the
slave to drive this line when changing direction the AW has a built in guard time before it starts to
drive the line. At reset this guard time is set to maximum (128 bit cycles), but can be lowered by
the master upon command.
The AW will release the line immediately after the stop character has been transmitted.
During the direction change there can be a period when the line is not driven. An external pullup
has to be added to RESET_N to keep the signal stable when neither master or slave is actively
driving the line.
31.6.6.2
The RESET_N pin
Normal reset functionality on the RESET_N pin is disabled when using aWire. However, the
user can reset the system through the RESET aWire command. During aWire operation the
RESET_N pin should not be connected to an external reset circuitry, but disconnected via a
switch or a jumper to avoid drive contention and speed problems.
Figure 31-11. Reset Circuitry and aWire.
MCU
aWire master connector
Board Reset
Circuitry
31.6.6.3
Jumper
AW Debug
Interface
RESET_N
Power Manager
Initializing the AW
To enable AW, the user has to send a 0x55 pattern with a baudrate of 1 kHz on the RESET_N
pin. The AW is enabled after transmitting this pattern and the user can start transmitting commands. This pattern is not the sync pattern for the first command.
After enabling the aWire debug interface the halt bit is set automatically to prevent the system
from running code after the interface is enabled. To make the CPU run again set halt to zero
using the HALT command.
31.6.6.4
Disabling the AW
To disable AW, the user can keep the RESET_N pin low for 100 ms. This will disable the AW,
return RESET_N to its normal function, and reset the device.
An aWire master can also disable aWire by sending the DISABLE command. After acking the
command the AW will be disabled and RESET_N returns to its normal function.
757
32099I–01/2012
AT32UC3L016/32/64
31.6.6.5
Resetting the AW
The aWire master can reset the AW slave by pulling the RESET_N pin low for 20 ms. This is
equivalent to disabling and then enabling AW.
31.6.6.6
2-pin Mode
To avoid using special hardware when using a normal UART device as aWire master, the aWire
slave has a 2-pin mode where one pin is used as input and on pin is used as output. To enable
this mode the 2_PIN_MODE command must be sent. After sending the command, all responses
will be sent on the DATAOUT pin instead of the RESET_N pin. Commands are still received on
the RESET_N pin.
31.6.6.7
Baud Rate Clock
The communication speed is set by the master in the sync field of the command. The AW will
use this to resynchronize its baud rate clock and reply on this frequency. The minimum frequency of the communication is 1 kHz. The maximum frequency depends on the internal clock
source for the AW (RC120M). The baud rate clock is generated by AW with the following
formula:
TUNE × f br
f aw = ----------------------------8
Where f br is the baud rate frequency and f aw is the frequency of the internal RC120M. TUNE is
the value returned by the BAUD_RATE response.
To find the max frequency the user can issue the TUNE command to the AW to make it return
the TUNE value. This value can be used to compute the f aw . The maximum operational frequency ( f brmax ) is then:
f aw
f brmax = ------4
31.6.6.8
Service Access Bus
The AVR32 architecture offers a common interface for access to On-Chip Debug, programming,
and test functions. These are mapped on a common bus called the Service Access Bus (SAB),
which is linked to the aWire through a bus master module, which also handles synchronization
between the aWire and SAB clocks.
For more information about the SAB and a list of SAB slaves see the Service Access Bus
chapter.
SAB Clock
When accessing the SAB through the aWire there are no limitations on baud rate frequency
compared to chip frequency, although there must be an active system clock in order for the SAB
accesses to complete. If the system clock (CLK_SYS) is switched off in sleep mode, activity on
the aWire pin will restart the CLK_SYS automatically, without waking the device from sleep.
aWire masters may optimize the transfer rate by adjusting the baud rate frequency in relation to
the CLK_SYS. This ratio can be measured with the MEMORY_SPEED_REQUEST command.
When issuing the MEMORY_SPEED_REQUEST command a counter value CV is returned. CV
can be used to calculate the SAB speed ( f sab ) using this formula:
758
32099I–01/2012
AT32UC3L016/32/64
3f aw
f sab = ----------------CV – 3
SAB Address Mode
The Service Access Bus uses 36 address bits to address memory or registers in any of the
slaves on the bus. The bus supports sized accesses of bytes (8 bits), halfwords (16 bits), or
words (32 bits). All accesses must be aligned to the size of the access, i.e. halfword accesses
must have the lowest address bit cleared, and word accesses must have the two lowest address
bits cleared.
Two instructions exist to access the SAB: MEMORY_WRITE and MEMORY_READ. These two
instructions write and read words, halfwords, and bytes from the SAB.
Busy Reporting
If the aWire master, during a MEMORY_WRITE or a MEMORY_READ command, transmit
another byte when the aWire is still busy sending the previous byte to the SAB, the AW will
respond with a MEMORY_READ_WRITE_STATUS error. See chapter Section 31.6.8.5 for
more details.
The aWire master should adjust its baudrate or delay between bytes when doing SAB accesses
to ensure that the SAB is not overwhelmed with data.
Error Reporting
If a write is performed on a non-existing memory location the SAB interface will respond with an
error. If this happens, all further writes in this command will not be performed and the error and
number of bytes written is reported in the MEMORY_READWRITE_STATUS message from the
AW after the write.
If a read is performed on a non-existing memory location, the SAB interface will respond with an
error. If this happens, the data bytes read after this event are not valid. The AW will include three
extra bytes at the end of the transfer to indicate if the transfer was successful, or in the case of
an error, how many valid bytes were received.
31.6.6.9
CRC Errors/NACK Response
The AW will calculate a CRC value when receiving the command, length, and data fields of the
command packets. If this value differs from the value from the CRC field of the packet, the AW
will reply with a NACK response. Otherwise the command is carried out normally.
An unknown command will be replied with a NACK response.
In worst case a transmission error can happen in the length or command field of the packet. This
can lead to the aWire slave trying to receive a command with or without length (opposite of what
the master intended) or receive an incorrect number of bytes. The aWire slave will then either
wait for more data when the master has finished or already have transmitted the NACK
response in congestion with the master. The master can implement a timeout on every command and reset the slave if no response is returned after the timeout period has ended.
759
32099I–01/2012
AT32UC3L016/32/64
31.6.7
aWire Command Summary
The implemented aWire commands are shown in the table below. The responses from the AW
are listed in Section 31.6.8.
Table 31-32. aWire Command Summary
COMMAND
Instruction
Description
0x01
AYA
“Are you alive”.
0x02
JTAG_ID
Asks AW to return the JTAG IDCODE.
0x03
STATUS_REQUEST
Request a status message from the AW.
0x04
TUNE
Tell the AW to report the current baud rate.
0x05
MEMORY_SPEED_REQUEST
Reports the speed difference between the aWire control and the SAB clock
domains.
0x06
CHIP_ERASE
Erases the flash and all volatile memories.
0x07
DISABLE
Disables the AW.
0x08
2_PIN_MODE
Enables the DATAOUT pin and puts the aWire in 2-pin mode, where all
responses are sent on the DATAOUT pin.
0x80
MEMORY_WRITE
Writes words, halfwords, or bytes to the SAB.
0x81
MEMORY_READ
Reads words, halfwords, or bytes from the SAB.
0x82
HALT
Issues a halt command to the device.
0x83
RESET
Issues a reset to the Reset Controller.
0x84
SET_GUARD_TIME
Sets the guard time for the AW.
All aWire commands are described below, with a summary in table form.
Table 31-33. Command/Response Description Notation
Command/Response
Description
Command/Response value
Shows the command/response value to put into the command/response field of the packet.
Additional data
Shows the format of the optional data field if applicable.
Possible responses
Shows the possible responses for this command.
31.6.7.1
AYA
This command asks the AW: “Are you alive”, where the AW should respond with an
acknowledge.
Table 31-34. AYA Details
Command
Details
Command value
0x01
Additional data
N/A
Possible responses
0x40: ACK (Section 31.6.8.1)
0x41: NACK (Section 31.6.8.2)
760
32099I–01/2012
AT32UC3L016/32/64
31.6.7.2
JTAG_ID
This command instructs the AW to output the JTAG idcode in the following response.
Table 31-35. JTAG_ID Details
31.6.7.3
Command
Details
Command value
0x02
Additional data
N/A
Possible responses
0xC0: IDCODE (Section 31.6.8.3)
0x41: NACK (Section 31.6.8.2)
STATUS_REQUEST
Asks the AW for a status message.
Table 31-36. STATUS_REQUEST Details
31.6.7.4
Command
Details
Command value
0x03
Additional data
N/A
Possible responses
0xC4: STATUS_INFO (Section 31.6.8.7)
0x41: NACK (Section 31.6.8.2)
TUNE
Asks the AW for the current baud rate counter value.
Table 31-37. TUNE Details
31.6.7.5
Command
Details
Command value
0x04
Additional data
N/A
Possible responses
0xC3: BAUD_RATE (Section 31.6.8.6)
0x41: NACK (Section 31.6.8.2)
MEMORY_SPEED_REQUEST
Asks the AW for the relative speed between the aWire clock (RC120M) and the SAB interface.
Table 31-38. MEMORY_SPEED_REQUEST Details
31.6.7.6
Command
Details
Command value
0x05
Additional data
N/A
Possible responses
0xC5: MEMORY_SPEED (Section 31.6.8.8)
0x41: NACK (Section 31.6.8.2)
CHIP_ERASE
This instruction allows a programmer to completely erase all nonvolatile memories in the chip.
This will also clear any security bits that are set, so the device can be accessed normally. The
command is acked immediately, but the status of the command can be monitored by checking
761
32099I–01/2012
AT32UC3L016/32/64
the Chip Erase ongoing bit in the status bytes received after the STATUS_REQUEST
command.
Table 31-39. CHIP_ERASE Details
31.6.7.7
Command
Details
Command value
0x06
Additional data
N/A
Possible responses
0x40: ACK (Section 31.6.8.1)
0x41: NACK (Section 31.6.8.2)
DISABLE
Disables the AW. The AW will respond with an ACK response and then disable itself.
Table 31-40. DISABLE Details
31.6.7.8
Command
Details
Command value
0x07
Additional data
N/A
Possible responses
0x40: ACK (Section 31.6.8.1)
0x41: NACK (Section 31.6.8.2)
2_PIN_MODE
Enables the DATAOUT pin as an output pin. All responses sent from the aWire slave will be sent
on this pin, instead of the RESET_N pin, starting with the ACK for the 2_PIN_MODE command.
Table 31-41. DISABLE Details
31.6.7.9
Command
Details
Command value
0x07
Additional data
N/A
Possible responses
0x40: ACK (Section 31.6.8.1)
0x41: NACK (Section 31.6.8.2)
MEMORY_WRITE
This command enables programming of memory/writing to registers on the SAB. The
MEMORY_WRITE command allows words, halfwords, and bytes to be programmed to a continuous sequence of addresses in one operation. Before transferring the data, the user must
supply:
1. The number of data bytes to write + 5 (size and starting address) in the length field.
2. The size of the transfer: words, halfwords, or bytes.
3. The starting address of the transfer.
762
32099I–01/2012
AT32UC3L016/32/64
The 4 MSB of the 36 bit SAB address are submitted together with the size field (2 bits). Then follows the 4 remaining address bytes and finally the data bytes. The size of the transfer is
specified using the values from the following table:
Table 31-42. Size Field Decoding
Size field
Description
00
Byte transfer
01
Halfword transfer
10
Word transfer
11
Reserved
Below is an example write command:
1. 0x55 (sync)
2. 0x80 (command)
3. 0x00 (length MSB)
4. 0x09 (length LSB)
5. 0x25 (size and address MSB, the two MSB of this byte are unused and set to zero)
6. 0x00
7. 0x00
8. 0x00
9. 0x04 (address LSB)
10. 0xCA
11. 0xFE
12. 0xBA
13. 0xBE
14. 0xXX (CRC MSB)
15. 0xXX (CRC LSB)
The length field is set to 0x0009 because there are 9 bytes of additional data: 5 address and size
bytes and 4 bytes of data. The address and size field indicates that words should be written to
address 0x500000004. The data written to 0x500000004 is 0xCAFEBABE.
Table 31-43. MEMORY_WRITE Details
31.6.7.10
Command
Details
Command value
0x80
Additional data
Size, Address and Data
Possible responses
0xC2: MEMORY_READWRITE_STATUS (Section 31.6.8.5)
0x41: NACK (Section 31.6.8.2)
MEMORY_READ
This command enables reading of memory/registers on the Service Access Bus (SAB). The
MEMORY_READ command allows words, halfwords, and bytes to be read from a continuous
sequence of addresses in one operation. The user must supply:
763
32099I–01/2012
AT32UC3L016/32/64
1. The size of the data field: 7 (size and starting address + read length indicator) in the
length field.
2. The size of the transfer: Words, halfwords, or bytes.
3. The starting address of the transfer.
4. The number of bytes to read (max 65532).
The 4 MSB of the 36 bit SAB address are submitted together with the size field (2 bits). The 4
remaining address bytes are submitted before the number of bytes to read. The size of the
transfer is specified using the values from the following table:
Table 31-44. Size Field Decoding
Size field
Description
00
Byte transfer
01
Halfword transfer
10
Word transfer
11
Reserved
Below is an example read command:
1. 0x55 (sync)
2. 0x81 (command)
3. 0x00 (length MSB)
4. 0x07 (length LSB)
5. 0x25 (size and address MSB, the two MSB of this byte are unused and set to zero)
6. 0x00
7. 0x00
8. 0x00
9. 0x04 (address LSB)
10. 0x00
11. 0x04
12. 0xXX (CRC MSB)
13. 0xXX (CRC LSB)
The length field is set to 0x0007 because there are 7 bytes of additional data: 5 bytes of address
and size and 2 bytes with the number of bytes to read. The address and size field indicates one
word (four bytes) should be read from address 0x500000004.
Table 31-45. MEMORY_READ Details
Command
Details
Command value
0x81
Additional data
Size, Address and Length
Possible responses
0xC1: MEMDATA (Section 31.6.8.4)
0xC2: MEMORY_READWRITE_STATUS (Section 31.6.8.5)
0x41: NACK (Section 31.6.8.2)
764
32099I–01/2012
AT32UC3L016/32/64
31.6.7.11
HALT
This command tells the CPU to halt code execution for safe programming. If the CPU is not
halted during programming it can start executing partially loaded programs. To halt the processor, the aWire master should send 0x01 in the data field of the command. After programming the
halting can be released by sending 0x00 in the data field of the command.
Table 31-46. HALT Details
31.6.7.12
Command
Details
Command value
0x82
Additional data
0x01 to halt the CPU 0x00 to release the halt and reset the
device.
Possible responses
0x40: ACK (Section 31.6.8.1)
0x41: NACK (Section 31.6.8.2)
RESET
This command resets different domains in the part. The aWire master sends a byte with the
reset value. Each bit in the reset value byte corresponds to a reset domain in the chip. If a bit is
set the reset is activated and if a bit is not set the reset is released. The number of reset domains
and their destinations are identical to the resets described in the JTAG data registers chapter
under reset register.
Table 31-47. RESET Details
31.6.7.13
Command
Details
Command value
0x83
Additional data
Reset value for each reset domain. The number of reset
domains is part specific.
Possible responses
0x40: ACK (Section 31.6.8.1)
0x41: NACK (Section 31.6.8.2)
SET_GUARD_TIME
Sets the guard time value in the AW, i.e. how long the AW will wait before starting its transfer
after the master has finished.
The guard time can be either 0x00 (128 bit lengths), 0x01 (16 bit lengths), 0x2 (4 bit lengths) or
0x3 (1 bit length).
Table 31-48. SET_GUARD_TIME Details
Command
Details
Command value
0x84
Additional data
Guard time
Possible responses
0x40: ACK (Section 31.6.8.1)
0x41: NACK (Section 31.6.8.2)
765
32099I–01/2012
AT32UC3L016/32/64
31.6.8
aWire Response Summary
The implemented aWire responses are shown in the table below.
Table 31-49. aWire Response Summary
RESPONSE
Instruction
Description
0x40
ACK
Acknowledge.
0x41
NACK
Not acknowledge. Sent after CRC errors and after unknown commands.
0xC0
IDCODE
The JTAG idcode.
0xC1
MEMDATA
Values read from memory.
0xC2
MEMORY_READWRITE_STATUS
Status after a MEMORY_WRITE or a MEMORY_READ command. OK, busy,
error.
0xC3
BAUD_RATE
The current baudrate.
0xC4
STATUS_INFO
Status information.
0xC5
MEMORY_SPEED
SAB to aWire speed information.
31.6.8.1
ACK
The AW has received the command successfully and performed the operation.
Table 31-50. ACK Details
31.6.8.2
Response
Details
Response value
0x40
Additional data
N/A
NACK
The AW has received the command, but got a CRC mismatch.
Table 31-51. NACK Details
31.6.8.3
Response
Details
Response value
0x41
Additional data
N/A
IDCODE
The JTAG idcode for this device.
Table 31-52. IDCODE Details
31.6.8.4
Response
Details
Response value
0xC0
Additional data
JTAG idcode
MEMDATA
The data read from the address specified by the MEMORY_READ command. The last 3 bytes
are status bytes from the read. The first status byte is the status of the command described in
the table below. The last 2 bytes are the number of remaining data bytes to be sent in the data
field of the packet when the error occurred. If the read was not successful all data bytes after the
failure are undefined. A successful word read (4 bytes) will look like this:
766
32099I–01/2012
AT32UC3L016/32/64
1. 0x55 (sync)
2. 0xC1 (command)
3. 0x00 (length MSB)
4. 0x07 (length LSB)
5. 0xCA (Data MSB)
6. 0xFE
7. 0xBA
8. 0xBE (Data LSB)
9. 0x00 (Status byte)
10. 0x00 (Bytes remaining MSB)
11. 0x00 (Bytes remaining LSB)
12. 0xXX (CRC MSB)
13. 0xXX (CRC LSB)
The status is 0x00 and all data read are valid. An unsuccessful four byte read can look like this:
1. 0x55 (sync)
2. 0xC1 (command)
3. 0x00 (length MSB)
4. 0x07 (length LSB)
5. 0xCA (Data MSB)
6. 0xFE
7. 0xXX (An error has occurred. Data read is undefined. 5 bytes remaining of the Data
field)
8. 0xXX (More undefined data)
9. 0x02 (Status byte)
10. 0x00 (Bytes remaining MSB)
11. 0x05 (Bytes remaining LSB)
12. 0xXX (CRC MSB)
13. 0xXX (CRC LSB)
The error occurred after reading 2 bytes on the SAB. The rest of the bytes read are undefined.
The status byte indicates the error and the bytes remaining indicates how many bytes were
remaining to be sent of the data field of the packet when the error occurred.
Table 31-53. MEMDATA Status Byte
status byte
Description
0x00
Read successful
0x01
SAB busy
0x02
Bus error (wrong address)
Other
Reserved
Table 31-54. MEMDATA Details
Response
Details
Response value
0xC1
Additional data
Data read, status byte, and byte count (2 bytes)
767
32099I–01/2012
AT32UC3L016/32/64
31.6.8.5
MEMORY_READWRITE_STATUS
After a MEMORY_WRITE command this response is sent by AW. The response can also be
sent after a MEMORY_READ command if AW encountered an error when receiving the
address. The response contains 3 bytes, where the first is the status of the command and the 2
next contains the byte count when the first error occurred. The first byte is encoded this way:
Table 31-55. MEMORY_READWRITE_STATUS Status Byte
status byte
Description
0x00
Write successful
0x01
SAB busy
0x02
Bus error (wrong address)
Other
Reserved
Table 31-56. MEMORY_READWRITE_STATUS Details
31.6.8.6
Response
Details
Response value
0xC2
Additional data
Status byte and byte count (2 bytes)
BAUD_RATE
The current baud rate in the AW. See Section 31.6.6.7 for more details.
Table 31-57. BAUD_RATE Details
31.6.8.7
Response
Details
Response value
0xC3
Additional data
Baud rate
STATUS_INFO
A status message from AW.
Table 31-58. STATUS_INFO Contents
Bit number
Name
Description
15-9
Reserved
8
Protected
The protection bit in the internal flash is set. SAB access is restricted. This bit
will read as one during reset.
7
SAB busy
The SAB bus is busy with a previous transfer. This could indicate that the CPU
is running on a very slow clock, the CPU clock has stopped for some reason
or that the part is in constant reset.
6
Chip erase ongoing
The Chip erase operation has not finished.
5
CPU halted
This bit will be set if the CPU is halted. This bit will read as zero during reset.
4-1
0
Reserved
Reset status
This bit will be set if AW has reset the CPU using the RESET command.
768
32099I–01/2012
AT32UC3L016/32/64
Table 31-59. STATUS_INFO Details
31.6.8.8
Response
Details
Response value
0xC4
Additional data
2 status bytes
MEMORY_SPEED
Counts the number of RC120M clock cycles it takes to sync one message to the SAB interface
and back again. The SAB clock speed ( f sab ) can be calculated using the following formula:
3f aw
f sab = ----------------CV – 3
Table 31-60. MEMORY_SPEED Details
31.6.9
Response
Details
Response value
0xC5
Additional data
Clock cycle count (MS)
Security Restrictions
When the security fuse in the Flash is programmed, the following aWire commands are limited:
• MEMORY_WRITE
• MEMORY_READ
Unlimited access to these instructions is restored when the security fuse is erased by the
CHIP_ERASE aWire command.
Note that the security bit will read as programmed and block these instructions also if the Flash
Controller is statically reset.
769
32099I–01/2012
AT32UC3L016/32/64
32. Electrical Characteristics
32.1
Absolute Maximum Ratings*
Table 32-1.
Absolute Maximum Ratings
Operating temperature..................................... -40°C to +85°C
*NOTICE:
Storage temperature...................................... -60°C to +150°C
Voltage on input pins (except for 5V pins) with respect to ground
.................................................................-0.3V to VVDD(2)+0.3V
Voltage on 5V tolerant(1) pins with respect to ground ...............
.............................................................................-0.3V to 5.5V
Total DC output current on all I/O pins - VDDIO ........... 120mA
Total DC output current on all I/O pins - VDDIN ............. 36mA
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.
Maximum operating voltage VDDCORE......................... 1.98V
Maximum operating voltage VDDIO, VDDIN .................... 3.6V
Notes:
1. 5V tolerant pins, see Section 3.2 ”Peripheral Multiplexing on I/O lines” on page 9
2. VVDD corresponds to either VVDDIN or VVDDIO, depending on the supply for the pin. Refer to Section 3.2 on page 9 for details.
32.2
Supply Characteristics
The following characteristics are applicable to the operating temperature range: TA = -40°C to 85°C, unless otherwise specified and are valid for a junction temperature up to T J = 100°C. Please refer to Section 6. ”Supply and Startup
Considerations” on page 36.
Table 32-2.
Supply Characteristics
Voltage
Symbol
Parameter
Min
Max
Unit
VVDDIO
DC supply peripheral I/Os
1.62
3.6
V
DC supply peripheral I/Os, 1.8V single
supply mode
1.62
1.98
V
DC supply peripheral I/Os and internal
regulator, 3.3V supply mode
1.98
3.6
V
VVDDCORE
DC supply core
1.62
1.98
V
VVDDANA
Analog supply voltage
1.62
1.98
V
VVDDIN
770
32099I–01/2012
AT32UC3L016/32/64
Table 32-3.
Supply Rise Rates and Order(1)
Rise Rate
Symbol
Parameter
Min
Max
Unit
VVDDIO
DC supply peripheral I/Os
0
2.5
V/µs
VVDDIN
DC supply peripheral I/Os
and internal regulator
0.002
2.5
V/µs
Slower rise time requires
external power-on reset
circuit.
VVDDCORE
DC supply core
0
2.5
V/µs
Rise before or at the same
time as VDDIO
VVDDANA
Analog supply voltage
0
2.5
V/µs
Rise together with
VDDCORE
Note:
32.3
Comment
1. These values are based on simulation and characterization of other AVR microcontrollers
manufactured in the same process technology. These values are not covered by test limits in
production.
Maximum Clock Frequencies
These parameters are given in the following conditions:
• VVDDCORE = 1.62V to 1.98V
• Temperature = -40°C to 85°C
Table 32-4.
32.4
Clock Frequencies
Symbol
Parameter
fCPU
Conditions
Min
Max
Units
CPU clock frequency
50
MHz
fPBA
PBA clock frequency
50
MHz
fPBB
PBB clock frequency
50
MHz
fGCLK0
GCLK0 clock frequency
DFLLIF main reference, GCLK0
pin
150
MHz
fGCLK1
GCLK1 clock frequency
DFLLIF dithering and ssg
reference, GCLK1 pin
150
MHz
fGCLK2
GCLK2 clock frequency
AST, GCLK2 pin
80
MHz
fGCLK3
GCLK3 clock frequency
PWMA, GCLK3 pin
110
MHz
fGCLK4
GCLK4 clock frequency
CAT, ACIFB, GCLK4 pin
110
MHz
fGCLK5
GCLK5 clock frequency
GLOC
80
MHz
Power Consumption
The values in Table 32-5 are measured values of power consumption under the following conditions, except where noted:
• Operating conditions internal core supply (Figure 32-1) - this is the default configuration
– VVDDIN = 3.0V
– VVDDCORE = 1.62V, supplied by the internal regulator
771
32099I–01/2012
AT32UC3L016/32/64
– Corresponds to the 3.3V supply mode with 1.8V regulated I/O lines, please refer to
the Supply and Startup Considerations section for more details
• Equivalent to the 3.3V single supply mode
• Consumption in 1.8V single supply mode can be estimated by subtracting the regulator static current
• Operating conditions external core supply (Figure 32-2) - used only when noted
– VVDDIN = VVDDCORE = 1.8V
– Corresponds to the 1.8V single supply mode, please refer to the Supply and Startup
Considerations section for more details
• TA = 25°C
• Oscillators
– OSC0 (crystal oscillator) stopped
– OSC32K (32KHz crystal oscillator) running with external 32KHz crystal
– DFLL running at 50MHz with OSC32K as reference
• Clocks
– DFLL used as main clock source
– CPU, HSB, and PBB clocks undivided
– PBA clock divided by 4
– The following peripheral clocks running
• PM, SCIF, AST, FLASHCDW, PBA bridge
– All other peripheral clocks stopped
• I/Os are inactive with internal pull-up
• Flash enabled in high speed mode
• POR33 disabled
772
32099I–01/2012
AT32UC3L016/32/64
Table 32-5.
Mode
Power Consumption for Different Operating Modes
Conditions
Active(1)
Measured on
Consumption Typ
-CPU running a recursive Fibonacci algorithm
260
-CPU running a division algorithm
165
Idle(1)
92
(1)
(1)
47
Stop
37
DeepStop
23
-OSC32K and AST stopped
-Internal core supply
Static
Shutdown
Amp0
10
µA
-OSC32K running
-AST running at 1KHz
-External core supply (Figure 32-2)
5.3
-OSC32K and AST stopped
-External core supply (Figure 32-2)
4.7
-OSC32K running
-AST running at 1KHz
600
nA
-AST and OSC32K stopped
Note:
µA/MHz
58
Frozen
Standby
Unit
9
1. These numbers are valid for the measured condition only and must not be extrapolated to other frequencies.
Figure 32-1. Measurement Schematic, Internal Core Supply
Amp0
VDDIN
VDDIO
VDDCORE
VDDANA
773
32099I–01/2012
AT32UC3L016/32/64
Figure 32-2. Measurement Schematic, External Core Supply
Amp0
VDDIN
VDDIO
VDDCORE
VDDANA
32.4.1
Peripheral Power Consumption
The values in Table 32-6 are measured values of power consumption under the following
conditions.
• Operating conditions internal core supply (Figure 32-1)
– VVDDIN = 3.0V
– VVDDCORE = 1.62V, supplied by the internal regulator
– Corresponds to the 3.3V supply mode with 1.8V regulated I/O lines, please refer to
the Supply and Startup Considerations section for more details
• TA = 25°C
• Oscillators
– OSC0 (crystal oscillator) stopped
– OSC32K (32KHz crystal oscillator) running with external 32KHz crystal
– DFLL running at 50MHz with OSC32K as reference
• Clocks
– DFLL used as main clock source
– CPU, HSB, and PB clocks undivided
• I/Os are inactive with internal pull-up
• Flash enabled in high speed mode
• POR33 disabled
Consumption active is the added current consumption when the module clock is turned on and
the module is doing a typical set of operations.
774
32099I–01/2012
AT32UC3L016/32/64
Table 32-6.
Peripheral
Typical Current Consumption by Peripheral(2)
Typ Consumption Active
ACIFB
14.0
ADCIFB(1)
14.9
AST
5.6
AW USART
6.8
CAT
12.4
EIC
1.3
FREQM
3.2
GLOC
0.4
GPIO
15.9
PWMA
2.5
SPI
7.6
TC
7.2
TWIM
5.1
TWIS
3.2
USART
12.3
WDT
2.3
Unit
µA/MHz
Notes:
1. Includes the current consumption on VDDANA and ADVREFP.
2. These numbers are valid for the measured condition only and must not be extrapolated to
other frequencies.
775
32099I–01/2012
AT32UC3L016/32/64
32.5
I/O Pin Characteristics
Table 32-7.
Normal I/O Pin Characteristics(1)
Symbol
Parameter
RPULLUP
Pull-up resistance
VIL
Input low-level voltage
VIH
Input high-level voltage
VOL
Output low-level voltage
VOH
Output high-level voltage
fMAX
Output frequency(2)
tRISE
Rise time(2)
Condition
Min
Typ
Max
Units
75
100
145
kOhm
VVDD = 3.0V
-0.3
0.3*VVDD
VVDD = 1.62V
-0.3
0.3*VVDD
VVDD = 3.6V
0.7*VVDD
VVDD + 0.3
VVDD = 1.98V
0.7*VVDD
VVDD + 0.3
VVDD = 3.0V, IOL = 3mA
0.4
VVDD = 1.62V, IOL = 2mA
0.4
VVDD = 3.0V, IOH = 3mA
VVDD - 0.4
VVDD = 1.62V, IOH = 2mA
VVDD - 0.4
V
VVDD = 3.0V, load = 10pF
45
VVDD = 3.0V, load = 30pF
23
VVDD = 3.0V, load = 10pF
4.7
VVDD = 3.0V, load = 30pF
11.5
VVDD = 3.0V, load = 10pF
4.8
VVDD = 3.0V, load = 30pF
12
1
MHz
ns
Fall time(2)
ILEAK
Input leakage current
Pull-up resistors disabled
TQFP48 package
1.4
CIN
Input capacitance, all
normal I/O pins except
PA05, PA07, PA17, PA20,
PA21, PB04, PB05
QFN48 package
1.1
TLLGA 48 package
1.1
TQFP48 package
2.7
QFN48 package
2.4
TLLGA 48 package
2.4
TQFP48 package
3.8
QFN48 package
3.5
TLLGA 48 package
3.5
Input capacitance, PA20
Input capacitance, PA05,
PA07, PA17, PA21, PB04,
PB05
CIN
Notes:
V
V
tFALL
CIN
V
µA
pF
1. VVDD corresponds to either VVDDIN or VVDDIO, depending on the supply for the pin. Refer to Section 3.2 on page 9 for details.
2. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
Table 32-8.
Symbol
RPULLUP
High-drive I/O Pin Characteristics(1)
Parameter
Pull-up resistance
Condition
Min
Typ
Max
PA06
30
50
110
PA02, PB01, RESET
75
100
145
PA08, PA09
10
20
45
Units
kOhm
776
32099I–01/2012
AT32UC3L016/32/64
Table 32-8.
High-drive I/O Pin Characteristics(1)
Symbol
Parameter
Condition
Min
Typ
Max
VIL
Input low-level voltage
VVDD = 3.0V
-0.3
0.3*VVDD
VVDD = 1.62V
-0.3
0.3*VVDD
VIH
Input high-level voltage
VVDD = 3.6V
0.7*VVDD
VVDD + 0.3
VVDD = 1.98V
0.7*VVDD
VVDD + 0.3
VOL
Output low-level voltage
VOH
Output high-level voltage
Output frequency, all Highdrive I/O pins, except
PA08 and PA09(2)
VVDD = 3.0V, load = 10pF
45
fMAX
VVDD = 3.0V, load = 30pF
23
Rise time, all High-drive
I/O pins, except PA08 and
PA09(2)
VVDD = 3.0V, load = 10pF
4.7
tRISE
VVDD = 3.0V, load = 30pF
11.5
VVDD = 3.0V, load = 10pF
4.8
tFALL
Fall time, all High-drive I/O
pins, except PA08 and
PA09(2)
VVDD = 3.0V, load = 30pF
12
Output frequency, PA08
and PA09(2)
VVDD = 3.0V, load = 10pF
52
fMAX
VVDD = 3.0V, load = 30pF
39
Rise time, PA08 and
PA09(2)
VVDD = 3.0V, load = 10pF
2.9
tRISE
VVDD = 3.0V, load = 30pF
4.9
VVDD = 3.0V, load = 10pF
2.5
tFALL
Fall time, PA08 and
PA09(2)
VVDD = 3.0V, load = 30pF
4.6
ILEAK
Input leakage current
Pull-up resistors disabled
1
CIN
Input capacitance, all
High-drive I/O pins, except
PA08 and PA09
VVDD = 3.0V, IOL = 6mA
0.4
VVDD = 1.62V, IOL = 4mA
0.4
Units
V
V
V
VVDD = 3.0V, IOH = 6mA
VVDD-0.4
VVDD = 1.62V, IOH = 4mA
VVDD-0.4
V
MHz
ns
MHz
ns
TQFP48 package
2,2
QFN48 package
2.0
TLLGA 48 package
2.0
TQFP48 package
7.0
QFN48 package
6.7
TLLGA 48 package
6.7
µA
pF
Input capacitance, PA08
and PA09
CIN
Notes:
1. VVDD corresponds to either VVDDIN or VVDDIO, depending on the supply for the pin. Refer to Section 3.2 on page 9 for details.
2. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
Table 32-9.
High-drive I/O, 5V Tolerant, Pin Characteristics(1)
Symbol
Parameter
RPULLUP
Pull-up resistance
VIL
Input low-level voltage
Condition
Min
Typ
Max
Units
30
50
110
kOhm
VVDD = 3.0V
-0.3
0.3*VVDD
VVDD = 1.62V
-0.3
0.3*VVDD
V
777
32099I–01/2012
AT32UC3L016/32/64
Table 32-9.
High-drive I/O, 5V Tolerant, Pin Characteristics(1)
Symbol
Parameter
VIH
Input high-level voltage
VOL
Output low-level voltage
VOH
Output high-level voltage
fMAX
Output frequency(2)
tRISE
Rise time(2)
(2)
tFALL
Fall time
ILEAK
Input leakage current
CIN
Notes:
Input capacitance
Condition
Min
Typ
Max
VVDD = 3.6V
0.7*VVDD
5.5
VVDD = 1.98V
0.7*VVDD
5.5
Units
V
VVDD = 3.0V, IOL = 6mA
0.4
VVDD = 1.62V, IOL = 4mA
0.4
V
VVDD = 3.0V, IOH = 6mA
VVDD-0.4
VVDD = 1.62V, IOH = 4mA
VVDD-0.4
V
VVDD = 3.0V, load = 10pF
87
VVDD = 3.0V, load = 30pF
58
VVDD = 3.0V, load = 10pF
2.3
VVDD = 3.0V, load = 30pF
4.3
VVDD = 3.0V, load = 10pF
1.9
VVDD = 3.0V, load = 30pF
3.7
5.5V, pull-up resistors disabled
10
MHz
ns
TQFP48 package
4.5
QFN48 package
4.2
TLLGA48 package
4.2
µA
pF
1. VVDD corresponds to either VVDDIN or VVDDIO, depending on the supply for the pin. Refer to Section 3.2 on page 9 for details.
2. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
Table 32-10. TWI Pin Characteristics(1)
Symbol
Parameter
RPULLUP
Pull-up resistance
VIL
Input low-level voltage
Input high-level voltage
VIH
Condition
Min
Typ
Max
Units
25
35
60
kOhm
VVDD = 3.0V
-0.3
0.3*VVDD
VVDD = 1.62V
-0.3
0.3*VVDD
VVDD = 3.6V
0.7*VVDD
VVDD + 0.3
VVDD = 1.98V
0.7*VVDD
VVDD + 0.3
Input high-level voltage, 5V
tolerant SMBUS compliant
pins
VVDD = 3.6V
0.7*VVDD
5.5
VVDD = 1.98V
0.7*VVDD
5.5
VOL
Output low-level voltage
IOL = 3mA
ILEAK
Input leakage current
Pull-up resistors disabled
IIL
Input low leakage
1
IIH
Input high leakage
1
CIN
Input capacitance
V
V
V
0.4
V
1
TQFP48 package
3.8
QFN48 package
3.5
TLLGA48 package
3.5
µA
pF
778
32099I–01/2012
AT32UC3L016/32/64
Table 32-10. TWI Pin Characteristics(1)
Symbol
Parameter
tFALL
Fall time
fMAX
Max frequency
Condition
Min
Typ
Cbus = 400pF, VVDD > 2.0V
250
Cbus = 400pF, VVDD > 1.62V
470
Cbus = 400pF, VVDD > 2.0V
Max
Units
ns
400
kHz
Note:
1. VVDD corresponds to either VVDDIN or VVDDIO, depending on the supply for the pin. Refer to Section 3.2 on page 9 for details.
32.6
Oscillator Characteristics
32.6.1
Oscillator 0 (OSC0) Characteristics
32.6.1.1
Digital Clock Characteristics
The following table describes the characteristics for the oscillator when a digital clock is applied
on XIN.
Table 32-11. Digital Clock Characteristics
Symbol
Parameter
fCPXIN
XIN clock frequency
tCPXIN
XIN clock duty cycle
tSTARTUP
Startup time
CIN
32.6.1.2
XIN input capacitance
Conditions
Min
Typ
Max
40
Units
50
MHz
60
%
0
TQFP48 package
7.0
QFN48 package
6.7
TLLGA48 package
6.7
cycles
pF
Crystal Oscillator Characteristics
The following table describes the characteristics for the oscillator when a crystal is connected
between XIN and XOUT as shown in Figure 32-3. The user must choose a crystal oscillator
where the crystal load capacitance CL is within the range given in the table. The exact value of CL
can be found in the crystal datasheet. The capacitance of the external capacitors (CLEXT) can
then be computed as follows:
C LEXT = 2 ( C L – C i ) – C PCB
where CPCB is the capacitance of the PCB and Ci is the internal equivalent load capacitance.
Table 32-12. Crystal Oscillator Characteristics
Symbol
Parameter
fOUT
Crystal oscillator frequency
CL
Crystal load capacitance
Ci
Internal equivalent load capacitance
Conditions
Min
Typ
Max
Unit
0.45
10
16
MHz
6
18
pF
2
779
32099I–01/2012
AT32UC3L016/32/64
Table 32-12. Crystal Oscillator Characteristics
Symbol
Parameter
tSTARTUP
Conditions
Min
(1)
Startup time
SCIF.OSCCTRL.GAIN = 2
Current consumption
Notes:
Max
(2)
30 000
Active mode, f = 0.45MHz,
SCIF.OSCCTRL.GAIN = 0
IOSC
Typ
Unit
cycles
30
µA
Active mode, f = 10MHz,
SCIF.OSCCTRL.GAIN = 2
170
1. Please refer to the SCIF chapter for details.
2. Nominal crystal cycles.
3. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
Figure 32-3. Oscillator Connection
CLEXT
XOUT
UC3L
Ci
CL
XIN
CLEXT
32.6.2
32KHz Crystal Oscillator (OSC32K) Characteristics
Figure 32-3 and the equation above also applies to the 32KHz oscillator connection. The user
must choose a crystal oscillator where the crystal load capacitance CL is within the range given
in the table. The exact value of CL can then be found in the crystal datasheet.
Table 32-13. 32 KHz Crystal Oscillator Characteristics
Symbol
Parameter
fOUT
Crystal oscillator frequency
tSTARTUP
Startup time
CL
Crystal load capacitance(2)
Ci
Internal equivalent load
capacitance
IOSC32
Current consumption
RS
Equivalent series resistance
Conditions
Min
Typ
Max
32 768
Hz
(1)
RS = 60kOhm, CL = 9pF
30 000
6
Unit
cycles
12.5
pF
2
0.9
(2)
32 768Hz
35
µA
85
kOhm
780
32099I–01/2012
AT32UC3L016/32/64
Notes:
1. Nominal crystal cycles.
2. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
32.6.3
Digital Frequency Locked Loop (DFLL) Characteristics
Table 32-14. Digital Frequency Locked Loop Characteristics
Symbol
Parameter
Conditions
Output frequency
fOUT
(2)
Reference frequency
fREF
(2)
FINE resolution
FINE > 100, all COARSE values
tSTARTUP
tLOCK
Notes:
Startup time
Lock time
(2)
Max
Unit
40
150
MHz
8
150
kHz
0.25
%
Fine lock, fREF = 32kHz, SSG disabled
0.1
0.5
Accurate lock, fREF = 32kHz, dither clk
RCSYS/2, SSG disabled
0.06
0.5
Fine lock, fREF = 8-150kHz, SSG
disabled
0.2
1
Accurate lock, fREF = 8-150kHz, dither
clk RCSYS/2, SSG disabled
0.1
1
Power consumption
IDFLL
Typ
See
Figure 324
Frequency drift over voltage
and temperature
Accuracy(2)
Min
%
22
Within 90% of final values
µA/MHz
100
fREF = 32kHz, fine lock, SSG disabled
600
fREF = 32kHz, accurate lock, dithering
clock = RCSYS/2, SSG disabled
1100
µs
1. Spread Spectrum Generator (SSG) is disabled by writing a zero to the EN bit in the SCIF.DFLL0SSG register.
2. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
781
32099I–01/2012
AT32UC3L016/32/64
Figure 32-4. DFLL Open Loop Frequency Variation(1)
DFLL Open Loop Frequency variation
160
150
Frequencies (MHz)
140
130
1,98V
120
1,8V
1.62V
110
100
90
80
-40
-20
0
20
40
60
80
Tem pera ture
Note:
1. The plot shows a typical behaviour for coarse = 99 and fine = 255 in open loop mode.
32.6.4
120MHz RC Oscillator (RC120M) Characteristics
Table 32-15. Internal 120MHz RC Oscillator Characteristics
Symbol
Parameter
Conditions
(1)
fOUT
Output frequency
IRC120M
Current consumption
tSTARTUP
Startup time
Note:
VVDDCORE = 1.8V
Min
Typ
Max
Unit
88
120
152
MHz
1.85
mA
3
µs
1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
782
32099I–01/2012
AT32UC3L016/32/64
32.6.5
32kHz RC Oscillator (RC32K) Characteristics
Table 32-16. 32kHz RC Oscillator Characteristics
Symbol
Parameter
Conditions
(1)
Min
Typ
Max
Unit
20
32
44
kHz
fOUT
Output frequency
IRC32K
Current consumption
0.6
µA
tSTARTUP
Startup time
100
µs
Note:
1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
32.6.6
System RC Oscillator (RCSYS) Characteristics
Table 32-17. System RC Oscillator Characteristics
Symbol
Parameter
Conditions
fOUT
Output frequency
Calibrated at 85°C
32.7
Min
Typ
Max
Unit
111.6
115
118.4
kHz
Flash Characteristics
Table 32-18 gives the device maximum operating frequency depending on the number of flash
wait states and the flash read mode. The FSW bit in the FLASHCDW FSR register controls the
number of wait states used when accessing the flash memory.
Table 32-18. Maximum Operating Frequency
Flash Wait States
Read Mode
Maximum Operating Frequency
1
50MHz
High speed read mode
0
25MHz
1
30MHz
Normal read mode
0
15MHz
Table 32-19. Flash Characteristics
Symbol
Parameter
tFPP
Page programming time
tFPE
Page erase time
tFFP
Fuse programming time
tFEA
Full chip erase time (EA)
tFCE
JTAG chip erase time (CHIP_ERASE)
Conditions
Min
Typ
Max
Unit
5
5
fCLK_HSB = 50MHz
1
ms
5
fCLK_HSB = 115kHz
170
783
32099I–01/2012
AT32UC3L016/32/64
Table 32-20. Flash Endurance and Data Retention
Symbol
Parameter
NFARRAY
Array endurance (write/page)
100k
NFFUSE
General Purpose fuses endurance (write/bit)
10k
tRET
Data retention
15
32.8
Conditions
Min
Typ
Max
Unit
cycles
years
Analog Characteristics
32.8.1
Voltage Regulator Characteristics
Table 32-21. VREG Electrical Characteristics
Symbol
Parameter
VVDDIN
Input voltage range
VVDDCORE
Output voltage, calibrated value
Min
Typ
Max
1.98
3.3
3.6
Units
V
Output voltage accuracy
IOUT
DC output current(1)
IVREG
Static current of internal regulator
Note:
Condition
VVDDIN >= 1.98V
1.8
IOUT = 0.1mA to 60mA,
VVDDIN > 2.2V
2
IOUT = 0.1mA to 60mA,
VVDDIN = 1.98V to 2.2V
4
%
Normal mode
60
Low power mode
1
mA
Normal mode
20
Low power mode
6
µA
1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
Table 32-22. Decoupling Requirements
Symbol
Parameter
CIN1
Input regulator capacitor 1
33
CIN2
Input regulator capacitor 2
100
CIN3
Input regulator capacitor 3
10
µF
COUT1
Output regulator capacitor 1
100
nF
COUT2
Output regulator capacitor 2
2.2
Note:
Condition
Typ
Techno.
Units
nF
Tantalum
0.5<ESR<10Ohm
µF
1. Refer to Section 6.1.2 on page 36.
784
32099I–01/2012
AT32UC3L016/32/64
32.8.2
Power-on Reset 18 Characteristics
Table 32-23. POR18 Characteristics
Symbol
Parameter
Condition
VPOT+
Voltage threshold on VVDDCORE rising
VPOT-
Voltage threshold on VVDDCORE falling
tDET
Detection time
Min
Typ
Max
1.45
1.58
Units
V
1.2
Time with VDDCORE < VPOTnecessary to generate a reset
signal
1.32
460
µs
VVDDCORE
Figure 32-5. POR18 Operating Principles
VPOT+
Reset
VPOT-
785
32099I–01/2012
AT32UC3L016/32/64
32.8.3
Power-on Reset 33 Characteristics
Table 32-24. POR33 Characteristics
Symbol
Parameter
Condition
Min
Typ
Max
1.49
1.58
Units
VPOT+
Voltage threshold on VVDDIN rising
VPOT-
Voltage threshold on VVDDIN falling
tDET
Detection time
Time with VDDIN < VPOTnecessary to generate a reset
signal
460
µs
IPOR33
Current consumption
After tRESET
15
µA
tSTARTUP
Startup time
400
µs
V
1.3
1.45
VVDDIN
Figure 32-6. POR33 Operating Principles
Reset
VPOT+
VPOT-
32.8.4
Brown Out Detector Characteristics
The values in Table 32-25 describe the values of the BODLEVEL in the flash General Purpose
Fuse register.
Table 32-25. BODLEVEL Values
BODLEVEL Value
Min
Typ
011111 binary (31) 0x1F
1.56
100111 binary (39) 0x27
1.65
Max
Units
V
786
32099I–01/2012
AT32UC3L016/32/64
Table 32-26. BOD Characteristics
Symbol
Parameter
Condition
VHYST
BOD hysteresis
T = 25°C
10
mV
tDET
Detection time
Time with VDDCORE <
BODLEVEL necessary to
generate a reset signal
1
µs
IBOD
Current consumption
16
µA
tSTARTUP
Startup time
5
µs
32.8.5
Min
Typ
Max
Units
Supply Monitor 33 Characteristics
Table 32-27. SM33 Characteristics
Symbol
VTH
Parameter
Voltage threshold
Condition
(1)
Calibrated , T = 25°C
Min
Typ
Max
Units
1.675
1.75
1.825
V
Step size, between adjacent values
in SCIF.SM33.CALIB
11
VHYST
Hysteresis
30
tDET
Detection time
Time with VDDIN < VTH
necessary to generate a reset
signal
280
µs
ISM33
Current consumption
Normal mode
15
µA
tSTARTUP
Startup time
Normal mode
140
µs
Note:
mV
1. Calibration value can be read from the SCIF.SM33.CALIB field. This field is updated by the flash fuses after a reset. Refer to
SCIF chapter for details.
787
32099I–01/2012
AT32UC3L016/32/64
32.8.6
Analog to Digital Converter Characteristics
Table 32-28. ADC Characteristics
Symbol
Parameter
Conditions
fADC
ADC clock frequency
tSTARTUP
Startup time
Return from Idle Mode
tCONV
Conversion time (latency)
fADC = 6MHz
Throughput rate
Min
Typ
10-bit resolution mode
6
8-bit resolution mode
6
15
11
Units
MHz
µs
26
VVDD > 3.0V, fADC = 6MHz,
10-bit resolution mode,
low impedance source
460
VVDD > 3.0V, fADC = 6MHz,
8-bit resolution mode,
low impedance source
460
cycles
kSPS
VADVREFP
Reference voltage range
VADVREFP = VVDDANA
IADC
Current consumption on VVDDANA
ADC Clock = 6MHz
300
IADVREFP
Current consumption on ADVREFP
pin
fADC = 6MHz
250
Note:
Max
1.62
1.98
V
µA
These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process
technology. These values are not covered by test limits in production.
32.8.6.1
Inputs and Sample and Hold Aquisition Time
Table 32-29. Analog Inputs
Symbol
Parameter
VADn
Input Voltage Range
CONCHIP
Internal Capacitance(1)
RONCHIP
Note:
Conditions
Internal Resistance
(1)
10-bit mode
8-bit mode
Min
0
Typ
Max
Units
VADVREFP
V
21.5
pF
VVDDIO = 3.0V to 3.6V,
VVDDCORE = 1.8V
2.55
VVDDIO = VVDDCORE = 1.62V to 1.98V
55.3
kOhm
1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
An analog voltage input must be able to charge the sample and hold (S/H) capacitor in the ADC
in order to achieve maximum accuracy. Seen externally the ADC input consists of a resistor
( R ONCHIP ) and a capacitor ( CONCHIP ). In addition the resistance ( R SOURCE ) and capacitance
( C SOURCE ) of the PCB and source must be taken into account when calculating the sample and
hold time. Figure 32-7 shows the ADC input channel equivalent circuit.
788
32099I–01/2012
AT32UC3L016/32/64
Figure 32-7. ADC Input
RSOURCE
Positive Input
RONCHIP
CSOURCE
VIN
CONCHIP
ADCVREFP/2
The minimum sample and hold time (in ns) can be found using this formula:
t SAMPLEHOLD ≥ ( R ONCHIP + R OFFCHIP ) × ( C ONCHIP + C OFFCHIP ) × ln ( 2
n+1
)
Where n is the number of bits in the conversion. t SAMPLEHOLD is defined by the SHTIM field in the
ADCIFB ACR register. Please refer to the ADCIFB chapter for more information.
32.8.6.2
Applicable Conditions and Derating Data
Table 32-30. Transfer Characteristics 10-bit Resolution Mode
Parameter
Conditions
Min
Resolution
Offset error
Max
10
Integral non-linearity
Differential non-linearity
Typ
Units
Bit
+/-2
ADC clock frequency = 6MHz
-0.9
1
+/-4
Gain error
LSB
+/-4
Table 32-31. Transfer Characteristics 8-bit Resolution Mode
Parameter
Conditions
Min
Resolution
Offset error
Gain error
Max
8
Integral non-linearity
Differential non-linearity
Typ
Units
Bit
+/-0.5
ADC clock frequency = 6MHz
-0.23
0.25
+/-1
LSB
+/-1
789
32099I–01/2012
AT32UC3L016/32/64
32.8.7
Temperature Sensor Characteristics
Table 32-32. Temperature Sensor Characteristics(1)
Symbol
Parameter
Condition
Min
Typ
Max
Units
Gradient
1
mV/°C
ITS
Current consumption
0.5
µA
tSTARTUP
Startup time
0
µs
Note:
1. The Temperature Sensor is not calibrated. The accuracy of the Temperature Sensor is governed by the ADC accuracy.
32.8.8
Analog Comparator Characteristics
Table 32-33. Analog Comparator Characteristics
Symbol
Parameter
Condition
Min
Typ
Max
Positive input
voltage range
-0.2
VVDDIO + 0.3
Negative input
voltage range
-0.2
VVDDIO - 0.6
Units
V
Statistical offset
VACREFN = 1.0V,
fAC = 12MHz,
filter length = 2,
hysteresis = 0(1)
Clock frequency for
GCLK4
fAC
Throughput rate(3)
fAC = 12MHz
Propagation delay
Delay from input
change to
Interrupt Status
Register Changes
IAC
Current
consumption
All channels,
VDDIO = 3.3V,
fA = 3MHz
tSTARTUP
Startup time
Input current per
pin
Notes:
20
1
⎛
+ 3⎞⎠ × t CLKACIFB
⎝ t---------------------------------------CLKACIFB × f AC
mV
12
MHz
12 000 000
Comparisons
per second
ns
420
µA
3
cycles
0.2
µA/MHz(2)
1. AC.CONFn.FLEN and AC.CONFn.HYS fields, refer to the Analog Comparator Interface chapter.
2. Referring to fAC.
3. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
790
32099I–01/2012
AT32UC3L016/32/64
32.8.9
Capacitive Touch Characteristics
32.8.9.1
Discharge Current Source
Table 32-34. DICS Characteristics
Symbol
Parameter
RREF
Internal resistor
120
kOhm
k
Trim step size
0.7
%
32.8.9.2
Min
Typ
Max
Unit
Strong Pull-up Pull-down
Table 32-35. Strong Pull-up Pull-down
Parameter
Min
Typ
Pull-down resistor
1
Pull-up resistor
1
Max
Unit
kOhm
791
32099I–01/2012
AT32UC3L016/32/64
32.9
Timing Characteristics
32.9.1
Startup, Reset, and Wake-up Timing
The startup, reset, and wake-up timings are calculated using the following formula:
t = t CONST + N CPU × t CPU
Where t CONST and N CPU are found in Table 32-36. t CPU is the period of the CPU clock. If
another clock source than RCSYS is selected as CPU clock the startup time of the oscillator,
t OSCSTART , must added to the wake-up time in the stop, deepstop, and static sleep modes.
Please refer to the source for the CPU clock in the ”Oscillator Characteristics” on page 779 for
more details about oscillator startup times.
Table 32-36. Maximum Reset and Wake-up Timing(1)
Max t CONST (in µs)
Max N CPU
Parameter
Measuring
Startup time from power-up, using
regulator
Time from VDDIN crossing the VPOT+ threshold of
POR33 to the first instruction entering the decode
stage of CPU. VDDCORE is supplied by the internal
regulator.
2210
0
Startup time from power-up, no
regulator
Time from VDDIN crossing the VPOT+ threshold of
POR33 to the first instruction entering the decode
stage of CPU. VDDCORE is connected to VDDIN.
1810
0
Startup time from reset release
Time from releasing a reset source (except POR18,
POR33, and SM33) to the first instruction entering
the decode stage of CPU.
170
0
Idle
0
19
Frozen
0
110
0
110
27 + t OSCSTART
116
Deepstop
27 + t OSCSTART
116
Static
97 + t OSCSTART
116
1180
0
Standby
Wake-up
Stop
Wake-up from shutdown
Note:
From wake-up event to the first instruction of an
interrupt routine entering the decode stage of the
CPU.
From wake-up event to the first instruction entering
the decode stage of the CPU.
1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
32.9.2
RESET_N Timing
Table 32-37. RESET_N Waveform Parameters(1)
Symbol
Parameter
tRESET
RESET_N minimum pulse length
Note:
Conditions
Min
10
Max
Units
ns
1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
792
32099I–01/2012
AT32UC3L016/32/64
32.9.3
USART in SPI Mode Timing
32.9.3.1
Master mode
Figure 32-8. USART in SPI Master Mode With (CPOL= CPHA= 0) or (CPOL= CPHA= 1)
SPCK
MISO
USPI0
USPI1
MOSI
USPI2
Figure 32-9. USART in SPI Master Mode With (CPOL= 0 and CPHA= 1) or (CPOL= 1 and
CPHA= 0)
SPCK
MISO
USPI3
USPI4
MOSI
USPI5
Table 32-38. USART in SPI Mode Timing, Master Mode(1)
Symbol
Parameter
Conditions
USPI0
MISO setup time before SPCK rises
USPI1
MISO hold time after SPCK rises
USPI2
SPCK rising to MOSI delay
USPI3
MISO setup time before SPCK falls
USPI4
MISO hold time after SPCK falls
USPI5
SPCK falling to MOSI delay
Notes:
Min
30.0+
VVDDIO from
3.0V to 3.6V,
maximum
external
capacitor =
40pF
Max
Units
tSAMPLE(2)
0
8.5
ns
25.5 + tSAMPLE(2)
0
13.6
1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
t SPCK
1⎞
2. Where: t SAMPLE = t SPCK – ⎛ -------------------------------------- × t CLKUSART
⎝ 2×t
2⎠
CLKUSART
793
32099I–01/2012
AT32UC3L016/32/64
Maximum SPI Frequency, Master Output
The maximum SPI master output frequency is given by the following formula:
1 f CLKSPI × 2
f SPCKMAX = MIN (f PINMAX,------------, -----------------------------)
SPIn
9
Where SPIn is the MOSI delay, USPI2 or USPI5 depending on CPOL and NCPHA. f PINMAX is
the maximum frequency of the SPI pins. Please refer to the I/O Pin Characteristics section for
the maximum frequency of the pins. f CLKSPI is the maximum frequency of the CLK_SPI. Refer
to the SPI chapter for a description of this clock.
Maximum SPI Frequency, Master Input
The maximum SPI master input frequency is given by the following formula:
f CLKSPI × 2
1
f SPCKMAX = MIN (------------------------------------,-----------------------------)
SPIn + t VALID
9
Where SPIn is the MISO setup and hold time, USPI0 + USPI1 or USPI3 + USPI4 depending
on CPOL and NCPHA. T VALID is the SPI slave response time. Please refer to the SPI slave
datasheet for T VALID . f CLKSPI is the maximum frequency of the CLK_SPI. Refer to the SPI
chapter for a description of this clock.
32.9.3.2
Slave mode
Figure 32-10. USART in SPI Slave Mode With (CPOL= 0 and CPHA= 1) or (CPOL= 1 and
CPHA= 0)
SPCK
MISO
USPI6
MOSI
USPI7
USPI8
Figure 32-11. USART in SPI Slave Mode With (CPOL= CPHA= 0) or (CPOL= CPHA= 1)
SPCK
MISO
USPI9
MOSI
USPI10
USPI11
794
32099I–01/2012
AT32UC3L016/32/64
Figure 32-12. USART in SPI Slave Mode NPCS Timing
USPI12
USPI13
USPI14
USPI15
SPCK, CPOL=0
SPCK, CPOL=1
NSS
Table 32-39. USART in SPI mode Timing, Slave Mode(1)
Symbol
Parameter
Conditions
USPI6
SPCK falling to MISO delay
Max
Units
27.6
USPI7
MOSI setup time before SPCK rises
USPI8
MOSI hold time after SPCK rises
USPI9
SPCK rising to MISO delay
USPI10
MOSI setup time before SPCK falls
USPI11
MOSI hold time after SPCK falls
USPI12
NSS setup time before SPCK rises
USPI13
NSS hold time after SPCK falls
USPI14
NSS setup time before SPCK falls
USPI15
NSS hold time after SPCK rises
Notes:
Min
tSAMPLE(2)
+ tCLK_USART
0
VVDDIO from
3.0V to 3.6V,
maximum
external
capacitor =
40pF
27.2
tSAMPLE(2)
+ tCLK_USART
ns
0
25.0
0
25.0
0
1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
t SPCK
1
2. Where: t SAMPLE = t SPCK – ⎛ -----------------------------------+ ---⎞ × t CLKUSART
⎝ 2×t
2⎠
CLKUSART
Maximum SPI Frequency, Slave Input Mode
The maximum SPI slave input frequency is given by the following formula:
f CLKSPI × 2 1
f SPCKMAX = MIN (-----------------------------,------------)
9
SPIn
Where SPIn is the MOSI setup and hold time, USPI7 + USPI8 or USPI10 + USPI11 depending
on CPOL and NCPHA. f CLKSPI is the maximum frequency of the CLK_SPI. Refer to the SPI
chapter for a description of this clock.
Maximum SPI Frequency, Slave Output Mode
795
32099I–01/2012
AT32UC3L016/32/64
The maximum SPI slave output frequency is given by the following formula:
f CLKSPI × 2
1
f SPCKMAX = MIN (-----------------------------, f PINMAX,------------------------------------)
9
SPIn + t SETUP
Where SPIn is the MISO delay, USPI6 or USPI9 depending on CPOL and NCPHA. T SETUP is
the SPI master setup time. Please refer to the SPI master datasheet for T SETUP . f CLKSPI is the
maximum frequency of the CLK_SPI. Refer to the SPI chapter for a description of this
clock. f PINMAX is the maximum frequency of the SPI pins. Please refer to the I/O Pin Characteristics section for the maximum frequency of the pins.
32.9.4
SPI Timing
32.9.4.1
Master mode
Figure 32-13. SPI Master Mode With (CPOL= NCPHA= 0) or (CPOL= NCPHA= 1)
SPCK
MISO
SPI0
SPI1
MOSI
SPI2
Figure 32-14. SPI Master Mode With (CPOL= 0 and NCPHA= 1) or (CPOL= 1 and NCPHA= 0)
SPCK
MISO
SPI3
SPI4
MOSI
SPI5
796
32099I–01/2012
AT32UC3L016/32/64
Table 32-40. SPI Timing, Master Mode(1)
Symbol
Parameter
SPI0
MISO setup time before SPCK rises
SPI1
MISO hold time after SPCK rises
SPI2
SPCK rising to MOSI delay
SPI3
MISO setup time before SPCK falls
SPI4
MISO hold time after SPCK falls
SPI5
SPCK falling to MOSI delay
Note:
Conditions
Min
Max
Units
28.4 + (tCLK_SPI)/2
VVDDIO from
3.0V to 3.6V,
maximum
external
capacitor =
40pF
0
7.1
ns
22.8 + (tCLK_SPI)/2
0
11.0
1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
Maximum SPI Frequency, Master Output
The maximum SPI master output frequency is given by the following formula:
1
f SPCKMAX = MIN (f PINMAX,------------)
SPIn
Where SPIn is the MOSI delay, SPI2 or SPI5 depending on CPOL and NCPHA. f PINMAX is the
maximum frequency of the SPI pins. Please refer to the I/O Pin Characteristics section for the
maximum frequency of the pins.
Maximum SPI Frequency, Master Input
The maximum SPI master input frequency is given by the following formula:
1
f SPCKMAX = -----------------------------------SPIn + t VALID
Where SPIn is the MISO setup and hold time, SPI0 + SPI1 or SPI3 + SPI4 depending on
CPOL and NCPHA. t VALID is the SPI slave response time. Please refer to the SPI slave
datasheet for t VALID .
32.9.4.2
Slave mode
Figure 32-15. SPI Slave Mode With (CPOL= 0 and NCPHA= 1) or (CPOL= 1 and NCPHA= 0)
SPCK
MISO
SPI6
MOSI
SPI7
SPI8
797
32099I–01/2012
AT32UC3L016/32/64
Figure 32-16. SPI Slave Mode With (CPOL= NCPHA= 0) or (CPOL= NCPHA= 1)
SPCK
MISO
SPI9
MOSI
SPI10
SPI11
Figure 32-17. SPI Slave Mode NPCS Timing
SPI12
SPI13
SPI14
SPI15
SPCK, CPOL=0
SPCK, CPOL=1
NPCS
Table 32-41. SPI Timing, Slave Mode(1)
Symbol
Parameter
SPI6
SPCK falling to MISO delay
SPI7
MOSI setup time before SPCK rises
SPI8
MOSI hold time after SPCK rises
SPI9
SPCK rising to MISO delay
SPI10
MOSI setup time before SPCK falls
SPI11
MOSI hold time after SPCK falls
SPI12
NPCS setup time before SPCK rises
SPI13
NPCS hold time after SPCK falls
0.2
SPI14
NPCS setup time before SPCK falls
2.2
SPI15
NPCS hold time after SPCK rises
Note:
Conditions
Min
Max
Units
30.8
0
4.1
VVDDIO from
3.0V to 3.6V,
maximum
external
capacitor =
40pF
29.9
0
ns
3.5
1.9
0
1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
Maximum SPI Frequency, Slave Input Mode
798
32099I–01/2012
AT32UC3L016/32/64
The maximum SPI slave input frequency is given by the following formula:
1
f SPCKMAX = MIN (f CLKSPI,------------)
SPIn
Where SPIn is the MOSI setup and hold time, SPI7 + SPI8 or SPI10 + SPI11 depending on
CPOL and NCPHA. f CLKSPI is the maximum frequency of the CLK_SPI. Refer to the SPI chapter for a description of this clock.
Maximum SPI Frequency, Slave Output Mode
The maximum SPI slave output frequency is given by the following formula:
1
f SPCKMAX = MIN (f PINMAX,------------------------------------)
SPIn + t SETUP
Where SPIn is the MISO delay, SPI6 or SPI9 depending on CPOL and NCPHA. t SETUP is the
SPI master setup time. Please refer to the SPI master datasheet for t SETUP . f PINMAX is the maximum frequency of the SPI pins. Please refer to the I/O Pin Characteristics section for the
maximum frequency of the pins.
32.9.5
TWIM/TWIS Timing
Figure 32-42 shows the TWI-bus timing requirements and the compliance of the device with
them. Some of these requirements (tr and tf) are met by the device without requiring user intervention. Compliance with the other requirements (tHD-STA, tSU-STA, tSU-STO, tHD-DAT, tSU-DAT-TWI, tLOWTWI, tHIGH, and fTWCK) requires user intervention through appropriate programming of the relevant
TWIM and TWIS user interface registers. Please refer to the TWIM and TWIS sections for more
information.
Table 32-42. TWI-Bus Timing Requirements
Minimum
Symbol
Parameter
Mode
Requirement
Standard(1)
tr
TWCK and TWD rise time
tf
TWCK and TWD fall time
tHD-STA
(Repeated) START hold time
tSU-STA
(Repeated) START set-up time
tSU-STO
STOP set-up time
tHD-DAT
Data hold time
Maximum
Device
Requirement
Device
-
1000
20 + 0.1Cb
300
-
300
20 + 0.1Cb
300
Unit
ns
Fast(1)
Standard
ns
Fast
Standard
4
Fast
0.6
Standard
4.7
Fast
0.6
Standard
4.0
Fast
0.6
Standard
Fast
0.3(2)
tclkpb
-
μs
tclkpb
-
μs
4tclkpb
-
μs
3.45()
2tclkpb
0.9()
15tprescaled + tclkpb
μs
799
32099I–01/2012
AT32UC3L016/32/64
Table 32-42. TWI-Bus Timing Requirements
Minimum
Symbol
Parameter
tSU-DAT-TWI Data set-up time
tSU-DAT
tLOW-TWI
Mode
Requirement
Standard
250
Fast
100
-
Device
-
Standard
4.7
Fast
1.3
TWCK LOW period
tLOW
-
tHIGH
TWCK HIGH period
fTWCK
TWCK frequency
-
Standard
4.0
Fast
0.6
Standard
Notes:
Maximum
Requirement
Unit
2tclkpb
-
ns
tclkpb
-
-
4tclkpb
-
μs
tclkpb
-
-
8tclkpb
-
μs
100
-
Fast
Device
400
1
-----------------------12t clkpb
kHz
1. Standard mode: f TWCK ≤ 100 kHz ; fast mode: f TWCK > 100 kHz .
2. A device must internally provide a hold time of at least 300 ns for TWD with reference to the falling edge of TWCK.
Notations:
Cb = total capacitance of one bus line in pF
tclkpb = period of TWI peripheral bus clock
tprescaled = period of TWI internal prescaled clock (see chapters on TWIM and TWIS)
The maximum tHD;DAT has only to be met if the device does not stretch the LOW period (tLOW-TWI)
of TWCK.
800
32099I–01/2012
AT32UC3L016/32/64
32.9.6
JTAG Timing
Figure 32-18. JTAG Interface Signals
JTAG2
TCK
JTAG0
JTAG1
TMS/TDI
JTAG3
JTAG4
JTAG7
JTAG8
TDO
JTAG5
JTAG6
Boundary
Scan Inputs
Boundary
Scan Outputs
JTAG9
JTAG10
Table 32-43. JTAG Timings(1)
Symbol
Parameter
JTAG0
TCK Low Half-period
23.2
JTAG1
TCK High Half-period
8.8
JTAG2
TCK Period
32.0
JTAG3
TDI, TMS Setup before TCK High
JTAG4
TDI, TMS Hold after TCK High
JTAG5
TDO Hold Time
JTAG6
TCK Low to TDO Valid
JTAG7
Boundary Scan Inputs Setup Time
JTAG8
Boundary Scan Inputs Hold Time
5.0
JTAG9
Boundary Scan Outputs Hold Time
8.7
JTAG10
TCK to Boundary Scan Outputs Valid
Note:
Conditions
VVDDIO from
3.0V to 3.6V,
maximum
external
capacitor =
40pF
Min
Max
Units
3.9
0.6
ns
4.5
23.2
0
17.7
1. These values are based on simulation and characterization of other AVR microcontrollers manufactured in the same process technology. These values are not covered by test limits in production.
801
32099I–01/2012
AT32UC3L016/32/64
33. Mechanical Characteristics
33.1
33.1.1
Thermal Considerations
Thermal Data
Table 33-1 summarizes the thermal resistance data depending on the package.
Table 33-1.
33.1.2
Thermal Resistance Data
Symbol
Parameter
Condition
Package
Typ
θJA
Junction-to-ambient thermal resistance
Still Air
TQFP48
63.2
θJC
Junction-to-case thermal resistance
TQFP48
21.8
θJA
Junction-to-ambient thermal resistance
QFN48
28.3
θJC
Junction-to-case thermal resistance
QFN48
2.5
θJA
Junction-to-ambient thermal resistance
TLLGA48
25.4
θJC
Junction-to-case thermal resistance
TLLGA48
12.7
Still Air
Still Air
Unit
°C/W
°C/W
°C/W
Junction Temperature
The average chip-junction temperature, TJ, in °C can be obtained from the following:
1.
T J = T A + ( P D × θ JA )
2.
T J = T A + ( P D × ( θ HEATSINK + θ JC ) )
where:
• θJA = package thermal resistance, Junction-to-ambient (°C/W), provided in Table 33-1.
• θJC = package thermal resistance, Junction-to-case thermal resistance (°C/W), provided in
Table 33-1.
• 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 32.4 on
page 771.
• 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.
802
32099I–01/2012
AT32UC3L016/32/64
33.2
Package Drawings
Figure 33-1. TQFP-48 Package Drawing
Table 33-2.
Device and Package Maximum Weight
140
Table 33-3.
mg
Package Characteristics
Moisture Sensitivity Level
Table 33-4.
MSL3
Package Reference
JEDEC Drawing Reference
MS-026
JESD97 Classification
E3
803
32099I–01/2012
AT32UC3L016/32/64
Figure 33-2. QFN-48 Package Drawing
Note:
The exposed pad is not connected to anything internally, but should be soldered to ground to increase board level reliability.
Table 33-5.
Device and Package Maximum Weight
140
Table 33-6.
mg
Package Characteristics
Moisture Sensitivity Level
Table 33-7.
MSL3
Package Reference
JEDEC Drawing Reference
M0-220
JESD97 Classification
E3
804
32099I–01/2012
AT32UC3L016/32/64
Figure 33-3. TLLGA-48 Package Drawing
Table 33-8.
Device and Package Maximum Weight
39.3
Table 33-9.
mg
Package Characteristics
Moisture Sensitivity Level
MSL3
Table 33-10. Package Reference
JEDEC Drawing Reference
N/A
JESD97 Classification
E4
805
32099I–01/2012
AT32UC3L016/32/64
33.3
Soldering Profile
Table 33-11 gives the recommended soldering profile from J-STD-20.
Table 33-11. Soldering Profile
Profile Feature
Green Package
Average Ramp-up Rate (217°C to Peak)
3°C/s max
Preheat Temperature 175°C ±25°C
150-200°C
Time Maintained Above 217°C
60-120s
Time within 5°C of Actual Peak Temperature
30s
Peak Temperature Range
260°C
Ramp-down Rate
6°C/s max
Time 25°C to Peak Temperature
8 minutes max
A maximum of three reflow passes is allowed per component.
806
32099I–01/2012
AT32UC3L016/32/64
34. Ordering Information
Table 34-1.
Device
Ordering Information
Ordering Code
Carrier Type
AT32UC3L064-AUTES
ES
AT32UC3L064-AUT
Tray
AT32UC3L064-AUR
Tape & Reel
Package
Package Type
Temperature Operating
Range
TQFP 48
JESD97 Classification E3
AT32UC3L064
AT32UC3L064-ZAUES
ES
AT32UC3L064-ZAUT
Tray
AT32UC3L064-ZAUR
Tape & Reel
AT32UC3L064-D3HES
ES
AT32UC3L064-D3HT
Tray
AT32UC3L064-D3HR
Tape & Reel
AT32UC3L032-AUT
Tray
AT32UC3L032-AUR
Tape & Reel
QFN 48
TLLGA 48
JESD97 Classification E4
TQFP 48
Industrial (-40°C to 85°C)
JESD97 Classification E3
AT32UC3L032-ZAUT
Tray
AT32UC3L032-ZAUR
Tape & Reel
AT32UC3L032-D3HT
Tray
AT32UC3L032-D3HR
Tape & Reel
AT32UC3L032
QFN 48
TLLGA 48
AT32UC3L016-AUT
Tray
AT32UC3L016-AUR
Tape & Reel
AT32UC3L016-ZAUT
Tray
AT32UC3L016-ZAUR
Tape & Reel
AT32UC3L016-D3HT
Tray
AT32UC3L016-D3HR
Tape & Reel
JESD97 Classification E4
TQFP 48
JESD97 Classification E3
AT32UC3L016
QFN 48
TLLGA 48
JESD97 Classification E4
807
32099I–01/2012
AT32UC3L016/32/64
35. Errata
35.1
35.1.1
Rev. E
Processor and Architecture
1. Hardware breakpoints may corrupt MAC results
Hardware breakpoints on MAC instructions may corrupt the destination register of the MAC
instruction.
Fix/Workaround
Place breakpoints on earlier or later instructions.
2. Privilege violation when using interrupts in application mode with protected system
stack
If the system stack is protected by the MPU and an interrupt occurs in application mode, an
MPU DTLB exception will occur.
Fix/Workaround
Make a DTLB Protection (Write) exception handler which permits the interrupt request to be
handled in privileged mode.
35.1.2
FLASHCDW
1. Flash self programming may fail in one wait state mode
Writes in flash and user pages may fail if executing code is located in address space
mapped to flash, and the flash controller is configured in one wait state mode (the Flash
Wait State bit in the Flash Control Register (FCR.FWS) is one).
Fix/Workaround
Solution 1: Configure the flash controller in zero wait state mode (FCR.FWS=0).
Solution 2: Configure the HMATRIX master 1 (CPU Instruction) to use the unlimited burst
length transfer mode (MCFG1.ULBT=0), and the HMATRIX slave 0 (FLASHCDW) to use
the maximum slot cycle limit (SCFG0.SLOT_CYCLE=255).
35.1.3
Power Manager
1. Clock sources will not be stopped in Static mode if the difference between CPU and
PBx division factor is larger than 4
If the division factor between the CPU/HSB and PBx frequencies is more than 4 when entering a sleep mode where the system RC oscillator (RCSYS) is turned off, the high speed
clock sources will not be turned off. This will result in a significantly higher power consumption during the sleep mode.
Fix/Workaround
Before going to sleep modes where RCSYS is stopped, make sure the division factor
between CPU/HSB and PBx frequencies is less than or equal to 4.
2. Clock Failure Detector (CFD) can be issued while turning off the CFD
While turning off the CFD, the CFD bit in the Status Register (SR) can be set. This will
change the main clock source to RCSYS.
Fix/Workaround
Solution 1: Enable CFD interrupt. If CFD interrupt is issues after turning off the CFD, switch
back to original main clock source.
Solution 2: Only turn off the CFD while running the main clock on RCSYS.
3. Sleepwalking in idle and frozen sleep mode will mask all other PB clocks
808
32099I–01/2012
AT32UC3L016/32/64
If the CPU is in idle or frozen sleep mode and a module is in a state that triggers sleep walking, all PB clocks will be masked except the PB clock to the sleepwalking module.
Fix/Workaround
Mask all clock requests in the PM.PPCR register before going into idle or frozen mode.
35.1.4
SCIF
1. PCLKSR.OSC32RDY bit might not be cleared after disabling OSC32K
In some cases the OSC32RDY bit in the PCLKSR register will not be cleared when OSC32K
is disabled.
Fix/Workaround
When re-enabling the OSC32K, read the PCLKSR.OSC32RDY bit. If this bit is:
0: Follow normal procedures.
1: Ignore the PCLKSR.OSC32RDY and ISR.OSC32RDY bit. Use the Frequency Meter
(FREQM) to determine if the OSC32K clock is ready. The OSC32K clock is ready when the
FREQM measures a non-zero frequency.
2. The RC32K output on PA20 is not always permanently disabled
The RC32K output on PA20 may sometimes re-appear.
Fix/Workaround
Before using RC32K for other purposes, the following procedure has to be followed in order
to properly disable it:
- Run the CPU on RCSYS
- Disable the output to PA20 by writing a zero to PM.PPCR.RC32OUT
- Enable RC32K by writing a one to SCIF.RC32KCR.EN, and wait for this bit to be read as
one
- Disable RC32K by writing a zero to SCIF.RC32KCR.EN, and wait for this bit to be read as
zero.
35.1.5
AST
1. Reset may set status bits in the AST
If a reset occurs and the AST is enabled, the SR.ALARM0, SR.PER0, and SR.OVF bits may
be set.
Fix/Workaround
If the part is reset and the AST is used, clear all bits in the Status Register before entering
sleep mode.
2. AST wake signal is released one AST clock cycle after the BUSY bit is cleared
After writing to the Status Clear Register (SCR) the wake signal is released one AST clock
cycle after the BUSY bit in the Status Register (SR.BUSY) is cleared. If entering sleep mode
directly after the BUSY bit is cleared the part will wake up immediately.
Fix/Workaround
Read the Wake Enable Register (WER) and write this value back to the same register. Wait
for BUSY to clear before entering sleep mode.
35.1.6
WDT
1. Clearing the Watchdog Timer (WDT) counter in second half of timeout period will
issue a Watchdog reset
If the WDT counter is cleared in the second half of the timeout period, the WDT will immediately issue a Watchdog reset.
Fix/Workaround
809
32099I–01/2012
AT32UC3L016/32/64
Use twice as long timeout period as needed and clear the WDT counter within the first half
of the timeout period. If the WDT counter is cleared after the first half of the timeout period,
you will get a Watchdog reset immediately. If the WDT counter is not cleared at all, the time
before the reset will be twice as long as needed.
2. WDT Control Register does not have synchronization feedback
When writing to the Timeout Prescale Select (PSEL), Time Ban Prescale Select (TBAN),
Enable (EN), or WDT Mode (MODE) fieldss of the WDT Control Register (CTRL), a synchronizer is started to propagate the values to the WDT clcok domain. This synchronization
takes a finite amount of time, but only the status of the synchronization of the EN bit is
reflected back to the user. Writing to the synchronized fields during synchronization can lead
to undefined behavior.
Fix/Workaround
-When writing to the affected fields, the user must ensure a wait corresponding to 2 clock
cycles of both the WDT peripheral bus clock and the selected WDT clock source.
-When doing writes that changes the EN bit, the EN bit can be read back until it reflects the
written value.
35.1.7
GPIO
1. Clearing GPIO interrupt may fail
Writing a one to the GPIO.IFRC register to clear an interrupt will be ignored if interrupt is
enabled for the corresponding port.
Fix/Workaround
Disable the interrupt, clear it by writing a one to GPIO.IFRC, then enable the interrupt.
35.1.8
SPI
1. SPI data transfer hangs with CSR0.CSAAT==1 and MR.MODFDIS==0
When CSR0.CSAAT==1 and mode fault detection is enabled (MR.MODFDIS==0), the SPI
module will not start a data transfer.
Fix/Workaround
Disable mode fault detection by writing a one to MR.MODFDIS.
2. Disabling SPI has no effect on the SR.TDRE bit
Disabling SPI has no effect on the SR.TDRE bit whereas the write data command is filtered
when SPI is disabled. Writing to TDR when SPI is disabled will not clear SR.TDRE. If SPI is
disabled during a PDCA transfer, the PDCA will continue to write data to TDR until its buffer
is empty, and this data will be lost.
Fix/Workaround
Disable the PDCA, add two NOPs, and disable the SPI. To continue the transfer, enable the
SPI and PDCA.
3. SPI disable does not work in SLAVE mode
SPI disable does not work in SLAVE mode.
Fix/Workaround
Read the last received data, then perform a software reset by writing a one to the Software
Reset bit in the Control Register (CR.SWRST).
4. SPI bad serial clock generation on 2nd chip_select when SCBR=1, CPOL=1, and
NCPHA=0
When multiple chip selects (CS) are in use, if one of the baudrates equal 1 while one
(CSRn.SCBR=1) of the others do not equal 1, and CSRn.CPOL=1 and CSRn.NCPHA=0,
then an additional pulse will be generated on SCK.
810
32099I–01/2012
AT32UC3L016/32/64
Fix/Workaround
When multiple CS are in use, if one of the baudrates equals 1, the others must also equal 1
if CSRn.CPOL=1 and CSRn.NCPHA=0.
5. SPI mode fault detection enable causes incorrect behavior
When mode fault detection is enabled (MR.MODFDIS==0), the SPI module may not operate
properly.
Fix/Workaround
Always disable mode fault detection before using the SPI by writing a one to MR.MODFDIS.
35.1.9
TWI
1. TWIM SR.IDLE goes high immediately when NAK is received
When a NAK is received and there is a non-zero number of bytes to be transmitted,
SR.IDLE goes high immediately and does not wait for the STOP condition to be sent. This
does not cause any problem just by itself, but can cause a problem if software waits for
SR.IDLE to go high and then immediately disables the TWIM by writing a one to CR.MDIS.
Disabling the TWIM causes the TWCK and TWD pins to go high immediately, so the STOP
condition will not be transmitted correctly.
Fix/Workaround
If possible, do not disable the TWIM. If it is absolutely necessary to disable the TWIM, there
must be a software delay of at least two TWCK periods between the detection of
SR.IDLE==1 and the disabling of the TWIM.
2. TWIM TWALM polarity is wrong
The TWALM signal in the TWIM is active high instead of active low.
Fix/Workaround
Use an external inverter to invert the signal going into the TWIM. When using both TWIM
and TWIS on the same pins, the TWALM cannot be used.
3. TWIS may not wake the device from sleep mode
If the CPU is put to a sleep mode (except Idle and Frozen) directly after a TWI Start condition, the CPU may not wake upon a TWIS address match. The request is NACKed.
Fix/Workaround
When using the TWI address match to wake the device from sleep, do not switch to sleep
modes