NSC CR16MBR5VJI1 Family of 16-bit can-enabled compactrisc microcontroller Datasheet

CR16HCS5/CR16HCS9/CR16MAR5/CR16MAS5
CR16MAS9/CR16MBR5/CR16MCS5/CR16MCS9
Family of 16-bit CAN-enabled CompactRISC
Microcontrollers
1.0
General Description
The family of 16-bit CompactRISC™ microcontroller is
based on a Reduced Instruction Set Computer (RISC) architecture. The device operates as a complete microcomputer with all system timing, interrupt logic, flash program
memory or ROM memory, RAM, EEPROM data memory,
and I/O ports included on-chip. It is ideally suited to a wide
range of embedded controller applications because of its
high performance, on-chip integrated features and low
power consumption resulting in decreased system cost.
plex Instruction Set Computer (CISC): compact code, onchip memory and I/O, and reduced cost. The CPU uses a
three-stage instruction pipeline that allows execution of up
to one instruction per clock cycle, or up to 25 million instructions per second (MIPS) at a clock rate of 24 MHz.
The device contains a FullCAN class, CAN serial interface
for low/high speed applications with 15 orthogonal message buffers, each supporting standard as well as extended message identifiers.
The device offers the high performance of a RISC architecture while retaining the advantages of a traditional Com-
Block Diagram
CR16B
RISC Core
Fast Clk
Processing
Unit
CR16CAN
FullCAN 2.0B
Clock Generator
Power-on-Reset
Core Bus
Peripheral
Bus
Controller
64k-Byte
Flash
Program
Memory
Slow Clk*
3k-Byte
RAM
2176-Byte 1.5k-Byte
ISP
EEPROM Memory
Data
Memory
Interrupt
Power
Management
12-ch
8-bit A/D
MIWU
Control
Timing and
Watchdog
Peripheral Bus
I/O
µWire/SPI
2x
USART
ACCESS
bus
4x
VTU
2x
MFT
2 Analog
Comparators
Please note that not all family members contain same peripheral modules and features.
TRI-STATE® is a registered trademark of National Semiconductor Corporation.
©2001 National Semiconductor Corporation
www.national.com
CR16HCS5/CR16HCS9/CR16MAR5/CR16MAS5 CR16MAS9/CR16MBR5/CR16MCS5/CR16MCS9 Family of 16-bit CAN-enabled CompactRISC Microcontrollers
January 2002
Table of Contents
1.0
2.0
3.0
4.0
5.0
6.0
7.0
8.0
9.0
10.0
11.0
12.0
General Description. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 1
Features . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 3
Device Overview . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.1
CR16B CPU Core . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.2
Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.3
Input/Output Ports . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.4
Bus Interface Unit . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.5
Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 5
3.6
Multi-Input Wake-up . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.7
Dual Clock and Reset . . . . . . . . . . . . . . . . . . . . . . . 6
3.8
Power Management . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.9
Multi-Function Timer . . . . . . . . . . . . . . . . . . . . . . . . 6
3.10 Versatile timer unit . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.11 Real-Time TIMER and Watchdog . . . . . . . . . . . . . . 6
3.12 USART . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.13 MICROWIRE/SPI . . . . . . . . . . . . . . . . . . . . . . . . . . . 6
3.14 CR16CAN . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.15 ACCESS.bus Interface . . . . . . . . . . . . . . . . . . . . . . 7
3.16 A/D Converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 7
3.17 Analog Comparators . . . . . . . . . . . . . . . . . . . . . . . . 7
3.18 Development Support . . . . . . . . . . . . . . . . . . . . . . . 7
Device Pinouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 8
4.1
Pin Description . . . . . . . . . . . . . . . . . . . . . . . . . . . . 10
System Configuration . . . . . . . . . . . . . . . . . . . . . . . . . . . 12
5.1
ENV0 and ENV1 Pins . . . . . . . . . . . . . . . . . . . . . . 12
5.2
Module Configuration (MCFG) Register . . . . . . . . 12
5.3
Module Status (MSTAT) Register . . . . . . . . . . . . . 12
Input/Output Ports . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
6.1
Port Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . . 13
6.2
Open-Drain Operation . . . . . . . . . . . . . . . . . . . . . . 14
CPU and Core Registers . . . . . . . . . . . . . . . . . . . . . . . . . 15
7.1
General-Purpose Registers . . . . . . . . . . . . . . . . . . 15
7.2
Dedicated Address Registers . . . . . . . . . . . . . . . . 15
7.3
Processor Status Register . . . . . . . . . . . . . . . . . . . 15
7.4
Configuration Register . . . . . . . . . . . . . . . . . . . . . . 16
7.5
Addressing Modes . . . . . . . . . . . . . . . . . . . . . . . . . 16
7.6
Stacks . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
7.7
Instruction Set . . . . . . . . . . . . . . . . . . . . . . . . . . . . 16
Bus Interface Unit . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
8.1
Bus Cycles . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 18
8.2
BIU Control Registers . . . . . . . . . . . . . . . . . . . . . . 18
8.3
Wait and Hold States Used . . . . . . . . . . . . . . . . . . 20
Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 22
9.1
Flash EEPROM Program Memory. . . . . . . . . . . . . 22
9.2
RAM Memory. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 25
9.3
Flash EEPROM Data Memory. . . . . . . . . . . . . . . . 25
9.4
ISP Memory . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 27
Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 31
10.1 Interrupt Operation . . . . . . . . . . . . . . . . . . . . . . . . . 31
10.2 Non-Maskable Interrupt . . . . . . . . . . . . . . . . . . . . . 32
10.3 Maskable Interrupts . . . . . . . . . . . . . . . . . . . . . . . . 32
10.4 Interrupt Registers . . . . . . . . . . . . . . . . . . . . . . . . . 33
10.5 Interrupt Programming Procedures . . . . . . . . . . . . 35
Power Management . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
11.1 Active Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
11.2 Power Save Mode . . . . . . . . . . . . . . . . . . . . . . . . . 36
11.3 Idle Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
11.4 Halt Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 36
11.5 Clock Inputs and Reset Configuration . . . . . . . . . . 36
11.6 Switching Between Power Modes . . . . . . . . . . . . . 36
Dual Clock and Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . 39
12.1 External Crystal Network . . . . . . . . . . . . . . . . . . . . 39
12.2 Main System Clock . . . . . . . . . . . . . . . . . . . . . . . . 40
12.3 Slow System Clock . . . . . . . . . . . . . . . . . . . . . . . . 40
12.4 Power-On Reset . . . . . . . . . . . . . . . . . . . . . . . . . . 41
12.5 External Reset . . . . . . . . . . . . . . . . . . . . . . . . . . . . 41
12.6 Dual Clock and Reset Registers . . . . . . . . . . . . . . 41
12.7 Slow Clock Prescaler Register (PRSSC) . . . . . . . . 41
www.national.com
13.0
14.0
15.0
16.0
17.0
18.0
19.0
20.0
21.0
22.0
23.0
24.0
25.0
26.0
2
12.8 Slow Clock Prescaler 1 Register (PRSSC1) . . . . .41
Multi-Input Wake-Up . . . . . . . . . . . . . . . . . . . . . . . . . . . .42
13.1 Wake-Up Edge Detection Register (WKEDG) . . . .42
13.2 Wake-Up Enable Register (WKENA) . . . . . . . . . . .42
13.3 Wake-Up Interrupt Control Register 1 (WKCTRL1) 43
13.4 Wake-Up Interrupt Control Register 1 (WKCTRL2) 43
13.5 Wake-Up Pending Register (WKPND) . . . . . . . . . .43
13.6 Wake-Up Pending Clear Register (WKPCL) . . . . .43
13.7 Programming Procedures . . . . . . . . . . . . . . . . . . .44
Real-Time Timer and WATCHDOG . . . . . . . . . . . . . . . . .45
14.1 TWM Structure . . . . . . . . . . . . . . . . . . . . . . . . . . . .45
14.2 Timer T0 Operation . . . . . . . . . . . . . . . . . . . . . . . .45
14.3 WATCHDOG Operation . . . . . . . . . . . . . . . . . . . . .46
14.4 TWM Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . .46
14.5 WATCHDOG Programming Procedure . . . . . . . . .47
Multi-Function Timer . . . . . . . . . . . . . . . . . . . . . . . . . . . .49
15.1 Timer Structure. . . . . . . . . . . . . . . . . . . . . . . . . . . .49
15.2 Timer Operating Modes . . . . . . . . . . . . . . . . . . . . .51
15.3 Timer Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . .54
15.4 Timer I/O Functions . . . . . . . . . . . . . . . . . . . . . . . .54
15.5 Timer Registers . . . . . . . . . . . . . . . . . . . . . . . . . . .56
Versatile-Timer-Unit (VTU) . . . . . . . . . . . . . . . . . . . . . . .58
16.1 VTU Functional Description . . . . . . . . . . . . . . . . . .58
16.2 VTU Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . .61
MICROWIRE/SPI . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .65
17.1 MICROWIRE Operation . . . . . . . . . . . . . . . . . . . . .65
17.2 Master Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . .66
17.3 Slave Mode . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .67
17.4 Interrupt Generation . . . . . . . . . . . . . . . . . . . . . . . .68
17.5 MICROWIRE Interface Registers. . . . . . . . . . . . . .68
USART . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .71
18.1 Functional Overview . . . . . . . . . . . . . . . . . . . . . . . .71
18.2 USART Operation . . . . . . . . . . . . . . . . . . . . . . . . .71
18.3 USART Registers . . . . . . . . . . . . . . . . . . . . . . . . . .75
18.4 Baud Rate Calculations . . . . . . . . . . . . . . . . . . . . .77
ACCESS.bus Interface . . . . . . . . . . . . . . . . . . . . . . . . . . .78
19.1 ACB Protocol Overview . . . . . . . . . . . . . . . . . . . . .78
19.2 ACB Functional Description . . . . . . . . . . . . . . . . . .79
19.3 ACB Registers . . . . . . . . . . . . . . . . . . . . . . . . . . . .82
19.4 Usage Hints . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .84
CR16CAN Module. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .85
20.1 Functional Description . . . . . . . . . . . . . . . . . . . . . .85
20.2 Basic CAN Concepts . . . . . . . . . . . . . . . . . . . . . . .87
20.3 Message Transfer . . . . . . . . . . . . . . . . . . . . . . . . .95
20.4 Acceptance Filtering . . . . . . . . . . . . . . . . . . . . . . . .96
20.5 Receive Structure . . . . . . . . . . . . . . . . . . . . . . . . . .97
20.6 Transmit Structure . . . . . . . . . . . . . . . . . . . . . . . . 100
20.7 Interrupts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 103
20.8 Time Stamp Counter . . . . . . . . . . . . . . . . . . . . . . 105
20.9 Memory Organization . . . . . . . . . . . . . . . . . . . . . .105
20.10 System Start-Up and Multi-Input Wake-Up . . . . .116
Analog Comparators . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
21.1 Analog Comparator Control/Status Register
(CMPCTRL). . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 118
21.2 Analog Comparator Usage . . . . . . . . . . . . . . . . . . 118
A/D Converter . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .119
22.1 Operating Modes . . . . . . . . . . . . . . . . . . . . . . . . . 119
22.2 A/D Converter Registers . . . . . . . . . . . . . . . . . . .120
22.3 A/D Converter Programming . . . . . . . . . . . . . . . . 122
Memory Map . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .123
Register Layouts . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
24.1 Register layout . . . . . . . . . . . . . . . . . . . . . . . . . . . 129
ELECTRICAL AND THERMAL CHARACTERISTICS . . 136
Appendix . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
26.1 CR16CAN. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
26.2 8/16-bit microwire/spi (MWSPI16) . . . . . . . . . . . . 154
26.3 Timing and watchdog module . . . . . . . . . . . . . . . 154
1.0
General Description (Continued)
The device has up to 64K bytes of reprogrammable flash EEPROM program memory or ROM memory, 1.5K bytes of
flash EEPROM In-System-Programming memory, 3K bytes
of static RAM, 2K bytes of non-volatile EEPROM data memory and 128 bytes with high endurance, two USARTs, two 16bit multi-function timers, one SPI/MICROWIRE-PLUS™ serial interface, a 12-channel A/D converter, two analog comparators, WATCHDOG™ protection mechanism, and up to 56
general-purpose I/O pins.
•
The device operates with a high-frequency crystal as the
main clock source and either the prescaled main clock
source or with a low frequency (32.768 kHz) oscillator in
Power Save mode. The device supports several Power Save
modes which are combined with multi-source interrupt and
wake-up capabilities.
•
•
•
This device also has a Versatile Timer Unit (VTU) with four
timer sub-systems, a CAN interface, and ACCESS.bus synchronous serial bus interface.
Powerful cross-development tools are available from National Semiconductor and third party suppliers to support the development and debugging of application software for the
device. These tools let you program the application software
in C and are designed to take full advantage of the CompactRISC architecture.
— FullCAN interface with 15 message buffers complaint
to CAN specification 2.0B active
— Versatile Timer Unit with four subsystems (VTU)
— Two analog comparators
— Integrated WATCHDOG logic
I/O Features
— Up to 56 general-purpose I/O pins (shared with on-chip
peripheral I/O pins)
— Programmable I/O pin characteristics: TRI-STATE output, push-pull output, weak pull-up input, high-impedance input
— Schmitt triggers on general purpose inputs
Power Supply
— 4.5V to 5.5V single-supply operation
Temperature Range
— –40°C to +85°C
— –40°C to +125°C
Development Support
— Real-time emulation and full program debug capabilities available
— CompactRISC tools provide C programming and debugging support
In the following text, device is always referred to the family of
16-bit CAN-enabled CompactRISC Microtroller.
2.0
Features
• CPU Features
— Fully static core, capable of operating at any rate from
0 to 24 MHz (4 MHz minimum in active mode)
— 50 ns instruction cycle time with a 20 MHz external
clock frequency
— Multi-source vectored interrupts (internal, external,
and on-chip peripheral)
— Dual clock and reset
• On-chip power-on reset
• On-Chip Memory
— Up to 64K bytes flash EEPROM program memory; can
be programmed, erased, and reprogrammed by software (100K cycles)
— 3K bytes of static RAM data memory
— For flash program memory devices, 1.5k bytes flash
EEPROM memory is available to store boot loader
code (100K cycles)
— 2K bytes of non-volatile EEPROM data memory with
low endurance (25K cycles) and 128 bytes with high
endurance (100K cycles)
• On-Chip Peripherals
— Two Universal Synchronous/Asynchronous Receiver/
Transmitter (USART) devices
— Programmable Idle Timer and real-time clock (T0)
— Two dual 16-bit multi-function timers (MFT1 and MFT2)
— 8/16-bit SPI/MICROWIRE-PLUS serial interface
— 12-channel, 8-bit Analog-to-Digital (A/D) converter
with external voltage reference, programmable sample-and-hold delay, and programmable conversion frequency
— ACCESS.bus synchronous serial bus
3
www.national.com
CR16 CompactRISC Microcontroller with CAN Interface Family Selection Guide
Programmable devices
NSID
Speed
(MHz)
Flash/
(kByte)
EEPROM
Data
Memory
(Bytes)
SRAM
(kBytes)
USART
Timer
I/Os
Temp.
Range
Peripherals
Package
Type
CR16MCS9VJEx
16
64
2176
3
2
2MFT,
VTU
56
E, I
ADC, CAN,
Comparators
80 PQFP
CR16MAS9VJEx
24
64
3
2
2MFT,
VTU
56
E, I
ADC, CAN,
Comparators
80 PQFP
I/Os
Temp.
Range
Peripherals
Package
Type
Factory Programmed devices
NSID
Speed Flash/
(MHz) (KByte)
EEPROM
Data
Memory
(Bytes)
SRAM
(kBytes)
USART Timer
CR16MCS9VJExy
16
64
2176
3
2
2MFT,
VTU
56
E, I
ADC, CAN,
Comparators
80 PQFP
CR16MCS9VJExy
24
64
2176
3
2
2MFT,
VTU
56
E, I
ADC, CAN,
Comparators
80 PQFP
NSID
Speed
(MHz)
Flash/
ROM
(KByte)
EEPROM
Data
Memory
(Bytes)
Timer
I/Os
Temp.
Range
Peripherals
Package
Type
CR16HCS9VJEx
24
64
2176
3
2
2MFT,
VTU
56
E, I
ADC, CAN,
Comparators
80 PQFP
CR16MCS5VJEx
24
64
2176
3
2
2MFT,
VTU
56
E, I
ADC, CAN,
Comparators
80 PQFP
CR16MBR5VJEx
24
32
2176
3
2
2MFT,
VTU
56
E, I
ADC, CAN,
Comparators
80 PQFP
CR16MAR5VJEx
24
32
3
2
2MFT,
VTU
56
E, I
CAN,
80 PQFP
CR16MAS5VJEx
24
64
3
2
2MFT,
VTU
56
E, I
CAN,
80 PQFP
ROM devices
SRAM
USART
(kBytes)
Note:
•
Suffix x in the NSID is defined below:
Temperature Ranges:
I = Industrial
E = Extended
•
-40 °C to +85°C is represented when x is 8
-40 °C to +125°C is represented when x is 7
Suffix y in the NSID defines the ROM code.
Note: All devices contains Access.bus (ACB), Clock and Reset, MICROWIRE/API, Multi-Input Wake-Up (MIWU), Power
Management (PMM), and the Real-Time Timer and Watchdog (TWM) modules. Access.bus is compatible with I2C bus
offered by Philips Semiconductor.
www.national.com
CR16 CompactRISC Microcontroller with CAN Interface
Family Devices
National Semiconductor currently offers a variety of the
CR16 CompactRISC Microcontrollers with CAN interface.
The CR16MCS offer complete functionality in an 80-pin
PQFP package.
4
3.0
Device Overview
The devices are complete microcomputers with all system
timing, interrupt logic, program memory, data memory, and I/
O ports included on-chip, making it well-suited to a wide
range of embedded controller applications.
3.1
The 3K bytes of static RAM are used for temporary storage
of data and for the program stack and interrupt stack. Read
and write operations can be byte-wide or word-wide, depending on the instruction executed by the CPU. Each memory
access requires one clock cycle; no wait cycles or hold cycles
are required.
CR16B CPU CORE
The device uses a CR16B CPU core module. This is the
same core used in other CompactRISC family member designs, like DECT or GSM chipsets.
There are two types of flash EEPROM data memory storage.
The 2K bytes of EEPROM data memory with low endurance
(25K cycles) and 128 bytes of flash EEPROM data memory
with high endurance (100K cycles) are used for non-volatile
storage of data, such as configuration settings entered by the
end-user.
The high performance of the CPU core results from the implementation of a pipelined architecture with a two-bytes-percycle pipelined system bus. As a result, the CPU can support
a peak execution rate of one instruction per clock cycle.
The 64K bytes of flash EEPROM program memory are used
to store the application program. It has security features to
prevent unintentional programming and to prevent unauthorized access to the program code. This memory can be programmed with a device external programming unit or with the
device installed in the application system (in-system programming).
Compared with conventional RISC processors, the device
differs in the following ways:
— The CPU core can use on-chip rather than external
memory. This eliminates the need for large and complex bus interface units.
— Most instructions are 16 bits, so all basic instructions
are just two bytes long. Additional bytes are sometimes
required for immediate values, so instructions can be
two or four bytes long.
— Non-aligned word access is allowed. Each instruction
can operate on 8-bit or 16-bit data.
— The device is designed to operate with a clock rate in
the 10 to 24 MHz range rather than 100 MHz or more.
Most embedded systems face EMI and noise constraints that limit clock speed to these lower ranges. A
lower clock speed means a simpler, less costly silicon
implementation.
— The instruction pipeline uses three stages. A smaller
pipeline eliminates the need for costly branch prediction mechanisms and bypass registers, while maintaining adequate performance for typical embedded
controller applications.
There is a factory programmed boot memory used to store
In-System-Programming (ISP) code. (This code allows programming of the program memory via one of the USART interfaces in the final application.)
For flash EEPROM program and data memory, the device internally generates the necessary voltages for programming.
No additional power supply is required.
3.3
The device has 56 software-configurable I/O pins, organized
into seven 8-pin ports called Port B, Port C, Port F, Port G,
Port H, Port I, and Port L. Each pin can be configured to operate as a general-purpose input or general-purpose output.
In addition, many I/O pins can be configured to operate as a
designated input or output for an on-chip peripheral module
such as the USART, timer, A/D converter, or MICROWIRE/
SPI interface.
For more information, please refer to the CR16B Programmer’s Reference Manual, Literature #: 633150.
3.2
INPUT/OUTPUT PORTS
The I/O pin characteristics are fully programmable. Each pin
can be configured to operate as a TRI-STATE output, pushpull output, weak pull-up input, or high-impedance input.
MEMORY
The CompactRISC architecture supports a uniform linear address space of 2 megabytes. The device implementation of
this architecture uses only the lowest 128K bytes of address
space. Four types of on-chip memory occupy specific intervals within this address space:
3.4
BUS INTERFACE UNIT
The Bus Interface Unit (BIU) controls the interface between
the on-chip modules to the internal core bus. It determines
the configured parameters for bus access (such as the number of wait states for memory access) and issues the appropriate bus signals for each requested access.
• 64K bytes of flash EEPROM program memory (100K cycles)
• 48K bytes ROM programm memory version available also
(100K cycles)
• 3K bytes of static RAM
• 2K bytes of EEPROM data memory with low endurance
(25K cycles)
• 128 bytes with high endurance (100K cycles)
• 1.5K bytes flash EEPROM memory for ISP code
The BIU uses a set of control registers to determine how
many wait states and hold states are to be used when accessing flash EEPROM program memory, ISP memory and
the I/O area (Port B and Port C). Upon start-up the configuration registers are set for slowest possible memory access.
To achieve fastest possible program execution, appropriate
values should be programmed. These settings vary with the
clock frequency and the type of on-chip device being accessed.
5
www.national.com
3.5
INTERRUPTS
3.9
The Interrupt Control Unit (ICU31L) receives interrupt requests from internal and external sources and generates interrupts to the CPU. An interrupt is an event that temporarily
stops the normal flow of program execution and causes a
separate interrupt service routine to be executed. After the interrupt is serviced, CPU execution continues with the next instruction in the program following the point of interruption.
The Multi-Function Timer (MFT16) module contains two independent timer/counter units called MFT1 and MFT2, each
containing a pair of 16-bit timer/counter registers. Each timer/
counter unit can be configured to operate in any of the following modes:
— Processor-Independent Pulse Width Modulation
(PWM) mode, which generates pulses of a specified
width and duty cycle, and which also provides a general-purpose timer/counter.
— Dual Input Capture mode, which measures the
elapsed time between occurrences of external events,
and which also provides a general-purpose timer/
counter.
— Dual Independent Timer mode, which generates system timing signals or counts occurrences of external
events.
— Single Input Capture and Single Timer mode, which
provides one external event counter and one system
timer.
Interrupts from the timers, USARTs, MICROWIRE/SPI interface, multi-input wake-up, and A/D converter are all
maskable interrupts; they can be enabled or disabled by the
software. There are 32 of these maskable interrupts, organized into 32 predetermined levels of priority.
The highest-priority interrupt is the Non-Maskable Interrupt
(NMI), which is generated by a signal received on the NMI input pin.
3.6
MULTI-INPUT WAKE-UP
The Multi-Input Wake-Up (MIWU16) module can be used for
either of two purposes: to provide inputs for waking up (exiting) from the HALT, IDLE, or Power Save mode; or to provide
general-purpose edge-triggered maskable interrupts from
external sources. This 16-channel module generates four
programmable interrupts to the CPU based on the signals received on its 16 input channels. Channels can be individually
enabled or disabled, and programmed to respond to positive
or negative edges.
3.7
3.10
DUAL CLOCK AND RESET
3.11
REAL-TIME TIMER AND WATCHDOG
The Timing and Watchdog Module (TWM) generates the
clocks and interrupts used for timing periodic functions in the
system. It also provides Watchdog protection against software errors. The module operates on the slow system clock.
This module also generates a slow system clock (32.768
kHz) from another external crystal network. The slow clock is
used for operating the device in power-save mode. Without a
32.768kHz external crystal network, the low speed system
clock can be derived from the high speed clock by a prescaler.
The real-time timer can generate a periodic interrupt to the
CPU at a software-programmed interval. This can be used
for real-time functions such as a time-of-day clock. The realtime timer can trigger a wake-up condition from power-save
mode via the Multi-Input Wake-Up module.
The Watchdog is designed to detect program execution errors such as an infinite loop or a “runaway” program. Once
Watchdog operation is initiated, the application program
must periodically write a specific value to a Watchdog register, within specific time intervals. If the software fails to do so,
a Watchdog error is triggered, which resets the device.
Also, two independent clocks divided down from the high
speed clock are available on output pins.
POWER MANAGEMENT
The Power Management Module (PMM) improves the efficiency of the device by changing the operating mode and
therefore the power consumption according to the required
level of activity.
3.12
USART
The USART supports a wide range of programmable baud
rates and data formats, and handles parity generation and
several error detection schemes. The baud rate is generated
on-chip, under software control.
The device can operate in any of four power modes:
— Active: The device operates at full speed using the
high-frequency clock. All device functions are fully operational.
— Power Save: The device operates at reduced speed
using the slow clock. The CPU and some modules can
continue to operate at this low speed.
— IDLE: The device is inactive except for the Power Management Module and Timing and Watchdog Module,
which continue to operate using the slow clock.
— HALT: The device is inactive but still retains its internal
state (RAM and register contents).
www.national.com
VERSATILE TIMER UNIT
The Versatile Timer Unit (VTU) module contains four independent timer subsystems, each operating in either dual 8-bit
PWM configuration, as a single 16-bit PWM timer, or a 16-bit
counter with two input capture channels. Each of the four timer subsystems offer an 8-bit clock prescaler to accommodate
a wide range of frequencies.
The Dual Clock and Reset (CLK2RES) module generates a
high-speed main system clock from an external crystal network. It also provides the main system reset signal and a
power-on reset function.
3.8
MULTI-FUNCTION TIMER
There are two independent USARTs in the device and they
offer a wake-up condition from the power-save mode via the
Multi-Input Wake-Up module.
3.13
MICROWIRE/SPI
The MICROWIRE/SPI (MWSPI) interface module supports
synchronous serial communications with other devices that
conform to MICROWIRE or Serial Peripheral Interface (SPI)
specifications. It supports 8-bit and 16-bit data transfers.
6
The MICROWIRE interface allows several devices to communicate over a single system consisting of four wires: serial
in, serial out, shift clock, and slave enable. At any given time,
the MICROWIRE interface operates as the master or a slave.
The support supports the full set of slave select for multislave implementation.
3.17
ANALOG COMPARATORS
In master mode, the shift clock is generated on chip under
software control. In slave mode, a wake-up out of powersave mode is triggered via the Multi-Input Wake-Up module.
The Dual Analog Comparator (ACMP2) module contains two
independent analog comparators with all necessary control
logic. Each comparator unit compares the analog input voltages applied to two input pins and determines which voltage
is higher. The CPU uses a memory-mapped register to control the comparator and to obtain the comparison results. The
comparison result can also be applied to comparator output
pins.
3.14
3.18
CR16CAN
The CR16CAN device contains a FullCAN class, CAN serial
bus interface for applications that require a high speed (up to
1MBits per second) or a low speed interface with CAN bus
master capability. The data transfer between CAN and the
CPU is established by 15 memory mapped message buffers,
which can be individually configured as receive or transmit
buffers. An incoming message is filtered by two masks, one
for the first 14 message buffers and another one for the 15th
message buffer to provide a basic CAN path. A priority decoder allows any buffer to have the highest or lowest transmit
priority. Remote transmission requests can be processed automatically by automatic reconfiguration to a receiver after
transmission or by automated transmit scheduling upon reception. In addition, a time stamp counter (16-bits wide) is
provided to support real time applications.
DEVELOPMENT SUPPORT
A powerful cross-development tool set is available from National Semiconductor and third parties to support the development and debugging of application software for the
CR16MCS9. The tool set lets you program the application
software in C and is designed to take full advantage of the
CompactRISC architecture.
There are In-System Emulation (ISE) devices available for
the device from iSYSTEM™, as well as lower-cost evaluation
boards. See your National Semiconductor sales representative for current information on availability and features of emulation equipment and evaluation boards.
The CR16CAN device is a fast core bus peripheral, which allows single cycle byte or word read/write access. A set of diagnostic features (such as loopback, listen only, and error
identification) support the development with the CR16CAN
module and provide a sophisticated error management tool.
The CR16CAN receiver can trigger a wake-up condition out
of the power-save modes via the Multi-Input Wake-Up module.
3.15
ACCESS.BUS INTERFACE
The ACCESS.bus interface module (ACB) is a two-wire serial interface with the ACCESS.bus physical layer. It is also
compatible with Intel’s System Management Bus (SMBus)
and Philips’ I2C bus. The ACB module can be configured as
a bus master or slave, and can maintain bi-directional communications with both multiple master and slave devices.
The ACCESS.bus receiver can trigger a wake-up condition
out of the power-save modes via the Multi-Input Wake-Up
module.
3.16
A/D CONVERTER
The A/D Converter (ADC) module is a 12-channel multiplexed-input analog-to-digital converter. The A/D Converter
receives an analog voltage signal on an input pin and converts the analog signal into an 8-bit digital value using successive approximation. The CPU can then read the result
from a memory-mapped register. The module supports four
automated operating modes, providing single-channel or 4channel operation in single or continuous mode.
The device has a separate pin, Vref, for the A/D reference
voltage.
7
www.national.com
4.0
Device Pinouts
Table 1
Package Pin Assignments
Pin Name
Alternate Function(s)
Pin Number
Type
PH4
PH5
MWCS
MD1D0
1
2
I/O
I/O
PH6
MD0D1
3
I/O
PH7
PB0
MSK
D0
4
5
I/O
I/O
PB1
D1
6
I/O
PB2
PB3
D2
D3
7
8
I/O
I/O
PB4
D4
9
I/O
PB5
PB6
D5
D6
10
11
I/O
I/O
PB7
D7
12
I/O
ENV0/CLKOUT1
SDA
13
14
I/O
I/O
SCL
15
I/O
GND
Vcc
16
17
PWR
PWR
GND
18
PWR
CANTx
CANRx
19
20
O
I
PC0
D8
21
I/O
PC1
PC2
D9
D10
22
23
I/O
I/O
PC3
D11
24
I/O
PC4
PC5
D12
D13
25
26
I/O
I/O
PC6
D14
27
I/O
PC7
PG7
D15
CKX1
28
29
I/O
I/O
PG6
TDX1
30
I/O
PG5
PG4
RDX1
TIO6
31
32
I/O
I/O
PG3
TIO5
33
I/O
PG2
PG1
CKX2
TDX2
34
35
I/O
I/O
PG0
RDX2
36
I/O
37
37
O
I/O
CLKOUT2
ENV1/CLK1
www.national.com
PF7
TIO4
38
I/O
PF6
PF5
TIO3
T2B
39
40
I/O
I/O
PF4
T2A
41
I/O
PF3
PF2
TIO2
TIO1
42
43
I/O
I/O
PF1
TIB
44
I/O
8
Table 1
Package Pin Assignments
Pin Name
Alternate Function(s)
Pin Number
Type
PF0
NMI
TIA
45
46
I/O
I
X1CKO
47
O
X1CKI
GND
48
49
I
PWR
Vcc
50
PWR
GND
X2CKO
51
52
PWR
O
X2CKI
53
I
2
RESET
PI0
ACH03
54
55
I
I/O
PI1
ACH13
56
I/O
PI2
PI3
ACN23
ACH33
57
58
I/O
I/O
PI4
ACH43
59
I/O
PI5
PI6
3
ACH5
ACH63
60
61
I/O
I/O
PI7
ACH73
62
I/O
63
64
PWR
PWR
65
PWR
Vref
AGND
AVcc
3
PH0
PH1
ACH8 , WUI4
ACH9 3 , WUI5
66
67
I/O
I/O
PH2
ACH103, WUI6
68
I/O
PH3
GND
ACH11 3 , WUI7
69
70
I/O
PWR
71
PWR
Vcc
Note 1:
Note 2:
Note 3:
GND
PL0
COMP1N3, WUI0
72
73
PWR
I/O
PL1
COMP1P 3, WUI1
74
I/O
PL2
PL3
COMP1O, WUI2
COMP2O, WUI3
75
76
I/O
I/O
PL4
COMP2P3
77
I/O
PL5
PL6
COMP2N
TIO7
3
78
79
I/O
I/O
PL7
TIO8
80
I/O
The ENV0 and ENV1 pins each have a weak pull-up to keep the input from floating.
The RESET input has a weak pulldown.
These functions are always enabled, due to the direct low-impedance path to these pins.
9
www.national.com
4.1
The following is a brief description of all device pins.
PIN DESCRIPTION
Some pins have alternate functions which may be enabled.
These pins can be individually configured as general purpose pins, even when the module they belong to is enabled.
Table 2
Signal
Type
Active
Input Pins
Pin (* for a shared pin)
Function
X1CKI
OSC
High
Main oscillator clock input.
X2CKI
OSC
High
32kHz oscillator clock input.
RESET CMOS
Low
Chip general reset pin. Schmitt trigger input, asynchronous.
ISE
CMOS
Low
T1B
CMOS
Prog.
*
Timer 1 input B. Shares pin with I/O port pin PF1.
T2B
CMOS
Prog.
*
Timer 2 input B. Shares pin with I/O port pin PF5.
RDX1
CMOS
High
*
USART 1 receive data input. Shares pin with I/O port pin PG5.
RDX2
CMOS
High
*
USART 2 receive data input. Shares pin with I/O port pin PG0.
ACH0
Analog
*
A2D converter channel 0. Shares pin with I/O port pin PI0
ACH1
Analog
*
A2D converter channel 1. Shares pin with I/O port pin PI1
ACH2
Analog
*
A2D converter channel 2. Shares pin with I/O port pin PI2
ACH3
Analog
*
A2D converter channel 3. Shares pin with I/O port pin PI3
ACH4
Analog
*
A2D converter channel 4. Shares pin with I/O port pin PI4
ACH5
Analog
*
A2D converter channel 5. Shares pin with I/O port pin PI5
ACH6
Analog
*
A2D converter channel 6. Shares pin with I/O port pin PI6
ACH7
Analog
*
A2D converter channel 7. Shares pin with I/O port pin PI7
ACH8
Analog
*
A2D converter channel 8. Shares pin with I/O port pin PH0
ACH9
Analog
*
A2D converter channel 9. Shares pin with I/O port pin PH1
ACH10
Analog
*
A2D converter channel 10. Shares pin with I/O port pin PH2
ACH11
Analog
*
A2D converter channel 11. Shares pin with I/O port pin PH3
MWCS
CMOS
Low
*
SPI/MICROWIRE slave select. Shares pin with I/O port pin PH4.
NMI
CMOS
Low
ENV0
CMOS
Low
*
Strap to select operating environment.
ENV1
CMOS
Low
*
Strap pin to select operating environment.
ENV2
CMOS
Low
Strap pin to select operating environment.
CANRx CMOS
High
CAN receive data input.
Interrupt input for development system.
External non-maskable interrupt.
Table 3
Signal
Type Active
Output Pins
Pin (* for
a shared pin)
Function
X1CKO
OSC
High
Main oscillator clock output.
X2CKO
OSC
High
32kHz oscillator clock output.
CLK
CMOS High
*
External reference clock for development environment (shared with ENV1).
CLKOUT
1
CMOS High
*
Clock output generated through prescaler (shared with ENV0).
CLKOUT
2
CMOS High
*
Clock output generated through prescaler (shared with ENV1).
www.national.com
10
Table 3 Output Pins
Signal
Type
Active
Pin (* for
a shared pin)
Function
TDX1
CMOS High
*
USART 1 transmit data output (shared with PG6).
TDX2
CMOS High
*
USART 2 transmit data output (shared with PG1).
CANTx
CMOS High
CAN output.
Table 4
Signal
Type
Input/Output Pins
Pin (* for a
shared pin)
Active
Function
PF[0:7]
CMOS High
*
Generic I/O port. Shared with T1A, T1B, TIO1, TIO2, T2A, T2B, TIO3, TIO4.
PG[0:7]
CMOS High
*
Generic I/O port. Shared with RDX2, TDX2, CKX2, TIO5, TIO6, RDX1, TDX1,
CKX1.
PB[0:7]
CMOS High
*
Generic I/O port.
PC[0:7]
CMOS High
*
Generic I/O port.
PL[0:7]
CMOS High
*
Generic I/O port. Shared with 6 comparator pins, MIWU16 on PL0:3.
PH[0:7]
CMOS High
*
Generic I/O port. Shared with ADC input channels 8-11, MWCS, MDIDO,
MDODI, MSK; MIWU16 on PH4:7.
PI[0:7]
CMOS High
*
Generic I/O port. Shared with ADC input channels 0-7.
T1A
CMOS Prog
*
Timer 1 input A. Shared with I/O port pin PF0.
T2A
CMOS Prog
*
Timer 2 input A. Shared with I/O port pin PF4.
TIO[0:7]
CMOS Prog
*
Versatile timer unit I/Os. Shared with PF2:3, PF6:7, PG3:4, PL6:7.
MDIDO
CMOS High
*
Master In/Slave Out port: SPI/Microwire. Shared with I/O pin PH5,
MDODI
CMOS High
*
Master Out/Slave In port: SPI/Microwire. Shared with I/O pin PH6.
MSK
CMOS Prog
*
SPI/Microwire clock. Shared with I/O pin PH7.
CKX1
CMOS High
*
USART 1 clock. Shared with I/O pin PG7.
CKX2
CMOS High
*
USART 2 clock. Shared with I/O pin PG2
SCL
CMOS High
ACCESS.bus clock I/O.
SDA
CMOS High
ACCESS.bus data I/O.
Table 5
Signal
Power Supply
Function
Vcc
Main digital power supply (4 total).
Vref
Voltage reference supply for analog to digital converter.
AVcc
Analog power supply for analog/digital converter.
AGND Analog reference ground supply.
GND
Main digital reference ground (8 total).
11
www.national.com
5.0
System Configuration
The device has two input pins, ENV0 and ENV1, which are
used to specify the operating environment of the device upon
reset. There are also two system configuration registers,
called the Module Configuration (MCFG) register and the
Module Status (MSTAT) register.
5.1
normal operating mode, the CLK pin operates
as a CPU clock output.
Generated Clock Output 1 Enable. When
cleared (0), the CLKOUT1 pin (ENV0) stays in
high impedance state. When set (1), the pin
outputs the clock from the prescaler controlled
by PRSSC1.SCDIV1.
Generated Clock Output 2 Enable. When this
bit is set (1) and CLKOE is cleared, the
CLKOUT2 pin (ENV1) outputs the clock from
the prescaler controlled by PRSSC1.SCDIV2.
Otherwise, the CLKOUT2 pin is in high impedance state.
CLK1OE
ENV0 AND ENV1 PINS
Upon reset, the operating mode of the device is determined
by the state of the ENV0 and ENV1 input pins, as indicated
in Table6.
CLK2OE
Table 6 Operating Environment Selection
ENV1
ENV0
0
0
Test Mode Flash Memory
5.3
0
1
Test Mode
1
0
In-System-Programming mode (ISP)
The MSTAT register is a byte-wide, read-only register that indicates the general status of the device.
1
Internal ROM enabled Mode (IRE), if
program memory is not empty; or ISPMode, if program memory is empty
1
Operating Environment
The MCFG register format is shown below.
7
4
Reserved
In the case where the ENV1 and ENV0 pins are both high,
the reset algorithm looks at the FLCTRL2.EMPTY bit to determine whether the program memory is empty, and sets the
operating mode accordingly.
OENV(1:0)
In the case where the ISP-Mode is selected, the chip starts
executing the ISP code residing in the on-chip ISP-Memory
area.
The test modes are Reserved for factory testing and for external programming of the flash EEPROM program memory.
They should not be invoked otherwise.
MODULE CONFIGURATION (MCFG)
REGISTER
The MCFG register is a byte-wide, read/write register that
sets the clock output features of the device.
Upon reset, the non-reserved bits of this register are cleared
to zero. The start-up software must write a specific value to
this register in order to configure the CLK output pin function.
When the software writes to this register, it must write a zero
to each reserved bit for the device to operate properly. The
register should be written in active mode only, not in power
save, HALT, or IDLE mode. However, the register contents
are preserved during all power modes.
The MCFG register format is shown below.
7
6
5
Reserved CLK2OE
CLKOE
4
Reserved
3
2
1
0
CLK1OE
CLKOE
Reserved
CPU Clock Output Enable. When this bit is
cleared (0), the CLK pin (ENV1) remains in the
high-impedance state. When this bit is set (1) in
www.national.com
3
PGMBUSY
2
Reserved
1
OENV1
0
OENV0
Operating Environment. These two bits contain
the values applied to the ENV1 and ENV0 pins
upon reset. These bit values are controlled by
the external hardware upon reset and are held
constant in the register until the next reset.
PGMBUSY Flash EEPROM Programming Busy. This bit is
automatically set to 1 when either the program
memory or the data memory is busy being programmed or erased. It is cleared to 0 when neither of the two flash EEPROM memories is
busy being programmed or erased. When this
bit is set, the software should not attempt any
write access to either of these two memories.
The ENV0 and ENV1 pins have on-chip pull-up devices that
are enabled during reset while the pins are being sampled.
Therefore, if they are left unconnected, the inputs are considered high and the normal operating mode (IRE-Mode) is selected and the CPU starts to execute code at address 0. To
enter any other operating mode, the external hardware must
drive the appropriate input low.
5.2
MODULE STATUS (MSTAT) REGISTER
12
6.0
Input/Output Ports
Each device has up to 56 software-configurable I/O pins, organized into seven ports of up to eight pins per port. The
ports are named Port B, Port C, Port F, Port G, Port H, Port
I, and Port L.
on-chip module where it is latched. A Schmitt-Trigger minimizes the effects of electrical noise.
Each pin can be configured to operate as a general-purpose
input or general-purpose output. In addition, many I/O pins
can be configured to operate as a designated input or output
for an on-chip peripheral module such as the USART or the
Multi-Input Wakeup. This is called the pin's “alternate function.” The alternate functions of all I/O pins are shown in the
pinout diagrams in Table1.
For some pins, a direct low-impedance path is provided between the pin and an internal analog function. These are the
input pins to the A/D converter and the analog comparators.
The electrical characteristics and drive capabilities of the input and output buffers are described in Section25.0.
6.1
PORT REGISTERS
Each port has an associated set of memory-mapped registers used for controlling the port and for holding the port data.
In general, there are five such registers:
The I/O pin characteristics are fully programmable. Each pin
can be configured to operate as a TRI-STATE output, pushpull output, weak pull-up input, or high-impedance input. Different pins within the same port can be individually configured to operate in different modes.
—
—
—
—
—
Figure1 is a diagram showing the functional features of an I/
O port pin. The register bits, multiplexers, and buffers allow
the port pin to be configured into the various operating
modes.The output buffer is a TRI-STATE buffer with weak
pull-up capability. The weak pull-up, if used, prevents the port
pin from going to an undefined state when it operates as an
input.
PxALT: Port alternate function register
PxDIR: Port direction register
PxDIN: Port data input register
PxDOUT: Port data output register
PxWKPU: Port weak pull-up register
In the descriptions of the ports and port registers, the lowercase letter “x” represents the port designation, either B, C, F,
G, H, I, or L. For example, “PxDIR register” means any one
of the port direction registers: PBDIR, PCDIR, PFDIR, and so
on.
All of the port registers are byte-wide read/write registers, except for the port data input registers, which are read-only registers. Each register bit controls the function of the
corresponding port pin. For example, PFDIR.2 (bit 2 of the
PFDIR register) controls the operation of port pin PF2.
The input buffer is disabled when it is not needed to prevent
leakage current caused by an input signal’s level between
Vcc -0.2 and Vss+0.2 [Volts]. When enabled, it buffers the input signal and sends the pin's logic level to the appropriate
Alternate Function
enable
{
{
Weak pull-up
MUX1
Weak pull-up
Register
Alt Device Direction
Direction
Register
Direction
Alt
Data Out
Register
{
PIN
MUX2
Alt Device Data Output
Data Out
Alt
Data Input
*
Alt
Data In Read Strobe
1
MUX3
Alternate Data Input
Alt
Figure 1. I/O Pin Functional Diagram
13
www.national.com
6.1.1
Port Alternate Function Register
6.2
Each port that supports an alternate function (any port other
than Port B or Port C) has an alternate function register (PxALT). This register determines whether the port pins are used
for general-purpose I/O or for the predetermined alternate
function. Each port pin can be controlled independently.
A port pin can be configured to operate as an inverting opendrain output buffer. To do this, the CPU should clear the bit in
the data output register (PxDOUT) and then use the port direction register (PxDIR) to set the value of the port pin. With
the direction register bit set to 1 (direction=out), the value
zero is forced on the pin. With the direction register bit
cleared to 0 (direction=in), the pin is placed in the TRI-STATE
mode. If desired, the internal weak pull-up can be enabled to
pull the signal high when the output buffer is in the TRISTATE mode.
A bit cleared to 0 in the alternate function register causes the
corresponding pin to be used for general-purpose I/O. In this
configuration, the output buffer is controlled by the direction
register and the data output register. The input buffer is routed to the data input register. The input buffer is blocked except when the buffer is actually being read.
A bit set to 1 in the alternate function register causes the corresponding pin to be used for its predetermined peripheral I/
O function. The output buffer data and TRI-STATE configuration are controlled by signals coming from the on-chip peripheral device. The input buffer is enabled continuously in this
case. To minimize power consumption, the input signal
should be held within 0.2 volts of the VCC or GND voltage.
A reset operation clears the port alternate function registers
to 0, which programs the pins to operate as general-purpose
I/O ports. This register must be enabled before the corresponding alternate function is enabled.
6.1.2
Port Direction Register
The port direction register (PxDIR) determines whether each
port pin is used for input or for output. A bit cleared to 0 causes the pin to operate as an input, which puts the output buffer
in the high-impedance state. A bit set to 1 causes the pin to
operate as an output, which enables the output buffer.
A reset operation clears the port direction registers to 0,
which programs the pins to operate as inputs.
6.1.3
Port Data Input Register
The data input register (PxDIN) is a read-only register that returns the current state of each port pin. The CPU can read
this register at any time even when the pin is configured as
an output.
6.1.4
Port Data Output Register
The data output register (PxDOUT) holds the data to be driven onto each port pin configured to operate as a general-purpose output. In this configuration, writing to the register
changes the output value. Reading the register returns the
last value written to the register.
A reset operation leaves the register contents unchanged.
Upon power-up, the registers contain unknown values.
6.1.5
Port Weak Pull-Up Register
The weak pull-up register (PxWKPU) determines whether
each port pin uses a weak pull-up on the output buffer. A bit
set to 1 causes the weak pull-up to be used, while a bit
cleared to 0 causes the causes the weak pull-up not to be
used.
The pull-up device, if enabled by the register bit, operates in
the general-purpose I/O mode whenever the port output buffer is in the TRI-STATE mode. In the alternate function mode,
the pull-ups are always disabled.
A reset operation clears the port weak pull-up registers to 0,
which disables all pull-ups.
www.national.com
OPEN-DRAIN OPERATION
14
7.0
CPU and Core Registers
The device uses the same CR16B CPU core as other CompactRISC family members. The core's Reduced Instruction
Set Computer (RISC) architecture allows a processing rate
of up to one instruction per clock cycle.
7.2.3
The INTBASE register holds the address of the Dispatch Table for interrupts and traps. The least significant bit of the register is always zero. Thus, the Dispatch Table starts at an
even address in the range of 0000 to FFFE.
The CPU core uses a set of internal registers:
—
—
—
—
General-purpose registers (R0-R13, RA, and SP)
Dedicated address registers (PC, ISP, and INTBASE)
Processor Status Register (PSR)
Configuration Register (CFG)
7.3
15 14 13 12 11 10 9
Reserved
I P E
Some register bits are designated as “reserved.” The CPU
must write a zero to each of these bit locations when it writes
to the register. Read operations from reserved bit locations
return undefined values.
C bit
GENERAL-PURPOSE REGISTERS
T bit
There are 16 general-purpose registers, designated R0
through R13, RA, and SP. Registers R0 through R13 can be
used for any purpose such as holding variables, addresses,
or index values. The RA register is usually used to store the
return address upon entry into a subroutine. The SP register
is usually used as the pointer to the program run-time stack.
L bit
If a general-purpose register is used for a byte-wide operation, only the low-order byte is referenced or modified. The
high-order byte is not used or affected by a byte-wide operation.
7.2
F bit
DEDICATED ADDRESS REGISTERS
There are three dedicated address registers: the Program
Counter (PC), the Interrupt Stack Pointer (ISP), and the Interrupt Base Register (INTBASE). Each of these registers is
21 bits wide.
7.2.1
Z bit
Program Counter
The PC register contains the address of the least significant
word currently being fetched. It is automatically incremented
or changed by the appropriate amount each time an instruction is executed.
N bit
The least significant bit of the PC is always zero, thus instructions must always be aligned to an even address in the range
of 0000 to 1FFFE hex.
E bit
Upon reset, the PC register is initialized to zero and program
execution starts at that address (if in IRE-Mode). When a reset signal is received, bits 1 through 16 of the PC register
(prior to initialization) are stored in register R0. This allows
the software to determine the point in the program at which
the reset occurred.
7.2.2
PROCESSOR STATUS REGISTER
The Processor Status Register (PSR) holds status information and selects the operating modes for the CPU core. The
format of the register is shown below.
All of these registers are 16 bits wide except for the three address registers, which are 21 bits wide.
7.1
Interrupt Base Register
Interrupt Stack Pointer
P bit
The ISP register points to the lowest address of the last item
stored on the interrupt stack. This stack is used by the hardware when an interrupt or trap service procedure is invoked.
15
8
0
7 6
N Z
5
F
4
0
3
0
2
L
1 0
T C
The Carry (C) bit indicates whether a carry or
borrow occurred after addition or subtraction. It
is set to 1 if a carry or borrow occurred, or
cleared to 0 otherwise.
The Trace (T) bit, when set, causes a Trace
(TRC) trap to be executed after every instruction. This bit is automatically cleared to 0 when
a trap or interrupt occurs.
The Low (L) bit is set by comparison operations. In a comparison of unsigned integers, the
bit is set to 1 if the second operand (Rdest) is
less than the first operand (Rsrc). Otherwise, it
is cleared to 0.
The Flag (F) bit is a general condition flag that
is set by various instructions. It may be used to
signal exception conditions or to distinguish the
results of an instruction. For example, integer
arithmetic instructions use this bit to indicate an
overflow condition after an addition or subtraction operation.
The Zero (Z) bit is set by comparison operations. In a comparison of integers, the bit is set
to 1 if the two operands are equal. Otherwise,
it is cleared to 0.
The Negative (N) bit is set by comparison operations. In a comparison of signed integers, the
bit is set to 1 if the second operand (Rdest) is
less than the first operand (Rsrc). Otherwise, it
is cleared to 0.
The Local Maskable Interrupt Enable (E) bit is
used to enable or disable maskable interrupts.
If this bit and the Global Maskable Interrupt Enable (I) bit are both set to 1, all maskable interrupts are accepted. Otherwise, only the nonmaskable interrupt is accepted. The E bit is set
to 1 by the Enable Interrupts (EI) instruction
and cleared to 0 by the Disable Interrupts (DI)
instruction.
The Trace Trap Pending (P) bit is used together with the Trace (T) bit to prevent a Trace
(TRC) trap from occurring more than once for
any instruction. The P bit may be cleared to 0
(no TRC trap pending) or set to 1 (TRC trap
pending).
www.national.com
I bit
The Global Maskable Interrupt Enable (I) bit is
used to enable or disable maskable interrupts.
If this bit and the Local Maskable Interrupt Enable (E) bit are both set to 1, all maskable interrupts are accepted. Otherwise, only the nonmaskable interrupt is accepted. This bit is automatically cleared to 0 when an interrupt occurs
and automatically set to 1 upon completion of
an interrupt service routine.
Absolute Mode The operand is located in memory. Its address is specified within the instruction.
For example:
LOADB 4000, R6
For additional information on the instruction set and instruction encoding, see the CompactRISC CR16B Programmer's
Reference manual.
7.6
Upon reset, all non-reserved bits of the register are cleared
to 0 except for the E bit (bit 9), which is set to 1. When a device reset occurs, the PSR contents prior to the reset are
stored into register R1, allowing the initialization software to
determine the state of the device prior to the reset operation.
7.4
A stack is a one-dimensional data buffer in which values are
entered and removed one at a time. The last valued entered
is the first one removed. A register called the stack pointer
contains the current address of the last item entered on the
stack. In the device, when an item is entered or “pushed”
onto the stack, the stack expands downward in memory (the
stack pointer is decremented). When an item is removed or
“popped” from the stack, the stack shrinks upward in memory
(the stack pointer is incremented).
CONFIGURATION REGISTER
The Configuration (CFG) register is a 16-bit core register that
determines the size of the INTBASE register. For the device,
the CFG register should always be left in its default state
(cleared to zero), resulting in a 16-bit INTBASE register.
7.5
The device uses two type of stacks: the program stack and
the interrupt stack.
ADDRESSING MODES
The program stack is used by the software to save and restore register values upon entry into and exit from a subroutine. The software can also use the program stack to store
local and temporary variables. The stack pointer for this stack
is the SP register.
Each instruction operates on one or more operands. An operand can be a register or a memory location.
Most instructions use one, two, or three device registers as
operands. The instruction opcode specifies the registers to
be operated on. Some instructions may use an immediate
value (a value provided in the instruction itself) instead of a
register.
The interrupt stack is used to save and restore the program
state when an exception occurs (an interrupt or software
trap). The on-chip hardware automatically pushes the program state information onto the stack before the exception
service procedure is executed. Upon exit from the exception
service procedure, the hardware pops this information from
the stack and restores the program state. The stack pointer
for this stack is the ISP register.
Memory locations are accessed only by the Load and Store
commands. The memory location to use for a particular instruction can be specified as an absolute, relative, or far-relative address.
The instruction set supports the following addressing modes:
7.7
Register Mode The operand is a general-purpose register: R0 through R13, RA, or SP. For example:
ADDB R1, R2
Immediate
A constant operand value is specified withMode
in the instruction. In a branch instruction,
the immediate operand is a displacement
from the program counter (PC). In the assembly language syntax, a dollar sign indicates an immediate value. For example:
MULW $4, R4
Relative Mode The operand is located in memory. Its address is obtained by adding the contents of
a general purpose register to the constant
value encoded into the displacement field
of the instruction. For example:
LOADW 12(R5), R6
Far-Relative
The operand is located in memory. Its adMode
dress is obtained by concatenating a pair
of adjacent general-purpose registers to
form a 21-bit value, and adding this value
to the constant value encoded into the displacement field of the instruction.
www.national.com
STACKS
INSTRUCTION SET
Table7 is a summary list of all instructions in the device instruction set. For each instruction, the table shows the mnemonic and a brief description of the operation performed.
In the Mnemonic column, the lower-case letter “i” is used to
indicate the type of integer that the instruction operates on,
either “B” for byte or “W” for word. For example, the notation
ADDi for the “add” instruction means that there are two forms
of this instruction, ADDB and ADDW, which operate on bytes
and words, respectively.
Similarly, the lower-case string “cond” is used to indicate the
type of condition tested by the instruction. For example, the
notation Jcond represents a class of conditional jump instructions: JEQ for Jump on Equal, JNE for Jump on Not Equal,
and so on.
16
For detailed information on all instructions, see the
CompactRISC CR16B Programmer's Reference manual.
Table 7 Device Instruction Set Summary
Mnemonic
Table 7 Device Instruction Set Summary
Mnemonic
Description
Description
SUBCi
Subtract Integer with Carry
TBIT
Test Bit
ADDi
Add Integer
WAIT
Wait for Interrupt
ADDUi
Add Unsigned Integer
XORi
Bitwise Logical Exclusive OR
ADDCi
Add Integer with Carry
ANDi
Bitwise Logical AND
ASHUi
Arithmetic Shift Unsigned
Bcond
Conditional Branch
Bcond0i
Compare Register to 0 and Branch
Bcond1i
Compare Register to 1and Branch
BAL
Branch and Link
BR
Unconditional Branch
CBITi
Clear Bit in Integer
CMPi
Compare Integer
DI
Disable Maskable Interrupts
EI
Enable Maskable Interrupts
EIWAIT
Enable Interrupts and Wait for Interrupt
EXCP
Exception
Jcond
Conditional Jump
JAL
Jump and Link
JUMP
Jump
LOADi
Load Integer
LOADM
Load Multiple Registers
LPR
Load Processor Register
LSHi
Logical Shift Integer
MOVi
Move Integer
MOVXB
Move with Sign-Extension
MOVZB
Move with Zero-Extension
MULi
Multiply Integer
MULSi
Multiply Signed
MULUW
Multiply Unsigned
NOP
No Operation
ORi
Bitwise Logical OR
POP
Pop Registers from Stack
POPRET
Pop and jump RA
PUSH
Push Registers on Stack
RETX
Return from Exception
Scond
Save Condition as Boolean
MULi
Multiply Integer
SBITi
Set Bit in Integer
STORi
Store Integer
STORM
Store Registers to Memory
SUBi
Subtract Integer
17
www.national.com
8.0
Bus Interface Unit
The Bus Interface Unit (BIU) controls the interface between
the internal core bus and those on-chip modules which are
mapped into BIU zones. These on-chip modules are the flash
EEPROM program memory, the ISP-memory and the I/Ozone. It determines the configured parameters for bus access (such as the number of wait states for memory access)
and issues the appropriate bus signals for the requested access.
8.2
The BIU has a set of control registers that determine how
many wait cycles and hold cycles are to be used for accessing memory. Upon start-up of the device, these registers
should be programmed with appropriate values so that the
minimum allowable number of cycles is used. This number
varies with the clock frequency used.
There are four applicable BIU registers: the BIU Configuration (BCFG) register, the I/O Configuration (IOCFG) register,
the Static Zone 0 Configuration (SZCFG0) register and the
Static Zone 1Configuration (SZCFG1) register. These registers control the bus cycle configuration used for accessing
the various on-chip memory types.
Note: The device is manufactured in a 224-pin version which
is used in emulation equipment. In the 224-pin device, the
BIU controls access to both on-chip and off-chip memory and
peripherals. Operation of the 224-pin device and the use of
chip-external memory is beyond the scope of this data sheet.
8.1
BUS CYCLES
Note: A system configuration register called the Module
Configuration (MCFG) register controls the number of wait
cycles used for accessing the EEPROM data memory. This
register is described in Section5.1.
There are four types of data transfer bus cycles:
—
—
—
—
Normal read
Fast read
Early write
Late write
8.2.1
BIU Configuration (BCFG) Register
The BIU Configuration (BCFG) Register is a byte-wide, read/
write register that selects either early write or late write bus
cycles. The register address is F900 hex. Upon reset, the
register is initialized to 07 hex. The register format is shown
below.
The type of data cycle used in a particular transaction depends on the type of CPU operation (a write or a read), the
type of memory or I/O being accessed, and the access type
programmed into the BIU control registers (early/late write or
normal/fast read).
7
For read operations, a basic normal read takes two clock cycles, whereas a fast read bus cycle takes one clock cycle.
Upon reset of the device, normal read bus cycles are enabled
by default.
EWR
For write operations, a basic late write bus cycle takes two
clock cycles, whereas a basic early write bus cycle takes
three clock cycles. Upon reset of the device, early write bus
cycles are enabled by default. However, late write bus cycles
are needed for ordinary write operations, so this configuration should be changed by the application software (see
Section8.2.1).
6 5 4
Reserved
3
2
Note 1
1
Note 1
0
EWR
Early Write. This bit is cleared to 0 for late write
operation (two clock cycles to write) or set to 1
for early write operation.
Note 1: These bits (bit 1 or bit 2) control the configuration of
the 224-pin device used in emulation equipment. The CPU
should set this bit to 1 when it writes to the register.
Upon reset, the BCFG register is initialized to 07 hex, which
selects early write operation. However, late write operation is
required for normal device operation, so the software should
change the register value to 06 hex.
In certain cases, one or more additional clock cycles are added to a bus access cycle. There are two types of additional
clock cycles for ordinary memory accesses, called internal
wait cycles (TIW) and hold (Thold) cycles.
A wait cycle is inserted in a bus cycle just after the memory
address has been placed on the address bus. This gives the
accessed memory more time to respond to the transaction
request. A hold cycle is inserted at the end of a bus cycle.
This holds the data on the data bus for an extended number
of clock cycles.
www.national.com
BIU CONTROL REGISTERS
8.2.2
I/O Zone Configuration (IOCFG) Register
The I/O Zone Configuration (IOCFG) register is a word-wide,
read/write register that sets the timing and bus characteristics of I/O Zone memory accesses. In the device implementation, the registers associated to Port B and Port C reside in
the I/O memory array. (These ports are used as a 16-bit data
port, if the device operates in development mode.)
18
The IOCFG register address is F902 hex. Upon reset, the
register is initialized to 069F hex. The register format is
shown below.
15
14
7
BW
WAIT
HOLD
BW
IPST
13
12
11
Reserved
6
5
Reserved
10
9
IPST
4
3
HOLD
2
HOLD
8
Reserved
1
WAIT
BW
0
Memory Wait cycles
This field specifies the number of TIW (internal
wait state) clock cycles added for each memory
access, ranging from 000 binary for no additional TIW wait cycles to 111 binary for seven
additional TIW wait cycles.
Memory Hold cycles
This field specifies the number of Thold clock
cycles used for each memory access, ranging
from 00 binary for no Thold cycles to 11 binary
for three Thold clock cycles.
Bus Width.
This bit defines the bus width of the zone.
If cleared to 0, a bus width of 8-bit is used.
if set to 1, a bus width of 16-bit is used.
For the device, a bus width of 16-bit needs to
be set.
Post Idle.
An idle cycle follows the current bus cycle,
when the next bus cycle accesses a different
zone.
If cleared to 0, no idle cycle is inserted.
If set to 1, one idle cycle is inserted.
The IPST bit can be cleared to 0, as no idle cycles are required for on-chip accesses.
FRE
IPST
IPRE
Note: Reserved bits must be cleared to 0 when the CPU
writes to the register.
Note: Reserved bits must be cleared to 0 when the CPU
writes to the register.
8.2.4
8.2.3
Static Zone 0 Configuration (SZCFG0) Register
The SCCFG1 register address is F906 hex. Upon reset, the
register is initialized to 069F hex. The register format is
shown below.
The SCCFG0 register address is F904 hex. Upon reset, the
register is initialized to 069F hex. The register format is
shown below.
14 13 12
Reserved
7
BW
WAIT
11
FRE
6
5
Reserved
10
IPRE
9
IPST
4
3
HOLD
2
15
8
Reserved
1
WAIT
Static Zone 1 Configuration (SZCFG1) Register
The Static Zone 1 Configuration (SZCFG1) register is a
word-wide, read/write register that sets the timing and bus
characteristics of Zone 1 memory accesses. In the device implementation of the CompactRISC architecture, Zone 1 is occupied by the boot ROM memory (ISP-Memory).
The Static Zone 0 Configuration (SZCFG0) register is a
word-wide, read/write register that sets the timing and bus
characteristics of Zone 0 memory accesses. In the device implementation of the CompactRISC architecture, Zone 0 is occupied by the flash EEPROM program memory.
15
Memory Hold cycles
This field specifies the number of Thold clock
cycles used for each memory access, ranging
from 00 binary for no Thold cycles to 11 binary
for three T hold clock cycles. These bits are ignored if the SZCFG0.FRE bit is set to 1.
Bus Width.
This bit defines the bus width of the zone.
If cleared to 0, a bus width of 8-bit is used.
if set to 1, a bus width of 16-bit is used.
For the devicedevice a bus width of 16-bit
needs to be set.
Fast Read Enable
This bit enables (1) or disables (0) fast read
bus cycles. A fast read operation takes one
clock cycle. A normal read operation takes at
least two clock cycles.
Post Idle.
An idle cycle follows the current bus cycle,
when the next bus cycle accesses a different
zone.
If cleared to 0, no idle cycle is inserted.
If set to 1, one idle cycle is inserted.
The IPST bit can be cleared to 0, as no idle cycles are required for on-chip accesses.
Preliminary Idle.
An idle cycle is inserted prior to the current bus
cycle, when the new bus cycle accesses a different zone.
If cleared to 0, no idle cycle is inserted.
If set to 1, one idle cycle is inserted.
The IPRE bit can be cleared to 0, as no idle cycles are required for on-chip accesses.
14 13 12
Reserved
7
BW
0
WAIT
Memory Wait cycles
This field specifies the number of TIW (internal
wait state) clock cycles added for each memory
access, ranging from 000 binary for no additional TIW wait cycles to 111 binary for seven
additional TIW wait cycles. These bits are ignored if the SZCFG0.FRE bit is set to 1.
HOLD
19
11
FRE
6
5
Reserved
10
IPRE
9
IPST
4
3
HOLD
2
8
Reserved
1
WAIT
0
Memory Wait cycles
This field specifies the number of TIW (internal
wait state) clock cycles added for each memory
access, ranging from 000 binary for no additional TIW wait cycles to 111 binary for seven
additional TIW wait cycles. These bits are ignored if the SZCFG0.FRE bit is set to 1.
Memory Hold cycles
This field specifies the number of Thold clock
www.national.com
BW
FRE
IPST
IPRE
cycles used for each memory access, ranging
from 00 binary for no T hold cycles to 11 binary
for three T hold clock cycles. These bits are ignored if the SZCFG0.FRE bit is set to 1.
Bus Width.
This bit defines the bus width of the zone.
If cleared to 0, a bus width of 8-bit is used.
if set to 1, a bus width of 16-bit is used.
For the device a bus width of 16-bit needs to be
set.
Fast Read Enable
This bit enables (1) or disables (0) fast read bus
cycles. A fast read operation takes one clock
cycle. A normal read operation takes at least
two clock cycles.
Post Idle.
An idle cycle follows the current bus cycle,
when the next bus cycle accesses a different
zone.
If cleared to 0, no idle cycle is inserted.
If set to 1, one idle cycle is inserted.
The IPST bit can be cleared to 0, as no idle cycles are required for on-chip accesses.
Preliminary Idle.
An idle cycle is inserted prior to the current bus
cycle, when the new bus cycle accesses a different zone.
If cleared to 0, no idle cycle is inserted.
If set to 1, one idle cycle is inserted.
The IPRE bit can be cleared to 0, as no idle cycles are required for on-chip accesses.
For a write operation normal read mode (SZCFG0.FRE=0),
the number of wait cycles is equal to the value written to the
SZCFG0. WAIT field plus one (in the late write mode) or two
(in the early write mode). The number of inserted hold cycles
is equal to the value written to the SCCFG0.HOLD field,
which can range from zero to three.
Writing to the flash EEPROM program memory is a Flash
programming operation that requires some additional steps,
as explained in Section9.3.
8.3.2
Read and write accesses to on-chip RAM is performed within
a single cycle, regardless of the BIU settings.
8.3.3
8.3.4
The IOCFG register determines the access timing for the address range FB00-FB16 hex (Ports B and Port C).
WAIT AND HOLD STATES USED
Flash EEPROM Program Memory
When the CPU accesses the flash EEPROM program memory (address ranges 0000-BFFF and 1C000-1FFFF), the
number of added wait and hold cycles depends on the type
of access and the BIU register settings.
In fast read mode (SZCFG0.FRE=1), a read operation is a
single cycle access. This limits the maximum CPU operating
frequency to either 10 MHz or 20 MHz (see Section9.1.5).
For a read operation in normal read mode (SZCFG0.FRE=0),
the number of inserted wait cycles is one plus the value written to the SZCFG0.WAIT field. The number in this field can
range from zero to seven, so the total number of wait cycles
can range from one to eight. The number of inserted hold cycles is equal to the value written to the SCCFG0.HOLD field,
which can range from zero to three.
For a write operation in fast read mode (SZCFG0.FRE=1),
the number of inserted wait cycles is one. No hold cycles are
used.
www.national.com
Accesses to Peripheral
When the CPU accesses on-chip peripherals in the range of
F800-FAFF hex and FC00-FFFF hex, one wait cycle and one
preliminary idle cycle is used. No hold cycles are used.
The number of wait cycles and hold cycles inserted into a bus
cycle depends on whether it is a read or write operation, the
type of memory or I/O being accessed, and the control register settings.
8.3.1
EEPROM Data Memory
There is either no wait state or one wait state used when the
CPU accesses the EEPROM data memory (address F000F27F hex). The number of required wait states (zero or one)
depends on the CPU clock frequency and operating mode,
and is controlled by programming of the DMCSR.ZEROWS
bit in the MCFG register, as explained in Section9.3. No hold
cycles are used.
Note: Reserved bits must be cleared to 0 when the CPU
writes to the register.
8.3
RAM Memory
20
8.3.5
Access Timing Summary Table
Table8 is a summary showing the number of access cycles
used for various address ranges.
Table 8
Address
Range (hex)
0000-BFFF
Access Timing Table
Access Cycles
Memory or
I/O Type
read
Flash EEPROM Program
Memory
write
SZCFG0.FRE=1:
1 cycle
SZCFG0.FRE=1:
1 cycle
+ BCFG.EWR
(+ programming time)
SZCFG0.FRE=0:
2 cycles
+ BCFG.EWR
+ SZCFG0.WAIT
+ SZCFG0.HOLD
(+ programming time)
1 cycle
MCFG.ZEROWS=1:
1 cycle
(+ programming time)
MCFG.ZEROWS=0:
2 cycles
(+ programming time)
2 cycles
SZCFG0.FRE=0:
2 cycles
+ SZCFG0.WAIT
+ SZCFG0.HOLD
C000-CBFF
F000-F27F
Static RAM Memory
EEPROM Data Memory
1 cycle
MCFG.ZEROWS=1:
1 cycle
MCFG.ZEROWS=0:
2 cycles
F900-FFFF
F800-F9FF
FC00-FFFF
FB00-FBFF
8.3.6
On-Chip Peripherals
2 cycles
Ports B and C
3 cycle
+ IOCFG.WAIT
+ IOCFG.HOLD
specific setup and hold requirements that can be met only by
using enough wait cycles and hold cycles.
Recommended Register Settings
Table9 shows the recommended register settings for various
clock rates. Different clock rates require different register settings because the flash EEPROM program memories have
Table 9
3 cycle
+ BCFG.EW
+ IOCFG.WAIT
+ IOCFG.HOLD
Between clock rates of 10 MHz and 20MHz, the number of
wait states required for memory access (either none or one)
depends on the desired power mode of the program memory.
Recommended Register Settings
Clock Rate
SZCFG0
SZCFG1
IOCFG
< 10 MHz,
0 wait state
0880 hex
0880 hex
0080 hex
10 to 20MHz,
0 wait state
0880 hex
0880 hex
0080 hex
10 to 20MHz,
1 wait state
0080 hex
0080 hex
0080 hex
> 20 MHz,
1 wait state
0080 hex
0080 hex
0080 hex
21
www.national.com
9.0
Memory
The CompactRISC architecture supports a uniform linear address space of 2 megabytes, addressed by 21 bits. The device implementation of this architecture uses only the lowest
128K bytes of address space. Each memory location contains a byte consisting of eight bits.
mode. At higher clock rates, memory read accesses can operate with one wait state.
The programmed number of wait cycles used (either zero or
one) is controlled by the BIU Configuration (BCFG) register
and the Static Zone 0 Configuration (SZCFG0) register.
These registers are described in Section8.0.
Various types of on-chip memory occupy specific intervals
within the address space: 64K bytes of flash EEPROM program memory, 3K bytes of static RAM, 2K bytes of low endurance EEPROM data memory, 128 bytes of high endurance
EEPROM data memory, and 1.5K bytes of ISP memory. All
of these memories are 16 bits wide, and their contents can
be accessed either as bytes (eight bits wide) or words (16
bits wide except for the program memory which only supports
word access).
9.1.2
The flash EEPROM program memory can be programmed
either with the device plugged into a flash EEPROM programmer unit (External Programming) or with the device already installed in the application system (In-SystemProgramming).
If the device is programmed using a flash EEPROM programmer, the device is set into an external programming mode. In
this mode the device operates as if it were a pure flash memory device. The flash memory is programmed without involving any CPU activity.
The CPU core uses the Load and Store instructions to access memory. These instructions can operate on bytes or
words. For a byte access, the CPU operates on a single byte
occupying a specified memory address. For a word access,
the CPU operates on two consecutive bytes. In that case, the
specified address refers to the least significant byte of the
data value; the most significant byte is located at the next
higher address. Thus, the ordering of bytes in memory is
from least to most significant byte, known as “little-endian” ordering. For more efficient data access operations, 16-bit variables should be stored starting at word boundaries (at even
address).
9.1
If the device is to be programmed within the user application,
it can either be done by an user written boot loader or by utilizing a pre-programmed in-system-programming code (ISPCode) residing in the boot ROM array of the device.
The device executes the pre-programmed in-system-programming code if it operates in the In-System-Programming
Mode (ISP-Mode). To enter the ISP-Mode the device must be
reset (or powered-up) with the ENV0-pin set to low level and
the ENV1-pin set to high level (or left open). Also if the flash
program memory is not programmed yet (FLCTRL2.EMPTY
bit is still set) the device automatically enters the ISP-Mode
after reset, even though both pins ENV0 and ENV1are at
high level (or left open). If the device enters the ISP-Mode it
starts execution at address E000 hex.
FLASH EEPROM PROGRAM MEMORY
The flash EEPROM program memory is used to store the application program. The 64K bytes of this memory reside in the
address range of 0000-BFFF hex and 1C000-1FFFF in Zone
0 of the CR16B address space. A normal CPU write operation to this memory has no effect.
In ISP-Mode the program code can be downloaded into the
device using one of the on-chip USARTs and written into the
flash program memory. For more detailed information on the
In-System-Programming features of the pre-programmed
ISP-Code please refer to the ISP-Monitor manual.
The flash EEPROM Program Memory module has the following features:
—
—
—
—
—
—
—
—
—
—
—
—
—
—
9.1.1
64K bytes arranged as 32K by 16 bits
Page size of 64 words
30 µs programming pulse per word
Page mode erase with a 1 ms pulse, mass erase with
4ms pulse
All erased flash EEPROM program memory bits read 1
Fast single cycle read access
Flexible software controlled In-System-Programming
(ISP) capability
Pipelined programming cycles through double-buffered data register, with write access disabled when the
register is full
Programming high voltage and timing generated onchip
Memory disabled when address is out of range
Requires valid key for program and erase to proceed
Provide busy status during programming and erase
Read accesses disabled during programming and
erase
Security features to limit read/write access
9.1.3
User-Coded Programming Routines
Instead of using a flash EEPROM programmer unit or the
conventional in-system programming mode, you can write
your own processor code to program and erase the flash
EEPROM program memory. User-written code is more flexible than using the other programming methods. Like the conventional in-system programming mode, the device is
programmed while it is installed in the system. It is not necessary to reset the device or use the ENV0/ENV1 pins to
configure the device.
User-written flash programming code must reside outside of
the flash program memory. This is because the entire program memory becomes unavailable while programming or
erasing any part of this memory.
9.1.4
Flash EEPROM Programming and Verify
The flash EEPROM program memory programming and
erase can be performed using different methods. It can be
done through user code that is stored in system RAM, or
through In-System-Programming mode, but should not be
programmed through the flash EEPROM program memory it-
Reading
Program memory read accesses can operate without wait cycles with a CPU clock rate of up to 20MHz in the normal
www.national.com
Conventional Programming Modes
22
self as no instruction or data can be fetched from it while it is
being programmed. All program and erase operations must
be preceded immediately by writing the proper key to the program memory key register PGMKEY.
separation of the program memory into rows is transparent to
the user, as the transition is handled by the flash program
memory interface. Figure 3 shows a flowchart for a programming sequence.
The flash EEPROM program memory is divided into 256
pages, each page containing 64 words (each 16 bits wide).
Each page is further divided into two adjacent rows. A page
erase will erase one page. Programming is done by writing to
all the words within a row, one word following another sequentially within one single high voltage pulse. This is supported through a double-buffered write-data buffer scheme.
Byte programming is not supported. Programming should be
done on erased rows.
start
MSTAT.PGMBUSY
=1?
A mass erase requires the following code sequence (assuming that this sequence will not be interrupted to do another
flash erase or programming):
Yes
No
1. Check for MSTAT.PGMBUSY not set.
2. Set up flash timing reload registers for mass erase operation.
3. Set FLCSR.MERASE = 1.
4. If interrupt was enabled, disable interrupt.
5. Write proper key value to PGMKEY.
6. Write to any valid location within the flash EEPROM program memory.
7. If interrupt was disabled in step 4, re-enable interrupt.
8. Wait for MSTAT.PGMBUSY to clear.
9. Set FLCSR.MERASE = 0.
10. Restore flash timing reload registers for normal operation.
disable interrupt
if necessary
write PGMKEY
write memory
A page erase requires the following code sequence (assuming that this sequence will not be interrupted to do another
flash erase or programming):
1.
2.
3.
4.
5.
6.
7.
re-enable interrupt
if necessary
Check for MSTAT.PGMBUSY not set.
Set FLCSR.ERASE = 1.
If interrupt was enabled, disable interrupt.
Write proper key value to PGMKEY.
Write to any valid location within the page to be erased.
If interrupt was disabled in step 3, re-enable interrupt.
Set FLCSR.ERASE = 0.
last word?
When programming, the data to be written into the flash EEPROM program memory is first written into a double-buffered
write-data buffer. When a piece of data is written to the page
while the flash EEPROM program memory is idle, the write
cycle will start. Due to the double-buffered nature of the writedata buffer, a second word can be written to the flash EEPROM program memory. This will then set FLCSR.PMLFULL flag indicating the buffer is now full. When the first write
is done, the memory address would be incremented, and the
second word would be written to that address while keeping
the high voltage pulse active; the FLCSR.PMLFULL flag is
cleared. Another word can then be written to the buffer, and
this programming will repeat until there are no more words to
be programmed. This allows pipelined writes to different
words on the same row within the same high voltage pulse.
If the programming sequence exceeds a row, the flash programming interface will automatically initiate a programming
pulse for the next row. The FLCSR.PMLFULL bit is also
cleared when programming of the last word of the current
row is completed, e.g. programming of the entire row is completed and MSTAT.PGMBUSY is cleared. This means, the
Yes
done
No
Yes
FLCSR.PMLFULL
=0?
No
Figure 2. Programming Sequence for
the Program Memory
9.1.5
Erase and Programming Timing
The internal hardware of the device handles the timing of
erase and programming operations. To drive the timing control circuits, the device divides the system clock by a programmable prescaler factor. You should select a prescaler
value to produce a program/erase clock of 200 kHz (or as
close as possible to 200 kHz without exceeding 200 kHz).
For the timing control circuit to operate correctly, you must
23
www.national.com
program the prescaler value in advance and leave it unchanged while a program or erase operation is in progress.
A similar (but separate) prescaler factor is applied to the EEPROM data memory. See Section9.1.7 and Section9.3.4 for
details.
9.1.6
12.5 MHz / (62+1) = 198.4 kHz. Do not modify this register
while a flash EEPROM program or erase operation is in
progress.
Upon reset, this register is programmed by default with the
value 63 hex (99 decimal), which is an appropriate setting for
a 20 MHz system clock.
Flash EEPROM Program Memory Control and
Status Register (FLCSR)
9.1.8
The Flash EEPROM Program Memory Control and Status
(FLCSR) register is a byte-wide, read/write register that contains several status and control bits related to the program
memory. All reserved bits must be written with 0 for the memory to operate properly when writing to this register. Upon reset, this register is cleared to zero when the flash memory on
the chip is in the idle state.
The FLSTART register is a byte-wide read/write register that
controls the program and erase start delay time. This value
is loaded into the lower 8 bits of the flash timing counter, and
at the same time, 002 is loaded into the upper 2 bits. Before
you program or erase the program memory for the first time,
program the FLSTART register with the proper prescaler value, FTSTART. The flash timing counter generates a delay of
(FTSTART+1) prescaler output clocks. The default value
provides a delay time of 10µs when the prescaler output
clock is 200kHz. Do not modify this register while a program
or erase operation is in progress.
The register format is shown below.
7
6
MERASE
PMER
PMBUSY
PMLFULL
MERASE
9.1.7
4
Reserved
3
2
1
0
PMLFULL
PMBUSY
PMER
Reserved
Flash EEPROM Program Memory page erase.
When set (1) with MERASE bit cleared, a valid
write to the flash EEPROM program memory
erases the entire flash EEPROM program
memory page pointed to by the write address
rather than performing a write to the addressed
memory location.
Program Memory Busy. This bit is automatically set to 1 when the flash EEPROM program
memory is busy being programmed, and
cleared to 0 at all other times. (The MSTAT.PGMBUSY is also set to 1 whenever the PMBUSY
bit is set to 1.)
Program Memory Write-Latch Buffer Full.
When set (1), the double-buffered data register
for program memory write operations is full.
When cleared (0), the double-buffered data
register is not full.
Mass Erase Flash EEPROM Program Memory
Array. When set (1) in ISP or test mode, a valid
write to the flash EEPROM program memory
performs an erase to the whole flash EEPROM
program memory rather than perform a write to
the addressed memory location. However, it is
necessary to enter new values into the
FLERASE and FLEND registers to adjust the
mass erase timing before starting the mass
erase.
Upon reset, this register resets to 0116 when the flash memory on the chip is in an idle state.
9.1.9
Program Memory Transition Time Reload
Register (FLTRAN)
The FLTRAN register is a byte-wide read/write register that
controls some program/erase transition times. This value is
loaded into the lower 8 bits of the flash timing counter, and at
the same time, 002 is loaded into the upper 2 bits. Before you
program or erase the program memory for the first time, you
should program the FLTRAM register with the proper prescaler value, FTTRAN. The flash timing counter generates a
delay of (FTTRAN + 1) prescaler output clocks. The default
value provides a delay time of 5µs when the prescaler output
clock is 200kHz. Do not modify this register while a program
or erase operation is in progress.
Upon reset, this register resets to 001 6 when the flash memory on the chip is in an idle state.
9.1.10
Program Memory Programming Time Reload
Register (FLPROG)
The FLPROG register is a byte-wide read/write register that
controls the programming pulse width. This value is loaded
into the lower 8 bits of the flash timing counter, and at the
same time, 002 is loaded into the upper 2 bits. Before you
program or erase the program memory for the first time, program the FLPROG register with the proper prescaler value,
FTPROG. The flash timing counter generates a programming
pulse width of (FTPROG + 1) prescaler output clocks. The
default value provides a delay time of 30µs when the prescaler output clock is 200kHz.
Program Memory Timing Prescaler Register
(FLPSLR)
The FLPSLR register is a byte-wide, read/write register that
selects the prescaler divider ratio for the flash EEPROM program memory programming clock. Before you program or
erase the program memory for the first time, you should program the FLPSLR register with the proper prescaler value,
an 8-bit value called FTDIV. The device divides the system
clock by (FTDIV+1) to produce the program memory programming clock.
Do not modify this register while program/erase operation is
in progress.
Upon reset, this register resets to 051 6 when the flash memory on the chip is in idle state.
You should choose a value of FTDIV to produce a clock of the
highest possible frequency that is equal to or just less than
200 kHz. For example, if the system clock frequency is 12.5
MHz, use the value 3E hex (62 decimal) for FTDIV, because
www.national.com
Program Memory Start Time Reload (FLSTART)
24
9.1.11
Program Memory Erase Time Reload Register
(FLERASE)
9.1.16
The PGMKEY register is a byte-wide, write-only register that
must be written with a key value (A316 ) immediately prior to
each write to the flash EEPROM program memory. Otherwise, the write operation to the program memory will fail. This
feature is intended to prevent unintentional programming of
the program memory.
The FLERASE register is a byte-wide read/write register that
controls the erase pulse width. This value is loaded into the
upper 8 bits of the flash timing counter, and at the same time,
112 is loaded into the lower 2 bits. Before you program or
erase the program memory for the first time, program the
FLERASE register with the proper prescaler value, FTER.
The flash timing counter generates a erase pulse width of
4×(FTER + 1) prescaler output clocks. The default value provides a delay time of 1ms when the prescaler output clock is
200kHz. Do not modify this register while a program or erase
operation is in progress.
Reading this register always returns FF hex.
Upon reset, the write enable status that is generated as a result of writing to this key register is cleared.
9.2
For mass erase, this value should be changed to C7 16 to
generate a pulse width that is four times as long as the page
erase.
Program Memory End Time Reload Register
(FLEND)
9.3
The FLEND register is a byte-wide read/write register that
controls the delay time after a program/erase operation. This
value is loaded into the lower 8 bits of the flash timing
counter, and at the same time, 00 2 is loaded into the upper 2
bits. Before you program or erase the program memory for
the first time, program the FLEND register with the proper
prescaler value, FTEND. The flash timing counter generates
a delay of (FTEND + 1) prescaler output clocks. The default
value provides a delay time of 5µs when the prescaler output
clock is 200kHz. Do not modify this register while program/
erase operation is in progress.
This memory also support flash memory test mode and there
is no read protection or permanent write protection for this
memory.
9.3.1
Program Memory Prescaler Count Register
(FLPCNT)
The programmed number of wait cycles used (either zero or
one) is controlled by a bit in the Data Memory Control Status
register (DMCSR.ZEROWS). This register is described in
Section9.3.3.
The FLPCNT register is a byte-wide read-only register that
returns the value of the program memory prescaler counter.
FPCNT contains the flash timing prescaler present count value.
9.3.2
Program Memory Timer Count Register 1
(FLCNT1)
Programming
Before you begin programming the flash EEPROM data
memory, you should set the value in the EEPROM Data
Memory Prescaler register. This register sets the prescaler
used to generate the data memory programming clock from
the system clock, as described in Section9.3.4.
The FLCNT1 register is a byte-wide read-only register that
returns the lower 8 bits of the program memory timing
counter value. FLCNTL is the lower 8 bits of the flash timer
present count value.
9.1.15
Reading
The flash EEPROM data memory read accesses can operate without wait cycles with a CPU clock rate of up to 20MHz
in the normal mode. At higher clock rates, read accesses can
operate with one wait state.
For mass erase, this value should be changed to 1316 to provide for a delay time twenty times that of the standard delay.
9.1.14
FLASH EEPROM DATA MEMORY
The flash EEPROM data memory is used for non-volatile
storage of data. The 2K bytes of low endurance memory reside in the address range of E800-EFFF hex and the 128
bytes of high endurance memory reside in the address range
of F000-F07F hex. The CPU reads or writes this memory by
using ordinary byte-wide or word-wide memory access commands. This memory shares the same array as the ISP flash
program memory.
Upon reset, this register resets to 0016 when the flash memory on the chip is in idle state.
9.1.13
RAM MEMORY
The static RAM memory is used for temporary storage of
data and for the program and interrupt stacks. The 3K bytes
of this memory reside in the address range of C000-CBFF
hex. Each memory access requires one clock cycle, for a
byte or word access. No wait cycles or hold cycles are required. For non-aligned word access, each memory access
requires multiple clock cycles.
Upon reset, this register resets to 31 1 6 when the flash memory on the chip is in idle state.
9.1.12
Program Memory Write Key Register (PGMKEY)
A code fetch from ISP flash EEPROM program memory is
not possible while flash EEPROM data memory is being programmed because they share the same memory array.
Program Memory Timer Count Register 2
(FLCNT2)
After the CPU performs a write to the flash EEPROM data
memory, the on-chip hardware completes the EEPROM programming in the background. When programming begins,
the on-chip hardware sets the DMCSR.DMBUSY bit to 1,
and also sets the MSTAT.PGMBUSY bit to 1. When programming is completed, it resets these status bits back to 0. Once
the software writes to the flash EEPROM data memory, it
should not attempt to access the EEPROM data memory
The FLCNT2 register is a byte-wide read-only register that
returns the upper 2 bits of the program memory timing
counter value and also the state of the key flash memory interface timing signals. The interface timing signals are only
used in special test modes. Their function is beyond the
scope of this document.
25
www.national.com
again until programming is completed and the status bit is reset to 0.
memory for the first time, you should program the DMPSLR
register with the proper prescaler value, an 8-bit value called
FTDIV. The device divides the system clock by (FTDIV+1) to
produce the data memory programming clock.
The device hardware internally generates the voltages and
timing signals necessary for programming. No additional
power supply is required, nor any software required except to
check the status bit for completion of programming. The minimum time required to erase and reprogram a byte or word is
1.1 ms. The programmed values can be verified by using normal memory read operations. The prescaler output drives a
10-bit counter to generate timing pulses and there are five reload registers to produce various pulse widths.
You should choose a value of FTDIV to produce a clock of the
highest possible frequency that is equal to or just less than
200 kHz. Upon reset, this register is programmed by default
with the value 63 hex (99 decimal), which is an appropriate
setting for a 20 MHz system clock.
9.3.5
If a reset occurs during a programming or erase operation,
the operation is terminated. The reset is extended until the
flash memory returns to the idle state. Therefore, the timing
logic and program or erase state machine is not cleared on
reset; they are cleared on power-up with the clear signal active until the bus signals are in a known state.
The DMSTART register is a byte-wide read/write register that
controls the program/erase start delay time. This value is
loaded into the lower 8 bits of the flash timing counter, and at
the same time, 002 is loaded into the upper 2 bits. Before you
write to the data memory for the first time, you should program the DMSTART register with the proper prescaler value,
an 8-bit value called FTSTART. The flash timing counter generates a delay of (FTSTART + 1) prescaler output clocks. The
default value provides a delay time of 10µs when the prescaler output clock is 200kHz. Do not modify this register while
program/erase operation is in progress.
The flash EEPROM data memory does not have permanent
read-protection or write-protection features like those available for the EEPROM program memory. However, the Data
Memory Write Key Register provides a way to “lock” the data
written to the data memory.
9.3.3
Data Memory Control and Status Register
(DMCSR)
Upon reset, this register resets to 0116 when the flash memory on the chip is in idle state.
The DMCSR register is a byte-wide, read/write register used
with the flash EEPROM data memory or ISP flash EEPROM
program memory. When writing to this register, all reserved
bits must be written with 0 for the memory to operate properly. There are two status/control bits, as shown in the register
format below.
7
6
5
4
Reserved
ZEROWS
DMBUSY
3
2
1
0
ERASE
DMBUSY
ZEROWS
Reserved
Data Memory Start Time Reload Register
(DMSTART)
9.3.6
Data Memory Transition Time Reload Register
(DMTRAN)
The DMTRAN register is a byte-wide read/write register that
controls some program/erase transition times. This value is
loaded into the lower 8 bits of the flash timing counter, and at
the same time, 002 is loaded into the upper 2 bits. Before you
write to the data memory for the first time, you should program the DMTRAN register with the proper prescaler value,
an 8-bit value called FTTRAN. The flash timing counter generates a delay of (FTTRAN + 1) prescaler output clocks. The
default value provides a delay time of 5 µs when the prescaler
output clock is 200kHz. Do not modify this register while program/erase operation is in progress.
Zero Wait-State Access. When cleared (0), the
flash EEPROM data memory will be read in two
cycles. When set (1), the flash EEPROM data
memory will be read in one cycle.
Data Memory Busy. This bit is automatically set
to 1 when the flash EEPROM data memory or
the ISP flash EEPROM program memory is
busy being programmed, and cleared to 0 at all
other times. (The MSTAT.PGMBUSY is also set
to 1 whenever the DMBUSY bit is set to 1.)
Erase ISP Flash Program Memory Page.
When set (1) a valid write to the ISP flash EEPROM program memory will erase the entire
ISP flash EEPROM program memory page
pointed to by the write address rather than performing a write to the addressed memory location. This bit should be cleared to 0 and remain
cleared after the write operation.
Upon reset, this register resets to 001 6 when the flash memory on the chip is in idle state.
9.3.7
Data Memory Programming Time Reload
Register (DMPROG)
Upon reset, the DMCSR register is cleared to zero when the
flash memory on the chip is in the idle state.
The DMPROG register is a byte-wide read/write register that
controls the programming pulse width. This value is loaded
into the lower 8 bits of the flash timing counter, and at the
same time, 002 is loaded into the upper 2 bits. Before you
write to the data memory for the first time, you should program the DMPROG register with the proper prescaler value,
an 8-bit value called FTPROG. The flash timing counter generates a programming pulse width of (FTPROG + 1) prescaler output clocks. The default value provides a delay time of
30µs when the prescaler output clock is 200kHz. Do not modify this register while program/erase operation is in progress.
9.3.4
Upon reset, this register resets to 051 6 when the flash memory on the chip is in idle state.
ERASE
Data Memory Prescaler Register (DMPSLR)
The DMPSLR register is a byte-wide, read/write register that
selects the prescaler divider ratio for the EEPROM data
memory programming clock. Before you write to the data
www.national.com
26
9.3.8
Data Memory Erase Time Reload Register
(DMERASE)
Note: Operation of this register is different in from the
PGMKEY register used with the program memory. It is not
necessary to write the key value to DMKEY every time you
write to the data memory.
The DMERASE register is a byte-wide read/write register
that controls the erase pulse width. This value is loaded into
the upper 8 bits of the flash timing counter, and at the same
time, 11 2 is loaded into the lower 2 bits. Before you write to
the data memory for the first time, you should program the
DMERASE register with the proper prescaler value, an 8-bit
value called FTER. The flash timing counter generates a
erase pulse width of 4×(FTER + 1) prescaler output clocks.
The default value provides a delay time of 1ms when the
prescaler output clock is 200kHz. Do not modify this register
while program/erase operation is in progress.
9.4
The In-System Program memory is part of the flash memory
array that contains the flash EEPROM data memory. It is not
possible to access the ISP memory while programming the
flash EEPROM data memory or access the flash EEPROM
data memory while programming the ISP memory. The 1.5K
bytes of ISP memory resides in the address range of E000E5FF and is used for storing the boot ROM. The ROM contains the code that performs in-system programming, and is
programmed at the factory. In ISP mode, code execution
starts at address E000.
Upon reset, this register resets to 31 1 6 when the flash memory on the chip is in idle state.
For mass erase, this value should be changed to C71 6 when
the flash EEPROM data memory goes to idle mode.
9.3.9
The ISP program memory and flash EEPROM data memory
share the same memory array, which makes it impossible to
access one type of memory while the other is being programmed.
Data Memory End Time Reload Register
(DMEND)
The ISP memory has the following features:
The DMEND register is a byte-wide read/write register that
controls the delay time after a program/erase operation. This
value is loaded into the lower 8 bits of the flash timing
counter, and at the same time, 00 2 is loaded into the upper 2
bits. Before you write to the data memory for the first time,
you should program the DMEND register with the proper
prescaler value, an 8-bit value called FTEND. The flash timing counter generates a delay of (FTEND + 1) prescaler output clocks. The default value provides a delay time of 5µs
when the prescaler output clock is 200kHz. Do not modify
this register while program/erase operation is in progress.
— 1.5K bytes flash EEPROM program memory
— Page size of 4 words, divided into two rows of 2 words
each
— Odd and even bytes within a page can be erased separately
— 30µs programming pulse width per word
— Page mode erase with 1ms pulse, mass erase with
4ms pulse
— All erased memory bits read 1
— Fast read access time
— Requires valid key for program and erase to proceed
— Provide memory protection and security features for
flash EEPROM program memory
— Security features may limit accesses to ISP memory
— Disable memory when address is out of range to prevent accessing data memory
— Mass erase only allowed in test modes
— Provide busy status during programming and erase
— Read/write accesses disabled during programming/
erase
— Programming high voltage and timing generated onchip
Upon reset, this register resets to 0016 when the flash memory on the chip is in idle state.
For mass erase, this value should be changed to 13 16 .
9.3.10
Data Memory Prescaler Count Register
(DMPCNT)
The DMPCNT register is a byte-wide read-only register that
returns the value of the data memory prescaler counter.
FPCNT is the flash timing prescaler present count value.
9.3.11
Data Memory Timer Count Register (DMCNT)
The DMCNT register is a word-wide read-only register that
returns the data memory timing counter value. The reserved
bits return 0000002.
9.4.1
Reading
The ISP flash EEPROM program memory read accesses
can operate without wait cycles with a CPU clock rate of up
to 20MHz in the normal mode. At higher clock rates, read accesses can operate with one wait state.
FTCNT[0:9] is the flash timer present count value.
9.3.12
ISP MEMORY
Data Memory Write Key Register (DMKEY)
The DMKEY register is a byte-wide, read/write register that
provides a way to “lock” the data contained in the EEPROM
data memory. Upon reset, the register is automatically set to
C9 hex, which is the key value. Writing to the EEPROM data
memory is allowed as long as the DMKEY register contains
this value. When the register contains any value other than
C9 hex, writing the EEPROM data memory is disallowed.
The programmed number of wait cycles used (either zero or
one) is controlled by BIU Configuration (BCFG) register and
the Static Zone 1 Configuration (SZCFG1) register. These
registers are described in Section8.0.
9.4.2
User-Coded Programming Routines
All program and erase operations must be preceded by writing the proper key to the program memory key register ISPKEY. The programming code can be in-system RAM, but
cannot be from ISP flash EEPROM program memory or flash
EEPROM data memory as accesses within these ranges are
To “lock” the current data stored in the data memory, write another value (such as 00 hex) to the DMKEY register. To “unlock” the data memory, write the value C9 hex to the DMKEY
register.
27
www.national.com
not permitted while ISP flash EEPROM program memory is
being programmed.
through a byte write instruction when the write instruction is
anywhere within the user boot ROM area (defined above) except for the last two words. When the user boot ROM area
has been disabled, this word cannot be programmed in the
IRE environment. Note that when this word is erased for reprogramming, the other words in the same page must first be
saved, and then re-programmed.
The ISP flash memory is divided into 192 pages, each page
containing 4 words (each 16 bits wide). Each page is further
divided into two rows. Erase is carried out one page at a time,
whereas programming is carried out one row (or one partial
row) at a time.
7
Once an erase or programming operation is started, the PGMBUSY bit in the MSTAT register is automatically set, and
then cleared when the operation is complete. All high-voltage
pulses and timing needed for programming and erasing are
provided internally. The program memory cannot be accessed while the PGMBUSY bit is set.
Erasing a page requires the following code sequence:
Verify that the MSTAT.PGMBUSY bit is cleared.
Set the DMCSR.ERASE bit to 1.
Locally disable interrupts.
Write proper key value to the ISPKEY register.
Write to any valid page to be erased.
Re-enable interrupts disabled in Step 3.
Set the DMCSR.ERASE bit to 0.
9.4.3
Programming Procedure
Programming is done by writing one byte or word at a time
and should be done on already erased memory.
0
E5FF Byte
Upon reset, the byte located in the E5FF address is read into
the FLSEC register. This byte cannot be written to in the IRE
environment. The format of the E5FF byte is shown below:
7
4
FROMWR
Programmed values can be verified through normal read operations.
3
0
FROMRD
The FROMRD and FROMWR fields in address location
E5FF respectively provide read and write security to the flash
EEPROM program memory array while executing instructions in all environments except IRE. The user should always
write 00002 to enable security feature.
If a reset occurs in the middle of an erase or programming
operation, the operation is terminated. The reset is extended
until the flash EEPROM memory returns to the idle state.
Erase and Programming Timing
0000, 0001, 0010, 0100, 1000: Security feature enabled
0011, 0101, 011x, 1001, 101x, 11xx: Security feature disabled
The program and erase timing are controlled by the flash EEPROM data memory logic.
9.4.5
1
CODEAREA[9:8]
011, 101, 11x: Program memory is empty, do not start up in IRE
Verify that the MSTAT.PGMBUSY bit is cleared.
Locally disable interrupts.
Write proper key value to the ISPKEY register.
Write a byte or word to the addressed location.
Re-enable interrupts disabled in Step 2.
9.4.4
2
Reserved
000, 001, 010, 100: Program memory contains user code
Programming the ISP flash EEPROM program memory requires the following code sequence:
1.
2.
3.
4.
5.
4
CODEAREA[9:8]
The 2 least significant bits in address E5FE
contains the two most significant bits of the 10bit CODEAREA field. The description of
CODEAREA is shown in the E5FC section.
EMPTY
The EMPTY status indicates if the flash EEPROM program memory array is empty or not.
It is located in the 3 most significant bits in address E5FE. When two or more bits in the
EMPTY field are set, the flash EEPROM program memory is empty. Upon reset of the device and the environment select pins are all
high, the device operates in ISP environment
rather than IRE environment. After the program
memory has been filled with user code, this
field should be cleared to 0002.
Erase Procedure
1.
2.
3.
4.
5.
6.
7.
5
EMPTY
FROMRD
Memory Control and Protection Features
The last 8 bytes of the ISP memory are reserved for special
functions and some of these bytes provide memory protection and security for the flash EEPROM program memory.
Read and various types of write protection are provided.
FROMWR
During the reset stretch period, bytes located at E5FE and
E5FF are read out to the FLCTRL2 and FLSEC registers respectively. Upon reset and before an instruction fetch, bytes
located at E5FC and E5FD are read out to the FLCTRL2 and
FLCTRL1 registers respectively. Parts of FLCTRL2 register
are loaded at different times.
Upon reset of the chip, read security is enabled
and 0000 is returned in all environments except
IRE. The internal program code can only be executed in the IRE environment when read security is activated.
Upon reset of the chip, write security is enabled
and program and erase operations to the flash
EEPROM program memory in either programming modes are prevented.
E5FE Byte
Once read/write security is enabled, the odd numbered bytes
from address E5F9 to E5FF cannot be erased. Once a security feature has been enabled, it cannot be undone. To prevent the security status from being erased, the ISP and data
memory array cannot be mass erased.
Upon reset of the chip, the byte located at E5FE is read into
the FLCTRL2 register. It can be written in the ISP or test environments. It can also be written in the IRE environment
Note: In flash memory test mode, this condition also prevents the odd numbered bytes of the high endurance flash
EEPROM data memory (F001 to F07F) from being erased;
www.national.com
28
however, the even numbered bytes of the high endurance
flash EEPROM data memory (F000 to F07E) and the ISP
flash EEPROM program memory (E000 to E5FE) can be
erased.
E5FD Byte
Upon the reset of the chip, the byte located at the E5FD address is read into the FLCTRL1 register. This byte can only
be written in the ISP or test environments but not in the IRE
environment. If this byte is erased for re-programming, the
user must first save the other bytes in the same page, and
then re-program those bytes. The format of the E5FD byte is
shown below:
Read/write is overridden through PADX.
E5FC Byte
Upon reset of the chip, E5FC is read into the FLCTRL2 register. The byte at E5FC is written in the ISP or test environments, or in the IRE environment through a byte-write
instruction when the write instruction is anywhere within the
user boot ROM area except for the last two words. When the
user boot ROM area has been disabled by having a value of
7F 16 in BOOTAREA, this word cannot be programmed in the
IRE environment. Note that when this word is erased for reprogramming, the other words in the same page must first be
saved, and then re-programmed also. The E5FC register format is shown below:
7
7
Reserved
6
0
BOOTAREA
BOOTAREA provides write protection to part of the program
memory, see Figure4. When the write security feature is not
enabled and BOOTAREA does not contain the value 7F 16 ,
then the program memory range from 0 to (BOOTAREA*128)+127 is considered as user boot ROM area and cannot be written to. The maximum protected memory range is
therefore 16K-127 bytes when BOOTAREA contains the value 7E16 .
0
CODEAREA[7:0]
1FFFFh
This byte contains the lowest 8 bits of the CODEAREA field.
When appended to the left with the lowest 2 bits in the address E5FE, it forms the complete CODEAREA field, which
provides write protection to all or part of the program memory, see Figure3. When write security is not enabled and
CODEAREA does not contain the value 3FF 16, the program
memory range from (CODEAREA×128) to 1FFFF is considered as protected user code area and cannot be written. The
minimum protected memory range is therefore 256 bytes
when CODEAREA contains the value 3FE. Note that the
C000-FFFF memory range is not considered as program
memory and is not protected by CODEAREA.
boot area maximum limit
3F80h
(BOOTAREA×128)+127
protected user boot
area
0000h
CR16MHR6
Address Map
Figure 4.
Memory Protection through BOOTAREA
1FFFFh
user
When BOOTAREA contains the value 7F 16 , write protection
is disabled. When write security has been enabled, the entire
program memory area is already write protected in all environments.
non-code area, not
protected
Note that when a new value is written into BOOTAREA, write
protection controlled by BOOTAREA is updated after the
next device reset.
protected
code area
10000h
C000h
protected
code area
user
CODEAREA×128
9.4.6
Test Mode
The ISP flash EEPROM program memory test mode allows
direct access to the flash memory from the device pins, and
bypasses the CR16B core. This test mode also accesses the
flash memory cells that are not used in data memory (three
out of four bytes in each page).
0000h
CR16MHR6
Address Map
Figure 3. Memory Protection through CODEAREA
9.4.7
When CODEAREA contains the value 3FF16 , write protection is disabled. When the user code area overlaps into the
user boot ROM area, the overlap area is governed by a more
restrictive write protection feature, which is the user boot
ROM area. When write security has been enabled, the entire
program memory area is already write protected in all environments.
Flash Program Memory Control Register 1
(FLCTRL1)
The FLCTRL1 register is a read-only byte-wide register. The
value of this register is loaded from memory address E5FD 16
when the chip comes out of reset. The BOOTAREA field defines a user boot ROM area to be write protected. The Flash
EEPROM Program Memory Control Register 1 format is
shown below:
Note that when a new value is written into CODEAREA, write
protection controlled by CODEAREA is updated after the
next device reset.
7
Reserved
29
6
0
BOOTAREA
www.national.com
When BOOTAREA has any value other than 7F1 6, then the
memory at 0 to (BOOTAREA×128)+15 is considered as user
boot ROM area and is write protected. When it has a value of
7F 16 , then there is no user boot ROM area to be write protected
9.4.8
9.4.10
The In-System-Programming Memory Write Key (ISPKEY)
register is a byte-wide, write-only register. It contains the enable key to enable writes to ISP flash EEPROM program
memory. A value of 6A16 must be written to this register immediately preceding every write to the ISP flash EEPROM
program memory for the flash write operation to proceed,
otherwise any other write operation will clear the key (the
only exception is that the subsequent write is another write to
this key register with the proper key, in which case the key is
still set). A read always returns FF 1 6. Engineering note: on
reset, the write enable status that is generated as a result of
a write to this key register is cleared. The ISP Memory Write
Key register format is shown below:
Flash Program Memory Control Register 2
(FLCTRL2)
The FLCTRL2 register is a read-only word-wide register. The
value of this register is loaded from memory addresses
E5FC1 6 and E5FE1 6 when the chip comes out of reset. When
the device starts execution, the EMPTY bit indicates whether
the flash EEPROM program memory is empty of not, and selects the chip to be in IRE or ISP environment if the external
environment pins are all high. The CODEAREA field defines
a user code area to be write protected. The Flash EEPROM
Program Memory Control Register 2 format is shown below:
15
13 12
EMPTY
10 9
Reserved
7
ISPKYVAL
0
When the bits are either 0112 , 1012, 110 2 , or
111 2 , and if the device’s environment select
pins are all high, the device will come out of reset in ISP environment instead of IRE environment.
CODEAREA When it has any value other than 3FF 16 , then
the memory (CODEAREA×128) to 1FFFF16 is
considered as user code area and is write protected. When it has a value of 3FF 1 6, then there
is no code protection area to be write protected.
ISPKYVAL is the ISP Flash Program Memory Write Enable
Key Value.
Flash Program Memory Security Register
(FLSEC)
The FLSEC register is a read-only byte-wide register. When
the chip comes out of reset, the value of this register is loaded from memory address E5FF 16 . The FROMRD and
FROMWR field control the read and write security of the flash
EEPROM program memory respectively. The Flash EEPROM Program Memory Security register format is shown
below:
7
4 3
FROMWR
0
FROMRD
0000, 0001, 0010, 0100, 1000: Security feature enabled
0011, 0101, 011x, 1001, 101x, 11xx: Security feature disabled
FROMRD
FROMWR
When read security feature is enabled, the
flash EEPROM program memory can only be
read in IRE environment, but will return 00001 6
in other environments; also, erase to odd numbered bytes from address E5F9 16 to E5FF 1 6
and mass erase to ISP and flash EEPROM
data memory array are ignored unless PADX is
activated (see security override below).
Unless PADX is activated (see override below),
when write security feature is enabled, all further writes and erases to flash EEPROM program memory, erase to odd numbered bytes
from address E5F9 16 to E5FF16 , and mass
erase to ISP and flash EEPROM data memory
array are ignored.
www.national.com
0
CODEAREA
EMPTY
9.4.9
ISP Memory Write Key Register (ISPKEY)
30
10.0 Interrupts
The Interrupt Control Unit (ICU31L) receives interrupt requests from internal and external sources and generates interrupts to the CPU. Interrupts from the timers, USARTs,
MICROWIRE/SPI interface, Multi-Input Wake-Up, and A/D
converter are all maskable interrupts. The highest-priority interrupt is the Non-Maskable Interrupt (NMI), which is triggered by a falling edge received on the NMI input pin. The
NMI pin is not available on the 44-pin packages.
10.1
Table 10
Dispatch Table Entries
0: Reserved
1: NMI
2: Reserved
3: Reserved
4: Reserved
5: SVC (Supervisor Call Trap)
INTERRUPT OPERATION
An exception is an event that temporarily stops the normal
flow of program execution and causes execution of a separate service routine. Upon completion of the service routine,
execution of the interrupted program continues from the point
at which it was stopped.
6: DVC (Divided by Zero Trap)
There are two kinds of exceptions, called traps and interrupts. A trap is the result of some action or condition in the
program itself, such as execution of an Exception (EXCP) instruction. An interrupt is a CPU-external event, such as a signal received on a Multi-Input Wake-Up input or a request
from an on-chip peripheral module for service.
10: UND (Undefined Instruction Trap)
7: FLG (Flag Trap)
8: BPT (Breakpoint Trap)
9: TRC (Trace Trap)
11: Reserved
12: Reserved
13: Reserved
14: Reserved
The operation of traps is beyond the scope of this data sheet.
For information on traps, and for additional detailed information on interrupts not provided in this data sheet, please refer
to the CompactRISC CR16B Programmer's Reference Manual.
10.1.1
15: Reserved
16: INT0 (Reserved)
17: INT1 (Flash EEPROM Program Memory)
18: INT2 (Reserved)
Interrupt Operation Summary
19: INT3 (Reserved)
When an interrupt occurs, the on-chip hardware performs the
following steps:
20: INT4 (Reserved)
21: INT5 (ADC)
1. Decrements the Interrupt Stack Point (ISP) by four.
2. Saves the contents of the Program Counter (PC) and
Processor Status Register (PSR) on the interrupt stack.
3. Clears the I, P, and T bits in the Processor Status Register (PSR). These are the Global Maskable Interrupt
Enable bit, Trace Trap Pending bit, and Trace bit, respectively.
4. Reads the interrupt vector from the Interrupt Vector Register (IVCT).
5. Combines the interrupt vector with the value in the Interrupt Base (INTBASE) register to obtain an address in
the Interrupt Dispatch Table, and loads the dispatch table entry into the Program Counter (PC).
22: INT6 (MIWU Interrupt 3)
23: INT7 (MIWU Interrupt 2)
24: INT8 (MIWU Interrupt 1)
25: INT9 (MIWU Interrupt 0)
26: INT10 (USART 2 Tx)
27: INT11 (USART 1Tx)
28: INT12 (Reserved)
29: INT13 (MICROWIRE/SPI Rx/TX)
30: INT14 (ACCESS.bus)
From this point onward, the CPU executes the interrupt service routine. The service routine ends with a Return from Exception (RETX) instruction. This returns the CPU to the
interrupted program. The CPU restores the contents of the
PC and PSR registers from the stack and increments the Interrupt Stack Pointer by four.
31: INT15 (USART 2 Rx)
10.1.2
36: INT20 (Reserved)
32: INT16 (USART 1 Rx)
33: INT17 (Reserved)
34: INT18 (CAN)
35: INT19 (Reserved)
Service Routine Addresses
37: INT21 (Reserved)
When an interrupt or trap occurs, the CPU executes a service routine. There are different service routines for different
interrupts and traps. Each service routine may reside anywhere in program memory. The starting addresses of the service routines are contained in a table called the Dispatch
Table. Entries in the table are organized in the order shown
in Table10.
31
www.national.com
Table 10
Dispatch Table Entries
another when the parameters are too large to easily fit into
the registers. A high-level language typically allocates the local (non-static) variables on the stack.
38: INT22 (Reserved)
39: INT23 (VTUD Interrupt Request 4)
The pointer to the program stack is the SP register, which
must be initialized prior to any register save/restore operations or data transfer operations. Using the program stack, an
interrupt routine needs to initially save the contests of all registers that it uses, and restore those register contents before
returning to the interrupted program.
40: INT24 (VTUD Interrupt Request 3)
41: INT25 (VTUD Interrupt Request 3)
42: INT26 (VTUD Interrupt Request 1)
43: INT27 (T2B Timer 2 Interrupt B)
44: INT28 (T2A Timer 2 Interrupt A)
45: INT29 (T1B Timer 1Interrupt B)
10.2
46: INT30 (T1A Timer 1Interrupt A)
A non-maskable interrupt is triggered by a falling edge on the
NMI input pin, which generates a software trap. The NMI pin
is an asynchronous input with Schmitt trigger characteristics
and an internal synchronization circuit. Therefore, no external synchronizing is needed.
47: INT31 (RTI Timer 0)
Each entry in the Dispatch Table consists of two bytes that
provide bits 1 through 16 of the starting address of the corresponding service routine. The full 21-bit address of a service
routine is reconstructed by adding a leading 0 and a trailing
0 to the 16-bit table entry.
Upon reset, the non-maskable interrupt is disabled and
should remain disabled until the software initializes the interrupt table, interrupt base, and interrupt stack pointer. It can
be enabled by setting either of two control bits in the External
NMI Control/Status (EXNMI) register. The two bits are called
the EN (Enable) bit and the ENLCK (Enable and Lock) bit.
The INTBASE register is a pointer to the Dispatch Table.
Upon reset, the initialization software must write the starting
address of the Dispatch Table to the INTBASE register, a 21bit register with the five most significant bits and the least significant bit always equal to 0. It is typically kept in the flash
EEPROM program memory. The Dispatch Table is 48 words
long.
The EN bit enables the NMI trap until an NMI trap event or a
reset occurs. An NMI trap automatically resets the EN bit. Using this bit to enable the NMI trap is intended for applications
where the NMI pin is toggled frequently but nested NMI traps
are not needed. The trap service routine should re-enable the
NMI trap by setting the EN bit before returning to the main
program.
Each interrupt or trap source has an associated vector number ranging from 0 to 31, as indicated in Table10. When an
interrupt occurs, the hardware multiplies the vector by 2,
adds the result to the contents of the INTBASE register, and
uses the resulting address to obtain the service routine starting address from the corresponding entry in the Dispatch Table. This address is placed in the Program Counter so that
the CPU begins executing the interrupt service routine.
The ENLCK bit enables the NMI trap and locks it in the enabled state. In other words, it leaves the NMI trap enabled
even after the trap occurs. It can be cleared only by a reset
operation. After the bit is set, an NMI trap is triggered by each
falling edge on the NMI pin, allowing nested NMI traps.
Figure5 summarizes the method used by the device to generate the starting address of a service routine.
10.1.3
To use the EN bit, the ENLCK must remain cleared to 0. Otherwise, the EN bit is ignored.
Stack Usage
10.3
When an interrupt occurs, the CPU automatically preserves
the contents of the Program Counter (PC) and Processor
Status Register (PSR) by pushing them on the interrupt stack
and decrementing the Interrupt Stack Pointer by four. The
service routine ends with a Return from Exception (RETX) instruction, which returns control to the interrupted program by
restoring the PC and PSR values and incrementing the Interrupt Stack Pointer (ISP) by four.
MASKABLE INTERRUPTS
Maskable interrupts can be enabled or disabled under software control. There are 31 level-triggered maskable interrupt
sources (including some reserved for future expansion), organized into levels of priority. If more than one interrupt event
occurs at any given time, the interrupt source with the highest
priority is serviced first. The others must wait until the highest-priority interrupt is serviced and is no longer pending.
Figure11 lists the maskable interrupt sources of the device
in order of priority, from the highest-priority interrupt (IRQ31)
to the lowest (IRQ0).
Prior to using any interrupts, the Interrupt Stack Pointer (ISP)
must be initialized so that it points to a space in RAM where
the interrupt stack will be kept. The stack grows downward in
memory (toward address zero) when an interrupt occurs and
items are pushed onto the stack. The stack shrinks upward
in memory when an interrupt service routine ends and items
are popped from the stack.
To enable a maskable interrupt, the enable bit must be set in
the applicable peripheral module and also in the appropriate
Interrupt and Enable Mask register, IENAM0 or IENAM1. In
addition, both the Global Maskable Interrupt Enable bit (I)
and the Local Maskable Interrupt Enable bit (E) must be set
to 1 in the PSR register. If either one of these bits is 0, then
all maskable interrupts are disabled. The CR16B core supports IRQ0, but ICU31L reserves IRQ0 so that it is not connected to any interrupt source.
Many routines need to use the general-purpose registers R0
through R13. To preserve the existing register contents, a
routine can save register contents on the program stack upon
start of the routine and restore the register contents prior to
completion of the routine. The software can also use the program stack to transfer data parameters from one routine to
www.national.com
NON-MASKABLE INTERRUPT
32
~
~
0
31
INTBASE
0
Reserved
1
NMI
2
Reserved
3
Reserved
4
Reserved
5
SVC
Supervisor Call Trap
6
DVZ
Divide By Zero Trap
7
FLG
Flag Trap
8
BPT
Breakpoint Trap
9
TRC
Trace Trap
10
UND
Undefined Instruction Trap
11
Reserved
12
Reserved
13
Reserved
14
DBG
Debug Trap
15
ISE
In-System Emulator Interrupt
16 to 127
INTn
Maskable Interrupts
~
Non-maskable Interrupt
~
Figure 5.
Table 11 Maskable Interrupt Priority List
Interrupt Request
Table 11 Maskable Interrupt Priority List
Source
Interrupt Request
Source
IRQ31
RTI (Timer 0), highest priority
IRQ8
MIWU16 Interrupt 1
IRQ30
T1A (Timer 1 input A)
IRQ7
MIWU16 Interrupt 2
IRQ29
T1B (Timer 1 input B)
IRQ6
MIWU16 Interrupt 3
IRQ28
T2A (Timer 2 input A)
IRQ5
ADC
IRQ27
T2B (Timer 2 input B)
IRQ4-IRQ2
Reserved
IRQ26
VTUA (VTU Interrupt Request 1)
IRQ1
Flash Program Memory
IRQ25
VTUB (VTU Interrupt Request 2)
IRQ0
Reserved, lowest priority
IRQ24
VTUC (VTU Interrupt Request 3)
IRQ23
VTUD (VTU Interrupt Request 4)
IRQ22-IRQ19
Reserved
IRQ18
CAN
IRQ17
Reserved
Both the E bit and I bit can be controlled with the Load Processor Register (LPR) instruction. In addition, the E bit is
easily changed by executing the Enable Interrupts (EI) or
Disable Interrupts (DI) instruction. Using the EI and DI instructions avoids the possibility of an interrupt occurring within a read-modify-write operation on the PSR register.
IRQ16
USART1 Rx
10.4
IRQ15
USART2 Rx
IRQ14
ACCESS.bus
The Interrupt Control Unit uses the following interrupt control
and status registers:
IRQ13
MICROWIRE/SPI Rx/Tx
IRQ12
Reserved
IRQ11
USART1 Tx
IRQ10
USART2 Tx
IRQ9
MIWU16 Interrupt 0
INTERRUPT REGISTERS
— Non-Maskable Interrupt Status Register (NMISTAT)
— Non-Maskable Interrupt Status Monitor Reg. (NMIMNTR)
— External NMI Control/Status Register (EXNMI)
— Interrupt Enable and Mask Register 0 (IENAM0)
— Interrupt Enable and Mask Register 1 (IENAM1)
— Interrupt Vector Register (IVCT)
33
www.national.com
— Interrupt Status Register 0 (ISTAT0)
— Interrupt Status Register 1 (ISTAT1)
— Interrupt Debug Register (IDBG)
10.4.3
The IVCT register is a byte-wide, read-only register that contains the encoded value of the enabled and pending
maskable interrupt with the highest priority. The on-chip hardware automatically updates this field whenever there is a
change in the highest-priority enabled and pending maskable
interrupt. The CPU reads this register during an interrupt acknowledge core bus cycle to determine where to begin executing the interrupt service routine. The register contents are
guaranteed to be valid at that time. The register is not guaranteed to contain valid data during a hardware update operation. The register format is shown below.
The following CPU core registers are also used in processing
interrupts:
— Interrupt Stack Pointer (ISP)
— Interrupt Base Register (INTBASE)
10.4.1
Non-Maskable Interrupt Status Register
(NMISTAT)
The NMISTAT register is a byte-wide, read-only register that
holds the current pending status of the Non-Maskable Interrupt (NMI). This register is cleared upon reset. It is also
cleared each time it is read. The register format is shown below.
7
6
EXT
10.4.2
5
4
3
Reserved
2
1
7
0
External Non-Maskable Interrupt Request.
When set to 1 by the hardware, it indicates an
external Non-Maskable Interrupt request has
occurred. See the description of the EXNMI
register below for more information.
10.4.4
External NMI Control/Status Register (EXNMI)
EN
PIN
ENLCK
6
5
4
Reserved
3
2
ENLCK
1
PIN
5
4
3
2
INTVECT
1
0
Interrupt Vector. This 6-bit field contains the encoded value of the enabled and pending
maskable interrupt with the highest priority. For
example, if interrupts IRQ1 and IRQ6 are both
enabled and pending, the higher-priority interrupt is IRQ6. As a result the 6 bit interrupt vector is 010110.
Interrupt Enable and Mask Register 0 (IENAM0)
The IENAM0 register is a word-wide, read/write register that
enables or disables the individual interrupts IRQ0 through
IRQ15. The register format is shown below.
15
0
IENA(15:0)
A bit set to 1 enables the corresponding interrupt. A bit
cleared to 0 disables the corresponding interrupt. Upon reset, this register is initialized to FFFF hex.
0
EN
10.4.5
Interrupt Enable and Mask Register 1 (IENAM1)
The IENAM0 register is a word-wide, read/write register that
enables or disables the individual interrupts IRQ16 through
IRQ31. The register format is shown below.
Enable NMI Trap. When set to 1, NMI traps are
enabled and falling edge on the NMI pin generates a NMI trap. Each occurrence of an NMI
trap automatically clears the EN bit. The trap
service routine should set the EN bit to 1 before
returning control to the interrupted program.
When EN is cleared to 0, NMI traps are disabled unless they are enabled with the ENLCK
bit. When the ENLCK bit is set to 1, the EN bit
is ignored.
NMI Pin. This bit shows the current state of the
NMI input pin (without logical inversion). A 1 indicates a high level and a 0 indicates a low level on the pin. This is a read-only bit. In a write
operation, the value written to this bit position is
ignored.
Enable and Lock NMI Trap. When set to 1, NMI
traps are enabled and locked in the enabled
state. Each falling edge on the NMI pin generates a NMI trap, even if a previous NMI trap has
occurred and is still being processed. When
ENLCK is cleared to 0, NMI traps are disabled
unless they are enabled with the EN bit.
www.national.com
6
0
INTVECT
0
EXT
The EXNMI register is a byte-wide, read/write register that
shows the current state of the NMI pin and also allows the
NMI trap to be enabled by setting either the EN bit or the ENLCK bit. Both of these bits are cleared upon reset. When the
software writes to this register, it must write 0 to all reserved
bit positions for the device to function properly. EN, ENLCK,
and TST are cleared upon reset. The register format is
shown below.
7
Interrupt Vector Register (IVCT)
15
0
IENA(31:16)
A bit set to 1 enables the corresponding interrupt. A bit
cleared to 0 disables the corresponding interrupt. Upon reset, this register is initialized to FFFF hex.
10.4.6
Interrupt Status Register 0 (ISTAT0)
The ISTAT0 register is a word-wide, read-only register that
indicates which maskable interrupt inputs to the ICU31L
(IRQ0 through IRQ15) are currently active. The register format is shown below.
15
0
IST(15:0)
IST(15:0)
34
Interrupt Status bits. Each bit indicates the current status of an interrupt input to the ICU31L,
corresponding to interrupts IRQ0 through
IRQ15. A bit set to 1 indicates an active interrupt input, even when the interrupt is masked
out by the IENAM0 register. A bit cleared to 0
indicates an inactive interrupt input.
10.5.2
10.4.7
Clearing an interrupt request before it is serviced may cause
a spurious interrupt because the CPU may detect an interrupt not reflected in the Interrupt Vector (IVCT) register. To
ensure reliable operation, clear interrupt requests only while
interrupts are disabled.
Interrupt Status Register 1 (ISTAT1)
The ISTAT1 register is a word-wide, read-only register that
indicates which maskable interrupt inputs to the ICU31L
(IRQ16 through IRQ31) are currently active. The register format is shown below.
15
Changing the polarity of an interrupt input (for example, in the
Multi-Input Wake-Up module) can cause a spurious interrupt,
and therefore should be done only while interrupts are disabled.
0
IST(31:16)
IST(31:16)
10.4.8
Interrupt Status bits. Each bit indicates the current status of an interrupt input to the ICU31L,
corresponding to interrupts IRQ16 through
IRQ31. A bit set to 1 indicates an active interrupt input, even when the interrupt is masked
out by the IENAM0 register. A bit cleared to 0
indicates an inactive interrupt input.
For the same reason, clearing an enable bit in a peripheral
module should be carried out only while the interrupt is disabled.
10.5.3
Interrupt Debug Register
Unless specifically enabled by the software, nested interrupts will not occur. When the CPU acknowledges an interrupt, the I bit in the PSR register is automatically cleared to 0
for the duration of the service routine, disabling any further
maskable interrupts.
INTERRUPT PROGRAMMING
PROCEDURES
To allow nested interrupts, an interrupt service routine should
first set or clear the respective interrupt enable bits to specify
which peripherals will be allowed to interrupt the current service routine. The present interrupt routine should be disabled
(or interrupt pending bit cleared). The service routine should
then set the PSR.I bit to 1, thus enabling maskable interrupts.
This bit can be controlled with the Store Processor Register
(SPR) and Load Processor Register (LPR) instructions.
The following subsections provide information on initializing
the device for interrupts, clearing interrupts, and nesting interrupts.
10.5.1
Nesting Interrupts
Interrupts may be nested, or in other words, an interrupt service routine can itself be interrupted by a different interrupt
source. There is no hardware limitation on the number of interrupt nesting levels. However, the interrupt stack must not
be allowed to overflow its allocated memory space.
The IDBG register is a word-wide read-only register, which
contains various status information of the ICU31L. The lowest 6 bits contain the INTVECT value during the last read
from address FE00. The next 6 bits contain the INTVECT
value when a maskable interrupt request is sent to the
CR16B core. Upon reset, this register is set to 0000 hex.
10.5
Clearing Interrupts
Initialization
Upon reset, all interrupts are disabled. To program the device
for interrupt operation and to enable interrupts, use the following procedure in the application software:
Note:
Clearing the pending bit of the current interrupt should not be
immediately followed by enabling further interrupts by setting
the I bit in the PSR register. Wait states must be inserted into
the software after clearing the interrupt pending bit and before another interrupt. Placing a NOP instruction will perform
this instruction. This is because the instruction which resets
the pending bit may not yet be finished when the interrupts
are already enabled again by setting the I bit in the PSR register. To avoid this situation the user has to make sure that
prior to enabling the interrupt an additional instruction is inserted. This could look like the example below:
1. Set the Interrupt Stack Pointer (ISP)
2. Load the INTBASE register so that it points to the base
of the Interrupt Dispatch Table.
3. Perform any required preparation steps for the interrupt
service routines.
4. Initialize the peripheral devices that can generate interrupts and set their respective interrupt enable bits.
5. Set the relevant bits in the interrupt mask registers
(IENAM0 and IENAM1)
Note: The MIWU16 interrupts have no local interrupt enable bits, which means you can only disable the
MIWU16 interrupts if you clear the specific bits in the IENAM register.
6. Use the Load Processor Register (LPR) instruction to
set I bit in the PSR register.
7. When the device is ready to execute interrupts, set the
E bit in the PSR register by executing the Enable Interrupts (EI) instruction.
SBITi $0, T1ICRL # clear pending bit
NOP
# NOP instruction
MOVW $0x0a00, r0 # enable further interrupts
LPR
r0, psr
A CBITi or SBITi instruction may be used to clear the interrupt
pending bit. In such cases, a spurious interrupt may occur.
Once maskable interrupts are enabled by setting the E and I
bits, you can disable and re-enable all maskable interrupts
locally by using the Enable Interrupts (EI) and Disable Interrupts (DI) instructions, which set and clear the E bit.
35
www.national.com
11.0 Power Management
The Power Management Module (PMM) improves the efficiency of the device by changing the operating mode (and
therefore the power consumption) according to the required
level of device activity.
by the slow clock rather than the normal high-speed clock. In
order to work properly in Power Save mode, modules that
perform real-time operations (such as a USART baud rate
generator) must be reprogrammed to use the slower clock.
The device can operate in any of four power modes:
To reduce power consumption as much as possible, the program should execute a WAIT instruction during periods of
CPU inactivity.
—
—
—
—
Active
Power Save
Idle
Halt
11.3
In the Idle mode, the clock is stopped for most of the device.
Only the Power Management Module and Timing and Watchdog Module continue to operate. Both of these modules use
the slow clock in this mode.
Table12 summarizes the main properties of the four operating modes: the state of the high-frequency oscillator (on or
off), the type of clock used by most modules, and the clock
used by the Timing and Watchdog Module (TWM).
Table 12
Mode
Active
11.4
Power Mode Operating Summary
Main Clock
Slow Clock
Power Save On or Off
Slow Clock
Slow Clock
Idle
On or Off
None
Slow Clock
Halt
Off
None
None
11.5
The low-frequency oscillator continues to operate in all four
modes and power must be provided continuously to the device power supply pins. In the Halt mode, however, the internal SLCLK does not toggle, and as a result, the TWM timer
and Watchdog Module do not operate. For the Power Save
and Idle modes, the high-frequency oscillator can be turned
on or off under software control, as long as the low-frequency
oscillator is used.
11.1
ACTIVE MODE
The low frequency clock is used in Power Save mode as the
system clock source. In Idle mode, it is used as the clock
source for the PMM and the TWM, both of which remain
clocked. The clock source may be a low frequency clock oscillator or the prescaler from the high frequency clock.
Power consumption in the Active mode can be reduced by
selectively disabling unused modules and/or by executing
the WAIT instruction. When WAIT is executed, the core stops
executing new instructions and waits for an interrupt.
The Oscillating Low Frequency Clock (OLFC) input indicates
to the PMM when the clock is stable and therefore usable.
When OLFC is set to 1, it indicates that the clock can be
used. When OLFC is set to 0, the PMM does not use the low
frequency clock. OLFC is generated by the “slow clock good”
output of the Dual Clock and Reset module (CLK2RES).
POWER SAVE MODE
In the Power Save mode, all device modules operate off the
low-frequency clock. If the low-frequency clock is generated
from an external crystal network, the high-frequency clock
oscillator can be turned off to further reduce power consumption.
While in reset (i.e., the reset signal is active), the PMM outputs the clock as long as the clock selected for use upon reset is stable (OHFC or OLFC are 1). If the clock selected is
not stable, the PMM clock output remains low.
All on-chip modules continue to operate in the Power Save
mode, with the SLCLK acting as their system clock. If this
mode is entered by using the WAIT command, the CPU is inactive and waits for an interrupt to wake up. Otherwise, CPU
continues to function normally at the lower frequency of the
slow clock.
11.6
SWITCHING BETWEEN POWER MODES
Switching from a higher to a lower power consumption mode
is accomplished by writing an appropriate value to the Power
Management Control/Status Register (PMCSR). Switching
from a lower power consumption mode to the Active mode is
usually triggered by a hardware interrupt. Figure6 shows the
four power consumption modes and the events that trigger a
transition from one mode to another.
The low frequency of the clock in Power Save mode limits the
operation of modules such as the USARTs, MICROWIRE interface, A/D Converter, and timers because they are driven
www.national.com
CLOCK INPUTS AND RESET
CONFIGURATION
The system uses a high frequency clock Active mode. The
source of this clock in the device is a high frequency crystal
oscillator. The Oscillating High Frequency Clock (OHFC) input indicates to the Power Management Module (PMM)
when this clock is stable and therefore usable. The clock can
be used when OHFC is set to 1. The PMM does not use the
high frequency clock when OHFC is set to 0. OHFC can be
the output of a clock monitor or a strapped input signal to this
module.
In the Active mode, all device modules are fully operational.
This is the operating mode upon reset. Most device modules
use the clock generated by the high-frequency clock oscillator. The clock rate is determined by the external crystal network.
11.2
HALT MODE
In the Halt mode, all device clocks are disabled and the highfrequency oscillator is shut off. In this mode, the device consumes the least possible power while maintaining the device
memory and register contents. The low-frequency oscillator
continues to operate in this mode, but with very low power
consumption due to its power-optimized design.
High-Frequency
Clock Used TWM Clock
Oscillator
On
IDLE MODE
36
Reset
IDLE
Active
PSM =1
Power Save
IDLE =1
HW event
or PSM =0
and WAIT
HALT =1
and WAIT
Idle
HW event
HALT
Halt
HW event
Figure 6.
Power Modes and Transitions
Some of the power-up transitions are based on the occurrence of a wake-up event. An event of this type can be either
a maskable interrupt or a non-maskable interrupt (NMI). All
of the maskable hardware wake-up events are gathered and
processed by the Multi-Input Wake-Up Module, which is active in all modes. Once a wake-up event is detected, it is
latched until an interrupt acknowledge cycle occurs or a reset
is applied.
WBPSM
A wake-up event causes a transition to the Active mode and
restores normal clock operation, but does not start execution
of the program. It is the interrupt service routine associated
with the wake-up source (MIWU16 or NMI) that causes actual program execution to resume.
11.6.1
Power Management Control/Status Register
(PMCSR)
The Power Management Control/Status Register (PMCSR)
is a byte-wide, read/write register that controls the operating
power mode (Active, Power Save, Idle, or Halt) and enables
or disables the high-frequency oscillator in the Power Save
and Idle modes. The two most significant bits, OLFC and
OHFC, are read-only status bits controlled by the hardware.
Upon reset, the non-reserved bits of this register are cleared.
The format of the register is shown below.
OHFC
7
6
5
4
3
2
1
0
OLFC OHFC WBPSM Reserved HALT IDLE DHF PSM
PSM
DHF
OLFC
Power Save Mode. When this bit is 0, the device operates in the Active mode. Writing a 1 to
this bit position puts the device into the Power
Save mode, either immediately or upon execution of the next WAIT instruction, depending on
the WBPSM bit.
The PSM bit can be set and cleared by the software. It is also cleared by the hardware when a
hardware wake-up event is detected.
Disable High-Frequency Oscillator. This bit enables (0) or disables (1) the high-frequency oscillator in the Power Save or Idle mode. (The
high-frequency oscillator is always enabled in
Active mode and always disabled in Halt mode,
regardless of this bit settings.) The DHF bit is
cleared automatically when a hardware wakeup event is detected.
37
Idle Mode. When this bit is set and the device
is in Power Save mode, the device enters the
Idle mode upon execution of a WAIT instruction. In order to enter the Idle mode directly
from the Active mode, the WBPSM bit must be
set before the WAIT instruction is executed.
The IDLE bit can be set and cleared by the software. When a hardware wake-up event is detected, this bit is cleared automatically and the
device returns to the Active mode.
Halt Mode. When this bit is set and the device
is in Idle mode, the device enters the Halt mode
upon execution of a WAIT instruction. In order
to enter the Halt mode directly from the Active
mode, the WBPSM bit must be set before the
WAIT instruction is executed.
The Halt bit can be set and cleared by the software. When a hardware wake-up event is detected, this bit is cleared automatically and the
device returns to the Active mode.
Wait Before Entering Power Save Mode. When
the CPU writes a 1 to the PSM bit, the WBPSM
determines when the transition from Active to
Power Save mode is done. If the WBPSM bit is
0, the switch to Power Save mode is initiated
immediately; the PSM bit in the register is set
to 1 upon completion of the switch to Power
Save mode. If the WBPSM bit is 1, the device
continues to operate in Active mode until the
next WAIT instruction, and then enters the
Power Save mode. In this case, the PSM bit is
set to 1 immediately, even if a WAIT instruction
has not yet been executed.
In the Active mode, the WBPSM bit must be set
in order to enter the Idle or Halt mode.
Oscillating High-Frequency Clock. This readonly bit indicates the status of the high-frequency clock. If this bit is 1, the high-frequency clock
is available and stable. If this bit is 0, the highfrequency clock is either disabled, not available
to the Power Management Module, or operating but not yet stable. The device can switch to
the Active mode only when this bit is 1.
Oscillating Low-Frequency Clock. This readonly bit indicates the status of the low-frequency (slow) clock. If this bit is 1, it indicates that
the slow clock is running and stable. The slow
clock can be either the prescaled fast clock (the
default) or the external oscillator (if selected).
The Dual Clock module will not allow a transition to the slow crystal mode unless the slow
crystal is operating, so this bit should be 1 under normal circumstances.
The device can switch from the Active mode to
the Power Save or Idle mode only if the OLFC
bit is 1. There is no such restriction on switching to the Halt mode.
www.national.com
11.6.2
Active to Power Save Mode
11.6.6
Software-Controlled Transition to Active Mode
A transition from the Active mode to the Power Save mode is
accomplished by writing a 1 to the PMCSR.PSM bit. The
transition to Power Save mode is either initiated immediately
or upon execution of the next WAIT instruction, depending on
the PMCSR.WBPSM bit.
A transition from the Power Save mode to the Active mode
can be accomplished by either a software command or a
hardware wake-up event. The software method is to write a
0 to the PMCSR.PSM bit. The value of the register bit changes only after the transition to the Active mode is completed.
For an immediate transition to Power Save mode (PMCSR.WBPSM=0), the CPU continues to operate using the lowfrequency clock. The PMCSR.PSM bit is set to 1 when the
transition to the Power Save mode is completed.
If the high-frequency oscillator is disabled for Power Save operation, the oscillator must be enabled and allowed to stabilize before the transition to Active mode. To enable the highfrequency oscillator, the software writes a 0 to the PMCSR.DHF bit. Before writing a 0 to the PMCSR.PSM bit, the
software should first monitor the PMCSR.OHFC bit to determine whether the oscillator has stabilized.
For a transition upon the next WAIT instruction (PMCSR.WBPSM=1), the CPU continues to operate in the Active
mode until it executes a WAIT instruction. Upon execution of
the WAIT instruction, the device enters the Power Save
mode and the CPU waits for the next interrupt event. In this
case, the PMCSR.PSM bit is set to 1 when it is written, even
before the WAIT instruction is executed.
11.6.3
11.6.7
A hardware wake-up event switches the device directly from
Power Save, Idle, or Halt mode to the Active mode.
Hardware wake-up events are:
Entering the Idle Mode
• a Non-Maskable Interrupt (NMI)
• a valid wake-up event on a Multi-Input Wake-Up channel
Entry into the Idle mode is accomplished by writing a 1 to the
PMCSR.IDLE bit and then executing a WAIT instruction.
When a wake-up event occurs, the on-chip hardware performs the following steps:
The Idle mode can be entered only from the Active or Power
Save mode. For entry from the Active mode, the PMCSR.WBPSM bit must be set before the WAIT instruction is executed.
11.6.4
1. Clears the PMCSR.DHF bit, thus enabling the high-frequency clock (if it was disabled).
2. Waits for the PMCSR.OHFC bit to be set, which indicates that the high-frequency clock is operating and is
stable.
3. Switches the device into the Active mode.
Disabling the High-Frequency Clock
In systems where the low-frequency crystal is available and
is used to generate the Slow Clock (SLCLK), power consumption can be reduced further in the Power Save or Idle
mode by disabling the high-frequency clock. This is accomplished by writing a 1 to the PMCSR.DHF bit before executing the WAIT instruction that puts the device in the Power
Save or Idle mode. The high-frequency clock is turned off
only after the device enters the Power Save or Idle mode.
11.6.8
Power Mode Switching Protection
The Power Management Module has several mechanisms to
protect the device from malfunctions caused by missing or
unstable clock signals.
The PMCSR.OHFC and PMCSR.OLFC bits indicate the current status of the high-frequency and low-frequency clock oscillators, respectively. The software can check the
appropriate bit before it changes to an operating mode that
requires the clock. A status bit set to 1 indicates an operating,
stable clock. A status bit cleared to 0 indicates a clock that is
disabled, not available, or not yet stable.
The CPU operates on the low-frequency clock in Power Save
mode. It can turn off the high-frequency clock at any time by
writing a 1 to the PMCSR.DHF bit.
The high-frequency oscillator is always enabled in Active
mode and always disabled in Halt mode, regardless of the
PMCSR.DHF bit setting.
During a power mode transition, if there is a request to switch
to a mode that uses clock with its status bit cleared to 0, the
switch is delayed until that bit is set to 1 by the hardware.
Immediately following power-up and entry into the Active
mode, the software must wait for the low-frequency clock to
become stable before it can put the device in the Power Save
mode. It should monitor the PMCSR.OLFC bit for this purpose. Once this bit is set to 1, the slow clock is stable and the
Power Save mode can be entered.
11.6.5
Wake-Up Transition to Active Mode
When the system is built without an external crystal network
for the low-frequency clock, the high-frequency clock is divided by a prescaler factor to produce the low-frequency clock.
In this situation, the high-frequency clock is disabled only in
the Halt mode, and cannot be disabled for the Power Save or
Idle mode, regardless of the software command issued.
Entering the Halt Mode
Entry into the Halt mode is accomplished by writing a 1 to the
PMCSR.HALT bit and then executing a WAIT instruction.
Without an external crystal network for the low-frequency
clock, the device comes out of the Halt or Idle mode and enters the Active mode with the high-speed oscillator used as
the clock. The device can still enter the Power Save from the
Active mode by using the high-frequency-clock divider to
generate the slow clock (PMCSR.DHF=0).
The Halt mode can be entered only from the Active or Power
Save mode. For entry from the Active mode, the PMCSR.WBPSM bit must be set before the WAIT instruction is executed.
Note: For correct operation in the absence of a low-frequency crystal, the X2CKI pin must be tied low (not left floating) so
that the hardware can detect the absence of the crystal.
www.national.com
38
12.0 Dual Clock and Reset
The Dual Clock and Reset module (CLK2RES) generates a
high-speed main system clock from an external crystal network and a slow clock (32.768 kHz or other rate) for operating the device in Power Save mode. It also provides the main
system reset signal, a power-on reset function, a main clock
prescaler to generate two additional low speed clocks, and
an 32kHz oscillator start-up delay.
Figure7 is block diagram of the Dual Clock and Reset module.
System
Reset
Reset
Power-On-Reset
Stop Main Osc.
Stop
Main Osc In
Preset
X1CKI
Start-Up-Delay
14-Bit Timer
X1CKO
Time-out
Main Clk
Main Osc.
4-Bit
Prescaler
2 Low
Speed Clk
Outputs
8-Bit
Prescaler
Div.
by-2
32kHz Osc.
Start-Up-Delay
6-Bit Timer
Mux
4-Bit
Prescaler
X2CKI
Time-out
Low Speed
Clk
Good Low
Speed Clk
Preset
X2CKO
Stop Low
Speed Clk
Stop 32kHz Osc.
Figure 7.
12.1
Good Main
Clk
Dual Clock and Reset Module Block Diagram
The crystals and other oscillator components should be
placed close to the X1CKI/X1CLO and X2CKI/X2CLO device
input pins to keep the printed trace lengths to an absolute
minimum.
EXTERNAL CRYSTAL NETWORK
An external crystal network is required at pins X1CKI and
X1CKO for the main clock. A similar external crystal network
may be used at pins X2CKI and X2CKO for the slow clock in
packages that have these pins. If an external crystal network
is not used for the slow clock, the clock is generated by dividing the fast main clock.
Figure8 shows the required crystal network at X1CKI/
X1CKO and optional crystal network at X2CKI/X2CKO.
Table13 shows the component specifications for the main
crystal network and Table14 shows the component specifications for the 32.768 kHz crystal network.
The crystal oscillator you choose may require external components different from the ones specified above. In that case,
consult with National’s engineer for the component specifications
39
www.national.com
X1CKI / X2CKI
XTAL
C1
R1
X1CKO / X2CKO
C2
R2
Figure 8. External Crystal Network
Table 13
Component
Oscillator
Crystal
Component Values of the High Frequency Crystal Circuit
Parameters
Resonance Frequency
Type
Max. Serial Resistance
Max. Shunt Capacitance
Load Capacitance
Resistor R1
Values
Values
Values
Values
Values
Tolerance
4 MHz
AT-Cut
75 Ω
4 pF
12 pF
12 MHz
AT-Cut
35 Ω
4 pF
15 pF
16 MHz
AT-Cut
35 Ω
4 pF
15 pF
20 MHz
AT-Cut
35 Ω
4 pF
20 pF
24 MHz
AT-Cut
35 Ω
4 pF
20 pF
N/A
1 MΩ
1 MΩ
1 MΩ
1 MΩ
1 MΩ
5%
0Ω
0Ω
0Ω
0Ω
0Ω
5%
22 pF
20 pF
20 pF
20 pF
20 pF
20%
Resistor R2
Capacitor C1, C2
Table 14 Component Values of the Low Frequency Crystal Circuit
Component
Oscillator
Parameters
Values
Tolerance
32.768kHz
Parallel
N-Cut or XY-bar
40 kΩ
2 pF
9-13 pF
N/A
Resistor R1
10-20 MΩ
5%
Resistor R2
4.7 kΩ
5%
Capacitor C1, C2
20 pF
20%
Resonance Frequency
Type
Maximum Serial Resistance
Maximum Shunt Capacitance
Load Capacitance
Crystal
Choose capacitor component values in the tables obtain the
specified load capacitance for the crystal when combined
with the parasitic capacitance of the trace, socket, and package (which can vary from 0 to 8 pF). As a guideline, the load
capacitance is:
This signal is an indicator that the main clock oscillator is stable.
The “Stop Main Osc” signal from the Power Management
Module stops and starts the main oscillator. When this signal
is asserted, it presets the 14-bit timer to 3FFF hex and stops
the main oscillator. When the signal goes inactive, the main
oscillator starts and the 14-bit timer counts down from its preset value. When the timer reaches zero, it stops counting and
asserts the “Good Main Clk” signal.
CL = (C1 * C 2)/(C1+C2) + C parasitic
C 2 > C1
C1 can be trimmed to obtain the desired load capacitance.
The start-up time of the 32.768 kHz oscillator can vary from
one to six seconds. The long start-up time is due to the high
“Q” value and high serial resistance of the crystal necessary
to minimize power consumption in Power Save mode.
12.2
12.3
MAIN SYSTEM CLOCK
The slow clock operates in a manner similar to the main
clock. The “Stop Slow Osc” signal from the Power Management Module stops and starts the slow oscillator. When this
signal is asserted, it presets a 6-bit timer to 3F hex and disables the slow oscillator. When the signal goes inactive, the
slow oscillator starts and the 6-bit timer counts down from its
preset value. When the timer reaches zero, it stops counting
The main system clock is generated by the main oscillator. It
can be stopped by the Power Management Module to reduce
power consumption during periods of reduced activity. When
the main clock is restarted, a 14-bit timer generates a “Good
Main Clk” signal after a start-up delay of 32,768 clock cycles.
www.national.com
SLOW SYSTEM CLOCK
The slow (32.768 kHz) clock is necessary for operating the
device in Power Save modes and to provide a clock source
for modules such as the Timing and Watchdog Module.
40
and asserts the “Good Low Speed Clk” signal, thus indicating
that the slow clock is stable.
For systems that do not require a reduced power consumption mode, the external crystal network may be omitted for
the slow clock. In that case, the slow clock can be created by
dividing the main clock by a prescaler factor. The prescaler
circuit consists of a fixed divide-by-2 counter and a programmable 8-bit prescaler register. This allows a choice of clock
divisors ranging from 2 to 512. The resulting slow clock frequency must not exceed 100 kHz.
POR
12.7
A software-programmable multiplexer selects either the
prescaled main clock or the 32.768 kHz oscillator as the slow
clock. Upon reset, the prescaled main clock is selected, ensuring that the slow clock is always present initially. Selection
of the 32.768 kHz oscillator as the slow clock disables the
clock prescaler, which allows the CLK1 oscillator to be turned
off during power-save operation, thus reducing power consumption and radiated emissions. This can be done only if
the module detects a togging low-speed oscillator. If the lowspeed oscillator is not operating, the prescaler remains available as the slow clock source.
12.4
7
SCDIV
POWER-ON RESET
12.8
6
5
4
3
SCDIV
2
1
0
Slow Clock Divisor. If the clock divider is enabled (CRCTRL.SCLK=0), the main clock is divided by (SCDIV+1)*2 to produce the slow
system clock. Upon reset, PRSSC register is
set to FF hex.
SLOW CLOCK PRESCALER 1 REGISTER
(PRSSC1)
The Slow Clock Prescaler 1 (PRSSC1) register is a bytewide read/write register that holds the clock divisor used to
generate the two additional slow clocks from the high-speed
clock. Upon reset, the register is set to 00. The format of the
register is shown below.
7
The circuit sets a power-on reset flag bit upon detection of a
power-on condition. The CPU can read this flag to determine
whether a reset was caused by a power-up or by the RESET
input.
4
SCDIV2
SCDIV1
Note: Power-On Reset circuit cannot be used to detect a
drop in the supply voltage.
SCDIV1
EXTERNAL RESET
An active-low reset input pin called RESET allows the device
to be reset at any time. When the signal goes low, it generates an internal system reset signal that remains active until
the RESET signal goes high again.
12.6
SLOW CLOCK PRESCALER REGISTER
(PRSSC)
The Slow Clock Prescaler (PRSSC) register is a byte-wide
read/write register that holds the clock divisor used to generate the slow clock from the main clock. The format of the register is shown below.
The Power-On Reset circuit generates a system reset signal
upon power-up and holds the signal active for a period of
time to allow the crystal oscillator to stabilize. The circuit detects a power turn-on condition, which presets the 14-bit timer to 3FFF hex. Once oscillation starts and the clock
becomes active, the timer starts counting down. When the
count reaches zero, the 14-bit timer stops counting and the
internal reset signal is deactivated (unless the RESET pin is
held low).
12.5
When this bit is cleared to 0, the prescaled
main clock is used for the slow clock. Upon reset, this bit is cleared to 0.
Power-On Reset. This bit is set to 1 by the
hardware when a power-on condition is detected, allowing the CPU to determine whether a
power-up has occurred. The CPU can clear
this bit to 0 but cannot set it to 1. Any attempt
by the CPU to set this bit is ignored.
3
0
SCDIV1
Slow Clock Divisor 1. The main clock is divided
by (SCDIV1+1) to obtain the first slow system
clock.
Slow Clock Divisor 2. The main clock is divided
by (SCDIV2+1) to obtain the second slow system clock.
DUAL CLOCK AND RESET REGISTERS
The Dual Clock and Reset module (CLK2RES) contains two
registers: the Clock and Reset Control register (CRCTRL)
and the Slow Clock Prescaler register (PRSSC).
12.6.1
Clock and Reset Control Register (CRCTRL)
Clock and Reset Control Register (CRCTRL) is a byte-wide
read/write register that contains the power-on reset flag and
selects the type of slow clock. The register format is shown
below.
7
SCLK
6
5
4
Reserved
3
2
1
POR
0
SCLK
Slow Clock Select. When this bit is set to 1, the
32.728 kHz oscillator is used for the slow clock.
41
www.national.com
13.0 Multi-Input Wake-Up
The Multi-Input Wake-Up (MIWU16) module monitors its 16
input channels for a software-selectable trigger condition.
Upon detection of a trigger condition, the module generates
an interrupt request and if enabled, a wake-up request. A
wake-up request can be used by the power management unit
to exit the Halt, Idle, or Power Save mode and return to the
active mode. An interrupt request generates an interrupt to
the CPU (interrupt IRQ2), allowing interrupt processing in response to external events.
13.1
The Wake-Up Edge Detection (WKEDG) register is a wordwide read/write register that controls the edge sensitivity of
the Multi-Input Wake-Up pins. Register bits 0 through 15 control input pins WUI0 through WUI15, respectively. A bit
cleared to 0 configures the corresponding input to trigger on
a rising edge (a low-to-high transition). A bit set to 1 configures the corresponding input to trigger on a falling edge (a
high-to-low transition).
The wake-up event only activates the clocks and CPU, but
does not by itself initiate execution of any code. It is the interrupt request associated with the MIWU16 that gets the CPU
to start executing code, by jumping to the proper interrupt
routine. Therefore, setting up the MIWU16 interrupt handler
is essential for any wake-up operation.
This register is cleared upon reset, which configures all 16 inputs to be triggered on rising edges.
The register format is shown below.
15
There are four interrupt requests that can be routed to the
ICU as shown in Figure9. Each of the 16 MIWU channels
can be programmed to activate one of these four interrupt requests.
13.2
WAKE-UP ENABLE REGISTER (WKENA)
The Wake-Up Enable (WKENA) register is a word-wide read/
write register that enables or disables each of the Multi-Input
Wake-Up channels. Register bits 0 through 15 control channels WUI0 through WUI15, respectively. A bit cleared to 0
disables the wake-up function and a bit set to 1 enables the
function.
PL0
PL1
PL2
PL3
PH0
PH1
PH2
PH3
TWM-T0OUT
ACCESS.bus
Canards
MWCS
RDX1
RDX2
Comparator 1
Comparator 2
This register is cleared upon reset, which disables all eight
wake-up/interrupt channels.
Each input can be configured to trigger on rising or falling
edges, as determined by the setting in the WKEDG register.
Each trigger event is latched into the WKPND register. If a
trigger event is enabled by its respective bit in the WKENA
register, an active wake-up/interrupt signal is generated. The
software can determine which channel has generated the active signal by reading the WKPND register.
The Multi-Input Wake-Up module is active at all times, including the Halt mode. All device clocks are stopped in this mode.
Therefore, detecting an external trigger condition and the
subsequent setting of the pending flag are not synchronous
to the system clock.
www.national.com
0
WKED15-WKED0
The input pins for the Multi-Input Wake-Up channels are
named WUI0 through WUI15.
WUI0
WUI1
WUI2
WUI3
WUI4
WUI5
WUI6
WUI7
WUI8
WUI9
WUI10
WUI11
WUI12
WUI13
WUI14
WUI15
WAKE-UP EDGE DETECTION REGISTER
(WKEDG)
42
Peripheral Bus
..........
15
0
WKICTL1-2
WKENA
WUI0
0
WUI15
4
Wake-Up Signal
To Power Mgt
15
WKEDG
EXINT3:0 to ICU
WKPND
Figure 9. Multi-Input Wake-Up Module Block Diagram
The register format is shown below.
15
quests outputs to the ICU31L are to be activated for the corresponding channel.
0
WKEN15-WKEN0
13.3
WAKE-UP INTERRUPT CONTROL
REGISTER 1 (WKCTL1)
The Wake-Up Interrupt Control Register 1 (WKICTL1) register is a word-wide read/write register that selects the interrupt
request signal for the associated channels WUI0 to WUI7.
Upon reset, WKICTL1 is set to 0, which selects MIWU Interrupt Request 0 for all eight channels. The register format is
shown below.
15
14 13
12 11
10
9
8
7
6
5
4 3
WKINTR WKINTR WKINTR WKINTR WKINTR WKINTR
7
6
5
4
3
2
2
WKINTR
1
1
enables MIWU Interrupt Request 0
01
enables MIWU Interrupt Request 1
10
enables MIWU Interrupt Request 2
11
enables MIWU Interrupt Request 3
13.5
WAKE-UP PENDING REGISTER (WKPND)
The Wake-Up Pending (WKPND) register is a word-wide
read/write register in which the Multi-Input Wake-Up module
latches any detected trigger conditions. Register bits 0
through 15 serve as latches for channels WUI0 through
WUI15, respectively. A bit cleared to 0 indicates that no trigger condition has occurred. A bit set to 1 indicates that a trigger condition has occurred and is pending on the
corresponding channel. This register is cleared upon reset.
0
WKINTR
0
WKINTR0:7 Wake-Up Interrupt Request Select. Each field
selects which of the following four interrupt requests outputs to the ICU31L are to be activated for the corresponding channel.
00
enables MIWU Interrupt Request 0
01
enables MIWU Interrupt Request 1
The CPU can only write a 1 to any bit position in this register.
If the CPU attempts to write a 0, it has no effect on that bit.
To clear a bit in this register, the CPU must use the WKPCL
register (described below). This implementation prevents a
potential hardware-software conflict during a read-modifywrite operation on the WKPND register.
10
enables MIWU Interrupt Request 2
The register format is shown below.
11
enables MIWU Interrupt Request 3
13.4
15
WAKE-UP INTERRUPT CONTROL
REGISTER 1 (WKCTL2)
14 13
12 11
10
9
8
7
6
5
4 3
WKINTR WKINTR WKINTR WKINTR WKINTR WKINTR
15
14
13
12
11
10
0
WKPD15-WKPD0
13.6
The Wake-Up Interrupt Control Register 2 (WKICTL2) register is a word-wide read/write register that selects the interrupt
request signal for the associated channels WUI8 to WUI15.
Upon reset, WKICTL2 is set to 0, which selects MIWU Interrupt Request 0 for all eight channels. The register format is
shown below.
15
00
2
WKINTR
9
1
WAKE-UP PENDING CLEAR REGISTER
(WKPCL)
The Wake-Up Pending Clear (WKPCL) register is a wordwide write-only register that lets the CPU clear bits in the WKPND register. Writing a 1 to a bit position in the WKPCL register clears the corresponding bit in the WKPND register.
Writing a 0 leaves the corresponding bit in the WKPND register unchanged.
0
WKINTR
8
Reading this register location returns unknown data. Therefore, do not use a read-modify-write sequence to set the individual bits. In other words, do not attempt to read the
WKINTR8:5 Wake-Up Interrupt Request Select. Each field
selects which of the following four interrupt re-
43
www.national.com
register and do a logical OR with the register value. Instead,
just write the mask directly to the register address.
The register format is shown below.
15
0
WKCL15-WKCL0
13.7
PROGRAMMING PROCEDURES
To set up and use the Multi-Input Wake-Up function, use the
following procedure. Performing the steps in the order shown
will prevent false triggering of a wake-up condition. This
same procedure should be used following a reset because
the wake-up inputs are left floating, resulting in unknown data
on the input pins.
1. Clear the WKENA register to disable the wake-up channels.
2. If the input originates from an I/O port (the usual case),
set the corresponding bit in the port direction register to
configure the I/O pin to operate as an input.
3. Write the WKEDG register to select the desired type of
edge sensitivity (clear to 0 for rising edge, set to 1 for falling edge).
4. Set all bits in the WKPCL register to clear any pending
bits in the WKPND register.
5. Set up the WKICTL1 and WKICTL2 registers to define
the interrupt request signal used for each channel.
6. Set the bits in the WKENA register corresponding to the
wake-up channels to be activated.
To change the edge sensitivity of a wake-up channel, use the
following procedure. Performing the steps in the order shown
will prevent false triggering of a wake-up/interrupt condition.
1. Clear the WKENA bit associated with the input to be reprogrammed.
2. Write the new value to the corresponding bit position in
the WKEDG register to reprogram the edge sensitivity of
the input.
3. Set the corresponding bit in the WKPCL register to clear
the pending bit in the WKPND register.
4. Set the same WKENA bit to re-enable the wake-up function.
www.national.com
44
14.0 Real-Time Timer and WATCHDOG
The Timing and WATCHDOG Module (TWM) generates the
clocks and interrupts used for timing periodic functions in the
system, and also provides Watchdog protection against software errors. The module operates off the slow clock either
generated by the external 32kHz oscillator or from the prescaled high speed system clock. The maximum operating
clock frequency is 100kHz.
All counting activities of the module are based on the slow
clock (SLCLK). A prescaler counter divides this clock to
make a slower clock. The prescaler factor is defined by a 3bit field in the Timer and WATCHDOG Prescaler register,
which selects either 1, 2, 4, 8, 16, or 32 and the divide-by factor. Thus, the prescaled clock period can be set to 1, 2, 4, 8,
16, or 32 times the slow clock period. The prescaled clock
signal is called T0IN.
The WATCHDOG is designed to detect program execution
errors. Once WATCHDOG operation is initiated, the software
must periodically write a specific value to a WATCHDOG register. If the software fails to do so, a WATCHDOG error is triggered, which resets the device.
14.2
The TWM is flexible in allowing selection of a variety of clock
ratios and clock sources for the WATCHDOG circuit. Once
the software configures the TWM, it can lock the configuration for a higher level of protection against erroneous software action. Once locked, the TWM can be released only by
a device reset.
14.1
TIMER T0 OPERATION
Timer T0 is a programmable 16-bit down counter that can be
used as the time base for real-time operations such as a periodic audible tick. It can also be used to drive the WATCHDOG circuit.
The timer starts counting from the value loaded into the
TWMT0 register and counts down on each rising edge of
T0IN. When the timer reaches zero, it is automatically reloaded from the TWMT0 register and continues counting down
from that value. Thus, the frequency of the timer is:
TWM STRUCTURE
f SLCLK / [(TWMT0+1) * prescaler]
Figure10 is a block diagram showing the internal structure of
the Timing and WATCHDOG module. There are two main
sections: the Real-Time Timer (T0) section at the top and the
WATCHDOG section on the bottom.
When an external crystal oscillator is used as the SLCLK
source or when the fast clock is divided accordingly, f SLCLK
is 32.768 kHz.
The value stored in TWMT0 can range from 0001 hex to
FFFF hex.
Peripheral Bus
REAL TIME TIMER (T0)
5-bit pre-scaler counter
(TWCP)
CLKIN1
slow clock from
dual clock and
reset module
T0IN
TWMT0 register
T0CSR Contrl. Reg.
T0LINT
(to ICU)
Restart
16-bit Timer (Timer0)
Underflow
T0OUT
(to Multi-InputWake-Up)
WATCHDOG Timer
Underflow
Restart
WDSDM
WATCHDOG
Service
Logic
WDCNT
WATCHDOG ERROR
WDERR
WATCHDOG
Figure 10. Timing and WATCHDOG Module Block Diagram
45
www.national.com
When the counter reaches zero, an internal timer signal
called T0OUT is set to 1 for one T0IN clock cycle. This signal
sets the TC bit in the TWMT0 Control and Status Register
(T0CSR). It also generates an interrupt called RTI (IRQ14) if
the interrupt is enabled by the T0CSR.T0INTE bit.
locking state of the TWCFG, TWCPR, TIMER0, T0CSR, and
WDCNT registers. A register that is locked cannot be read or
written. A write operation is ignored and a read operation returns unpredictable results.
If the TWCFG register is itself locked, it remains locked until
the device is reset. Any other locked registers also remain
locked until the device is reset. This feature prevents a runaway program from tampering with the programmed
WATCHDOG function.
If the software loads TWMT0 with a new value, the timer uses
that value the next time that it reloads the 16-bit timer register
(in other words, after reaching zero). The software can restart
the timer at any time (on the very next edge of the T0IN clock)
by setting the Restart (RST) bit in the T0CSR register. The
T0CSR.RST bit is cleared automatically upon restart of the
16-bit timer.
14.3.2
The Timer and WATCHDOG Module is active in both the
Power Save and Idle modes. The clocks and counters continue to operate normally in these modes. The WDSDM register is accessible in the Power Save and Idle modes, but the
other TWM registers are accessible only in the Active mode.
Therefore, WATCHDOG servicing must be carried out using
the WDSDM register in the Power Save or Idle mode.
Note: If the user wishes to switch to power save or idle mode
after setting T0CSR.RST, the user must wait for reset operation to complete before doing the switch.
14.3
WATCHDOG OPERATION
The WATCHDOG is an 8-bit down counter that operates on
the rising edge of a specified clock source. Upon reset, the
WATCHDOG is disabled; it does not count and no WATCHDOG signal is generated. A write to either the WATCHDOG
Count (WDCNT) register or the WATCHDOG Service Data
Match (WDSDM) register starts the counter. The WATCHDOG counter counts down from the value programmed in to
the WDCNT register. Once started, only a reset can stop the
WATCHDOG from operating.
In the Halt mode, the entire device is frozen, including the
Timer and WATCHDOG Module. Upon return to the Active
mode, operation of the module resumes at the point at which
it was stopped.
Note: After a restart or WATCHDOG service through WDCNT, do not enter Power Save mode for a period equivalent
to 5 slow clock cycles.
14.4
The WATCHDOG can be programmed to use either T0OUT
or T0IN as its clock source (the output and input of Timer T0,
respectively). The TWCFG.WDCT0I bit controls this clock
selection.
— Timer and WATCHDOG Configuration Register
(TWCFG)
— Timer and WATCHDOG Clock Prescaler Register
(TWCP)
— TWM Timer 0 Register (TWMT0)
— TWMT0 Control and Status Register (T0CSR)
— WATCHDOG Count Register (WDCNT)
— WATCHDOG Service Data Match Register (WDSDM)
If TWCFG.WDSDME bit is cleared to 0, the WATCHDOG is
serviced by writing a value to the WDCNT register. The value
written to the register is reloaded into the WATCHDOG
counter. The counter then continues counting down from that
value.
The WDSDM register is accessible in both Active and Power
Save mode. The other TWM registers are accessible only in
Active mode.
If TWCFG.WDSDME bit is set to 1, the WATCHDOG is serviced by writing the value 5C hex to the WATCHDOG Service
Data Match (WDSDM) register. This reloads the WATCHDOG counter with the value previously programmed into the
WDCNT register. The counter then continues counting down
from that value.
14.4.1
Timer and WATCHDOG Configuration Register
(TWCFG)
The TWCFG register is a byte-wide, read/write register that
selects the WATCHDOG clock input and service method,
and also allows the WATCHDOG registers to be selectively
locked. Once a bit is set, that bit cannot be cleared until the
device resets. Upon reset, the non-reserved bits of the register are all cleared to 0. The register format is shown below.
A WATCHDOG error signal is generated by any of the following events:
— The WATCHDOG serviced too late.
— The WATCHDOG serviced too often.
— The WDSDM register is written with a value other than
5C hex when WDSDM type servicing is enabled
(TWCFG.WDSDME=1).
7
6
5
4
3
2
1
0
Reserved WDSDME WDCT0I LWDCNT LTWMT0 LTWCP LTWCFG
LTWCFG
A WATCHDOG error condition resets the device.
Register Locking
The Timer and WATCHDOG Configuration (TWCFG) register is used to set the WATCHDOG configuration. It controls
the WATCHDOG clock source (T0IN or T0OUT), the type of
WATCHDOG servicing (using WDCNT or WDSDM), and the
www.national.com
TWM REGISTERS
The TWM registers controls the operation of the Timing and
WATCHDOG Module. There are six such registers:
The software must periodically “service” the WATCHDOG.
There are two ways to service the WATCHDOG, the choice
depending on the programmed value of the WDSDME bit in
the Timer and WATCHDOG Configuration (TWCFG) register.
14.3.1
Power Save Mode Operation
46
Lock TWCFG Register. When cleared to 0, access to the TWCFG register is allowed. When
set to 1, the TWCFG register is locked. A
locked register cannot be read or written; a
read operation returns unpredictable values
and a write operation is ignored. Locking the
TWCFG register remains in effect until the device is reset.
LTWCP
Lock TWCP Register. When cleared to 0, access to the TWCP register is allowed. When
set to 1, the TWCP register is locked.
Lock TWMT0 Register. When cleared to 0, access to the TWMT0 and T0CSR registers are
allowed. When set to 1, the TWMT0 and
T0CSR registers are locked.
Lock LDWCNT Register. When cleared to 0,
access to the LDWCNT register is allowed.
When set to 1, the LDWCNT register is locked.
WATCHDOG Clock from T0IN. When cleared
to 0, the T0OUT signal (the output of Timer T0)
is used as the WATCHDOG clock. When set to
1, the T0IN signal (the prescaled slow clock) is
used as the WATCHDOG clock.
WATCHDOG Service Data Match Enable.
When cleared to 0, WATCHDOG servicing is
accomplished by writing a count value to the
WDCNT register; write operations to the
WATCHDOG Service Data Match (WDSDM)
register are ignored. When set to 1, WATCHDOG servicing is accomplished by writing the
value 5C hex to the WDSDM register.
LTWMT0
LWDCNT
WDCT0I
WDSDME
14.4.2
T0IN divided by (PRESET+1). The allowed values of PRESET are 0001 hex through FFFF
hex.
14.4.4
The T0CSR register is a byte-wide, read/write register that
controls Timer T0 and shows its current status. Upon reset,
the non-reserved bits of the register are cleared to 0. The
register format is shown below.
7
RST
TC
Timer and WATCHDOG Clock Prescaler
Register (TWCP)
T0INTE
The TWCP register is a byte-wide, read/write register that
defines the prescaler value used for dividing the low frequency clock to generate the T0IN clock. Upon reset, the non-reserved bits of the register are cleared to 0. The register
format is shown below.
7
MDIV
14.4.3
6
5
4
Reserved
3
2
1
MDIV
14.4.5
0
14.4.6
TWM Timer 0 Register (TWMT0)
PRESET
5
4
3
2
1
5
4
Reserved
3
2
T0INTE
1
TC
0
RST
Restart. When this bit is set to 1, it forces the
timer to reload the value in the TWMT0 register
on the next rising edge of the selected input
clock. The RST bit is reset automatically by the
hardware on the same rising edge of the selected input clock. Writing a 0 to this bit position
has no effect. Upon reset, the non-reserved
bits of the register are cleared to 0.
Terminal Count. This bit is set to 1 by the hardware when the Timer T0 count reaches zero
and is cleared to 0 when the software reads the
T0CSR register. It is a read-only bit. Any data
written to this bit position is ignored.
Timer T0 Interrupt Enable. When this bit is set
to 1, it enables an interrupt to the CPU each
time the Timer T0 count reaches zero. When
this bit is cleared to 0, Timer T0 interrupts are
disabled.
WATCHDOG Count Register (WDCNT)
WATCHDOG Service Data Match Register
(WDSDM)
The WSDSM register is a byte-wide, write-only register used
for servicing the WATCHDOG. When this type of servicing is
enabled (TWCFG.WDSDME=1), the WATCHDOG is serviced by writing the value 5C hex to the WSDSM register.
Each such servicing reloads the WATCHDOG counter with
the value previously written to the WDCNT register. Writing
any data other than 5C hex triggers a WATCHDOG error.
Writing to the register more than once in one WATCHDOG
clock cycle also triggers a WATCHDOG error signal. If this
type of servicing is disabled (TWCFG.WDSDME=0), any
write to the WSDSM register is ignored.
The TWMT0 register is a word-wide, read/write register that
defines the T0OUT interrupt rate. Upon reset, TWMT0 register is initialized to FFFF hex. The register format is shown below.
8 7 6
PRESET
6
The WDCNT register is a byte-wide, write-only register that
holds the value that is loaded into the WATCHDOG counter
each time the WATCHDOG is serviced. The WATCHDOG is
started by the first write to this register. Each successive write
to this register restarts the WATCHDOG count with the written value. Upon reset, this register is initialized to 0F hex.
Main Clock Divide. This 3-bit field defines the
prescaler factor used for dividing the low speed
device clock to create the T0IN clock. The allowed 3-bit values and the corresponding clock
divisors and clock rates are listed below.
MDIV
Clock Divisor TOIN Frequency
(fSCLK=32.768 kHz)
000
1
32.768 kHz
001
2
16.384 kHz
010
4
8.192 kHz
011
8
4.096 kHz
100
16
2.056 kHz
101
32
1.024 kHz
other
Reserved
N/A
15 14 13 12 11 10 9
TWMT0 Control and Status Register (T0CSR)
14.5
0
WATCHDOG PROGRAMMING
PROCEDURE
The highest level of protection against software errors is
achieved by programming and then locking the WATCHDOG
registers and using the WDSDM register for servicing. This is
the procedure:
Timer T0 Preset. Timer T0 is reloaded with this
value on each underflow. Thus, the frequency
of the Timer T0 interrupt is the frequency of
47
www.national.com
1. Write the desired values into the TWM Clock Prescaler
register (TWCP) and the TWM Timer 0 register
(TWMT0) to control the T0IN and T0OUT clock rates.
The frequency of T0IN can be programmed to any of six
frequencies ranging from 1/32*fSLCLK to fSLCLK . The frequency of T0OUT is equal to the frequency of T0IN divided by (1+PRESET), where PRESET is the value
written to the TWMT0 register.
2. Configure the WATCHDOG clock to use either T0IN or
T0OUT by setting or clearing the TWCFG.WDCT0I bit.
3. Write the initial value into the WDCNT register. This
starts operation of the WATCHDOG and specifies the
maximum allowed number of WATCHDOG clock cycles
between service operations.
4. Lock the WATCHDOG registers and enable the
WATCHDOG Service Data Match Enable function by
setting bits 0, 1, 2, 3, and 5 in the TWCFG register.
5. Service the WATCHDOG by periodically writing the value 5C hex to the WDSDM register at an appropriate rate.
Servicing must occur at least once per period programmed into the WDCNT register, but no more than
once in a single WATCHDOG input clock cycle.
www.national.com
48
15.0 Multi-Function Timer
The Multi-Function Timer (MFT16) module contains two independent timer/counter units called MFT1 and MFT2, each
containing a pair of 16-bit timer/counters. Each timer/counter
unit offers a choice of clock sources for operation and can be
configured to operate in any of the following modes:
The two timer units, MFT1 and MFT2, are identical in operation and separately programmable. Each timer unit uses two
I/O pins, called T1A and T1B (for Timer MFT1) or T2A and
T2B (for Timer MFT2). The timer I/O pins are alternate functions of the Port F I/O pins.
• Processor-Independent Pulse Width Modulation (PWM)
mode, which generates pulses of a specified width and
duty cycle, and which also provides a general-purpose
timer/counter
• Dual Input Capture mode, which measures the elapsed
time between occurrences of external events, and which
also provides a general-purpose timer/counter
• Dual Independent Timer mode, which generates system
timing signals or counts occurrences of external events
• Single Input Capture and Single Timer mode, which provides one external event counter and one system timer
In the description of the timers, the lower-case letter “n” represents the timer number, either 1 or 2. For example, “TnA”
means I/O pin T1A or T2A.
15.1
Figure11 is a block diagram showing the internal structure of
each timer. There are two main functional blocks: a Timer/
Counter and Action block and a Clock Source block. The
Timer/Counter and Action block contains two separate timer/
counter units, called Timer/Counter I and Timer/Counter II (a
total of four timer/counter unit in both MFT1 and MFT2).
Timer/Counter
Clock Source
TIMER STRUCTURE
Action
Toggle/Capture/Interrupt
System
Clock
Clock Prescaler/Selector
Reload/Capture
A
Timer/Counter
1
Reload/Capture
B
Timer/Counter
2
External Event
TnA
Interrupt A
Interrupt B
TnB
PWM/Capture/Counter
Mode Select + Control
Figure 11. Multi-Function Timer Block Diagram
15.1.1
Timer/Counter Block
Counter Clock Source Select
The Timer/Counter block contains the following functional
blocks:
There are two clock source selectors that allow the software
to independently select the clock source for each of the two
16-bit counters from any one of the following sources:
— two 16-bit counters, Timer/Counter I (TnCNT1) and
Timer/Counter II (TnCNT2)
— two 16-bit reload/capture registers, TnCRA and
TnCRB
— control logic necessary to configure the timer to operate in any of the four operating modes
— interrupt control and I/O control logic
—
—
—
—
—
In a power-saving mode that uses the low-frequency (32.768
kHz) clock as the system clock, the synchronization circuit
requires that the slow clock operate at no more than onefourth the speed of the 32.768 kHz system clock.
15.1.2
no clock (which stops the counter)
prescaled system clock
external event count based on TnB
pulse accumulate mode based on TnB
slow clock (derived from the low-frequency oscillator or
divided from the high-speed oscillator)
Prescaler
The 5-bit clock prescaler allows the software to run the timer
with a prescaled clock signal. The prescaler consists of a 5bit read/write prescaler register (TnPRSC) and a 5-bit down
counter. The system clock is divided by the value contained
in the prescaler register plus 1. Thus, the timer clock period
can be set to any value from 1 to 32 divisions of the system
clock period. The prescaler register and down counter are
both cleared upon reset.
Clock Source Block
The Clock Source block generates the signals used to clock
the two timer/counter registers. The internal structure of the
Clock Source block is shown in Figure12.
49
www.national.com
Prescaler Register
TnPRSC
System
Clock
TnB
Prescaled
Clock
5-bit
Reset
No Clock
Counter I
Clock
Select
Counter I
Clock
Prescaler Counter
Pulse
Accumulate
Counter II
Clock
Select
External
Event
Synchr.
Counter II
Clock
Figure 12. Clock Source Block Diagram
External Event Clock
Pulse Accumulate Mode
The TnB I/O pin can be configured to operate as an external
event input clock for either of the two 16-bit counters. This input can be programmed to detect either rising or falling edges. The minimum pulse width of the external signal is one
system clock cycle. This means that the maximum frequency
at which the counter can run in this mode is one-half of the
system clock frequency. This clock source is not available in
the capture modes (modes 2 and 4) because the TnB pin is
used as one of the two capture inputs.
The counter can also be configured to count prescaler output
clock pulses when the TnB is high and not count when TnB
is low, as illustrated in Figure13. The resulting count is an indicator of the cumulative time that TnB is high. This is called
the “pulse accumulate” mode. In this mode, an AND gate
generates a clock signal for the counter whenever a prescaler clock pulse is generated and TnB input is high. (The polarity of the TnB signal is programmable, so the counter can
count when TnB is low rather than high.) The pulse accumulate mode is not available in the capture modes (modes 2 and
4) because the TnB pin is used as one of the two capture inputs.
Prescaler Output
TnB
Counter Clock
Figure 13. Pulse Accumulate Mode Operation
Slow Clock
clock to input clock ratio needed for the synchronization cannot be maintained. However, the External Event Clock and
Pulse Accumulate Mode will still work, as long as the external
event pulses are at least the size of the whole slow-clock period. Using the prescaled system clock will also work, but at
a much slower rate than the original system clock.
The slow clock is generated by the Dual Clock and Reset
(CLK2RES) module. The clock source is either the divided
fast clock or the external 32.768 kHz clock crystal (if available
and selected). The slow clock can be used as the clock
source for the two 16-bit counters. Because the slow clock
can be asynchronous to the system clock, a circuit is provided to synchronize the clock signal to the high-frequency system clock before it is used for clocking the counters. The
synchronization circuit requires that the slow clock operate at
no more than one-fourth the speed of the system clock.
Some Power Save modes stops the system clock (the highfrequency and/or low-frequency clock) completely. If the system clock is stopped, the timer stops counting until the system clock resumes operation.
In the Idle or Halt mode, the system clock stops completely,
which stops the operation of the timers. In that case, the timers stop counting until the system clock resumes operation.
Limitations in Low-Power Modes
The Power Save mode uses the low-frequency clock as the
system clock. In this mode, the slow clock cannot be used as
a clock source for the timers because both CLK and SLCLK
are driven then at the same frequency, and the 2:1 systemwww.national.com
50
15.2
15.2.1
TIMER OPERATING MODES
Mode 1 is the Processor-Independent Pulse Width Modulation (PWM) mode, which generates pulses of a specified
width and duty cycle, and which also provides a separate
general-purpose timer/counter.
Each timer/counter unit can be configured to operate in any
of the following modes:
— Processor-Independent Pulse Width Modulation
(PWM) mode
— Dual Input Capture mode
— Dual Independent Timer mode
— Single Input Capture and Single Timer mode
Figure14 is a block diagram of the Multi-Function Timer configured to operate in Mode 1. Timer/Counter I (TnCNT1)
functions as the time base for the PWM timer. It counts down
at the clock rate selected for the counter. When an underflow
occurs, the timer register is reloaded alternately from the
TnCRA and TnCRB register, and counting proceeds downward from the loaded value.
Upon reset, the timers are disabled. To configure and start
the timers, the software must write a set of values to the registers that control the timers. The registers are described in
Section15.5.
Reload A = Time 1
TnCRA
TnAPND
Underflow
Timer I
Clock
Mode 1: Processor-Independent PWM
Timer/Counter I
TnCNT1
TnAIEN
TnA
TnAEN
Underflow
TnBIEN
Reload B = Time 2
TnCRB
Timer II
Clock
Timer
Interrupt A
Timer
Interrupt B
TnBPND
Timer/Counter II
TnCNT2
TnDIEN
Timer
Interrupt D
TnDPND
Clock
Selector
TnB
Figure 14. Mode 1: Processor-Independent PWM Block Diagram
On the first underflow, the timer is loaded from TnCRA, then
from TnCRB on the next underflow, then from TnCRA again
on the next underflow, and so on. Every time the counter is
stopped and restarted, it always obtains its first reload value
from TnCRA. This is true whether the timer is restarted upon
reset, after entering Mode 1 from another mode, or after
stopping and restarting the clock with the Timer/Counter I
clock selector.
determine the cause of each interrupt by looking at the
TnAPND and TnBPND flags, which are set by the hardware
upon each occurrence of a timer reload.
In Mode 1, Timer/Counter II (TnCNT2) can be used either as
a simple system timer, an external event counter, or a pulse
accumulate counter. The clock counts down using the clock
selected with the Timer/Counter II clock selector. It generates
an interrupt upon each underflow if the interrupt is enabled
with the TnDIEN bit.
The timer can be configured to toggle the TnA output bit upon
each underflow. This generates a clock signal on TnA with
the width and duty cycle determined by the values stored in
the TnCRA and TnCRB registers. This is a “processor-independent” PWM clock because once the timer is set up, no
more action is required from the CPU to generate a continuous PWM signal.
15.2.2
Mode 2: Dual Input Capture
Mode 2 is the Dual Input Capture mode, which measures the
elapsed time between occurrences of external events, and
which also provides a separate general-purpose timer/
counter.
Figure15 is a block diagram of the Multi-Function Timer configured to operate in Mode 2. The time base of the capture
timer depends on Timer/Counter I, which counts down using
the clock selected with the Timer/Counter I clock selector.
The timer can be configured to generate separate interrupts
upon reload from TnCRA and TnCRB. The interrupts can be
enabled or disabled under software control. The CPU can
51
www.national.com
The TnA and TnB pins function as capture inputs. A transition
received on the TnA pin transfers the timer contents to the
TnCRA register. Similarly, a transition received on the TnB
pin transfers the timer contents to the TnCRB register. Each
input pin can be configured to sense either rising or falling
edges.
TnAIEN
Timer
Interrupt I
TnAPND
Capture A
TnCRA
TnA
Preset
TnAEN
Timer I
Clock
Timer/Counter I
TnCNT1
TnCPND
Underflow
TnCIEN
Preset
Timer
Interrupt I
TnBEN
Capture B
TnCRB
TnB
TnBPND
TnBIEN
Timer II
Clock
Timer/Counter II
TnCNT2
TnDPND
Underflow
TnDIEN
Figure 15. Mode 2: Dual Input Capture Block Diagram
The TnA and TnB inputs can be configured to preset the
counter to FFFF hex upon reception of a valid capture event.
In this case, the current value of the counter is transferred to
the corresponding capture register and then the counter is
preset to FFFF hex. Using this approach allows the software
to determine the on-time and off-time and period of an external signal with a minimum of CPU overhead.
Timer
Interrupt II
an interrupt upon each underflow if the interrupt is enabled
with the TnDIEN bit.
Neither Timer/Counter I (TnCNT1) nor Timer/Counter II
(TnCNT2) can be configured to operate as an external event
counter or to operate in the pulse accumulate mode because
the TnB input is used as a capture input. Attempting to select
one of these configurations will cause one or both counters
to stop.
The values captured in the TnCRA register at different times
reflect the elapsed time between transitions on the TnA pin.
The same is true for the TnCRB register and the TnB pin. The
input signal on TnA or TnB must have a pulse width equal to
or greater than one system clock cycle.
15.2.3
Mode 3: Dual Independent Timer/Counter
Mode 3 is the Dual Independent Timer mode, which generates system timing signals or counts occurrences of external
events.
There are three separate interrupts associated with the capture timer, each with its own enable bit and pending flag. The
three interrupt events are reception of a transition on TnA, reception of a transition on TnB, and underflow of the TnCNT1
counter. The enable bits for these events are TnAIEN, TnBIEN, and TnCIEN, respectively.
Figure16 is a block diagram of the Multi-Function Timer configured to operate in Mode 3. The timer is configured to operate as a dual independent system timer or dual external
event counter. In addition, Timer/Counter I can generate a
50% duty cycle PWM signal on the TnA pin. The TnB pin can
be used as an external event input or pulse accumulate input
and can be used as the clock source for either Timer/Counter
I or Timer/Counter II. Both counters can also be clocked by
the prescaled system clock.
In Mode 2, Timer/Counter II (TnCNT2) can be used as a simple system timer. The clock counts down using the clock selected with the Timer/Counter II clock selector. It generates
www.national.com
Timer
Interrupt I
52
Reload A
TnCRA
TnAPND
Underflow
Timer I
Clock
Timer/Counter I
TnCNT1
TnAIEN
Timer
Interrupt I
TnA
TnAEN
Reload B
TnCRB
Underflow
Timer II
Clock
Timer/Counter II
TnCNT2
TnDIEN
TnDPND
Clock
Selector
TnB
Figure 16.
Mode 3: Dual Independent Timer/Counter Block Diagram
Timer/Counter I (TnCNT1) counts down at the rate of the selected clock. Upon underflow, it is reloaded from the TnCRA
register and counting proceeds down from the reloaded value. In addition, the TnA pin is toggled on each underflow if
this function is enabled by the TnAEN bit. The initial state of
the TnA pin is software-programmable. When the TnA pin is
toggled from low to high, it sets the TnCPND interrupt pending flag and also generates an interrupt if the interrupt is enabled by the TnAIEN bit.
Timer/Counter I (TnCNT1) operates the same as in Mode 3.
It counts down at the rate of the selected clock. Upon underflow, it is reloaded from the TnCRA register and counting proceeds down from the reloaded value. The TnA pin is toggled
on each underflow if this function is enabled by the TnAEN
bit. When the TnA pin is toggled from low to high, it sets the
TnCPND interrupt pending flag and also generates an interrupt if the interrupt is enabled by the TnAIEN bit. A 50% duty
cycle PWM signal can be generated on TnA without any further action from the CPU once the pulse train is initiated.
Because TnA toggles on every underflow, a 50% duty cycle
PWM signal can be generated on TnA without any further action from the CPU once the pulse train is initiated.
Timer/Counter II (TnCNT1) counts down at the rate of the selected clock. The TnB pin functions as the capture input. A
transition received on TnB transfers the timer contents to the
TnCRB register. The input pin can be configured to sense either rising or falling edges.
Timer/Counter II (TnCNT2) counts down at the rate of the selected clock. Upon underflow, it is reloaded from the TnCRB
register and counting proceeds down from the reloaded value. In addition, each underflow sets the TnDPND interrupt
pending flag and generates an interrupt if the interrupt is enabled by the TnDIEN bit.
15.2.4
Timer
Interrupt II
The TnB input can be configured to preset the counter to
FFFF hex upon reception of a valid capture event. In this
case, the current value of the counter is transferred to the
capture register and then the counter is preset to FFFF hex.
Mode 4: Input Capture Plus Timer
Mode 4 is the Single Input Capture and Single Timer mode,
which provides one external event counter and one system
timer.
The values captured in the TnCRB register at different times
reflect the elapsed time between transitions on the TnA pin.
The input signal on TnB must have a pulse width equal to or
greater than one system clock cycle.
Figure17 is a block diagram of the Multi-Function Timer configured to operate in Mode 4. This mode offers a combination
of Mode 3 and Mode 2 functions. Timer/Counter I is used as
a system timer as in Mode 3 and Timer/Counter II is used as
a capture timer as in Mode 2, but with a single input rather
than two inputs.
There are two separate interrupts associated with the capture timer, each with its own enable bit and pending flag. The
two interrupt events are reception of a transition on TnB and
underflow of the TnCNT2 counter. The enable bits for these
events are TnBIEN and TnDIEN, respectively.
Neither Timer/Counter I (TnCNT1) nor Timer/Counter II
(TnCNT2) can be configured to operate as an external event
53
www.national.com
Reload A
TnCRA
TnAPND
Underflow
Timer I
Clock
TnAIEN
Timer/Counter I
TnCNT1
Timer
Interrupt I
TnA
TnATEN
TnBIEN
Timer
Interrupt I
TnBPND
Capture B
TnCRB
TnB
Preset
TnBEN
TnDPND
Timer II
Clock
Timer/Counter II
TnCNT2
TnDIEN
Timer
Interrupt II
Figure 17. Mode 4: Input Capture Plus Timer Block Diagram
counter or to operate in the pulse accumulate mode because
the TnB input is used as a capture input. Attempting to select
one of these configurations will cause one or both counters
to stop. In this mode, Timer/Counter II must be enabled at all
times.
15.3
shows the functions of the pins in each operating mode, and
for each combination of enable bit settings.
When pin TnA is configured to operate as a PWM output
(TnAEN = 1), the state of the pin is toggled on each underflow
of the TnCNT1 counter. In this case, the initial value on the
pin is determined by the TnAOUT bit. For example, to start
with TnA high, the software should set the TnAOUT bit to 1
prior to enabling the timer clock. This option is available only
when the timer is configured to operate in Mode 1, 3, or 4 (in
other words, when TnCRA is not used in Capture mode).
TIMER INTERRUPTS
Each Multi-Function Timer unit has four interrupt sources,
designated A, B, C, and D. Interrupt sources A, B, and C are
mapped into a single system interrupt called Timer Interrupt
I, while interrupt source D is mapped into a system interrupt
called Timer Interrupt II. Each of the four interrupt sources
has its own enable bit and pending flag. The enable flags are
named TnAIEN, TnBIEN, TnCIEN, and TnDIEN. The pending flags are named TnAPND, TnBPND, TnCPND, and TnDPND.
For Multi-Function Timer unit MFT1, Timer Interrupts I and II
are system interrupts T1A and T1B (IRQ13 and IRQ12), respectively. For Multi-Function Timer unit MFT2, Timer Interrupts I and II are system interrupts T2A and T2B (IRQ11 and
IRQ10), respectively.
Table15 shows the events that trigger interrupts A, B, C, and
D in each of the four operating modes. Note that some interrupt sources are not used in some operating modes, as indicated by the notation “N/A” (Not Applicable) in the table.
15.4
TIMER I/O FUNCTIONS
Each Multi-Function Timer unit uses two I/O pins, called T1A
and T1B (for Timer MFT1) or T2A and T2B (for Timer MFT2).
The function of each pin depends on the timer operating
mode and the TnAEN and TnBEN enable bits. Table16
www.national.com
54
Table 15
Sys. Int.
Interrupt
pending
flag
Timer
TnAPND
Int. I
(TnA Int.) TnBPND
TnCPND
Timer
TnDPND
Int. II
(TnB Int.)
Timer Interrupts Overview
Mode 1
Mode 2
Mode 3
Mode 4
PWM + Counter
Dual Input Capture +
counter
Dual Counter
Single Capture +
counter
TnCNT1 reload from
TnCRA
Input capture on TnA
transition
TnCNT1 reload from
TnCRA
TnCNT1 reload from
TnCRA
TnCNT1 reload from
TnCRB
Input Capture on TnB
transition
N/A
Input Capture on TnB
transition
N/A
TnCNT1 underflow
N/A
N/A
TnCNT2 underflow
TnCNT2 underflow
TnCNT2 reload from
TnCRB
TnCNT2 underflow
Table 16
I/O
TnA
TnB
TnAEN
TnBEN
Timer I/O Functions
Mode 1
Mode 2
Mode 3
Mode 4
PWM + Counter
Dual Input Capture +
counter
Dual Counter
Single Capture +
counter
TnAEN=0
TnBEN=X
No Output
TnAEN=1
TnBEN=X
Capture TnCNT1 into
TnCRA
No Output toggle
No Output toggle
Toggle Output on
Capture TnCNT1 into
underflow of TnCNT1 TnCRA and preset
TnCNT1
Toggle Output on
underflow of TnCNT1
Toggle Output on
underflow of TnCNT1
TnAEN=X
TnBEN=0
Ext. Event or Pulse
Accumulate Input
Capture TnCNT1 into
TnCRB
Ext. Event or Pulse
Accumulate Input
Capture TnCNT2 into
TnCRB
TnAEN=X
TnBEN=1
Ext. Event or Pulse
Accumulate Input
Capture TnCNT1 into
TnCRB and preset
TnCNT1
Ext. Event or Pulse
Accumulate Input
Capture TnCNT2 into
TnCRB and preset
TnCNT2
55
www.national.com
15.5
* Operation of the slow clock is determined by the CRCTRL.SCLK control bit, as described in Section12.6.1.
TIMER REGISTERS
The following CPU-accessible registers are used to control
the Multi-Function Timers:
—
—
—
—
—
—
—
—
—
15.5.3
Clock Prescaler Register (TnPRSC)
Clock Unit Control Register (TnCKC)
Timer/Counter I Register (TnCNT1)
Timer/Counter II Register (TnCNT2)
Reload/Capture A Register (TnCRA)
Reload/Capture B Register (TnCRB)
Timer Mode Control Register (TnCTRL)
Timer Interrupt Control Register (TnICTL)
Timer Interrupt Clear Register (TnICLR)
15.5.1
The Timer/Counter I (TnCNT1) register is a word-wide, read/
write register that holds the current count value for Timer/
Counter I. The register contents are not affected by a reset
and are unknown upon power-up.
15.5.4
Clock Prescaler Register (TnPRSC)
6
5
Reserved
CLKPS
15.5.2
4
3
2
CLKPS
15.5.5
1
0
15.5.6
15.5.7
C1CSEL
C2CSEL
3
2
1
C1CSEL
Timer Mode Control Register (TnCTRL)
The Timer Mode Control (TnCTRL) register is a byte-wide,
read/write register that sets the operating mode of the timer/
counter and the TnA and TnB pins. This register is cleared
upon reset. The register format is shown below.
The Clock Unit Control (TnCKC) register is a byte-wide, read/
write register that selects the clock source for each timer/
counter. Selecting the clock source also starts the counter.
This register is cleared upon reset, which disables the timer/
counters. The register format is shown below.
4
C2CSEL
Reload/Capture B Register (TnCRB)
The Reload/Capture B (TnCRB) register is a word-wide,
read/write register that holds the reload or capture value for
Timer/Counter II. The register contents are not affected by a
reset and are unknown upon power-up.
Clock Unit Control Register (TnCKC)
5
Reload/Capture A Register (TnCRA)
The Reload/Capture A (TnCRA) register is a word-wide,
read/write register that holds the reload or capture value for
Timer/Counter I. The register contents are not affected by a
reset and are unknown upon power-up.
Clock Prescaler. When the timer is configured
to use the prescaled clock, the system clock is
divided by CLKPS+1 to produce the timer
clock. Thus, the system clock divide-by factor
can range from 1 to 32.
7
6
Reserved
Timer/Counter II Register (TnCNT2)
The Timer/Counter II (TnCNT2) register is a word-wide, read/
write register that holds the current count value for Timer/
Counter II. The register contents are not affected by a reset
and are unknown upon power-up.
The Clock Prescaler (TnPRSC) register is a byte-wide, read/
write register that holds the current value of the 5-bit clock
prescaler (CLKPS). This register is cleared upon reset. The
register format is shown below.
7
Timer/Counter I Register (TnCNT1)
7
6
Reserved TnAOUT
0
MDSEL
Counter I Clock Select. This 3-bit field defines
the clock mode for Timer/Counter I as follows:
000 = no clock (timer/counter I stopped)
001 = prescaled system clock
010 = external event on TnB (modes 1 and 3
only)
011 = pulse accumulate mode based on TnB
(modes 1 and 3 only)
100 = slow clock *
other values = undefined
Counter II Clock Select. This 3-bit field defines
the clock mode for Timer/Counter II as follows:
TnAEDG
TnBEDG
000 = no clock (Timer/Counter II stopped
modes 1, 2, and 3 only)
001 = prescaled system clock
010 = external event on TnB (modes 1 and 3
only)
011 = pulse accumulate mode based on TnB
(modes 1 and 3 only)
100 = slow clock *
other values = undefined
TnAEN
56
5
4
TnBEN
TnAEN
3
2
1
TnBEDG TnAEDG
0
MDSEL
Mode Select. This 2-bit field sets the operating
mode of the timer/counter as follows:
00 = Mode 1: PWM plus system timer
01 = Mode 2: Dual Input Capture plus system
timer
10 = Mode 3: Dual Timer/Counter
11 = Mode 4: Single Input Capture and Single
Timer
TnA Edge Polarity. When cleared (0), input pin
TnA is sensitive to falling edges (high to low
transitions). When set (1), input pin TnA is sensitive to rising edges (low to high transitions).
TnB Edge Polarity. When cleared (0), input pin
TnB is sensitive to falling edges (high to low
transitions). When set (1), input pin TnB is sensitive to rising edges (low to high transitions). In
pulse accumulate mode, when this bit is set (1),
the counter is enabled only when TnB is high;
when this bit is cleared (0), the counter is enabled only when TnB is low.
TnA Enable. When set (1), the TnA pin is enabled to operate as a preset input or as a PWM
output, depending on the timer operating
mode. In Mode 2 (Dual Input Capture), a transition on the TnA pin presets the TnCNT1
counter to FFFF hex. In the other modes, TnA
functions as a PWM output. When this bit is
www.national.com
cleared (0), operation of the pin for the timer/
counter is disabled.
TnB Enable. When set (1), the TnB pin in enabled to operate in Mode 2 (Dual Input Capture) or Mode 4 (Single Input Capture and
Single Timer). A transition on the TnB pin presets the corresponding timer/counter to FFFF
hex (TnCNT1 in Mode 2 or TnCNT2 in Mode
4). When this bit is cleared (0), operation of the
pin for the timer/counter is disabled. This bit
setting has no effect in Mode 1 or Mode 3.
TnA Output Data. This is a status bit that indicates the current state of the TnA pin when the
pin is used as a PWM output. When set (1), the
TnA pin is high; when cleared (0), the TnA pin
is low. The hardware sets and clears this bit,
but the software can also read or write this bit
at any time and thus control the state of the output pin. In case of conflict, a software write has
precedence over a hardware update. This bit
setting has no effect when TnA is used as an
input.
TnBEN
TnAOUT
15.5.8
TnCIEN
Timer Interrupt C Enable. See the description
of TnAIEN.
Timer Interrupt D Enable. See the description
of TnAIEN.
TnDIEN
15.5.9
Timer Interrupt Clear Register (TnICLR)
The Timer Interrupt Clear (TnICLR) register is a byte-wide,
write-only register that allows the software to clear the TnAPND, TnBPND, TnCPND, and TnDPND bits in the Timer Interrupt Control (TnICTRL) register. The register format is shown
below.
7 6 5 4
Reserved
TnACLR
TnBCLR
TnCCLR
Timer Interrupt Control Register (TnICTL)
TnDCLR
The Timer Interrupt Control (TnICTL) register is a byte-wide,
read/write register that contains the interrupt enable bits and
interrupt pending bits for the four timer interrupt sources,
designated A, B, C, and D. The condition that causes each
type of interrupt depends on the operating mode, as shown
in Table15.
3
TnDCLR
2
TnCCLR
1
TnBCLR
0
TnACLR
Timer Pending A Clear. When written with a 1,
the Timer Interrupt Source A Pending bit
(TnAPND) is cleared in the Timer Interrupt
Control register (TnICTL). Writing a 0 to the
TnACLR bit has no effect.
Timer Pending B Clear. See the description of
TnACLR.
Timer Pending C Clear. See the description of
TnACLR.
Timer Pending D Clear. See the description of
TnACLR.
This register is cleared upon reset. The register format is
shown below.
7
6
5
4
3
TnDIEN
TnCIEN
TnBIEN
TnAIEN
TnDPND
TnAPND
TnBPND
TnCPND
TnDPND
TnAIEN
TnBIEN
2
1
0
TnCPND TnBPND TnAPND
Timer Interrupt Source A Pending. When this
bit is set (1), it indicates that timer interrupt condition “A” has occurred. When this bit is cleared
(0), it indicates that the interrupt condition has
not occurred. For an explanation of interrupt
conditions A, B, C, and D, see Table15
This bit can be set by the hardware or by the
software. To clear this bit, the software must
use the Timer Interrupt Clear Register (TnICLR). Any attempt by the software to directly
write a 0 to this bit is ignored.
Timer Interrupt Source B Pending. See the description of TnAPND.
Timer Interrupt Source C Pending. See the description of TnAPND.
Timer Interrupt Source D Pending. See the description of TnAPND.
Timer Interrupt A Enable. When set (1), this bit
enables an interrupt on each occurrence of interrupt condition “A.” When cleared (0), an occurrence of interrupt condition “A” does not
generate an interrupt to the CPU, but still sets
the associated pending flag (TnAPND). For an
explanation of interrupt conditions A, B, C, and
D, see Table15.
Timer Interrupt B Enable. See the description
of TnAIEN.
57
www.national.com
16.0 Versatile-Timer-Unit (VTU)
The Versatile Timer Unit (VTU) contains four fully independent 16-bit timer subsystems. Each timer subsystem can operate either as dual 8-bit PWM timers, as a single 16-bit
PWM timer, or as a 16-bit counter with 2 input capture channels. These timer subsystems offers an 8-bit clock prescaler
to accommodate a wide range of system frequencies.
• The Versatile-Timer-Unit controls a total of eight I/O pins,
each of which can function as either:
— PWM output with programmable output polarity
— Capture input with programmable event detection and
timer reset
• A flexible interrupt scheme with
— four separate system level interrupt requests
— a total of 16 interrupt sources each with a separate interrupt pending flag and interrupt enable bit
The Versatile Timer Unit offers the following features:
• The Versatile Timer Unit (VTU) can be configured to provide:
— Eight fully independent 8-bit PWM channels
— Four fully independent 16-bit PWM channels
— Eight 16-bit input capture channels
• The VTU consists of four timer subsystems, each of which
contains:
— a 16-bit counter
— two 16-bit capture / compare registers
— an 8-bit fully programmable clock prescaler
• Each of the four timer subsystems can operate in the following modes:
— low power mode, i.e. all clocks are stopped
— dual 8-bit PWM mode
— 16-bit PWM mode
— dual 16-bit input capture mode
16.1
VTU FUNCTIONAL DESCRIPTION
The Versatile-Timer-Unit (VTU) is comprised of four timer
subsystems. Each timer subsystem contains an 8-bit clock
prescaler, a 16-bit up-counter and two 16-bit registers. Each
timer subsystem controls two I/O pins which either function
as PWM outputs or capture inputs depending on the mode of
operation. There are four system level interrupt requests,
one for each timer subsystem. Each system level interrupt request is controlled by four interrupt pending flags with associated enable/disable bits. All four timer subsystems are fully
independent and each may operate as a dual 8-bit PWM timer, a 16-bit PWM timer or as a dual 16-bit capture timer. Figure 18 illustrates the main elements of the Versatile-TimerUnit (VTU).
15
0
MODE
15
0
15
15
0
IO1CTL
0
INTCTL
IO2CTL
15
0
INTPND
Timer Subsystem 1
7
Timer Subsystem 2
0
7
Timer Subsystem 3
0
7
Timer Subsystem 4
0
7
0
C1PRSC
C2PRSC
C3PRSC
C4PRSC
==
Prescaler
Counter
==
Prescaler
Counter
==
Prescaler
Counter
==
Prescaler
Counter
15
0
COUNT1
PERCAP1
15
DTYCAP1
compare - capture
PERCAP3
compare - capture
TIO2
I/O control
TIO3
I/O control
PERCAP4
compare - capture
DTYCAP4
I/O control
TIO5
58
compare - capture
DTYCAP3
Figure 18. VTU Block Diagram
www.national.com
0
compare - capture
I/O control
TIO4
15
COUNT4
compare - capture
DTYCAP2
I/O control
0
COUNT3
PERCAP2
compare - capture
TIO1
0
COUNT2
compare - capture
I/O control
15
TIO6
I/O control
TIO7
I/O control
TIO8
16.1.1
Dual 8-bit PWM Mode
The period of the PWM output waveform is determined by
the value of the PERCAPx register. The TIOx output starts at
the default value as pro-grammed via the IOxCTL.PxPOL bit.
Once the counter value reaches the value of the period register PERCAPx, the counter is reset to 0016 upon the next
counter increment. Upon the following increment from 00 16
to 0116 , the TIOx output will change to the opposite of the default value.
Each timer subsystem may be configured to generate two
fully independent PWM waveforms on the respective TIOx
pins. In this mode, the counter COUNTx is split and operates
as two independent 8-bit counters. Each counter increments
at the rate determined by the clock prescaler.
Each of the two 8-bit counters may be started and stopped
separately via the associated TxRUN bits. Once either of the
two 8-bit timers is running the clock prescaler starts counting.
Once the clock prescaler counter value matches the value of
the associated CxPRSC register field, COUNTx is incremented.
The duty cycle of the PWM output waveform is controlled by
the DTYCAPx register value. Once the counter value reaches the value of the duty cycle register DTYCAPx, the PWM
output TIOx changes back to its default value upon the next
counter increment. Figure19 illustrates this concept.
COUNTx
0A
0A
PERCAPx
09
09
08
08
07
07
06
06
05
05
04
04
DTYCAPx
03
03
02
02
01
01
00
00
TxRUN=1
TIOx (PxPOL=0)
TIOx (PxPOL=1)
Figure 19. VTU PWM generation
The period time is determined by the following formula:
register while either of the two PWM channels is enabled, the
new value will not take effect until the counter value matches
the previous period value or the timer is stopped.
PWMperiod = (PERCAPx + 1) * (CxPRSC + 1) * TCLK
The duty cycle in percent is calculated as follows:
Reading the PERCAPx or DTYCAPx register will always return the most recent value written to it.
DutyCycle[%] = (DTYCAPx / (PERCAPx+1)) *100
If the duty cycle register (DTYCAPx) holds a value which is
greater then the value held in the period register (PERCAPx)
the TIOx output will remain at the opposite of its default value
which corresponds to a duty cycle of 100%. If the duty cycle
register (DTYCAPx) register holds a value of 00 16 , the TIOx
output will remain at the default value which corresponds to
a duty cycle of 0%. In that case the value contained in the
PERCAPx register is irrelevant. This scheme allows the duty
cycle to be programmed in a range from 0% to 100%.
The counter registers can be written if both 8-bit counters are
stopped. This allows the user to preset the counters before
starting and therefore generate PWM output waveforms with
a phase shift relative to one another. If the counter is written
with a value other then 001 6 it will start incrementing from that
value while TIOx remains at its default value until the first
0016 to 01 1 6 transition of the counter value occurs. If the
counter is preset to values which are smaller or equal then
the value held in the period register (PERCAPx) the counter
will count up until a match between the counter value and the
PERCAPx register value occurs. The counter will then be reset to 0016 and continue counting up. Alternatively the
counter may be written with a value which is greater then the
In order to allow fully synchronized updates of the period and
duty cycle compare values, the PERCAPx and DTYCAPx
registers are double buffered when operating in PWM mode.
Therefore if the user writes to either the period or duty cycle
59
www.national.com
value held in the period register. In that case the counter will
count up to FF 1 6 and then roll over to 0016 . In any case the
TIOx pin always changes its state at the 0016 to 011 6 transition of the counter.
Figure21 illustrates the configuration of a timer subsystem
while operating in 16-bit PWM mode. The numbering in
Figure21 refers to timer subsystem 1 but equally applies to
the other three timer subsystems.
The user software may only write to the COUNTx register if
both TxRUN bits of a timer subsystem are cleared. Any
writes to the counter register while either timer is running will
be ignored.
7
T1RUN
[15:0]
compare
PERCAP1[15:0]
compare
DTYCAP1[15:0]
S
16.1.3
0
[15:8]
Res
compare
compare
PERCAP1[15:8]
PERCAP1[7:0]
compare
compare
DTYCAP1[15:8]
DTYCAP1[7:0]
S
[7:0]
COUNT1[7:0]
R
P2POL
P1POL
TIO1
Dual 16-Bit Capture Mode
In capture mode the counter COUNTx operates as a 16-bit
up-counter while the two TIOx pins associated with a timer
subsystem operate as capture inputs. A capture event on the
TIOx pins causes the contents of the counter register
(COUNTx) to be copied to the PERCAPx or DTYCAPx registers respectively.
Starting the counter is identical to the 16-bit PWM mode, i.e.
setting the lower of the two MODE.TxRUN bits will start the
counter and the clock prescaler. In addition, the capture
event inputs are enabled once the MODE.TxRUN bit is set.
TIO1
The TIOx capture inputs can be independently configured to
detect a capture event on either a positive transition, a negative transition or both a positive and a negative transition. In
addition, any capture event may be used to reset the counter
COUNTx and the clock prescaler counter. This avoids the
need for the user software to keep track of timer overflow
conditions and greatly simplifies the direct frequency and
duty cycle measurement of an external signal.
Figure 20. VTU Dual 8-bit PWM Mode
16-Bit PWM Mode
Each of the four timer subsystems may be independently
configured to provide a single 16-bit PWM channel. In this
case the lower and upper bytes of the counter are concatenated to form a single 16-bit counter.
Figure22 illustrates the configuration of a timer subsystem
while operating in capture mode. The numbering in Figure22
Operation in 16-bit PWM mode is conceptually identical to
the dual 8-bit PWM operation as outlined under Dual 8-bit
PWM Mode on page 59. The 16-bit timer may be started or
stopped with the lower MODE.TxRUN bit, i.e. T1RUN for timer subsystem 1.
The two TIOx outputs associated with a timer subsystem can
be used to produce either two identical PWM waveforms or
two PWM waveforms of opposite polarities. This can be accomplished by setting the two PxPOL bits of the respective
timer subsystem to either identical or opposite values.
www.national.com
P1POL
Q
R
TIO2
16.1.2
P2POL
In addition to the two PWM modes, each timer subsystem
may be configured to operate in an input capture mode which
provides two 16-bit capture channels. The input capture
mode can be used to precisely measure the period and duty
cycle of external signals.
T1RUN
7
Q
Q
R
Figure 21. VTU 16-bit PWM Mode
TMOD1=01
T2RUN
S
S
TIO2
==
Prescaler
Counter
Res
Q
R
0
C1PRSC
COUNT1[15:8]
COUNT1[15:0]
Restart
Figure20 illustrates the configuration of a timer subsystem
while operating in dual 8-bit PWM mode. The numbering in
Figure20 refers to timer subsystem 1 but equally applies to
the other three timer subsystems.
8
0
15
— The associated TIOx pin will return to its default value
as defined by the IOxCTL.PxPOL bit.
— The counter will stop and will retain its last value.
— Any pending updates of the PERCAPx and DTYCAPx
register will be completed.
— The prescaler counter will be stopped and reset if both
MODE.TxRUN bits are cleared.
15
TMOD1=10
==
Prescaler
Counter
The two I/O pins associated with a timer subsystem function
as independent PWM outputs in the dual 8-bit PWM mode. If
a PWM timer is stopped via its associated MODE.TxRUN bit
the following actions result:
7
0
C1PRSC
60
refers to timer subsystem 1 but equally applies to the other
three timer subsystems.
7
ed with them. All interrupt pending flags are denoted IxAPD
through IxDPD where “x” relates to the specific timer subsystem. There is one system level interrupt request for each
of the four timer subsystems.
0
C1PRSC
TMOD1=11
Figure23 illustrates the interrupt structure of the versatile
timer module.
==
Prescaler
Counter
I1AEN
T1RUN
I1BEN
0
15
[15:0]
I1CEN
COUNT1[15:0]
Restart
I1DEN
System
Interrupt
Request 1
capture
I1APD
PERCAP1[15:0]
I1BPD
capture
DTYCAP1[15:0]
I1CPD
I1DPD
cap
cap
rst
rst
2
0
2
C1EDG
0
C2EDG
TIO1
I4AEN
TIO2
I4BEN
Figure 22. VTU Dual 16-bit Capture Mode
16.1.4
I4CEN
I4DEN
Low Power Mode
In case a timer subsystem is not used, the user can place it
in a low-power-mode. All clocks to a timer subsystem are
stopped and the counter and prescaler contents are frozen
once low-power-mode is entered. The user may continue to
write to the MODE, INTCTL, IOxCTL and CLKxPS registers.
Write operations to the INTPND register are allowed; but if a
timer subsystem is in low power mode, its associated interrupt pending bits cannot be cleared. The user cannot write to
the COUNTx, PERCAPx and DTYCAPx registers of a timer
subsystem while it is in low-power-mode. All registers can be
read at any time.
16.1.5
I4BPD
I4CPD
I4DPD
Figure 23. VTU Interrupt Request Structure
Each of the timer pending flags - IxAPD through IxDPD - is
set by a specific hardware event depending on the mode of
operation, i.e., PWM or Capture mode. Table17 outlines the
specific hardware events relative to the operation mode
which cause an interrupt pending flag to be set.
Interrupts
The Versatile-Timer-Unit (VTU) has a total of 16 interrupt
sources, four for each of the four timer subsystems. All interrupt sources have a pending flag and an enable bit associat-
Table 17
Pending Flag
System
Interrupt
Request 4
I4APD
VTU Interrupt Sources
Dual 8-bit PWM Mode
16-bit PWM Mode
Capture Mode
IxAPD
Low Byte Duty Cycle match
Duty Cycle match
Capture to DTYCAPx
IxBPD
Low Byte Period match
Period match
Capture to PERCAPx
IxCPD
High Byte Duty Cycle match
N/A
Counter Overflow
IxDPD
High Byte Period match
N/A
N/A
16.1.6
ISE Mode operation
16.2
The VTU supports breakpoint operation of the In-SystemEmulator (ISE). If FREEZE is asserted, all timer counter
clocks will be inhibited and the current value of the timer registers will be frozen; in capture mode, all further capture
events are disabled. Once FREEZE becomes inactive,
counting will resume from the previous value and the capture
input events are re-enabled.
VTU REGISTERS
The Versatile-Timer-Unit contains a total of 19 user accessible registers. All registers are word-wide and are initialized to
a known value upon reset. All software accesses to the VTU
registers must be word accesses.
61
www.national.com
16.2.1
Mode Control Register (MODE)
value of this three bit field has no effect while
operating in PWM mode.
The Mode Control (MODE) registries a word-wide read/write
register which controls the mode selection of all four timer
subsystems. The register is cleared (00001 6) upon reset.
15
14
TMOD4
7
13
12
T8RUN
T7RUN
6
TMOD2
TxRUN
5
4
T4RUN
T3RUN
TMOD3
3
00:
000
rising edge
No
T6RUN
T5RUN
001
falling edge
No
1
0
010
rising edge
Yes
T2RUN
T1RUN
011
falling edge
Yes
100
both edges
No
101
both edges
rising edge
110
both edges
falling edge
111
both edges
both edges
PxPOL
Low-Power-Mode enabled. All clocks to
the counter subsystem are stopped. The
counter is stopped regardless of the value of the TxRUN bits. Read operations
to the Timer Subsystem will return the
last value; the user shall not perform any
write operations to the Timer Subsystem
while it is disabled since those will be ignored.
01:
16.2.3
15
P4POL
12
Capture Mode enabled. Both 8-bit
counters are concatenated and operate
as a single 16-bit counter. The counter
may be started or stopped with the lower
of the two TxRUN bits, i.e., T1RUN,
T3RUN, T5RUN and T7RUN. The TIOx
pins will function as capture inputs.
11
C4EDG P3POL
10
8
7
C3EDG P2POL
6
4
3
C2EDG P1POL
2
16.2.4
I/O Control Register 2 (IO2CTL)
14
12
11
10
8
7
6
4
3
2
0
Interrupt Control Register (INTCTL)
15
14
13
12
11
10
9
8
I4DEN I4CEN I4BEN I4AEN I3DEN I3CEN I3BEN I3AEN
7
6
5
4
3
2
1
0
I2DEN I2CEN I2BEN I2AEN I1DEN I1CEN I1BEN I1AEN
0
C1EDG
Capture Edge Control. Defines the polarity of a
capture event and the reset of the counter. The
www.national.com
The PWM output is reset (0) upon the
001 6 to 0116 transition of the counter and
will be set (1) once the counter value
matches the duty cycle value.
The Interrupt Control (INTCTL) register is a word-wide read/
write register. It contains the interrupt enable bits for all 16 interrupt sources of the Versatile-Timer-Unit. Each interrupt enable bit corresponds to an interrupt pending flag located in
the Interrupt Pending Register (INTPND). All INTCTL register bits are solely under software control. The register is
cleared (00001 6) upon reset..
IxAEN
CxEDG
1=
The functionality of the bit fields of the IO2CTL register is
identical to the ones described in the IO1CTL register section.
I/O Control Register 1 (IO1CTL)
14
The PWM output is set (1) upon the 0016
to 01 1 6 transition of the counter and will
be reset (0) once the counter value
matches the duty cycle value.
P8POL C8EDG P7POL C7EDG P6POL C6EDG P5POL C5EDG
The I/O Control Register 1 (IO1CTL) is a word-wide read/
write register. The register controls the functionality of the
I/O pins TIO1 through TIO4 depending on the selected mode
of operation. The register is cleared (00001 6) upon reset.
15
0=
The I/O Control Register 2 (IO2CTL) is a word-wide read/
write register. The register controls the functionality of the
I/O pins TIO5 through TIO8 depending on the selected mode
of operation. The register is cleared (0000) upon reset.
16-bit PWM mode enabled. The two 8bit counters are concatenated to form a
single 16-bit counter. The counter may
be started or stopped with the lower of
the two TxRUN bits, i.e. T1RUN,
T3RUN, T5RUN and T7RUN. The TIOx
pins will function as PWM outputs.
11:
PWM Polarity. While operating in PWM mode
the bit defines the output polarity of the corresponding PWM output (TIOx).
Once a counter is stopped, the output will assume the value
of PxPOL, i.e., its initial value. The PxPOL bit has no effect
while operating in capture mode.
Dual 8-bit PWM mode enabled. Each 8bit counter may individually be started or
stopped via its associated TxRUN bit.
The TIOx pins will function as PWM outputs.
10:
Counter Reset
8
2
TMOD1
Capture
9
10
Timer start/stop. If set (1), the associated
counter and clock prescaler is started depending on the mode of operation. Once set, the
clock to the clock prescaler and the counter are
enabled and the counter will increment each
time the clock prescaler counter value matches
the value defined in the associated clock prescaler field (CxPRSC).
Timer System Operating Mode. This 2-bit wide
field enables or disables the Timer Subsystem
and defines it’s operating mode.
TMODx
16.2.2
11
CxEDG
62
Timer x interrupt A enable. Enable/Disable an
interrupt request based on the corresponding
IxAPD flag being set. The associated IxAPD
flag will be updated regardless of the value of
the IxAEN bit.
IxBEN
0
Enable system interrupt request for the
IxAPD pending flag
1
Disable system interrupt request for the
IxAPD pending flag
Timer x interrupt B enable. Enable/Disable an
interrupt request based on the corresponding
IxBPD flag being set. The associated IxBPD
flag will be updated regardless of the value of
the IxBEN bit.
0
IxCEN
IxDPD
Enable system interrupt request for the
IxBPD pending flag
1
16.2.6
Disable system interrupt request for the
IxBPD pending flag
IxDEN
15
C1PRSC
Disable system interrupt request for the
IxCPD pending flag
Timer x interrupt D enable. Enable/Disable an
interrupt request based on the corresponding
IxDPD flag being set. The associated IxDPD
flag will be updated regardless of the value of
the IxDEN bit.
0
Enable system interrupt request for the
IxDPD pending flag
1
Disable system interrupt request for the
IxDPD pending flag
C2PRSC
Interrupt Pending Register (INTPND)
16.2.7
The Interrupt Pending (INTPND) register is a word-wide
read/write register which contains all 16 interrupt pending
flags. There are four interrupt pending flags called IxAPD
through IxDPD per timer subsystem. Each interrupt pending
flag is set by a hardware event and can be cleared if the user
software writes a 1 to the bit position. The value will remain
unchanged if a 0 is written to the bit position. All interrupt
pending flags are cleared (0) upon reset.
15
14
13
12
11
10
9
6
5
4
3
2
1
IxBPD
Clock Prescaler 1 compare value. Holds the 8bit prescaler value for timer subsystem 1. The
counter of timer subsystem is incremented
each time when the clock prescaler compare
value matches the value of the clock prescaler
counter. The divide-by-ratio is equal to
C1PRSC+1 i.e. a value of 001 6 results in a divide by 1 whereas the maximum divide-by ratio
is 256 for a C1PRSC value of FF1 6.
Clock Prescaler 2 compare value. Holds the 8bit prescaler value for timer subsystem 2. The
functionality of this field is identical to the one
described for C1PRSC in the previous paragraph.
Clock Prescaler Register 2 (CLK2PS)
15
8 7
C4PRSC
8
C3PRSC
0
I2DPD I2CPD I2BPD I2APD I1DPD I1CPD I1BPD I1APD
IxAPD
0
C1PRSC
The Clock Prescaler Register 2 (CLK2PS) is a word-wide
read/write register. The register is split into two 8-bit wide
fields called C3PRSC and C4PRSC. Each field holds the 8bit clock prescaler compare value for timer subsystems 3 and
4 respectively. The register is cleared upon reset.
I4DPD I4CPD I4BPD I4APD I3DPD I3CPD I3BPD I3APD
7
8 7
C2PRSC
Enable system interrupt request for the
IxCPD pending flag
1
Clock Prescaler Register 1 (CLK1PS)
CLK1PS is a word-wide read/write register. The register is
split into two 8-bit wide field called C1PRSC and C2PRSC.
Each field holds the 8-bit clock prescaler compare value for
timer subsystems 1 and 2 respectively. The register is
cleared upon reset.
Timer x interrupt C enable. Enable/Disable an
interrupt request based on the corresponding
IxCPD flag being set. The associated IxCPD
flag will be updated regardless of the value of
the IxCEN bit.
0
16.2.5
IxCPD
subsystem has occurred. Table 17 on page 61
lists the hardware condition which causes this
bit to be set.
Timer x interrupt C pending. If set (1), indicates
that an interrupt condition for the related timer
subsystem has occurred. Table 17 on page 61
lists the hardware condition which causes this
bit to be set.
Timer x interrupt D pending. If set (1), indicates
that an interrupt condition for the related timer
subsystem has occurred. Table 17 on page 61
lists the hardware condition which causes this
bit to be set.
C4PRSC
Timer x interrupt A pending. If set (1), indicates
that an interrupt condition for the related timer
subsystem has occurred. Table 17 on page 61
lists the hardware condition which causes this
bit to be set.
Timer x interrupt B pending. If set (1), indicates
that an interrupt condition for the related timer
16.2.8
0
C3PRSC
Clock Prescaler 3 compare value. Holds the 8bit prescaler value for timer subsystem 3. The
functionality of this field is identical to the one
described for C1PRSC on page 63.
Clock Prescaler 4 compare value. Holds the 8bit prescaler value for timer subsystem 4. The
functionality of this field is identical to the one
described for C1PRSC on page 63.
Counter Registers (COUNTx)
The Counter (COUNTx) registers are word wide read/write
registers. There are a total of four registers called COUNT1
through COUNT4, one for each of the four timer subsystems.
The user software may read the registers at any time. Read-
63
www.national.com
ing the register will return the current value of the counter.
The register may only be written if the counter is stopped i.e.
if both TxRUN bits associated with a timer subsystem are
cleared. The registers are cleared upon reset (0000).
15
0
CNTx
16.2.9
Period/Capture Registers (PERCAPx)
The Period/Capture (PERCAPx) registers are word-wide
read/write registers. There are a total of four registers called
PERCAP1 through PERCAP4, one for each timer subsystem. The register hold the period compare value in PWM
mode of the counter value at the time the last associated capture event occurred. In PWM mode the register is double
buffered. If a new period compare value is written while the
counter is running, the write will not take effect until counter
value matches the previous period compare value or until the
counter is stopped. Reading may take place at any time and
will return the most recent value which was written. The PERCAPx registers are reset to 0000 upon reset.
15
0
PCAPx
16.2.10 Duty Cycle / Capture Registers (DTYCAPx)
The Duty Cycle/Capture (DTYCAPx) registers are word-wide
read/write registers. There are a total of four registers called
DTYCAP1 through DTYCAP4, one for each timer subsystem. The registers hold the period compare value in PWM
mode or the counter value at the time the last associated
capture event occurred. In PWM mode the register is double
buffered. If a new duty cycle compare value is written while
the counter is running, the write will not take effect until the
counter value matches the previous period compare value or
until the counter is stopped. In other words, the update takes
effect on period boundaries only. Reading may take place at
any time and will return the most recent value which was written. The DTYCAPx registers are reset to 000016 upon reset.
15
0
DCAPx
www.national.com
64
17.0 MICROWIRE/SPI
MICROWIRE/PLUS is a synchronous serial communications
protocol, originally implemented in National Semiconductor's
COPS™ and HPC™ families of microcontrollers to minimize
the number of connections, and therefore the cost, of communicating with peripherals.
5
The device has an enhanced MICROWIRE/SPI interface
module (MWSPI) that can communicate with all peripherals
that conform to MICROWIRE or Serial Peripheral Interface
(SPI) specifications. This enhanced MICROWIRE interface
is capable of operating as either a master or slave and in 8or 16-bit mode. Figure24 shows a typical enhanced MICROWIRE interface application.
Chip Select Lines
MCS
MCS
CS
Master
8-Bit
A/D
CS
1K Bit
EEPROM
I/O
Lines
DO SK DI
CS
CS
LCD
Display
driver
VF
Display
Driver
SK DI
DO
DO SK DI
Slave
I/O
Lines
SK DI
MDIDO
MDIDO
MDODI
MDODI
MSK
MSK
Figure 24. MICROWIRE Interface
The enhanced MICROWIRE interface module includes the
following features:
—
—
—
—
—
—
—
—
—
—
data out signal (MDODI for master mode, MDIDO for slave
mode) and the serial clock (MSK).
Programmable operation as a Master or Slave
Programmable shift-clock frequency (master only)
Programmable 8- or 16-bit mode of operation
8- or 16-bit serial I/O data shift register
Two modes of clocking data
Serial clock can be low or high when idle
16-bit read buffer
Busy flag, Read Buffer Full flag, and Overrun flag for
polling and as interrupt sources
Supports multiple masters
Maximum bit rate of 10M bits/second (master mode)
5M bits/second (slave mode) at 20MHz system clock
Supports very low-end slaves with the Slave Ready
output
Echo back enable/disable (Slave only)
In slave mode, an optional fourth signal (MCS) may be used
to enable the slave transmit. At any given time, only one
slave can respond to the master. Each slave device has its
own chip select signal (MCS) for this purpose.
The MICROWIRE interface allows the device to operate either as a master or slave transferring 8- or 16-bits of data.
This is configured via the MMNS bit.
Figure25 shows a block diagram of the enhanced MICROWIRE serial interface in the device.
17.1.1
Shifting
The MICROWIRE interface allows several devices to be connected on one three-wire system. At any given time, one of
these devices operates as the master while all other devices
operate as slaves.
The MICROWIRE interface is a full duplex transmitter/receiver. A 16-bit shifter, which can be split into a low and high byte,
is used for both transmitting and receiving. In 8-bit mode,
only the lower 8-bits are used to transfer data. The transmitted data is shifted out through MDODI pin (master mode) or
MDIDO pin (slave mode), starting with the most significant
bit. At the same time, the received data is shifted in through
MDIDO pin (master mode) or MDODI pin (slave mode), also
starting with the most significant bit first.
The master device supplies the synchronous clock (MSK) for
the serial interface and initiates the data transfer. The slave
devices respond by sending (or receiving) the requested data. Each slave device uses the master’s clock for serially
shifting data out (or in), while the master shifts the data in (or
out).
The shift in and shift out are controlled by the MSK clock. In
each clock cycle of MSK, one bit of data is transmitted/received. The 16-bit shifter is accessible via the MWDAT register. Reading the MWDAT register returns the value in the
read buffer. Writing to the MWDAT register updates the 16bit shifter.
The three-wire system includes: the serial data in signal
(MDIDO for master mode, MDODI for slave mode), the serial
17.1.2
—
—
17.1
MICROWIRE OPERATION
Reading
The enhanced MICROWIRE interface implements a double
buffer on read. As illustrated in Figure25, the double read
65
www.national.com
Interrupt Request
Control + Status
MCS
Read Data
16-bit Read Buffer
Write Data
8
8
MWDAT
Slave
16-bit Shift Register
Data Out
Master
MDODI
Slave
Data In
Master
MDIDO
MSK
MSK
Clock Prescaler + Select
Master
System Clock
Figure 25. MICROWIRE Block Diagram
buffer consists of the 16-bit shifter and a buffer, called the
read buffer.
(master mode) or the MDODI pin (slave mode), is sampled
on the rising edge of MSK.
The 16-bit shifter loads the read buffer with new data when
the data transfer sequence is completed and previous data in
the read buffer has been read. In master mode, an Overrun
error occurs when the read buffer is full, the 16-bit shifter is
full and a new data transfer sequence starts.
In the alternate mode, the output data is shifted out on the rising edge of MSK on the MDODI pin (master mode) or MDIDO
pin (slave mode). The input data, which is received via MDIDO pin (master mode) or MDODI pin (slave mode), is sampled on the falling edge of MSK.
When 8-bit mode is selected, the lower byte of the shift register is loaded into the lower byte of the read buffer and the
read buffer’s higher byte remains unchanged.
The clocking modes are selected with the MSKM bit. The
MIDL bit allows selection of the value of MSK when it is idle
(when there is no data being transferred). Various MSK clock
frequencies can be programmed via the MCDV bits. Figures
27, 28, 29, and 30 show the data transfer timing for the normal and the alternate modes with the MIDL bit equal to 0 and
equal to 1.
The “Receive Buffer Full” (MRBF) bit indicates if the MWDAT
register holds valid data. The MOVR bit indicates that an
overrun condition has occurred.
17.1.3
Writing
Note that when data is shifted out on MDODI (master mode)
or MDIDO (slave mode) on the leading edge of the MSK
clock, bit 14 (16-bit mode) is shifted out on the second leading edge of the MSK clock. When data are shifted out on
MDODI (master mode) or MDIDO (slave mode) on the trailing edge of MSK, bit 14 (16-bit mode) is shifted out on the first
trailing edge of MSK.
The “MICROWIRE Busy” (MBSY) bit indicates whether the
MWDAT register can be written. All write operations to the
MWDAT register update the shifter while the data contained in
the read buffer is not affected. Undefined results will occur if
the MWDAT register is written to while the MBSY bit is set to 1.
17.1.4
Clocking Modes
17.2
Two clocking modes are supported: the normal mode and the
alternate mode.
In Master mode, the MSK pin is an output for the shift clock,
MSK. When data is written to the (MWnDAT register), eight
or sixteen MSK clocks, depending on the mode selected, are
generated to shift the eight or sixteen bits of data and then
In the normal mode, the output data, which is transmitted on
the MDODI pin (master mode) or the MDIDO pin (slave
mode), is clocked out on the falling edge of the shift clock
MSK. The input data, which is received via the MDIDO pin
www.national.com
MASTER MODE
66
End of Transfer
MSK
Data Out
MSB
msb-1
msb-2 5
Bit 1
Bit 0
(lsb)
Data In
MSB
msb-1
msb-2
Bit 1
Bit 0
(lsb)
Shift Out
Figure 26.
Sample Point
Normal Mode, MIDL Bit = 0
End of Transfer
MSK
Data Out
msb
msb-1
msb-2
Bit 1
Bit 0 (lsb)
Data In
msb
msb-1
msb-2
Bit 1
Bit 0 (lsb)
Shift Out
Figure 27.
Sample Point
Normal Mode, MIDL Bit = 1
End of Transfer
MSKn
Data Out
msb
msb-1
msb-2
Bit 1
Bit 0 (lsb)
Data In
msb
msb-1
msb-2
Bit 1
Bit 0 (lsb)
Shift Out
Sample Point
Figure 28. Alternate Mode, MIDL Bit = 0
MSK goes idle again. The MSK idle state can be either high
or low, depending on the MIDL bit.
17.3
active. Data transfer is enabled when MCS is active.
The slave starts driving MDIDO when MCS is activated. The
most significant bit (lower byte in 8-bit mode or upper byte in
16-bit mode) is output onto the MDIDO pin first. After eight or
sixteen clocks (depending on the selected mode), the data
transfer is completed.
SLAVE MODE
In Slave mode, the MSK pin is an input for the shift clock
MSK. MDIDO is placed in TRI-STATE mode when MCS is in-
67
www.national.com
End of Transfer
MSKn
Data Out
msb
msb-1
msb-2
Bit 1
Bit 0 (lsb)
Data In
msb
msb-1
msb-2
Bit 1
Bit 0 (lsb)
Shift Out
Sample Point
Figure 29. Alternate Mode, MIDL Bit = 1
17.5
If a new shift process starts before MWDAT was written, i.e.,
while MWDAT does not contain any valid data, and the “Echo
Enable” (MECH) bit is set to 1, the data received from MDODI is transmitted on MDIDO in addition to being shifted to
MWDAT. If the MECH bit is cleared to 0, the data transmitted
on MDIDO is the data held in the MWDAT register, regardless of its validity. The master may negate the MCS signal to
synchronize the bit count between the master and the slave.
In the case that the slave is the only slave in the system, MCS
can be tied to VSS .
17.4
The software interacts with the MICROWIRE interface by accessing the MICROWIRE registers. There are five such registers:
— MICROWIRE Data Register (MWDAT)
— MICROWIRE Control Register (MWCTL)
— MICROWIRE Status Register (MWSTAT)
17.5.1
INTERRUPT GENERATION
— When the read buffer is full (MRBF=1) and the “Enable
Interrupt for Read” bit is set (MEIR=1).
— Whenever the shifter is not busy, i.e. the MBSY bit is
cleared (MBSY=0) and the “Enable Interrupt for Write”
bit is set (MEIW=1).
— When an overrun condition occurs (MOVR is set to 1)
and the “Enable Interrupt on Overrun” bit is set
(MEIO=1). This usage is restricted to master mode.
Figure30 illustrates the various interrupt capabilities of this
module.
MEIO
MOVR = 1
MWSPI
Interrupt
MEIR
MBSY = 0
MEIW
Figure 30. MWSPI Interrupts
www.national.com
MICROWIRE Data Register (MWDAT)
The MWDAT register is a word-wide, read/write register used
to transmit and receive data through the MDODI and MDIDO
pins. Figure31 shows the hardware structure of the register.
An interrupt is generated in any of the following cases:
MRBF = 1
MICROWIRE INTERFACE REGISTERS
68
write
DIN
Shift Register
1
0
High-Byte
Low-Byte
DOUT
(store & MWMOD)
(store)
MWMOD
Read Buffer
Low-Byte
High-Byte
read
MWDAT
Figure 31.
17.5.2
MWDAT Register Structure
MICROWIRE Control Register (MWCTL)
MECH
Upon reset, all non-reserved bits are cleared to 0. The register format is shown below.
15
9
8
7
6
5
4
3
2
1
0
MCDV MIDL MSKM MEIW MEIR MEIO MECH MMOD MMNS MEN
[6:0]
MEN
MICROWIRE Enable. This bit enables (1) or
disables (0) the MICROWIRE interface module. Clearing this bit disables the module,
clears the status bits in the MICROWIRE status
register (the MBSY, MRBF, and MOVR flags in
MWSTAT), and places the MICROWIRE interface pins in the states described in Table18.
MEIO
Table 18 Pin Values with MICROWIRE
Disabled
MSK
Master: MnIDL Bit
Slave: input
MCS
Input
MDIDO
Master: input
Slave: TRI-STATE
MDODI
MMNS
MMOD
MEIR
Master: known Value
Slave: input
MICROWIRE Master/Slave Select. When
cleared to 0, the device operates as a slave.
When set to 1, the device operates as the master.
MICROWIRE Mode Select (8- or 16-bit). When
set to 0, the device operates in 8-bit mode.
When set to 1, the device operates in 16-bit
mode. This bit should only be changed when
the module is disabled or the MICROWIRE interface is idle (MWSTAT.MBSY=0).
MEIW
MSKM
69
MICROWIRE Echo Back. This bit enables (1)
or disables (0) the echo back function in slave
mode. This bit should be written only when the
MICROWIRE interface is idle (MWSTAT.MBSY=0). The MECH bit is ignored in master
mode. The MWDAT register is valid from the
time the register has been written until the end
of the transfer.
In the echo back mode, MDODI is transmitted
(echoed back) on MDIDO if MWDAT does not
contain any valid data. With the echo back
function disabled, the data held in the MWDAT
register is transmitted on MDIDO, whether or
not the data is valid.
MICROWIRE Enable Interrupt on Overrun.
This bit enables or disables the overrun error
interrupt. When set to 1, an interrupt is generated when the Receive Overrun Error flag
(MWSTAT.MOVR) is set. Otherwise, no interrupt is generated when an overrun error occurs. This bit should only be enabled in master
mode.
MICROWIRE Enable Interrupt for Read. When
set to 1, an interrupt is generated when the
Read Buffer Full flag (MWSTAT.MRBF) is set.
Otherwise, no interrupt is generated when the
read buffer is full.
MICROWIRE Enable Interrupt for Write. When
set to 1, an interrupt is generated when the
Busy bit (MWSTAT.MBSY) is cleared, which indicates that a data transfer sequence has been
completed and the read buffer is ready to receive the new data. Otherwise, no interrupt is
generated when the Busy bit is cleared.
MICROWIRE Clocking Mode. When cleared to
0, the device uses the normal clocking mode.
When set to 1, the device uses the alternate
www.national.com
MIDL
MCDV
17.5.3
clocking mode. In the normal mode, the output
data is clocked out on the falling edge of MSK
and the input data is sampled on the rising
edge of MSK. In the alternate mode, the output
data is clocked out on the rising edge of MSK
and the input data is sampled on the falling
edge of MSK.
MICROWIRE Idle. This bit sets the value of the
MSK output when the MICROWIRE interface is
idle: 0 for low or 1 for high. This bit should be
changed only when the MICROWIRE interface
module is disabled (MEN=0) or when no bus
transaction is in progress (MWSTAT.MBSY=0).
MICROWIRE Clock Divider Value. This 7-bit
field specifies the divide-by factor used for generating the MSK shift clock from the system
clock. The divide-by factor is 2*(MCDV[6:0]+1).
This allows selection of a divide-by ratio from 2
to 256. This field is ignored in slave mode
(MWCTL1.MMNS=0).
MOVR
MICROWIRE Status Register (MWSTAT)
The MICROWIRE Status Register is a word-wide, read-only
register that shows the current status of the MICROWIRE interface module. Upon reset, all non-reserved bits are cleared
to 0. The register format is shown below.
15
3
Reserved
MBSY
MRBF
2
MOVR
1
MRBF
0
MBSY
MICROWIRE Busy. This bit, when set to 1, indicates that the MICROWIRE shifter is busy.
In master mode, MBSY is set to 1 when the
MWDAT register is written. In slave mode, this
bit is set to 1 on the first leading edge of MSK
when MCS is asserted or when the MWDAT
register is written, whatever occurs first.
In both master and slave modes, this bit is
cleared to 0 when the MICROWIRE data transfer sequence is completed and the read buffer
is ready to receive the new data; in other
words, when the previous data held in the read
buffer has already been read.
If the previous data in the read buffer has not
been read and a new data has been received
into the shift register, the MBSY will not be
cleared, as the transfer could not be completed. This is because the contents of the shift
register could not be copied into the read buffer.
MICROWIRE Read Buffer Full. This bit, when
set to 1, indicates that the MICROWIRE read
buffer is full and ready to be read by the software. It is set to 1 when the shifter loads the
read buffer, which occurs upon completion of a
transfer sequence if the read buffer is empty.
The MRBF bit is updated when the MWDAT
register is read. At that time, the MRBF bit is
www.national.com
70
cleared to 0 if the shifter does not contain any
new data (in other words, the shifter is not receiving data or has not yet received a full byte
of data). The MRBF bit remains set to 1 if the
shifter already holds new data at the time that
MWDAT is read. In that case, MWDAT is immediately reloaded with the new data and is ready
to be read by the software.
MICROWIRE Receive Overrun Error. This bit,
when set to 1 in master mode, indicates that a
receive overrun error has occurred. This error
occurs when the read buffer is full, the 8-bit
shifter is full, and a new data transfer sequence
starts. This bit is undefined in slave mode.
The MOVR bit, once set, remains set until
cleared by the software. The software clears
this bit by writing a 1 to its bit position. Writing
a 0 to this bit position has no effect. No other
bits in the MWSTAT register are affected by a
write operation to the register.
18.0 USART
The USART module is a full-duplex Universal Synchronous/
Asynchronous Receiver/Transmitter that supports a wide
range of software-programmable baud rates and data formats. It handles automatic parity generation and several error detection schemes. There are one or two independent
USART modules in each device, depending on the package
type.
the first stage based on the programmed baud rate divisor to
create the baud rate clock.
The Control and Error Detection block contains the USART
control registers, control logic, error detection circuit, parity
generator/checker, and interrupt generation logic. The control registers and control logic determine the data format,
mode of operation, clock source, and type of parity used. The
error detection circuit generates parity bits and checks for
parity, framing, and overrun errors.
Each USART module offers the following features:
— Full-duplex double-buffered receiver/transmitter
— Synchronous or asynchronous operation
— programmable
baud
rate
from
SYS_CLK/
[2*(1+2^11)*16] up to SYSCLK/2 for USART configured to run in synchronous mode
— programmable
baud
rate
from
SYS_CLK/
[16*(1+2^11)*16] up to SYSCLK/16 for USART configured to run in asynchronous mode
— Programmable framing formats: seven, eight, or nine
data bits; one or two stop bits; and odd, even, mark,
space, or no parity
— Hardware parity generation for data transmission and
parity check for data reception
— Interrupts on “transmit ready” and “receive ready” conditions, separately enabled
— Software-controlled break transmission and detection
— Internal diagnostic capability
— Automatic detection of parity, framing, and overrun errors
18.2
USART OPERATION
The USART has two basic modes of operation: synchronous
and asynchronous. In addition, there are two specialpurpose synchronous and asynchronous modes, called attention and diagnostic. This section describes the operating
modes of the USART.
18.2.1
Asynchronous Mode
The asynchronous mode of the USART enables the device
to communicate with other devices using just two communication signals: transmit and receive.
— Transmitter
— Receiver
— Baud Rate Generator
— Control and Error Detection
Note: In the description of the USART, the lower-case letter
“n” represents the USART number. For example, TDXn
means TDX1 or TDX2.
In the asynchronous mode, the transmit shift register (TSFT)
and the transmit buffer (UnTBUF) double-buffer the data for
transmission. To transmit a character, a data byte is loaded
in the UnTBUF register. The data is then transferred to the
TSFT register. While the TSFT is shifting out the current character (LSB first) on the TDXn pin, the UnTBUF register is
loaded by software with the next byte to be transmitted.
When TSFT finishes transmission of the last stop bit of the
current frame, the contents of UnTBUF are transferred to the
TSFT register and the Transmit Buffer Empty flag (UnTBE) is
set. The UnTBE flag is automatically reset by the USART
when the software loads a new character into the UnTBUF
register. During transmission, the UnXMIP bit is set high by
the USART. This bit is reset only after the USART has sent
the last stop bit of the current character and the UnTBUF register is empty. The UnTBUF register is a read/write register.
The TSFT register is not user accessible.
The Transmitter block consists of an 8-bit transmit shift register and an 8-bit transmit buffer. Data bytes are loaded in
parallel from the buffer into the shift register and then shifted
out serially on the TDXn pin.
In asynchronous mode, the input frequency to the USART is
16 times the baud rate. In other words, there are 16 clock cycles per bit time. In asynchronous mode the baud rate generator is always the USART clock source.
The Receiver block consists of an 8-bit receive shift register
and an 8-bit receive buffer. Data is received serially on the
RDXn pin and shifted into the shift register. Once eight bits
have been received, the contents of the shift register are
transferred in parallel to the receive buffer.
The receive shift register (RSFT) and the receive buffer (UnRBUF) double buffer the data being received. The USART
receiver continuously monitors the signal on the RDXn pin for
a low level to detect the beginning of a start bit. Upon sensing
this low level, the USART waits for seven input clock cycles
and samples again three times. If all three samples still indicate a valid low, then the receiver considers this to be a valid
start bit, and the remaining bits in the character frame are
each sampled three times, around the mid-bit position. For
any bit following the start bit, the logic value is found by majority voting, i.e. the two samples with the same value define
the value of the data bit. Figure33 illustrates the process of
start bit detection and bit sampling.
18.1
FUNCTIONAL OVERVIEW
Figure32 is a block diagram of the USART module showing
the basic functional units in the USART:
The Transmitter and Receiver blocks both contain extensions for 9-bit data transfers, as required by the 9-bit and
loopback operating modes.
The Baud Rate Generator generates the clock for the synchronous and asynchronous operating modes. It consists of
two registers and a two-stage counter. The registers are
used to specify a prescaler value and a baud rate divisor. The
first stage of the counter divides the USART clock based on
the value of the programmed prescaler to create a slower
clock. The second stage of the counter divides the output of
Serial data input on the RDXn pin is shifted into the RSFT
register. Upon receiving the complete character, the contents
of the RSFT register are copied into the UnRBUF register
and the Receive Buffer Full flag (UnRBF) is set. The UnRBF
71
www.national.com
flag is automatically reset when software reads the character
from the UnRBUF register. The RSFT register is not user accessible.
Transmitter
TDXn
Baud clock
Sys_clk
Internal Bus
Control and
Error Detection
CKXn
Baud Rate Generator
Parity
Generator/Checker
Baud Clock
Receiver
RDXn
Figure 32. USART Block Diagram
16
1
2
3
4
5
6
7
8
Sample
9
10
11
12
13
14
15
1
2
1
Sample
DATA (LSB)
STARTBIT
16
16
3
4
5
6
7
8
9
10
11
12
13
14
15
16
1
Sample
DATABIT
Figure 33. USART Asynchronous Communication
18.2.2
Synchronous Mode
Data bytes are transmitted and received least significant bit
(LSB) first.
The synchronous mode of the USART enables the device to
communicate with other devices using three communication
signals: transmit, receive, and clock. In this mode, data bits
are transferred synchronously with the USART clock signal.
Data bits are transmitted on the rising edges and received on
the falling edges of the clock signal, as shown in Figure34.
www.national.com
In the synchronous mode, the transmit shift register (TSFT)
and the transmit buffer (UnTBUF) double-buffer the data for
transmission. To transmit a character, a data byte is loaded
in the UnTBUF register. The data is then transferred to the
TSFT register. The TSFT register shifts out one bit of the current character, LSB first, on each rising edge of the clock.
72
to the receive buffer. The UnRBF flag is set and an interrupt
(if enabled) is generated. The UnATN bit is automatically reset to zero, and the USART begins receiving all subsequent
characters. The software must examine the contents of the
UnRBUF register and respond by accepting the subsequent
characters (by leaving the UnATN bit reset) or waiting for the
next address character (by setting the UnATN bit again).
CKX
TDX
The operation of the USART transmitter is not affected by the
selection of this mode. The value of the ninth bit to be transmitted is programmed by setting or clearing a bit called
UnXB9 in the USART Frame Select register. The value of the
ninth bit received is read from UnRB9 in the USART Status
Register.
RDX
Sample Input
18.2.4
The Diagnostic mode is available for testing of the USART. In
this mode, the TDXn and RDXn pins are internally connected
together, and data that is shifted out of the transmit shift register is immediately transferred to the receive shift register.
This mode supports only the 9-bit data format with no parity.
The number of start and stop bits is programmable.
Figure 34. USART Synchronous Communication
While the TSFT is shifting out the current character on the
TDXn pin, the UnTBUF register may be loaded by the software with the next byte to be transmitted. When the TSFT finishes transmission of the last stop bit within the current
frame, the contents of UnTBUF are transferred to the TSFT
register and the Transmit Buffer Empty flag (UnTBE) is set.
The UnTBE flag is automatically reset by the USART when
the software loads a new character into the UnTBUF register.
During transmission, the UnXMIP bit is set high by the
USART. This bit is reset only after the USART has sent the
last frame bit of the current character and the UnTBUF register is empty.
18.2.5
Frame Format Selection
The format shown in Figure35 consists of a start bit, seven
data bits (excluding parity), and one or two stop bits. If parity
bit generation is enabled by setting the UnPEN bit, a parity
bit is generated and transmitted following the seven data bits.
1
The receive shift register (RSFT) and the receive buffer
(UnRBUF) double-buffer the data being received. Serial data
received on the RDXn pin is shifted into the RSFT register at
the first falling edge of the clock. Each subsequent falling
edge of the clock causes an additional bit to be shifted into
the RSFT register. The USART assumes a complete character has been received after the correct number of rising edges on CKXn (based on the selected frame format) have been
detected. Upon receiving a complete character, the contents
of the RSFT register are copied into the UnRBUF register
and the Receive Buffer Full flag (UnRBF) is set. The UnRBF
flag is automatically reset when the software reads the character from the UnRBUF register.
1a
1b
1c
START
BIT
7 BIT DATA
S
START
BIT
7 BIT DATA
START
BIT
7 BIT DATA
PA
START
BIT
7 BIT DATA
PA
2S
S
2S
Figure 35. Seven Data Bit Frame Options
The format shown in Figure36 consists of one start bit, eight
data bits (excluding parity), and one or two stop bits. If parity
bit generation is enabled by setting the UnPEN bit, a parity
bit is generated and transmitted following the eight data bits.
The transmitter and receiver may be clocked from either an
external source provided to the CKXn pin or by the internal
baud rate generator. In the latter case, the clock signal is
placed on the CKXn pin as an output.
18.2.3
Diagnostic Mode
2
Attention Mode
2a
The Attention mode is available for networking this device
with other processors. This mode requires the 9-bit data format with no parity. The number of start bits and number of
stop bits are programmable. In this mode, two types of 9-bit
characters are sent on the network: address characters consisting of 8 address bits and a 1 in the ninth bit position and
data characters consisting of 8 data bits and a 0 in the ninth
bit position.
2b
2c
START
BIT
8 BIT DATA
START
BIT
8 BIT DATA
START
BIT
8 BIT DATA
PA
START
BIT
8 BIT DATA
PA
S
2S
S
2S
Figure 36. Eight Data Bit Frame Options
While in Attention mode, the USART receiver monitors the
communication flow but ignores all characters until an address character is received. Upon the receipt of an address
character, the contents of the receive shift register are copied
The format shown in Figure37 consists of one start bit, nine
data bits, and one or two stop bits. This format also supports
the USART attention feature. When operating in this format,
all eight bits of UnTBUF and UnRBUF are used for data. The
73
www.national.com
ninth data bit is transmitted and received using two bits in the
control registers, called UnXB9 and UnRB9. Parity is not
generated or verified in this mode.
3
3a
START
BIT
9 BIT DATA
START
BIT
9 BIT DATA
Figure 37.
18.2.6
Table 19
S
2S
Nine Data Bit Frame Options
Prescaler
Select
Prescaler
Factor
Prescaler
Select
Prescaler
Factor
01100
6.5
11100
14.5
01101
7
11101
15
01110
7.5
11110
15.5
01111
8
11111
16
A prescaler factor of zero corresponds to “no clock.” The “no
clock” condition is the USART power down mode, in which
the USART clock is turned off to reduce power consumption.
The application program should select the “no clock” condition before entering a new baud rate. Otherwise, it could
cause incorrect data to be received or transmitted. The
UnPSR register must contain a value other than zero when
an external clock is used at CKXn.
Baud Rate Generator
The Baud Rate Generator creates the basic baud clock from
the system clock. The system clock is passed through a twostage divider chain consisting of a 5-bit baud rate prescaler
(UnPSC) and an 11-bit baud rate divisor (UnDIV).
The relationship between the 5-bit prescaler select (UnPSC)
setting and the prescaler factors is shown in Table19.
In asynchronous mode, the baud rate is calculated by:
SYS_CLK
BR = ------------------------------( 16 × N × P )
Table 19 Prescaler Factors
Prescaler
Select
Prescaler
Factor
Prescaler
Select
Prescaler
Factor
00000
1
10000
8.5
00001
1
10001
9
00010
1.5
10010
9.5
00011
2
10011
10
00100
2.5
10100
10.5
Prescaler Factors
where BR is the baud rate, SYS_CLK is the system clock, N
is the value of the baud rate divisor + 1, and P is the prescaler
divide factor selected by the value in the UnPSR register.
The divide by 16 is performed because in the asynchronous
mode, the input frequency to the USART is 16 times the baud
rate. In synchronous mode, the input clock to the USART
equals the baud rate.
18.2.7
Interrupts
The USART is capable of generating interrupts on:
00101
3
10101
11
00110
3.5
10110
11.5
00111
4
10111
12
01000
4.5
11000
12.5
01001
5
11001
13
01010
5.5
11010
13.5
01011
6
11011
14
• Receive Buffer Full
• Receive Error
• Transmit Buffer Empty
Figure38 shows a diagram of the interrupt sources and associated enable bits.
UnFE
UnEEI
UnDOE
UnPE
UnERR
RX
Interrupt
UnRBF
UnERI
TX
UnTBE
Interrupt
UnETI
Figure 38. USART Interrupts
www.national.com
74
The interrupts can be individually enabled or disabled using
the Enable Transmit Interrupt (UnETI), Enable Receive Interrupt (UnERI) and Enable Receive Error Interrupt (UnEER)
bits in the UnICTRL register.
the upper three bits of the baud rate divisor. This register is
cleared upon reset. The register format is shown below.
7
A transmit interrupt is generated when both the UnTBE and
UnETI bits are set. To remove this interrupt, software must either disable the interrupt by clearing the UnETI bit or write to
the UnTBUF register (thus clearing the UnTBE bit).
18.3.4
A line break is detected if RDXn remains low for 10 bit times
or longer after a missing stop bit is detected.
Parity Generation and Detection
18.3.5
5
4
3
2
1
0
Baud Rate Divisor (bits 7-0). This field contains
the eight lowest-order bits of the USART baud
rate divisor used in the second stage of the
two-stage divider chain. The three highest-order bits are contained in the UnPSR register.
The divisor value used is the 11-bit UnDIV value plus 1.
USART Frame Select Register (UnFRS)
The USART Frame Select Register is a byte-wide, read/write
register that controls the frame format, including the number
of data bits, number of stop bits, and parity type. This register
is cleared upon reset. The register format is shown below.
USART Receive Data Buffer (UnRBUF)
USART Transmit Data Buffer (UnTBUF)
USART Baud Rate Prescaler Register (UnPSR)
USART Baud Rate Divisor Register (UnBAUD)
USART Frame Select Register (UnFRS)
USART Mode Select Register (UnMDSL)
USART Status Register (UnSTAT)
USART Interrupt Control Register (UnICTRL)
7
6
5
4
3
2
1
0
Reserved UnPEN UnPSEL UnXB9 UnSTP UnCHAR
UnCHAR
Character Frame Format. This 2-bit field selects the number of data bits per frame, not including the parity bit, as follows:
00 = eight data bits per frame
01 = seven data bits per frame
10 = nine data bits per frame
11 = loopback mode; nine data bits per frame
USART Receive Data Buffer (UnRBUF)
The USART Receive Data Buffer is a byte-wide, read/write
register used to receive each data byte.
USART Transmit Data Buffer (UnTBUF)
UnSTP
The USART Transmit Data Buffer is a byte-wide, read/write
register used to transmit each data byte.
18.3.3
6
UnDIV[7:0]
The software interacts with the USART by accessing the USART registers. There are eight such registers:
18.3.2
USART Baud Rate Divisor (UnBAUD)
7
USART REGISTERS
18.3.1
0
UnDIV8
UnDIV7 UnDIV6 UnDIV5 UnDIV4 UnDIV3 UnDIV2 UnDIV1 UnDIV0
Parity is only generated or checked with the 7-bit and 8-bit
data formats. It is not generated or checked in the diagnostic
loopback mode, the attention mode, or in the normal mode
with the 9-bit data format. Parity generation and checking are
enabled and disabled via the PEN bit in the UnFRS register.
The UnPSEL bits in the UnFRS register are used to select
odd, even, mark, or space parity.
—
—
—
—
—
—
—
—
1
UnDIV9
The USART Baud Rate Divisor Register is a byte-wide, read/
write register that contains the lower eight bits of the baud
rate divisor. This register contents are unknown upon powerup and are left unchanged by a reset operation. The register
format is shown below.
Break Generation and Detection
A line break is generated when the BRK bit is set in the UnMDSL register. The TDXn line remains low until the program
resets the BRK bit.
18.3
2
UnDIV10
Prescaler. This 5-bit field specifies the prescaler value used for dividing the system clock in
the first stage of the two-stage divider chain.
For the prescaler factors corresponding to
each 5-bit value, see Table19.
UnDIV[10:8] Baud Rate Divisor (bits 10-8). This field contains the three highest-order bits (bits 10, 9,
and 8) of the USART baud rate divisor used in
the second stage of the two-stage divider
chain. The remaining bits of the baud rate divisor are contained in the UnBAUD register.
1. Both the UnRBF and UnERI bits are set. To remove this
interrupt, software must either disable the interrupt by
clearing the UnERI bit or read from the UnRBUF register
(thus clearing the UnRBF bit).
2. Both the UnERR and the UnEEI bits are set. To remove
this interrupt the software must either disable it by clearing the UnEEI bit or read the UnSTAT register (thus
clearing the UnERR bit).
18.2.9
3
UnPSC
A receive interrupt is generated on two conditions:
18.2.8
6 5 4
UnPSC
USART Baud Rate Prescaler (UnPSR)
UnXB9
The USART Baud Rate Prescaler Register is a byte-wide,
read/write register that contains the 5-bit clock prescaler and
75
Number of Stop Bits. This bit sets the number
of stop bits transmitted in each frame. If this bit
is 0, one stop bit is transmitted. If this bit is 1,
two stop bits are transmitted.
Transmit 9th Data Bit. This bit is the value of
the ninth data bit, either 0 or 1, transmitted
when the USART is configured to transmit nine
data bits per frame. It has no effect when the
USART is configured to transmit seven or eight
data bits per frame.
www.national.com
UnPSEL
Parity Select. This 2-bit field selects parity type
as follows:
UnFE
00 = odd parity
01 = even parity
10 = mark (0)
11 = space (1)
UnDOE
When the USART is configured to transmit nine
data bits per frame, the parity bit is omitted and
the UnPSEL field is ignored.
Parity Enable. This bit enables (1) or disables
(0) parity bit generation and parity checking.
When the USART is configured to transmit nine
data bits per frame, there is no parity bit and the
UnPEN bit is ignored.
UnPEN
18.3.6
UnERR
USART Mode Select Register (UnMDSL)
The USART Mode Select Register is a byte-wide, read/write
register that selects the clock source, synchronization mode,
attention mode, and line break generation. This register is
cleared upon reset. When the software writes to this register,
the reserved bits must be cleared to 0 for proper operation.
The register format is shown below.
7
6 5 4
Reserved
UnMOD
2
UnBRK
1
UnATN
0
UnMOD
Mode of Operation. Set to 0 for asynchronous
operation or 1 for synchronous operation.
Attention Mode. When set to 1, this bit selects
the attention mode of operation for the USART.
When cleared to 0, the attention mode is disabled. The hardware clears this bit after an address frame is received. An address frame is a
9-bit character with a 1 in the ninth bit position.
Force Transmission Break. Setting this bit to 1
causes the TDXn pin to go low. TDXn remains
low until the UnBRK bit is cleared to 0 by the
software.
Synchronous Clock Source. This bit controls
the clock source when the USART operates in
the synchronous mode (UnMOD=1). If the
UnCKS bit is set to 1, the USART operates
from an external clock provided on the CKXn
pin. If the UnCKS bit is cleared to 0, the USART
operates from the baud rate clock produced by
the USART on the CKXn pin. This bit is ignored
when the USART operates in the asynchronous mode.
UnATN
UnBRK
UnCKS
18.3.7
3
UnCKS
UnBKD
UnRB9
UnXMIP
18.3.8
7
7
UnEEI
UnTBE
USART Status Register (UnSTAT)
6
UnPE
5
4
3
UnRB9
UnBKD
UnERR
2
1
UnRBF
0
UnDOE UnFE UnPE
Parity Error. This bit is set to 1 when a parity error is detected within a received character. This
www.national.com
USART Interrupt Control Register (UnICTRL)
The USART Interrupt Control Register is a byte-wide register
that contains the receive and transmit interrupt status flags
(read-only bits) and the interrupt enable bits (read/write bits).
The register is set to 01 hex upon reset. The register format
is shown below.
The USART Status Register is a byte-wide, read-only register that contains the receive and transmit status bits. This
register is cleared upon reset. Any attempt by the software to
write to this register is ignored. The register format is shown
below.
Reserved UnXMIP
bit is automatically cleared to 0 by the hardware when the UnSTAT register is read.
Framing Error. This bit is set to 1 when the USART fails to receive a valid stop bit at the end
of a frame. This bit is automatically cleared to 0
by the hardware when the UnSTAT register is
read.
Data Overrun Error. This bit is set to 1 when a
new character is received and transferred to
the UnBUF register before the software has
read the previous character from UnBUF. This
bit is automatically cleared to 0 by the hardware when the UnSTAT register is read.
Error Status Flag. This bit is set when a parity,
framing, or overrun error occurs (any time that
the UnPE, UnFE, or UnDOE bit is set). It is automatically cleared to 0 by the hardware when
the UnPE, UnFE, and UnDOE bits are all 0.
Break Detect. This bit is set to 1 when a line
break condition occurs. This condition is detected if RDXn remains low for at least ten bit
times after a missing stop bit has been detected at the end of a frame.
The hardware automatically clears the UnBKD
bit upon read of the UnSTAT register, but only
if the break condition on RXDn no longer exists. If reading the UnSTAT register does not
clear the UnBKD bit because the break is still
actively driven on the line, the hardware clears
the bit as soon as the break condition no longer
exists (when RXDn returns to a high level).
Received 9th Data Bit. With the USART configured to operate in the 9-bit data format, this is
equal to the ninth data bit of the last frame received.
Transmit In Progress. The hardware sets this
bit to 1 when the USART is transmitting data
and clears it to 0 at the end of the last frame bit.
76
6
UnERI
5
UnETI
4 3 2
Reserved
1
UnRBF
0
UnTBE
Transmit Buffer Empty. This read-only bit is set
to 1 by the hardware when the USART transfers data from the UnTBUF register to the
transmit shift register for transmission. It is automatically cleared to 0 by the hardware on the
next write to the UnTBUF register.
Receive Buffer Full. This read-only bit is set by
the hardware when the USART has received a
complete data frame and has transferred the
data from the receive shift register to the UnRBUF register. It is automatically cleared to 0 by
the hardware when the UnRBUF register is
read.
UnETI
Enable Transmitter Interrupt. This read/write
bit, when set to 1, enables generation of an interrupt when the hardware sets the UnTBE bit.
Enable Receiver Interrupt. This read/write bit,
when set to 1, enables generation of an interrupt when the hardware sets the UnRBF bit.
Enable Receive Error Interrupt. This read/write
bit, when set to 1, enables generation of an interrupt when the hardware sets the UnERR bit
in the UnSTAT register.
UnERI
UnEEI
18.4
System
Clock
Desired
Baud Rate
20 MHz
19200
18.4.2
N
P
5 13
Actual
Baud Rate
Percent
Error
19230.769
0.16
Baud Rate in Synchronous Mode
The equation for calculating the baud rate in synchronous
mode is:
SYS_CLK
BR = ---------------------------(2 × N × P)
BAUD RATE CALCULATIONS
where BR is the baud rate, SYS_CLK is the system clock, N
is the value of the baud rate divisor + 1, and P is the prescaler
divide factor selected by the value in the UnPSR register.
The USART baud rate is determined by the system clock frequency and the values programmed into the UnPSR and UnBAUD registers. Unless the system clock frequency is an
exact multiple of the desired baud rate, there will be a small
amount of error in the resulting baud rate clock.
Use the same procedure to determine the values of N and P
as in the asynchronous mode. In this case, however, only integer prescaler values are allowed.
The method of baud rate calculation depends on whether the
USART is configured to operate in the asynchronous or synchronous mode.
18.4.1
Baud Rate in Asynchronous Mode
The equation for calculating the baud rate in asynchronous
mode is:
SYS_CLK
BR = ------------------------------( 16 × N × P )
where BR is the baud rate, SYS_CLK is the system clock, N
is the value of the baud rate divisor + 1, and P is the prescaler
divide factor selected by the value in the UnPSR register.
Assuming a system clock of 5 MHz and a desired baud rate
of 9600, the NxP term according to the equation above is:
( 5 ×106 )
N × P = ----------------------------- = 32.552
( 16 × 9600 )
The NxP term is then divided by each Prescaler Factor from
Table 19 to obtain a value closest to an integer. The factor for
this example is 6.5.
32.552
N = ---------------- = 5.008 (N = 5)
6.5
The baud rate register is programmed with a baud rate divisor of 4 (N = baud rate divisor +1). This produces a baud
clock of:
6
( 5×10 )
BR = --------------------------------- = 9615.385
(16 × 5 × 6.5 )
9615.385 – 9600) = 0.16
%error = (--------------------------------------------9600
Note that the percent error is much lower than would be possible without the non-integer prescaler factor. Refer to the table below for more examples.
System
Clock
Desired
Baud Rate
Actual
Baud Rate
Percent
Error
4 MHz
9600
2 13
9615.385
0.16
5 MHz
9600
5 6.5
9615.385
0.16
10 MHz
19200
5 6.5
19230.769
0.16
N
P
77
www.national.com
19.0 ACCESS.bus Interface
The ACCESS.bus interface module (ACB) is a two wire serial
interface compatible with the ACCESS.bus physical layer. It
permits easy interfacing to a wide range of low-cost memories and I/O devices, including: EEPROMs, SRAMs, timers,
A/D converters, D/A converters, clock chips and peripheral
drivers. It is also compatible with Intel’s SMBus and Philips’
I2 C bus. The module can be configured as a bus master or
slave, and can maintain bi-directional communications with
both multiple master and slave devices.
SDA
SCL
Data Line
Stable:
Data Valid
This section presents an overview of the bus protocol, and its
implementation by the module.
ACCESS.bus, SMBus and I2 C compliant
ACCESS.bus master and slave
Supports polling and interrupt controlled operation
Generate a wake-up signal on detection of a Start Condition, while in power-down mode
— Optional internal pull-up on SDA and SCL pins
—
—
—
—
19.1
Figure 39. Bit Transfer
Each data transaction is composed of a Start Condition, a
number of byte transfers (set by the software), and a Stop
Condition to terminate the transaction. Each byte is transferred with the most significant bit first, and after each byte (8
bits), an Acknowledge signal must follow.
ACB PROTOCOL OVERVIEW
At each clock cycle, the slave can stall the master while it
handles the previous data, or prepares new data. This can be
done for each bit transferred or on a byte boundary by the
slave holding SCL low to extend the clock-low period. Typically, slaves extend the first clock cycle of a transfer if a byte
read has not yet been stored, or if the next byte to be transmitted is not yet ready. Some microcontrollers with limited
hardware support for ACESS.bus extend the access after
each bit, thus allowing the software time to handle this bit.
The ACCESS.bus protocol uses a two-wire interface for bidirectional communications between the ICs connected to
the bus. The two interface lines are the Serial Data Line
(SDA), and the Serial Clock Line (SCL). These lines should
be connected to a positive supply, via a pull-up resistor, and
remain HIGH even when the bus is idle.
The ACCESS.bus protocol supports multiple master and
slave transmitters and receivers. Each IC has a unique address and can operate as a transmitter or a receiver (though
some peripherals are only receivers).
Start and Stop
The ACCESS.bus master generates Start and Stop Conditions (control codes). After a Start Condition is generated the
bus is considered busy and it retains this status until a certain
time after a Stop Condition is generated. A high-to-low transition of the data line (SDA) while the clock (SCL) is high indicates a Start Condition. A low-to-high transition of the SDA
line while the SCL is high indicates a Stop Condition
(Figure40).
During data transactions, the master device initiates the
transaction, generates the clock signal and terminates the
transaction. For example, when the ACB initiates a data
transaction with an attached ACCESS.bus compliant peripheral, the ACB becomes the master. When the peripheral responds and transmits data to the ACB, their master/slave
(data transaction initiator and clock generator) relationship is
unchanged, even though their transmitter/receiver functions
are reversed.
19.1.1
Change
of Data
Allowed
Data Transactions
SDA
One data bit is transferred during each clock pulse. Data is
sampled during the high state of the serial clock (SCL). Consequently, throughout the clock’s high period, the data should
remain stable (see Figure 39). Any changes on the SDA line
during the high state of the SCL and in the middle of a transaction aborts the current transaction. New data should be
sent during the low SCL state. This protocol permits a single
data line to transfer both command/control information and
data using the synchronous serial clock.
SCL
S
P
Start
Condition
Stop
Condition
Figure 40.
Start and Stop Conditions
In addition to the first Start Condition, a repeated Start Condition can be generated in the middle of a transaction. This
allows another device to be accessed, or a change in the direction of the data transfer.
Acknowledge Cycle
The Acknowledge Cycle consists of two signals: the acknowledge clock pulse the master sends with each byte
www.national.com
78
transferred, and the acknowledge signal sent by the receiving device (Figure 41).
The address is the first seven bits after a Start Condition. The
direction of the data transfer (R/W) depends on the bit sent
after the address — the eighth bit. A low-to-high transition
during a SCL high period indicates the Stop Condition, and
ends the transaction (Figure 43).
Acknowledgment
Signal From Receiver
SDA
MSB
SCL
S
1
2 3-6
7
8
9
ACK
1
2
3-8
9
ACK
Start
Condition
SDA
P
Stop
Condition
Byte Complete
Clock Line Held
Low by Receiver
While Interrupt
is Serviced
I nterrupt Within
Receiver
Figure 41.
SCL
1-7
8
9
Address R/ W ACK
1-7
Data
8
9
ACK
1-7
8
Data
9
P
ACK
Stop
Condition
Start
Condition
ACCESS.bus Data Transaction
Figure 43. A Complete ACCESS.bus Data Transaction
The master generates the acknowledge clock pulse on the
ninth clock pulse of the byte transfer. The transmitter releases the SDA line (permits it to go high) to allow the receiver to
send the acknowledge signal. The receiver must pull down
the SDA line during the acknowledge clock pulse, thus signalling the correct reception of the last data byte, and its
readiness to receive the next byte. Figure 42 illustrates the
acknowledge cycle.
When the address is sent, each device in the system compares this address with its own. If there is a match, the device
considers itself addressed and sends an acknowledge signal. Depending upon the state of the R/W bit (1:read,
0:write), the device acts as a transmitter or a receiver.
The I2C bus protocol allows sending a general call address
to all slaves connected to the bus. The first byte sent specifies the general call address (00 16 ) and the second byte
specifies the meaning of the general call (for example, “Write
slave address by software only”). Those slaves that require
the data acknowledge the call and become slave receivers;
the other slaves ignore the call.
Data Output
by
Transmitter
Transmitter Stays Off
the Bus During the
Acknowledgment Clock
Data Output
by
Receiver
S
Arbitration on the Bus
Acknowledgment
Signal From Receiver
SCL
S
1
2 3-6
7
8
Multiple master devices on the bus, require arbitration between their conflicting bus-access demands. Control of the
bus is initially determined according to address bits and clock
cycle. If the masters are trying to address the same IC, data
comparisons determine the outcome of this arbitration. In
master mode, the device immediately aborts a transaction if
the value sampled on the SDA lines differs from the value
driven by the device. (Exceptions to this rule are SDA while
receiving data; in these cases the lines may be driven low by
the slave without causing an abort).
9
Start
Condition
Figure 42. ACCESS.bus Acknowledge Cycle
The master generates an acknowledge clock pulse after
each byte transfer. The receiver sends an acknowledge signal after every byte received.
The SCL signal is monitored for clock synchronization purpose and allow the slave to stall the bus. The actual clock period will be the one set by the master with the longest clock
period or by the slave stall period. The clock high period is
determined by the master with the shortest clock high period.
There are two exceptions to the “acknowledge after every
byte” rule.
1. When the master is the receiver, it must indicate to the
transmitter an end of data by not-acknowledging (“negative acknowledge”) the last byte clocked out of the
slave. This “negative acknowledge” still includes the acknowledge clock pulse (generated by the master), but
the SDA line is not pulled down.
2. When the receiver is full, otherwise occupied, or a problem has occurred, it sends a negative acknowledge to
indicate that it can not accept additional data bytes.
When an abort occurs during the address transmission, the
master that identify the conflict, give-up the bus and should
switch to slave mode and continue to sample SDA to see if it
is being addressed by the winning master on the ACCESS.bus.
19.2
Addressing Transfer Formats
ACB FUNCTIONAL DESCRIPTION
The ACB module provides the physical layer for an ACCESS.bus compliant serial interface. The module is configurable as either a master or slave device. As a slave device,
the ACB module may issue a request to become the bus
master.
Each device on the bus has a unique address. Before any
data is transmitted, the master transmits the address of the
slave being addressed. The slave device should send an acknowledge signal on the SDA line, once it recognizes its address.
79
www.national.com
19.2.1
Master Mode
6. Check that both ACBST.BER and ACBST.NEGACK are
cleared. If the ACBCTL1.INTEN bit is set, an interrupt is
generated when either ACBST.BER or ACBST.NEGACK is set.
An ACCESS.bus transaction starts with a master device requesting bus mastership. It sends a Start Condition, followed
by the address of the device it wants to access. If this transaction is successfully completed, the software can assume
that the device has become the bus master.
Master Transmit
After becoming the bus master, the device can start transmitting data on the ACCESS.bus.
For a device to become the bus master, the software should
perform the following steps:
To transmit a byte, the software should:
1. Set ACBCTL1.START, and configure ACBCTL1.INTEN
to the desired operation mode (Polling or Interrupt). This
causes the ACB to issue a Start Condition on the ACCESS.bus, as soon as the ACCESS.bus is free
(ACBCST.BB=0). It then stalls the bus by holding SCL
low.
2. If a bus conflict is detected, (i.e., some other device pulls
down the SCL signal before this device does), ACBST.BER is set.
3. If there is no bus conflict, ACBST.MASTER and ACBST.SDAST are set.
4. If ACBCTL1.INTEN is set, and either ACBST.BER or
ACBST.SDAST is set, an interrupt is sent to the ICU.
1. Check that the BER and NEGACK bits in ACBST are
cleared and ACBST.SDAST is set. Also, if
ACBCTL1.STASTRE is set, check that ACBST.STASTR
is cleared.
2. Write the data byte to be transmitted to the ACBSDA
register.
When the slave responds with a negative acknowledge, the
ACBST.NEGACK bit is set and the ACBST.SDAST bit remains cleared. In this case, if ACBCTL1.INTEN is set, an interrupt is sent to the core.
Master Receive
Sending the Address Byte
After becoming the bus master, the device can start receiving
data on the ACCESS.bus.
Once this device is the active master of the ACCESS.bus
(ACBST.MASTER is set), it can send the address on the bus.
To receive a byte, the software should:
1. Check that ACBST.SDAST is set and ACBST.BER is
cleared. Also, if ACBCTL1.STASTRE is set, check that
ACBST.STASTR is cleared.
2. Set the ACBCTL1.ACK bit to 1, if the next byte is the last
byte that should be read. This causes a negative acknowledge to be sent.
3. Read the data byte from the ACBSDA register.
The address sent should not be this device’s own address as
defined in ACBADDR.ADDR if ACBADDR.SAEN is set, nor
should it be the global call address if ACBST.GCMTCH is set.
To send the address byte use the following sequence:
1. Configure the ACBCTL1.INTEN bit according to the desired operation mode. For a receive transaction where
the software wants only one byte of data, it should set
the ACBCTL1.ACK bit.
If only an address needs to be sent, set (1) the
ACBCTL1.STASTRE bit.
2. Write the address byte (7-bit target device address), and
the direction bit, to the ACBSDA register. This causes
the module to generate a transaction. At the end of this
transaction, the acknowledge bit received is copied to
ACBST.NEGACK. During the transaction the SDA and
SCL lines are continuously checked for conflict with other devices. If a conflict is detected, the transaction is
aborted, ACBST.BER is set, and ACBST.MASTER is
cleared.
3. If ACBCTL1.STASTRE is set, and the transaction was
successfully completed (i.e., both ACBST.BER and
ACBST.NEGACK are cleared), ACBST.STASTR is set.
In this case, the ACB stalls any further ACCESS.bus operations (i.e., holds SCL low). If ACBCTL1.INTE is set, it
also sends an interrupt to the core.
4. If the requested direction is transmit, and the start transaction was completed successfully (i.e., neither ACBST.NEGACK nor ACBST.BER is set, and no other
master has accessed the device), ACBST.SDAST is set
to indicate that the module awaits attention.
5. If the requested direction is receive, the start transaction
was completed successfully and ACBCTL1.STASTRE is
cleared, the module starts receiving the first byte automatically.
www.national.com
Master Stop
A Stop Condition may be issued only when this device is the
active bus master (ACBST.MASTRER=1). To end a transaction, set (1) ACBCTL1.STOP before clearing the current stall
flag (i.e., ACBST.SDAST, ACBST.NEGACK or ACBST.STASTR). This causes the module to send a Stop Condition immediately, and clear ACBCTL1.STOP.
Master Bus Stall
The ACB module can stall the ACCESS.bus between transfers while waiting for the core’s response. The ACCESS.bus
is stalled by holding the SCL signal low after the acknowledge cycle. Note that this is interpreted as the beginning of
the following bus operation. The user must make sure that
the next operation is prepared before the flag that causes the
bus stall is cleared.
The flags that can cause a stall in master mode are:
— Negative acknowledge after sending a byte (ACBSTNEGACK=1).
— ACBST.SDAST bit is set.
— If ACBCTL1.STASTRE=1, after a successful start
(ACBST.STASTR=1).
Repeated Start
A repeated start is performed when this device is already the
bus master (ACBST.MASTER is set). In this case the ACCESS.bus is stalled and the ACB is awaiting the core handling due to: negative acknowledge (ACBST.NEGACK=1),
80
empty buffer (ACBST.SDAST=1) and/or a stop after start
(ACBST.STASTR=1).
ACBST.SDAST is set to indicate that the buffer is empty.
— If ACBCTL1.INTEN is set, an interrupt is generated if
both the INTEN and NMINTE bits in ACBCTL1 registers are set.
— The software then reads the ACBST.XMIT bit to identify the direction requested by the master device. It
clears the ACBST.NMATCH bit so future byte transfers
are identified as data bytes.
For a repeated start:
— Set the ACBCTL1.START bit.
— In master receive mode, read the last data item from
ACBSDA.
— Follow the address send sequence, as described in
“Sending the Address Byte” on page 80.
— If the ACB was awaiting handling due to ACBST.STASTR=1, clear it only after writing the requested address
and direction to ACBSDA.
Slave Receive and Transmit
Slave Receive and Transmit are performed after a match is
detected and the data transfer direction is identified. After a
byte transfer the ACB extend the acknowledge clock until the
software reads or writes the ACBSDA register. The receive
and transmit sequence are identical to those used in the
master routine.
Master Error Detections
The ACB detects illegal Start or Stop Conditions (i.e., a Start
or Stop Condition within the data transfer, or the acknowledge cycle) and a conflict on the data lines of the ACCESS.bus. If an illegal action is detected, BER is set, and the
MASTER mode is exited (MASTER is cleared).
Slave Bus Stall
When operating as a slave, this device stalls the ACCESS.bus by extending the first clock cycle of a transaction
in the following cases:
Bus Idle Error Recovery
When a request to become the active bus master or a restart
operation fails, the ACBST.BER bit is set to indicate the error.
In some cases, both this device and the other device may
identify the failure and leave the bus idle. In this case, the
start sequence may not be completed and the ACCESS.bus
may remain deadlocked forever.
— ACBST.SDAST is set.
— ACBST.NMATCH, and ACBCTL1.NMINTE are set.
Slave Error Detections
The ACB detects illegal Start and Stop Conditions on the ACCESS.bus (i.e., a Start or Stop Condition within the data
transfer or the acknowledge cycle). When an illegal Start or
Stop Condition is detected, the BER bit is set and MATCH
and GMATCH are cleared, setting the module to be an unaddressed slave.
To recover from deadlock, use the following sequence:
1. Clear the ACBST.BER bit and ACBCST.BB bit.
2. Wait for a time-out period to check that there is no other
active master on the bus (i.e., ACBCST.BB remains
cleared).
3. Disable, and re-enable the ACB to put it in the non-addressed slave mode.
4. At this point some of the slaves may not identify the bus
error. To recover, the ACB becomes the bus master by
issuing a Start Condition and sends an address field;
then issue a Stop Condition to synchronize all the
slaves.
19.2.2
Power Down
When this device is in Power Save, Idle, or Halt mode, the
ACB module is not active but retains its status. If the ACB is
enabled (ACBCTL2.ENABLE=1) on detection of a Start Condition, a wake-up signal is issued to the MIWU module. Use
this signal to switch this device to Active mode.
The ACB module cannot check the address byte following
the start condition that has awaken this device for a match.
The ACB responds with a negative acknowledge, and the device should re-send both the Start Condition and the address
after this device has had time to wake up.
Slave Mode
A slave device waits in Idle mode for a master to initiate a bus
transaction. Whenever the ACB is enabled, and it is not acting as a master (i.e., ACBST.MASTER is cleared), it acts as
a slave device.
Check that the ACBCST.BUSY bit is inactive before entering
Power Save, Idle or Halt mode. This guarantees that this device does not acknowledge an address sent, and stop responding later.
Once a Start Condition on the bus is detected, this device
checks whether the address sent by the current master
matches either:
— The ACBADDR.ADDR value if ACBADDR.SAEN is
set.
— The general call address if ACBCTL1.GCM is set.
19.2.3
SDA and SCL Pins Configuration
The SDA and SCL are open-drain signals. For more information, see the I/O configuration section.
This match is checked even when ACBST.MASTER is set. If
a bus conflict (on SDA or SCL) is detected, ACBST.BER is
set, ACBST.MASTER is cleared and this device continues to
search the received message for a match.
19.2.4
ACB Clock Frequency Configuration
The ACB module permits the user to set the clock frequency
used for the ACCESS.bus clock. The clock is set by the
ACBCTL2.SCLFRQ field. This field determines the SCL
clock period used by this device. This clock low period may
be extended by stall periods initiated by the ACB module or
by another ACCESS.bus device. In case of a conflict with another bus master, a shorter clock high period may be forced
by the other bus master until the conflict is resolved.
If an address match, or a global match, is detected:
— This device asserts its data pin during the acknowledge cycle.
— The ACBCST.MATCH and ACBST.NMATCH bits are
set. If ACBST.XMIT is set (i.e., slave transmit mode),
81
www.national.com
19.3
The ACCESS.bus Interface uses the following registers:
—
—
—
—
—
—
cleared when the module is disabled. Writing 0
to NEGACK is ignored.
Bus Error. BER is set by the hardware when a
Start or Stop Condition is detected during data
transfer (i.e., Start or Stop Condition during the
transfer of bits 2 through 8 and acknowledge
cycle), or when an arbitration problem is detected. Writing 1 to BER clears it. It is also
cleared when the module is disabled. Writing 0
to BER is ignored.
SDA Status. When set, this bit indicates that
the SDA data register is waiting for data (transmit - master or slave) or holds data that should
be read (receive - master or slave). This bit is
cleared when reading from the ACBSDA register during a receive, or when written to during a
transmit. When ACBCTL1.START is set, reading ACBSDA register does not clear SDAST.
This enables the ACB to send a repeated start
in master receive mode.
Slave Stop. If set, SLVSTP indicates that a
Stop Condition was detected after a slave
transfer (i.e., after a slave transfer in which
MATCH or GCMATCH is set). Writing 1 to
SLVSTP clears it. It is also cleared when the
module is disabled. Writing 0 to SLVSTP is ignored.
ACB REGISTERS
BER
ACB Serial Data Register (ACBSDA)
ACB Status Register (ACBST)
ACB Status Control Register (ACBCST)
ACB Control 1 Register (ACBCTL1)
ACB Control 2 Register (ACBCTL2)
ACB Own Address Register (ACBADDR)
19.3.1
ACB Serial Data Register (ACBSDA)
SDAST
The ACB Serial Data Register (ACBSDA) is a byte-wide,
read/write shift register used to transmit and receive data.
The most significant bit is transmitted (received) first and the
least significant bit is transmitted (received) last. Reading or
writing to the ACBSDA register is allowed when ACBST.SDAST is set; or for repeated starts after setting the
START bit. An attempt to access the register in other cases
produces unpredictable results.
7
0
SLVSTP
DATA
19.3.2
ACB Status Register (ACBST)
The ACB Status Register (ACBST) is a byte-wide, read-only
register that maintains current ACB status. Upon reset, and
when the module is disabled, ACBST is cleared (0016 ).
7
SLVST
P
6
5
4
3
SDAST BER NEGACK STASTR
XMIT
MASTER
NMATCH
STASTR
NEGACK
2
1
0
NMATC
H
MASTER
XMIT
19.3.3
ACB Control Status Register (ACBCST) is a byte-wide, read/
write register that maintains current ACB status. Upon reset
and when the module is disabled, the non-reserved bits of
ACBCST are cleared (0).
Direction Bit. The XMIT bit is set when the ACB
module is currently in master/slave transmit
mode. Otherwise it is cleared.
MASTER. When set, the MASTER bit indicates
that the module is currently in master mode. It
is set when a request for bus mastership succeeds. It is cleared upon arbitration loss (BER
is set) or the recognition of a Stop Condition.
New match. The NMATCH bit is set when the
address byte following a Start Condition, or repeated starts, causes a match or a global-call
match. NMATCH is cleared when 1 is written to
it. Writing 0 to NMATCH is ignored. If
ACBCTL1.INTEN is set, an interrupt is sent
when this bit is set.
Stall After Start. The STASTR bit is set by the
successful completion of an address sending
(i.e., a Start Condition sent without a bus error,
or negative acknowledge) if ACBCTL1.STASTRE is set. This bit is ignored in slave mode.
When STASTR is set, it stalls the ACCESS.bus
by pulling down the SCL line, and suspends
any other action on the bus (e.g., receives first
byte in master receive mode). In addition, if
ACBCTL1.INTEN is set, it also sends an interrupt to the core. Writing 1 to STASTR clears it.
It is also cleared when the module is disabled.
Writing 0 to STASTR has no effect.
Negative acknowledge. This bit is set by hardware when a transmission is not acknowledged
on the ninth clock. (In this case SDAST is not
set.) Writing 1 to NEGACK clears it. It is also
www.national.com
ACB Control Status Register (ACBCST)
7
6
Reserved
BUSY
BB
82
5
4
TGSCL
TSDA
3
2
GCMTCH MATCH
1
0
BB
BUSY
BUSY. When BUSY is set, it indicates that the
ACB module is:
• Generating a Start Condition
• In Master mode (ACBST.MASTER is set)
• In Slave mode (ACBCST.MATCH or
ACBCST.GCMTCH is set)
• In the period between detecting a Start and
completing the reception of the address
byte. After this, the ACB either becomes
not busy or enters slave mode.
The BUSY bit is cleared by the completion of
any of the above states, and by disabling the
module. BUSY is a read only bit. It should always be written with 0.
Bus Busy When set, BB indicates the bus is
busy. It is set when the bus is active (i.e., a low
level on either SDA or SCL), or by a Start Condition. It is cleared when the module is disabled, upon detection of a Stop Condition, or
when writing 1 to this bit. See “Usage Hints” on
page 84 for a description of the use of this bit.
This bit should be set when either SDA or SCL
are low. This should be done by sampling the
SDA and SCL lines continuously and, setting
the bit if one of them is low. The bit remains set
MATCH
GCMTCH
TSDA
TGSCL
19.3.4
until cleared by a STOP condition or a one is
written to it.
Address Match. In slave mode, MATCH is set
when ACBADDR.SAEN is set and the first seven bits of the address byte (the first byte transferred after a Start Condition) matches the 7-bit
address in the ACBADDR register. It is cleared
by Start Condition, repeated start and Stop
Condition (including illegal Start or Stop Condition).
Global Call Match bit. In slave mode, GCMTCH
is set when ACBCTL1.GCMEN is set and the
address byte (the first byte transferred after a
Start Condition) is 0016 . It is cleared by Start
Condition, repeated Start and Stop Condition
(including illegal Start or Stop Condition).
Test SDA Line. Reads the current value of the
SDA line. This bit can be used while recovering
from an error condition in which the SDA line is
constantly pulled low by a slave that went out
of synch. This bit is a read-only bit. Data written
to it is ignored.
Toggle SCL Line. This bit enables toggling the
SCL line during the process of error recovery.
When the SDA line is low, writing 1 to this bit
toggles the SCL line for one cycle. Writing 1 to
TGSCL when SDA is high is ignored. The bit is
cleared when the clock toggle is completed.
STOP
INTEN
ACK
ACB Control 1 Register (ACBCTL1)
ACB Control 1 Register (ACBCTL1) is a byte-wide, read/
write register that configures and controls the ACB module.
Upon reset and while the module is disabled (ACBCTL2.ENABLE=0), the ACBCTL1 is cleared (00 16 ).
7
6
STASTRE
NMINTE
START
5
4
3
2
GCMEN ACK Reserved INTEN
1
0
STOP
START
GCMEN
NMINTE
START. This bit is set when a Start Condition
needs to be generated on the ACCESS.bus.
The START bit is cleared when the Start Condition is sent, or upon detection of a Bus Error
(ACBST.BER=1). This bit should be set only
when in Master mode, or when requesting
Master mode.
If this device is not the active master of the bus
(ACBST.MASTER=0), setting START generates a Start Condition as soon as the
ACCESS.bus is free (ACBCST.BB=0). An address send sequence should then be performed.
If this device is the active master of the bus
(ACBST.MASTER=1), when START is set, a
write to the ACBSDA register generates a Start
Condition, then the ACBSDA data is transmitted as the slave’s address and the requested
transfer direction.
This case is a repeated Start Condition. It may
be used to switch the direction of the data flow
between the master and the slave, or to choose
another slave device without using a Stop Condition in between.
STASTRE
19.3.5
STOP. In master mode, setting this bit generates a Stop Condition that completes or aborts
the current message transfer. This bit clears itself after the STOP is issued.
Interrupt Enable. When INTEN is cleared ACB
interrupt is disabled. When INTEN is set, interrupts are enabled. An interrupt is generated
(the interrupt signals to the ICU is high) upon
one of the following events:
• An address MATCH is detected (ACBST.NMATCH=1) and NMINTE is set.
• A Bus Error occurs (ACBST.BERR=1).
• Negative acknowledge after sending a byte
(ACBST.NEGACK=1).
• An interrupt is generated upon acknowledge of each transaction (same as the
hardware set of the ACBST.SDAST bit).
• In master mode if ACBCTL1.STASTRE=1,
after a successful start (ACBST.STASTR=1).
• Detection of a Stop Condition while in slave
receive mode (ACBST.SLVSTP=1).
Acknowledge bit. When acting as a receiver
(slave or master), this bit holds the value this
device sends during the next acknowledge cycle. Setting this bit to 1 instructs the transmitting device to stop sending data, since the
receiver either does not need, or cannot receive, any more data. This bit is cleared after
the first acknowledge cycle.
This bit is ignored when in transmit mode.
Global Call Match enable. When this bit is set,
it enables the match of an incoming address
byte to the general call address (Start Condition followed by address byte of 001 6) while the
ACB is in slave mode. When cleared, the ACB
does not respond to a global call.
New Match Interrupt Enable. Set NMINTE to
enable the interrupt on a new match (i.e., when
ACBST.NMATCH is set). The interrupt is issued only if ACBCTL1.INTEN is set.
Stall After Start Enable. When set enables the
stall after start mechanism. In such a case, the
ACB is stalled after the address byte. When
STASTRE is cleared, ACBST.STASTR is always cleared.
ACB Control 2 Register (ACBCTL2)
The ACB Control 2 register (ACBCTL2) is a byte-wide, read/
write register that enables/disables the module and determines ACB clock rate. Upon reset ACBCTL2 is set to 00 16 .
7
1
SCLFRQ
ENABLE
SCLFRQ
83
0
ENABLE
Enable. When this bit is set, the ACB module is
enabled. When the Enable bit is cleared, the
ACB module is disabled, ACBCTL1, ACBST
and ACBCST are cleared, and the clocks are
halted.
SCL Frequency. This field defines the SCL’s
period (low time and high time) when this de-
www.national.com
vice serves as a bus master. The clock low time
and high time are defined as follows:
4. In some cases the bus may get stuck with the SCL and/
or SDA lines active. A possible cause to this is an erroneous Start or Stop Conditions that occur in the middle
of a slave receive session.
tSCLl = tSCLh = 2*SCLFRQ*tCLK
Where t CLK is this device’s clock cycle when in
Active mode.
SCLFRQ may be programmed to values in the
range of 00010002 (810 ) through 11111112
(12710 ). Using any other value has unpredictable results.
19.3.6
When the SCL line is stuck active, there is nothing that
can be done, and it is the responsibility of the module
that holds the bus to release it.
In case of SDA line is stuck active, the ACB module enable the release of the bus by using the following sequence. Note that in normal cases SCL may be toggled
only by the bus master. This protocol is a recovery
scheme which is an exception that should be used only
in the case where there is no other master on the bus.
The recovery scheme is as follows:
ACB Own Address Register (ACBADDR)
ACB Own Address Register (ACBADDR) is a byte-wide,
read/write register that holds the module’s ACCESS.bus address. Reset value is undefined.
7
6
SAEN
ADDR
SAEN
19.4
0
ADDR
a. Disable and re-enable the module to set it into the
not addressed slave mode.
Own Address. Holds the 7-bit ACCESS.bus
address of this device. When in slave mode,
the first seven bits received after a Start Condition are compared to this field (first bit received
to bit-6, and the last to bit-0). If the address field
matches the received data and SAEN is set, a
match is declared.
Slave Address Enable. When set SAEN indicates that the ADDR field holds a valid address
and enables the match of ADDR to an incoming address byte. When cleared, the ACB does
not check for an address match.
b Set the ACBCTL1.START bit to make an attempt to
issue a Start Condition.
c. Check if the SDA line is active (low) by reading
ACBCST.TSDA bit. If it is active, issue a single SCL
cycle by writing 1 to ACBCST.TGSCL bit. If the SDA
line is not active, continue from step ‘e’.
d. Check if ACBST.MASTER is set, which indicates
that the Start Condition was sent. If not, repeat step
c and d until the SDA is released.
e. Clear the BB bit. This enables the START bit to be
executed. Continue according to “Bus Idle Error Recovery” on page 81.
USAGE HINTS
1. When the ACB is disabled the ACBCST.BB bit is
cleared. After enabling the ACB (ACBCTL2.ENABLE is
set to 1) in systems with more then one master, the bus
may be in the middle of a transaction with another device, which is not reflected by BB.
There is a need to allow the ACB to synchronize to the
bus activity status before issuing a request to become
the bus master, to prevent bus errors. Thus, before issuing a request to become the bus master for the first time,
the software should check that there is no activity on the
bus by checking the BB bit after the bus allowed time-out
period.
2. When waking up from power down, before checking
ACBCST.MATCH, use ACBCST.BUSY to make sure
that the address transaction is over.
3. The BB bit is intended to solve a deadlock in which two,
or more, devices detect a usage conflict on the bus and
both devices cease being bus masters at the same time.
In this situation, the BB bits of both devices are active
(because each deduces that there is another master
currently performing a transaction, while in fact no device is executing a transaction), and the bus would stay
locked until some device sends a ACBCTL1.STOP condition.
The ACBCST.BB bit allows the software to monitor bus
usage, so it can avoid sending a STOP signal in the middle of the transaction of some other device on the bus.
This bit detects whether the bus remains unused over a
certain period, while the BB bit is set.
www.national.com
84
20.0 CR16CAN Module
The CR16CAN device contains a FULL-CAN class, CAN
(Controller Area Network) serial bus interface for low/high
speed applications. It supports the reception and transmission of extended frames with 29-bit identifier, standard
frames with 11-bit identifier, applications that require a high
speed (up to 1MBit/s), and a low speed CAN interface with
CAN master capability. The data transfer between CAN and
the CPU is established by 15 message buffers, which can be
individually configured as receive or transmit buffers. Every
message buffer includes a status/control register which provides information about its current status and capabilities to
configure the buffer. All message buffers are able to generate
an interrupt upon the reception of a valid frame or the successful transmission of a frame. In addition, an interrupt on
bus errors can be generated.
•
•
•
•
•
•
An incoming message is only accepted if the message identifier passes one of two acceptance filtering masks. The filtering mask can be configured to receive a single message ID
per buffer or a group of IDs per receive buffer. One of the
buffers uses a separate message filtering procedure. This
provides the capability to establish a BASIC-CAN path. Remote transmission requests can be processed automatically
by automatic reconfiguration to a receiver after transmission
or by automated transmit scheduling upon reception. A priority decoder allows any buffer to have one of 16 transmit priorities including the highest or lowest absolute priority,
totaling 240 different transmit priorities.
— two filtering capabilities: global acceptance mask & individual buffer identifiers
— one of the buffers uses an independent acceptance filtering procedure
Programmable transmit priority
Interrupt capability
— one interrupt vector for all message buffers (receive/
transmit/error)
— each interrupt source can be enabled/disabled
16-bit counter with time stamp capability on successful reception or transmission of a message
Power Save capabilities with programmable Wake-Up
over the CAN bus (alternate source for the Multi-Input
Wake-Up module)
Push-Pull capability of the input/output pins
Diagnostic functions
— error identification
— loopback and listen-only features for test and initialization purposes
20.1
FUNCTIONAL DESCRIPTION
As shown in Figure44, the CR16CAN module is separated
into three blocks: the CAN core, the interface management
and a dual ported RAM containing the message buffers.
There are two dedicated device pins for the CR16CAN interface, CANTX as the transmit output and CANRX as the receive input.
A decided bit time counter (16-bit wide) is provided to support
real time applications. The contents of this counter is captured into the message buffer RAM upon reception or transmission. The counter can be synchronized via the CAN
network. This synchronization feature allows a reset of the
counter after the reception or transmission of a message in
buffer 0.
The CAN Core implements the basic CAN protocol features
such as bit-stuffing, CRC calculation/checking and error
management. It controls the transceiver logic and creates error signals according to the bus rules. In addition, it converts
the data stream from the CPU (parallel data) to the serial
CAN bus data.
The Interface Management is divided into the register block
and the interface management processor. The register block
provides the CAN Interface with control information from the
CPU and in turn provides the CPU with status information
from the CAN module. Additionally it generates the interrupt
to the CPU.
The CR16CAN is a fast core bus peripheral which allows single cycle byte or word read/write access. The CPU controls
the CR16CAN by modifying the various registers in the
CR16CAN register block. This includes the initialization of
the CAN baud rate, the CAN pin logic level, and the enable/
disable of the CR16CAN. A set of diagnostic features, such
as loopback, listen only and error identification, support the
development with the CR16CAN module and provide a sophisticated error management tool.
The interface management processor is a state machine executing the CPU’s transmission and reception commands
and controlling the data transfer between several message
buffers and RX/TX shift registers.
The CR16CAN implements the following features:
Fifteen Message Buffers are memory mapped into RAM to
transmit/receive data via the CAN bus. Eight 16-bit registers
belong to each buffer. One of the registers contains control
and status information about the message buffer configuration and the current state of the buffer. The other registers are
used for the message identifier, a maximum of up to eight
data bytes and the time stamp information. During the receive process the incoming message will be stored at first in
a hidden receive buffer until the message is valid. Then the
buffer contents will be copied into the first message buffer
which accepts the ID of the received message.
• CAN specification 2.0B
— standard data and remote frames
— extended data and remote frames
— 0 - 8 bytes data length
— programmable bit rate up to 1 Mbit/s
• 15 message buffers, each configurable as receive or
transmit buffers
— message buffers are 16-bit wide dual-port RAM
— one buffer may be used as BASIC-CAN path
• Remote Frame support
— automatic transmission after reception of a Remote
Transmission Request (RTR)
— auto receive after transmission of a RTR
• Acceptance filtering
85
www.national.com
CANTX
CANRX
wakeup
CTX
2:1
0 1
0 1
2:1
CAN CORE
Transceiver Logic
BTL, RX shift, TX shift, CRC
Error Management Logic
Bit Stream Processor
control
data
status
INTERFACE MANAGEMENT
Interface Management
Processor
RAM
control
Acceptance Filtering
TX/RX
Message Buffer 0
STATUS REGISTER
TX/RX
Message Buffer 1
BTL CONFIG
CAN PRESCALER
CONTROL
TX/RX
Message Buffer 14
ACCEPTANCE
MASKS
data
core bus
Figure 44.
www.national.com
Block Diagram CR16CAN Interface
86
CRX
20.2
remote frame, to transmit a data frame after the remote
frame has been completed.
BASIC CAN CONCEPTS
This section provides a generic overview of the basic concepts of the Controller Area Network (CAN).
Additional modules can be added to an existing network without a configuration change. These modules can either perform completely new functions requiring new data, or
process existing data to perform a new functionality.
The CAN protocol is a message based protocol that allows a
total of 2032 ( = 211-16) different messages in the standard
format and 512 million ( = 2 29 -16) different messages in the
extended frame format.
As the CAN network is message oriented, a message can be
used as a variable which is automatically updated by the controlling processor. If any module cannot process information,
it can send an overload frame.
Every CAN Frame is broadcasted on the common bus. Each
module receives every frame and filters out the frames which
are not required for the module's task. For example, if a
dashboard sends a request to switch on headlights, the CAN
module responsible for brake lights must not process this
message.
The CAN protocol allows several transmitting modules to
start a transmission at the same time as soon as they monitor
the bus to be idle. During the start of transmission, every
node monitors the bus line to detect whether its message is
overwritten by a message with a higher priority. As soon as a
transmitting module detects another module with a higher
priority accessing the bus, it stops transmitting its own frame
and switches to receive mode. For illustration, see Figure45.
A CAN master module has the ability to set a specific bit
called the “remote data request bit” (RTR) in a frame. Such a
message is also called “Remote Frame”. It causes another
module, either another master or a slave which accepts this
TxPIN
MODULE A
RxPIN
TxPIN
MODULE B
RxPIN
BUS LINE
RECESSIVE
DOMINANT
MODULE A SUSPENDS TRANSMISSION
Figure 45.
CAN message arbitration
If a data or remote frame loses arbitration on the bus due to
a higher-prioritized data or remote frame, or if it is destroyed
by an error frame, the transmitting module will automatically
retransmit it until the transmission was successful or the user
has canceled the transmit request.
Data and remote frames can be used in both standard and
extended frame format. If no message is being transmitted,
i.e., the bus is idle, the bus is kept at the ‘recessive’ level.
Remote and data frames are non-return to zero (NRZ) coded
with bit-stuffing in every bit field, which holds computable information for the interface, i.e., start of frame, arbitration field,
control field, data field (if present) and CRC field.
If a transmitted message loses arbitration, the CR16CAN will
restart transmission at the next possible time with the message which has the highest internal transmit priority.
20.2.1
Error and overload frames are also NRZ coded but without
bit-stuffing.
CAN Frame Formats
After five consecutive bits of the same value (including inserted stuff bits so that the stuffed bit stream will not have more
than five consecutive bits of the same value), a stuff bit of the
inverted value is inserted into the bit stream by the transmitter and deleted by the receiver. The following shows the
stuffed and destuffed bit stream for consecutive ones and zeros.
Communication via the CAN bus is basically established by
means of four different frame types:
—
—
—
—
data frame
remote frame
error frame
overload frame
87
www.national.com
original or destuffed bit
stream
stuffed bit stream
10000011111x
a
1000001111101x
The ACK field is two bits long and contains the ACK slot and
the ACK delimiter. The ACK slot is filled with a ‘recessive’ bit
by the transmitter. This bit is overwritten with a ‘dominant’ bit
by every receiver that has received a correct CRC sequence.
The second bit of the ACK field is a ‘recessive’ bit called the
acknowledge delimiter.
01111100000x
0111110000010x
a. x = {0,1}
The End of Frame field closes a data and a remote frame.
It consists of seven ‘recessive’ bits.
Frame Fields
Data and remote frames consist of the following different bit
fields:
—
—
—
—
—
—
—
Data Frame
The structure of a standard and extended data frame is
shown in Figure46.
Start of Frame
Arbitration Field
Control Field
Data Field
CRC Field
ACK Field
EOF Field
A CAN data frame consists of the following fields as previously described:
—
—
—
—
—
—
—
The Start of Frame indicates the beginning of data and remote frames. It consists of a single 'dominant' bit. A node is
only allowed to start transmission when the bus is idle. All
nodes have to synchronize to the leading edge (first edge after the bus was idle) caused by SOF of the node which starts
transmission first.
Remote Frame
Figure47 shows the structure of a standard and extended remote frame.
The Arbitration field consists of the identifier field and the
RTR (Remote Transmission Request) bit. For extended
frames there is also a SRR (Substitute Remote Request) and
a IDE (ID Extension) bit inserted between ID18 and ID17 of
the identifier field. The value of the RTR bit is 'dominant' in a
data frame and 'recessive' in a remote frame.
A remote frame is comprised of the following fields sections,
which is the same as a data frame (see Frame Fields on
page 88) except for the data field, which is not present.
—
—
—
—
—
—
The Control field consists of six bits. For standard frames it
starts with the ID Extension bit (IDE) and a reserved bit
(RB0). For extended frames the control field starts with two
reserved bits (RB1, RB0). These bits are followed by the 4bit Data Length Code (DLC).
Start of Frame (SOF)
Arbitration field + Extended Arbitration
Control field
Cyclic Redundancy Check field (CRC)
Acknowledgment field (ACK)
End of Frame (EOF)
Note that the DLC must have the same value as the corresponding data frame to prevent contention on the bus. The
RTR bit is ‘recessive’.
The CR16CAN receiver accepts all possible combinations of
the reserved bits (RB1, RB0). The transmitter must be configured to send only '0' bits.
The DLC indicates the number of bytes in the data field. It
consists of four bits. The data field can be of length zero. The
admissible number of data bytes for a data frame ranges
from 0 to 8.
The Data field consists of the data to be transferred within a
data frame. It can contain 0 to 8 bytes. A remote frame has
no data field.
The CRC field consists of the CRC sequence followed by the
CRC delimiter. The CRC sequence is derived by the transmitter from the modulo 2 division of the preceding bit fields,
starting with the SOF up to the end of the data field, excluding
stuff-bits, by the generator polynomial:
x1 5 + x14 + x10 + x 8 + x 7 + x 4 + x 3 + 1
The remainder of this division is the CRC sequence transmitted over the bus. On the receiver side, the module divides all
bit fields up to the CRC delimiter excluding stuff-bits, and
checks if the result is zero. This will then be interpreted as a
valid CRC. After the CRC sequence a single ‘recessive’ bit is
transmitted as the CRC delimiter.
www.national.com
Start of Frame (SOF)
Arbitration field + Extended Arbitration
Control field
Data field
Cyclic Redundancy Check field (CRC)
Acknowledgment field (ACK)
End of Frame (EOF)
88
Control Field
4
d
16
DATA FIELD
CRC FIELD
8
8
15
CRC
DLC0
11
≤ N ≤ 8)
CRC DEL
ACKNOWLEDGEMENT
ACK DEL
8N ( 0
Arbitration Field
ID0
RTR
IDE
RB0
DLC3
START OF FRAME
ID 10
STANDARD DATA FRAME (number of bits = 44 + 8N)
ddd
IDENTIFIER
10 ... 0
r
END OF
FRAME
r rrr rr rr
DATA
LENGTH
CODE
Bit Stuffing
18
d
IDENTIFIER
28 ... 18
rr
dd d
IDENTIFIER
17 ... 0
16
CRC FIELD
8
8
DLC0
4
ID0
RTR
RB1
RB0
DLC3
11
≤ N ≤ 8)
DATA FIELD
15
CRC
END OF
FRAME
CRC DEL
ACK
ACK DEL
8N ( 0
Control Field
Arbitration Field
ID18
SRR
IDE
ID17
START OF FRAME
ID28
EXTENDED DATA FRAME (number of bits = 64 + 8N)
r
rrr rr rrr
DATA
LENGTH
CODE
Bit Stuffing
Note:
d = dominant
r = recessive
Figure 46. CAN Data Frame (standard and extended)
89
www.national.com
CRC DEL
ACKNOWLEDGEMENT
ACK DEL
16
Arbitration Field
Control Field
CRC FIELD
4
d
15
CRC
DLC0
ID0
RTR
IDE
RB0
DLC3
11
ID3
START OF FRAME
ID 10
STANDARD REMOTE FRAME (number of bits = 44)
r dd
IDENTIFIER
10 ... 0
r
END OF
FRAME
r rrr rr r r
DATA
LENGTH
CODE
16
18
d
IDENTIFIER
28 ... 18
rr
rdd
IDENTIFIER
17 ... 0
15
DLC0
4
ID0
RTR
RB1
RB0
DLC3
11
CRC
r
DATA
LENGTH
CODE
Note:
d = dominant
r = recessive
Figure 47. CAN Remote Frame (standard and extended)
www.national.com
90
END OF
FRAME
CRC FIELD
CRC DEL
ACK
ACK DEL
Control Field
Arbitration Field
ID18
SRR
IDE
ID17
START OF FRAME
ID28
EXTENDED REMOTE FRAME (number of bits = 64)
r rr rrr r r
Error Frame
CRC error is detected, transmission of the error flag starts at
the bit following the acknowledge delimiter, unless an error
flag for a previous error condition has already been started.
As shown in Figure48, the Error Frame consists of the error
flag and the error delimiter bit fields. The error flag field is
built up from the various error flags of the different nodes.
Therefore, its length may vary from a minimum of six bits up
to a maximum of twelve bits depending on when a module
has detected the error. Whenever a bit error, stuff error, form
error, or acknowledgment error is detected by a node, this
node starts transmission of an error flag at the next bit. If a
If a device is in the error active state, it can send a ‘dominant’
error flag, whereas a error passive device is only allowed to
transmit ‘recessive’ error flags. This is done to prevent the
CAN bus from getting stuck due to a local defect. For the various CAN device states, please refer to Error Detection and
Management on page 92.
ERROR FRAME
DATA FRAME OR
REMOTE FRAME
6
≤6
8
ERROR
FLAG
ECHO
ERROR FLAG
ERROR
DELIMITER
d d d d d d d
INTER-FRAME SPACE OR
OVERLOAD FRAME
d d r r r r r r r r d
An error frame can start anywhere within a frame.
Note:
d = dominant
r = recessive
Figure 48. CAN Error Frame
Overload Frame
dition and start the transmission of an overload flag. After an
overload flag has been transmitted, the overload frame is
closed by the overload delimiter.
As shown in Figure49, an overload frame consists of the
overload flag and the overload delimiter bit fields. The bit
fields have the same length as the error frame field: six bits
for the overload flag and eight bits for the delimiter. The overload frame can only be sent after the end of frame (EOF) field
and in this way destroys the fixed form of the intermission
field. As a result, all other nodes also detect an overload con-
Note: The CR16CAN never initiates an overload frame due
to its inability to process an incoming message. However, it
is able to recognize and respond to overload frames initiated
by other devices.
OVERLOAD FRAME
6
END OF FRAME OR
OVERLOAD
ERROR DELIMITER OR
FLAG
OVERLOAD DELIMITER
8
OVERLOAD
DELIMITER
INTER-FRAME SPACE OR
ERROR FRAME
d d d d d d d r r r r r r r r
An overload frame can only start at the end of a frame.
Figure 49.
Note:
d = dominant
r = recessive
CAN Overload Frame
91
www.national.com
Interframe Space
not preceded by an interframe space; they can be transmitted as soon as the condition occurs. The interframe space
consists of a minimum of three bit fields depending on the error state of the node.
Data and remote frames are separated from every preceding
frame (data, remote, error and overload frames) by the interframe space (see Figure50). Error and overload frames are
3
8
SUSPEND
TRANSMIT
INT
BUS IDLE
ANY FRAME
r r r r r r r r r r r r r r r r r r r r
START OF FRAME
INTERFRAME SPACE
DATA FRAME OR
REMOTE FRAME
r r r d
Note:
d = dominant
r = recessive
INT = Intermission
Suspend Transmission is only for error passive nodes.
Figure 50.
20.2.2
CAN Interframe Space
Error Detection and Management
— Stuff Error
A stuff error is detected if the bit level after 6 consecutive bit times has not changed in a message field that
has to be coded according to the bit stuffing method.
— Form Error
A form error is detected, if a fixed frame bit (e.g., CRC
delimiter, ACK delimiter) does not have the specified
value. For a receiver, a ‘dominant’ bit during the last bit
of End of Frame does not constitute a frame error.
— Bit CRC Error
A CRC error is detected if the remainder of the CRC
calculation of a received CRC polynomial is non-zero.
— Acknowledgment Error
An acknowledgment error is detected whenever a
transmitting node does not get an acknowledgment
from any other node (i.e., when the transmitter does
not receive a ‘dominant’ bit during the ACK frame)
There are multiple mechanisms in the CAN protocol to detect
errors and inhibit erroneous modules from disabling all bus
activities. Each CAN module includes two error counters, a
receive and a transmit error counter, for error management.
Error Types
The following errors can be detected:
— Bit Error
A CAN device which is currently transmitting also monitors the bus. If the monitored bit value is different from
the transmitted bit value, a bit error is detected. However, the reception of a ‘dominant’ bit instead of a ‘recessive’ bit during the transmission of a passive error
flag, during the stuffed bit stream of the arbitration field
or during the acknowledge slot is not interpreted as a
bit error.
SYNC
external RESET or
enable CR16CAN
11 consecutive ‘recessive’ bits
received
(TEC OR REC) > 95
ERROR
ACTIVE
(TEC AND REC) < 96
(TEC OR REC) > 127
ERROR
WARNING
(TEC AND REC) < 128
ERROR
PASSIVE
TEC > 255
128 occurrences of
11 consecutive ‘recessive’ bits
BUS
OFF
Figure 51. CR16CAN Bus States
www.national.com
92
— Synchronize
Once the CR16CAN is enabled, it goes into a synchronization state to synchronize with the bus by waiting for
11 consecutive recessive bits. After that the CR16CAN
becomes error active and can participate in the bus
communication. This state must also be entered after
waking-up the device via the Multi-Input Wake-Up feature. See System Start-Up and Multi-Input Wake-Up
on page 116.
— Error active
An error active unit can participate in bus communication and may send an active (‘dominant’) error flag.
— Error Warning
The Error Warning state is a sub-state of Error Active
to indicate a heavily disturbed bus. The CR16CAN behaves as in Error Active mode. The device is reset into
the Error Active mode if the value of both counters is
less than 96.
— Error passive
An error passive unit can participate in bus communication. However, if the unit detects an error it is not allowed to send an active error flag. The unit sends only
a passive (‘recessive’) error flag. A device is error passive when the transmit error counter or the receive error counter is greater than 127. A device becoming
Table 20
error passive will send an active error flag. An error
passive device becomes error active again when both
transmit and receive error counter are less than 128.
— Bus off
A unit that is bus off has the output drivers disabled,
i.e., it does not participate in any bus activity. A device
is bus off when the transmit error counter is greater
than 255. A bus off device will become error active
again after monitoring 128*11 ‘recessive’ bits (including bus idle) on the bus. When the device goes from
‘bus off’ to ‘error active’, both error counters will have
the value ‘0’.
Error Counters
The CR16CAN module contains two error counters to perform the error management. The receive error counter (REC)
and the transmit error counter (TEC) are 8-bits wide, located
in the 16-bit wide CANEC register. The counters are modified
by the CR16CAN according to the rules listed in Table20 “Error Counter Handling”.
The Error counters can be read by the users software as described under CAN Error Counter Register (CANEC) on
page 115.
Error Counter Handling
Conditiona
Action
Receive Error Counter Conditions b
A receiver detects a Bit Error during sending an active error flag.
increment by 8
A receiver detects a ‘dominant’ bit as the first bit after sending an error flag
increment by 8
After detecting the 14th consecutive ‘dominant’ bit following an active error flag or overload
flag, or after detecting the 8th consecutive ‘dominant’ bit following a passive error flag.
After each sequence of additional 8 consecutive ‘dominant’ bits.
increment by 8
Any other error condition (stuff, frame, CRC, ACK)
increment by 1
A valid reception or transmission
decrement by 1 unless
counter is already 0
Transmit Error Counter Conditions
A transmitter detects a Bit Error during sending an active error flag
increment by 8
After detecting the 14th consecutive ‘dominant’ bit following an active error flag or overload flag increment by 8
or after detecting the 8th consecutive ‘dominant’ bit following a passive error flag.
After each sequence of additional 8 consecutive ‘dominant’ bits.
Any other error condition (stuff, frame, CRC, ACK)
increment by 8
A valid reception or transmission
decrement by 1 unless
counter is already 0
a. This table provides an overview of the CAN error conditions and the behavior of the CR16CAN; for a detailed
description of the error management and fault confinement rules, please refer to the CAN Specification 2.0B
b. If the MSB (bit 7) of the REC is set, the node is error passive and the REC will not increment any further.
Special error handling for the TEC counter is performed in
the following situations:
— An ACK-error occurs in an error passive device and no
‘dominant’ bits are detected while sending the passive
error flag. This does not lead to an increment of the
TEC.
— A stuff error occurs during arbitration, when a transmitted ‘recessive’ stuff bit is received as a ‘dominant’ bit.
This does not lead to an increment of the TEC.
93
www.national.com
— If only one device is on the bus and this device transmits a message, it will get no acknowledgment. This
will be detected as an error and the message will be repeated. When the device goes ‘error passive’ and detects an acknowledge error, the TEC counter is not
incremented. Therefore the device will not go from ‘error passive’ to the ‘bus off’ state due to such a condition.
20.2.3
CR16CAN divides a nominal bit time into three time segments: synchronization segment, time segment 1 (TSEG1)
and time segment 2 (TSEG2). Figure52 shows the various
elements of a CAN bit time.
CAN Bit Time
The number of time quanta in a CAN bit (CAN Bit Time) lies
between 4 and 25. The sample point is positioned between
TSEG1 and TSEG2 and the transmission point is positioned
at the end of TSEG2.
Bit Time Logic
In the Bit Time Logic (BTL), the CAN bus speed and the Synchronization Jump Width can be configured by the user.
INTERNAL
TIME QUANTA
CLOCK
ONE TIME QUANTUM
4 to 25 tq
A
1 tq
TIME SEGMENT 1 (TSEG1)
TIME SEGMENT 2 (TSEG2)
2 to 16 tq
1 to 8 tq
SAMPLE
POINT
A = synchronization segment (Sync)
TRANSMISSION
POINT
Figure 52. Bit Timing
The time segment 1 includes the propagation segment and
the phase segment 1 as specified in the CAN specification
2.0.B. The length of the time segment 1 in time quantas (tq)
is defined by the TSEG1[3:0] bits.
with the internal bit timing. The internal bit timing can be
adapted by either hard or soft synchronization (re-synchronization).
Hard synchronization is done at the beginning of a new
frame with the falling edge on the bus while the bus is idle.
This is interpreted as the SOF. It restarts the internal logic.
The time segment 2 represents the phase segment 2 as
specified in the CAN specification 2.0.B. The length of the
time segment 2 in time quantas (tq) is defined by the
TSEG2[2:0] bits.
Soft synchronization is used during the reception of a bit
stream to lengthen or shorten the internal bit time. Depending
on the phase error (e), the time segment 1 may be increased
or the time segment 2 may be decreased by a specific value,
the re-synchronization jump width (SJW).
The Synchronization Jump Width (SJW) defines the maximum number of time quanta (tq) by which a received CAN
bit can be shortened or lengthened in order to achieve resynchronization on ‘recessive’ to ‘dominant’ data transitions
on the bus. In the CR16CAN implementation the SJW has to
be configured less or equal to TSEG1 or TSEG2, whatever is
smaller.
The phase error is given by the deviation of the edge to the
SYNC segment, measured in CAN clocks. The value of the
phase error is defined as:
e = 0, if the edge occurs within the SYNC segment.
e > 0, if the edge occurs within TSEG1
e < 0, if the edge occurs within TSEG2 of the previous bit.
Synchronization
A CAN device expects the transition of the data signal to be
within the synchronization segment of each CAN bit time.
This segment has the fixed length of one time quantum.
Re-synchronization is performed according to the following
rules:
However, two CAN nodes never operate at exactly the same
clock rate and furthermore the bus signal may deviate from
the ideal waveform due to the physical conditions of the network (bus length and load). In order to compensate for the
various delays within a network, the sample point can be positioned by programming the length of time segments 1 and
2 (see Figure52).
• If the magnitude of e is less or equal to the programmed
value of SJW, re-synchronization will have the same effect
as hard synchronization.
• If e > SJW, the time segment 1 will be lengthened by the
value of the SJW (see Figure53).
• If e < -SJW, the time segment 2 will be shortened by the
value SJW (see Figure 54).
In addition to that, two types of synchronization are supported. The BTL logic compares the incoming edge of a CAN bit
www.national.com
94
e
BUS SIGNAL
CAN
CLOCK
PREVIOUS
BIT
A
TSEG1
NEXT BIT
TSEG2
“NORMAL” BIT TIME
PREVIOUS
BIT
A
TSEG1
TSEG2
SJW
NEXT BIT
BIT TIME LENGTHENED BY SJW
Figure 53.
Re-synchronization (e > SJW)
e
BUS SIGNAL
CAN
CLOCK
PREVIOUS A
BIT
TSEG1
TSEG2
“NOMINAL” BIT TIME
PREVIOUS A
BIT
TSEG1
NEXT BIT
TSEG2
BIT TIME SHORTENED BY SJW
Figure 54. Re-synchronization (e < -SJW)
20.2.4
Clock Generator
PSC = PSC[5:0] + 2
TSEG1 = TSEG1[3:0] + 1
TSEG2 = TSEG2 [2 : 0] + 1
The CAN prescaler (PSC) is shown is Figure55. It divides
the CKI input clock by the value defined in the CTIM register.
The resulting clock is called time quanta clock and defines
the length of one time quanta (tq).
CKI
Please refer to CAN Timing Register (CTIM) on page 112 for
a detailed description of the CTIM register.
:-
-:
PSC
(1+TSEG1+TSEG2)
bit rate
internal time
quanta clock (1/tq)
Note: PSC is the value of the clock prescaler. TSEG1 and
TSEG2 are the length of time segment 1 and 2 in tq.
The resulting bus clock can be calculated by the equation:
Figure 55.
CKI
busclock = ------------------------------------------------------------------------------------( PSC)x ( 1 + TSEG 1 + TSEG 2 )
20.3
Bit Rate Generation
MESSAGE TRANSFER
The CR16CAN has access to 15 independent message buffers, memory mapped in RAM. Each message buffer consists
of 8 different 16-bit RAM locations and can be individually
configured as a receive message buffer or as a transmit message buffer.
The values of PSC and TSEG 1 and 2 are specified by the
contents of the registers PSC, TSEG1 and TSEG2 as follows:
95
www.national.com
A dedicated acceptance filtering procedure enables the user
to configure each buffer to receive only a single message ID
or a group of messages. One buffer uses an independent filtering procedure, which provides the possibility to establish a
BASIC-CAN path.
the hidden buffer are copied into the first buffer with matching
filtering mask.
Bits holding a “1” in the global filtering mask (GMASK) can be
represented as a “don’t care” of the associated bit of each
buffer identifier, regardless of whether the buffer identifier bit
is “1” or “0”.
For reception of data frame or remote frames, the CR16CAN
follows a “receive on first match” rule which means that a given message is only received by one buffer — the first one
which matches to the received message ID.
This provides the capability to accept only a single ID per
buffer or to accept a group of IDs. The following two examples illustrate the difference.
The transmission of a frame can be initiated by the user software writing to the transmit status and priority register. An alternate way to schedule a transmission is the automatic
answer to remote frames. In the latter case, the CR16CAN
will schedule every buffer for transmission to respond to remote frames with a given identifier if the acceptance mask
matches. This implies that a single remote frame is able to
poll multiple matching buffers configured to respond to the
triggering remote transmission request.
20.4
Example 1: Acceptance of a Single Identifier
If the global mask is set to 00 16 the acceptance filtering of an
incoming message is only determined by the individual buffer
ID. This means that only one message ID is accepted per
buffer.
GMASK1
00000000
ACCEPTANCE FILTERING
GMASK2
00000000
BUFFER_ID1
Two 32-bit masks are used to filter unwanted messages from
the CAN bus GMASK and BMASK. Figure56 shows the
mask and the buffers controlled by the masks.
10101010
10101010
10101010
10101010
10101
Accepted ID
BUFFER_ID
Figure 57.
GMASK1
10101010
10101
Acceptance of a Single Identifier
Example 2: Reception of an Identifier Group
Buffer13
Bits in the global mask register set to ‘1’ change the corresponding bit status within the buffer ID to “don’t care” (“X”).
Therefore all messages which match the non-“don’t care”
bits are accepted.
BUFFER_ID
GMASK1
00000000
Buffer14
BMASK1
00000
BUFFER_ID2
10101010
Buffer 0
GMASK2
00000000
BUFFER_ID
GMASK2
11111111
BUFFER_ID1
BMASK2
10101010
00000000
00000
BUFFER_ID2
10101010
10101010
10101
Accepted ID group
Figure 56.
Acceptance Filtering Structure
10101010
The acceptance filtering of the incoming messages for the
buffers 0...13 is done by means of a global filtering mask
(GMASK) and by the buffer ID of each buffer.
10101010
10101
Figure 58. Acceptance of a Group of Identifiers
A separate filtering path is used for buffer 14. For this buffer
the acceptance filtering is established by the buffer ID in conjunction with the basic filtering mask. This basic mask uses
the same method as the global mask. Setting a bit to “1”
changes the associated bit in the buffer ID to a “don’t care”
bit.
The acceptance filtering of incoming messages for buffer 14
is done via a separate filtering mask (BMASK) and by the
buffer ID of each that buffer.
Once a received message is waiting in the hidden buffer (see
Receive Buffer Structure on page 98) to be copied into a buffer, CR16CAN scans all buffer configured as receive buffers
for a matching filtering mask. The buffers 0 to 13 are checked
in ascending order beginning with buffer 0. The contents of
www.national.com
XXXXXXXX
Therefore the basic mask allows a large number of infrequent
messages to be received by this buffer.
96
Note: If the BMASK register is equal to the GMASK register,
the buffer 14 can be used the same way as the buffers 0 to
13.
a buffer with the second highest priority will receive a message if the buffer with the highest priority has already received a message and is now locked (provided that both
buffers use the same acceptance filtering mask).
The buffers 0 to 13 are scanned prior to buffer 14. Subsequently, the buffer 14 will not be checked for a matching ID
when one of the buffers 0 to 13 has already received a message.
As shown in Figure59, several messages with the same ID
are received while BUFFLOCK is enabled. The filtering mask
of the buffers 0, 1, 13 and 14 is set to accept this message.
The first incoming frame will be received by buffer 0. As buffer 0 is now locked the next frame will be received by buffer 1,
and so on. If all matching receive buffers are full and locked,
a further incoming message will not be received by any buffer.
By setting the BUFFLOCK bit in the configuration register,
the receiving buffer is automatically locked after a reception
of one valid frame. The buffer will be unlocked again after the
CPU has read the data and has written RX_READY in the
buffer status field. With this lock function, the user has the capability to save several messages with the same identifier or
same identifier group into more than one buffer. For example,
received ID
01010
10101010
10101010
10101010
GMASK
00000
11111111
00000000
00000000
BUFFER0_ID
01010
XXXXXXXX
10101010
10101010
saved when buffer
is empty
BUFFER1_ID
01010
XXXXXXXX
10101010
10101010
saved when buffer
is empty
BUFFER13_ID
01010
XXXXXXXX
10101010
10101010
saved when buffer
is empty
BMASK
00000
BUFFER14_ID
01010
11111111
XXXXXXXX
00000000
00000000
10101010
10101010
saved when buffer
is empty
Figure 59. Message Storage with BUFFLOCK Enabled
20.5
RECEIVE STRUCTURE
All received frames will initially be buffered in a hidden receive buffer until the frame is valid. (The validation point for a
received message is the penultimate bit of EOF.) The received identifier is then compared to every buffer ID together
with the respective mask and the status. As soon as the validation point is reached, the whole contents of the hidden
buffer is copied into the matching message buffer as shown
in Figure60.
Note: The hidden receive buffer must not be accessed by
the CPU.
The following section gives an overview of the reception of
the different types of frames.
97
www.national.com
All contents of the hidden receive buffer are always copied
into the respective receive buffer. This includes the received
message ID as well as the received Data Length Code
(DLC); therefore when some mask bits are set to don’t care,
the ID field will get the received message ID which could be
different from the previous ID. The DLC of the receiving buffer will be updated by the DLC of the received frame. Note
that the DLC of the received message is not compared with
the DLC already present in the CNSTAT register of the message buffer. This implies that the DLC code of the CNSTAT
register indicates how may data bytes actually belong to the
latest received message.
Buffer 0
BUFFER_ID
Buffer 13
CR16CAN
HIDDEN
RECEIVE
BUFFER
BUFFER_ID
The remote frames are handled by the CR16CAN interface
in two different ways. Firstly, remote frames can be received
like data frames by configuring the buffer to be RX_READY
and setting the ID bits including the RTR bit. In that case the
same procedure applies as described for Data Frames. Secondly, a remote frame can trigger one or more message buffer to transmit a data frame upon reception. This procedure is
described under To answer Remote Frames on page 100.
Buffer 14
BUFFER_ID
20.5.1
As soon as CR16CAN receives a dominant bit on the CAN
bus, the receive process is started. The received ID and data
will be stored in the hidden receive buffer if the global or basic
acceptance filtering matches. After the reception of the data,
CR16CAN tries to match the buffer ID of buffer 0...14. The
data will be copied into the buffer after the reception of the 6th
EOF bit as a message is valid at this time. The copy process
of every frame, regardless of the length, takes at least 17 CKI
cycles (see also CPU Access to CR16CAN Registers/Memory on page 105). Figure61 illustrates the receive timing.
Figure 60. Receive Buffer Structure
The received data frame will be stored in the first matching
receive buffer beginning with buffer 0. For example, if the
message is accepted by buffer 5, then at the time the message will be copied, the RX request is cleared and CR16CAN
will not try to match the frame to any subsequent buffer.
ARBITRATION FIELD
SOF + CONTROL
BUS IDLE 1 BIT 12/29 BIT+ 6 BIT
Receive Timing
DATA FIELD
(IF PRESENT)
n * 8 BIT
CRC
FIELD
16 BIT
ACK
FIELD
2 BIT
EOF
7 BIT
IFS
3 BIT
rx_start
copy to buffer
BUSY
Figure 61.
Receive Timing
In order to indicate that a frame is waiting in the hidden buffer,
the BUSY bit ST[0] of the selected buffer is set during the
copy procedure. The BUSY bit will be cleared by CR16CAN
right after the data bytes are copied into the buffer. After the
copy process is finished, CR16CAN changes the status field
to RX_FULL. In turn the CPU should change the status field
to RX_READY when the data is processed. When a new
message has been received by the same buffer, before the
CPU changed the status to RX_READY, the CR16CAN will
change the status to RX_OVERRUN to indicate that at least
one frame has been overwritten by a new one. Table21 sum-
marizes the current status and the resulting update from the
CR16CAN.
Table 21 Writing to Buffer Status Code During
RX_BUSY
Current Status
Resulting Status
RX_READY
RX_FULL
RX_NOT_ACTIVE
RX_NOT_ACTIVE
RX_FULL
RX_OVERRUN
During the assertion of the BUSY bit, all writes to the receiving buffer are disabled with the exception of the status field.
www.national.com
98
If the status is changed during BUSY being active, the status
is updated by the CR16CAN as shown in Table21.
The buffer states are indicated and controlled by the ST[3:0]
bits in the CNSTAT register (see Buffer Status/Control Register (CNSTAT) on page 106. The various receive buffer
states are explained in RX Buffer States on page 100.
20.5.2
read buffer
read CNSTAT
Receive Procedure
The user has to execute the following procedure to initialize
a message buffer for the reception of a CAN message.
Y
1. Configure the receive masks (GMASK or BMASK, respectively).
2. Configure the buffer ID.
3. Configure the message buffer status as RX_READY.
RX_READY?
N
RX_BUSYx?
In order to read the out of a received message, the CPU has
to execute the following steps (see Figure62):
Y
N
Interrupt Entry Point
RX_OVERRUN?
(optional, for information)
write RX_READY
read buffer (id/data/cntrl)
A new message has
been received while
reading data from the
receive buffer
read CNSTAT
RX_BUSYx?
Y
N
RX_FULL? or
RX_OVERRUN?
Y
N
clear RX_PND
exit
Figure 62. Buffer Read Routine (BUFFLOCK Disabled)
The first step is only applicable if polling is used to get the
status of the receive buffer. It can be deleted for an interrupt
driven receive routine.
1. Read the status (CNSTAT) of the receive buffer. If the
status is RX_READY, no was the message received, exit. If the status is RX_BUSY, copy process from hidden
receive buffer is not completed yet, read CNSTAT again.
If a buffer is configured to RX_READY and its interrupt
is enabled, it will generate an interrupt as soon as the
99
www.national.com
2.
3.
4.
5.
6.
7.
buffer has received a message and entered the
RX_FULL state (see also Interrupts on page 104). In that
case the procedure described below should be followed.
Read the status to determine if a new message has
overwritten the one originally received which triggered
the interrupt.
Write RX_READY into CNSTAT.
Read the ID/data and message control (DLC/RTR) from
the message buffer.
Read the buffer status again and check it is not
RX_BUSYx. If it is, repeat this step until RX_BUSYx has
gone away.
If the buffer status is RX_FULL or RX_OVERRUN, one
or more messages were copied. In that case, start over
with step 2.
If status is still RX_READY (as set by the CPU at step
2), clear interrupt pending bit and exit.
When the BUFFLOCK function is enabled (see BUFFLOCK
on page 97), it is not necessary to check for new messages
received during the read process from the buffer, as this buffer is locked after the reception of the first valid frame. A read
from a locked receive buffer can be performed as shown in
Figure63.
cally change to RX_OVERRUN to indicate that at least one
message was lost. During the copy process the buffer will
again be RX_BUSYx for a short time, but in this case the CNSTAT status section will be 0101 2 , as the buffer was
RX_FULL (01002) before. After finally reading the last received message, the CPU can reset the buffer to
RX_READY.
20.6
TRANSMIT STRUCTURE
In order to transmit a CAN message, the user has to configure the message buffer by changing the buffer status to
TX_NOT_ACTIVE. The buffer is configured for transmission
if the ST[3] bit of the buffer status code (CNSTAT) is set to ‘1’.
In TX_NOT_ACTIVE status, the buffer is ready to receive
data from the CPU. After receiving all transmission data (ID,
data bytes, DLC and PRI), the CPU can start the transmission by writing TX_ONCE into the buffer status register. During the transmission the status of the buffer is TX_BUSYx.
After successful transmission CR16CAN will reset the buffer
status to TX_NOT_ACTIVE. When the transmission process
fails, the buffer condition will remain TX_BUSYx for re-transmission until the frame was successfully transmitted or the
CPU has canceled the transmission request.
In order to Send a Remote Frame (Remote Transmission
Request) to other CAN nodes, the user needs to set the RTR
bit of the message identifier to “1” (see Storage of Remote
Messages on page 109) and change the status of the message buffer to TX_ONCE. After this remote frame has been
transmitted successfully, this message buffer will automatically enter the RX_READY state and is ready to receive the
appropriate answer. Note that the mask bits RTR/XRTR need
to be set to receive a data frame (RTR = 0) in a buffer which
was configured to transmit a remote frame (RTR = 1).
Interrupt Entry Point
read buffer (id/data/cntrl)
write RX_READY
To answer Remote Frames if the CPU writes TX_RTR in the
buffer status register, the buffer will wait for a remote frame.
When a remote frame passes the acceptance filtering mask
of one or more buffers, the buffer status will change to
TX_ONCE_RTR, the contents of the buffer will be transmitted and afterwards CR16CAN will write TX_RTR in the status
code register again.
clear RX_PND
exit
Figure 63. Buffer Read Routine (BUFFLOCK Enabled)
If the CPU writes TX_ONCE_RTR in the buffer status, the
contents of the buffer will be transmitted, and the successful
transmission the buffer goes into the “wait for Remote
Frame” condition TX_RTR.
For simplicity only the applicable interrupt routine is shown:
20.6.1
1. Read the ID/data and message control (DLC/RTR) from
the message buffer.
2. Write RX_READY into CNSTAT.
3. Clear interrupt pending bit and exit.
20.5.3
RX Buffer States
As shown in Figure64, a receive procedure starts as soon as
the user has set the buffer from the RX_NOT_ACTIVE state
into the RX_READY state. The status section of CNSTAT
register is set from 00002 to 00102 . When a message is received, the buffer will be RX_BUSYx during the copy
process from the hidden receive buffer into the message
buffer. Afterwards this buffer is RX_FULL. Now the CPU can
read the buffer data and either reset the buffer status to
RX_READY or receive a new frame before the CPU reads
the buffer. In the second case, the buffer state will automatiwww.national.com
Transmit Scheduling
After writing TX_ONCE in the buffer status, the transmission
process begins and the BUSY-bit is set. As soon as a buffer
gets the TX_BUSY status, the buffer is not accessible anymore by the CPU except for the ST[3:1] bits of the CNSTAT
register. Starting with the beginning of the CRC field of the
current frame, CR16CAN looks for another buffer transmit request and selects the buffer with the highest priority for the
next transmission by changing the buffer state from
TX_ONCE to TX_BUSY. This transmit request can be canceled by the CPU or can be overwritten by another transmit
request of a buffer with a higher priority as long as the transmission of the next frame has not yet started. This means
that between the beginning of the CRC field of the current
frame and the transmission start of the next frame, two buffers, the current buffer and the buffer scheduled for the next
100
transmission, are in the BUSY status. In order to cancel the
transmit request of the next frame, the CPU has to change
the buffer state to TX_NOT_ACTIVE. When the transmit request has been overwritten by another request of a higher
priority buffer, CR16CAN changes the buffer state from
TX_BUSY to TX_ONCE. Thus, the transmit request remains
pending. Figure64 further illustrates the transmit timing.
ARBITRATION FIELD DATA FIELD
(IF PRESENT)
BUS IDLE SOF + CONTROL
1 BIT 12/29 BIT + 6 BIT n * 8 BIT
CRC
FIELD
16 BIT
ACK
FIELD
2 BIT
EOF
7 BIT
IFS
3 BIT
TX_BUSY
current buffer
TX_BUSY
next buffer
CPU write TX_ONCE
in buffer status
begin selection of next buffer
if new tx_request
Figure 64. Data Transmission
If the transmit process fails or the arbitration is lost, the transmission process will be stopped and will continue after the interrupting reception or the error signaling has finished (see
Figure65). In that case a new buffer select follows and the
TX process is executed again.
Table23 shows the transmit priority configuration if TXPRI is
different from the buffer number:
Table 23 Transmit Priority (TXPRI not 0)
Note: The canceled message can be delayed by a TX request of a buffer with a higher priority. During TX_BUSY high,
the user cannot change the contents of the message buffer.
In all cases writing to the BUSY bit will be ignored.
20.6.2
Transmit Priority
CR16CAN is able to generate a stream of scheduled messages without releasing the bus between two messages so
that an optimized performance can be achieved. It will arbitrate for the bus right after sending the previous message
and will only release the bus due to a lost arbitration.
If more than one buffer is scheduled for transmission, the priority is built by the message buffer number and the priority
code in the CNSTAT register. The 8-bit value of the priority is
formed by combining the 4-bit TXPRI value and the 4-bit buffer number (0...14) as shown below. The lowest resulting
number results in the highest transmit priority.
TXPRI
BUFFER #
Table22 shows the transmit priority configuration if the priority is set to TXPRI = 0 for all transmit buffers:
Table 22
Transmit Priority (TXPRI=0)
TXPRI
Buffer
Number
PRI
TX Priority
0
0
0
highest
0
1
1
:
:
:
:
:
:
:
:
0
14
14
lowest
TXPRI
Buffer
Number
PRI
TX Priority
14
0
224
lowest
13
1
209
12
2
194
11
3
179
10
4
164
9
5
149
8
6
134
7
7
119
6
8
104
5
9
89
4
10
74
3
11
59
2
12
44
1
13
29
0
14
14
highest
Note: If two buffers have the same priority (PRI), the buffer
with the lower buffer number will have the higher priority.
101
www.national.com
20.6.3
Transmit Procedure
The transmission of a CAN message has to be executed as
follows (see also Figure65)
write_buffer
1. Configure CNSTAT status field as TX_NOT_ACTIVE. If
the status is TX_BUSY, a previous transmit request is
still pending and the user has no access to the data contents of the buffer. In that case the user may choose to
wait until the buffer becomes available again as shown.
Other options are to exit from the update routine until the
buffer has been transmitted with an interrupt generated,
or the transmission is aborted by an error.
2. Load buffer identifier & data registers. (For remote
frames the RTR bit of the identifier needs to be set and
loading data bytes can be omitted.)
3. Configure CNSTAT status field to the desired value:
— TX_ONCE to trigger the transmission process of a single frame.
— TX_ONCE_RTR to trigger the transmission of a single
data frame and then wait for a received remote frame
to trigger consecutive data frames.
— TX_RTR waits for a remote frame to trigger the transmission of a data frame.
write
TX_NOT_ACTIVE
TX_BUSYx?
Y
(see text)
N
write ID/data
write
TX_ONCE
or
TX_ONCE_RTR
or
Writing TX_ONCE or TX_ONCE_RTR in the CNSTAT status
field will set the internal transmit request for the CR16CAN.
TX_RTR
If a buffer is configured as TX_RTR and a remote frame is received, the data contents of the addressed buffer will be
transmitted automatically without further CPU activity.
exit
Figure 65. Buffer Write Routine
www.national.com
102
20.6.4
TX Buffer States
The transmission process can be started after the user has
loaded the buffer registers (data, ID, DLC, PRI) and set the
buffer status from TX_NOT_ACTIVE to TX_ONCE, TX_RTR
or TX_ONCE_RTR.
When the CPU writes TX_ONCE, the buffer will be
TX_BUSY as soon as CR16CAN has scheduled this buffer
for the next transmission. After the frame could be successfully transmitted, the buffer status will be automatically reset
to TX_NOT_ACTIVE when a data frame was transmitted or
to RX_READY when a remote frame was transmitted.
If the CPU configures the message buffer to
TX_ONCE_RTR, it will transmit its data contents. During the
transmission the buffer state is 1111 2 as the CPU wrote 11102
into the status section of the CNSTAT register. After the successful transmission the buffer enters the TX_RTR state and
waits for a remote frame. When it receives a remote frame, it
will go back into the TX_ONCE_RTR state, transmit its data
bytes and return to TX_RTR. If the CPU writes 1010 2 into the
buffer status section, it will only enter the TX_RTR state. But
it will not send its data bytes before it waits for a remote
frame. Figure 66 illustrates the possible transmit buffer
states.
TX_ONCE_RTR
1110
CAN
schedules TX
RTR
received
TX request
CPU writes 1110
TX_BUSY2
1111
transmit
request cancelled
CPU writes 1000
transmit failed
TX done
CPU writes 1010
TX_NOT_ACTIVE
1000
TX_RTR
1010
TX request
CPU writes 1100
TX done
*1: TX request delayed
by a TX request of higher
priority message
TX_ONCE
1100
CAN
schedules TX
transmit
request cancelled
CPU writes 1000
*1
TX_BUSY0
1101
RX_READY
0010
Remote transmission
request sent - now wait
to receive a data frame
transmit failed
Figure 66. Transmit Buffer States
103
www.national.com
20.7
Figure67 illustrates the CR16CAN interrupt management.
INTERRUPTS
CR16CAN has access to one interrupt vector in the CR16
CPU. The interrupt process can be initiated from the following sources.
CIEN
• CAN data transfer
— Reception of a valid data frame in the buffer. (Buffer
state changes from RX_READY to RX_FULL or
RX_OVERRUN).
— Successful transmission of a data frame. (Buffer state
changes from TX_ONCE to TX_NOT_ACTIVE or
RX_READY)
— Successful response to a remote frame. (Buffer state
changes from TX_ONCE_RTR to TX_RTR).
— Transmit scheduling. (Buffer state changes from
TX_RTR to TX_ONCE_RTR).
• CAN error conditions is the detection of an CAN error.
(The CEIPND bit in the CIPND register will be set as well
as the corresponding bits in the error diagnostic register
CEDIAG).
CICLR
clear interrupt flags of every
message buffer individually
CIPND
CICEN
ICODE
IRQ
The receive/transmit interrupt access to every message buffer can be individually enabled/disabled in the CIEN register.
The pending flags of the message buffer are located in the
CIPND register (read only) and can be cleared by resetting
the flags in the CICLR registers.
20.7.1
IST2
IST1
IST0
Figure 67. CR16CAN Interrupt Management
The highest priority interrupt source is translated into the bits
IRQ and IST[3:0] as shown in Table24.
Table 24
Highest Priority Interrupt Code
In order to reduce decoding time of the CIPND register, the
buffer interrupt request with the highest priority is placed as
interrupt status code into the IST[3:0] section of the CSTPND
register.
Each of the buffer interrupts as well as the error interrupt can
be individually enabled or disabled in the CAN Interrupt Enable register (CIEN). As soon as an interrupt condition occurs, every interrupt request is indicated by a flag in the CAN
Interrupt Pending register (CIPND). When the interrupt code
logic for the present highest priority interrupt request is enabled, this interrupt will be translated into the IST[3:0] bits of
the CAN Status Pending register (CSTPND). An interrupt request can be cleared by setting the corresponding bit in the
CAN Interrupt Clear register (CICLR) to ‘1’.
www.national.com
IST3
104
Highest Priority Interrupt Code (ICEN=FFFF)
CAN interrupt
request
IRQ
IST3
IST2
IST1
IST0
no request
0
0
0
0
0
Error interrupt
1
0
0
0
0
Buffer 0
1
0
0
0
1
Buffer 1
1
0
0
1
0
Buffer 2
1
0
0
1
1
Buffer 3
1
0
1
0
0
Buffer 4
1
0
1
0
1
Buffer 5
1
0
1
1
0
Buffer 6
1
0
1
1
1
Buffer 7
1
1
0
0
0
Buffer 8
1
1
0
0
1
Buffer 9
1
1
0
1
0
Buffer 10
1
1
0
1
1
Buffer 11
1
1
1
0
0
Buffer 12
1
1
1
0
1
Buffer 13
1
1
1
1
0
Buffer 14
1
1
1
1
1
20.7.2
Usage Hints
20.9
The interrupt code IST[3:0] can be used within the interrupt
handler as a displacement in order to jump to the relevant
subroutine.
CR16CAN occupies 144 words in the memory address
space. This space is separated into 15*8 + 8(reserved)
words for the message buffers and 14 + 2(reserved) words
for control and status.
The CAN Interrupt Code Enable (CICEN) register is used in
the CAN interrupt handler if the user wants to service all receive buffer interrupts first followed by all transmit buffer interrupts. In this case, the user can first enable only all receive
buffer interrupts to be coded, scan and service all pending interrupt requests in the order of their priority. Then, the user
changes the CICEN register to disable all receive buffers, but
enable all transmit buffers and service all pending transmit
buffer interrupt requests according to their priorities.
20.8
20.9.1
Both word and byte access to the buffer RAM are allowed. If
a buffer is busy during the reception of a message (copy process from the hidden receive buffer) or is scheduled for transmission, the CPU has no write access to the data contents of
the buffer. Write to the status/control byte and read access to
the whole buffer is always enabled.
CR16CAN features a free running 16-bit timer (CTMR) incrementing every bit time recognized on the CAN bus. The value of this timer during the ACK slot is captured into the TSTP
register of a message buffer after a successful transmission
or reception of a message. Figure68 shows a simplified
block diagram of the Time Stamp counter.
All configuration and status registers can either be accessed
by CR16CAN or the CPU only. These registers provide single
cycle word and byte access without any potential wait state.
CAN bits on the bus
+1
Reset
CPU Access to CR16CAN Registers/Memory
All memory locations occupied by the message buffers are
shared by the CPU and CR16CAN (dual ported RAM). The
CR16CAN and the CPU normally have single cycle access
to this memory. However, if an access contention occurs, the
access to the memory is altered every cycle until the contention is resolved. This internal access arbitration is transparent to the user.
TIME STAMP COUNTER
16-bit counter
MEMORY ORGANIZATION
ACK slot & buffer 0 active
All register descriptions within the next sections utilize the following layout:
ACK slot
bit 15
... bit number ...
TSTP register
Figure 68.
bit 0
... bit name ...
... reset value ...
... CPU access ...
r = register bit is read only
w = register bit is write only
r/w = register bit is read/write
Time Stamp Counter
The timer can be synchronized over the CAN network by receiving or transmitting a message to/from buffer 0. In that
case the TSTP register of buffer 0 captures the current
CTMR value during the ACK slot of a message (as above)
and afterwards the CTMR is reset to 00002. Synchronization
can be enabled or disabled via the CGCR.TSTPEN bit.
Table 25
20.9.2
Message Buffer Organization
The message buffers are the communication interfaces between CAN and the CPU for the transmission and the reception of CAN frames. There are 15 message buffers located at
fixed addresses in the RAM location. As shown in Table25,
each buffer consists of two words reserved for the identifiers,
4 words reserved for up to eight CAN data bytes, one word is
reserved for time stamp and one word for data length code,
transmit priority code and the buffer status code.
Message Buffer Organization
ADDR
BUFFER
register
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
xxxE16
ID1
XI28
ID10
XI27
ID9
XI26
ID8
XI25
ID7
XI24
ID6
XI23
ID5
XI22
ID4
XI21
ID3
XI20
ID2
XI19
ID1
XI18
ID0
SRR
IDE
XI17
XI16
XI15
xxxC 16
ID0
XI14
XI13
XI12
XI11
XI10
XI9
XI8
XI7
XI6
XI5
XI4
XI3
XI2
XI1
XI0
RTR
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
xxxA 16
xxx8 1 6
xxx6 1 6
DATA0
DATA1
DATA2
RTR
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1.0
2.7
2.6
2.5
2.4
2.3
2.2
2.1
2.0
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
3.7
3.6
3.5
3.4
3.3
3.2
3.1
3.0
4.7
4.6
4.5
4.4
4.3
4.2
4.1
4.0
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
5.7
5.6
5.5
5.4
5.3
5.2
5.1
5.0
6.7
6.6
6.5
6.4
6.3
6.2
6.1
6.0
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
7.5
7.4
7.3
7.2
xxx4 1 6
DATA3
7.7
7.6
xxx2 1 6
TSTP
TSTP15
TSTP14
xxx0 1 6
CNTSTAT
DLC3
DLC2
TSTP13 TSTP12
DLC1
DLC0
TSTP11 TSTP10
7.1
7.0
8.7
8.6
8.5
8.4
8.3
8.2
8.1
8.0
TSTP
9
TSTP
8
TSTP
7
TSTP
6
TSTP
5
TSTP
4
TSTP
3
TSTP
2
TSTP
1
TSTP
0
PRI3
PRI2
PRI1
PRI0
ST3
ST2
ST1
ST0
Reserved
105
www.national.com
20.9.3
Buffer Status/Control Register (CNSTAT)
The buffer status, the buffer priority and the data length code
are controlled by manipulating the contents of the Buffer Status/Control Register (CNSTAT). CPU and CR16CAN have
access to this register.
15
DLC[3:0]
ST[3:0]
Table 26
12 11
8 7
4
Reserved
PRI[3:0]
0
r/w
3
0
ST[3:0]
Buffer Status — The CNSTAT register has a
status section, which contains the status information of the buffer as shown in Table26. This
section can be modified by CR16CAN.
The ST0 bits acts as a buffer busy indication.
When the BUSY bit is set, any write access to
the buffer is disabled with the exception of the
lower byte of the CNTSTAT register. The
CR16CAN sets this bit if the buffer data is currently copied from the hidden buffer or if a message is scheduled for transmission or is
currently transmitting. The CR16CAN will always reset this bit on a status update.
Buffer Status Section of the CNSTAT Register
ST3 (DIR)
ST2
ST1
ST0 (BUSY)
Buffer Status
0
0
0
0
RX_NOT_ACTIVE
0
0
0
1
0
0
1
0
RX_READY
0
0
1
1
RX_BUSY0
0
1
0
0
RX_FULL
0
1
0
1
RX_BUSY1
0
1
1
0
0
1
1
1
1
0
0
0
TX_NOT_ACTIVE
1
0
0
1
Reserved for TX_BUSY c
1
1
0
0
TX_ONCE
1
1
0
1
TX_BUSY0 d
1
0
1
0
TX_RTR
(automatic response to a remote frame)
1
0
1
1
Reserved for TX_BUSY1 e
1
1
1
0
TX_ONCE_RTR
(changes to TX_RTR after transmission)
Reserved for RX_BUSY
a
b
b
RX_OVERRUN
RX_BUSY2
b
1
1
1
1
TX_BUSY2 d
a. This condition indicates that the user wrote RX_NOT_ACTIVE to a buffer when the data copy process is still
active.
b. RX_BUSYx indicates that coping is in progress at three possible times
- data is copied for the first time RX_READY → RX_BUSY0
- data is copied for the second time RX_FULL → RX_BUSY1
- data is copied for the third or more time RX_OVERRUN → RX_BUSY2
c. This state indicates that the user wrote TX_NOT_ACTIVE to a transmit buffer which is scheduled for
transmission or is currently transmitting.
d. TX_BUSYx indicates that a buffers is scheduled for transmission or is actively transmitting; it can be due to
one of two cases:
- a message is pending for transmission or is currently transmitting
- an automated answer is pending for transmission or is currently transmitting
e. This condition does not occur
www.national.com
106
PRI[3:0]
Transmit Priority Code. The PRI[3:0] bits contain the user defined transmit priority code for
the message buffer.
Data Length Code. The DLC[3:0] bits determine the number of data bytes within a received/transmitted frame. For transmission,
these bits need to be set according to the number of data bytes to be transmitted. For reception, these bits indicate the number of valid
received data bytes available in the message
buffer. Table27 shows the possible bit combinations for DLC[3:0] for data lengths from 0 to
8 bytes.
DLC[3:0]
Table 27
Table 27
DLC3
DLC2
DLC1
DLC0
7
0
1
1
1
8
1
0
0
0
20.9.4
DLC3
DLC2
DLC1
DLC0
0
0
0
0
0
1
0
0
0
1
2
0
0
1
0
3
0
0
1
1
4
0
1
0
0
5
0
1
0
1
6
0
1
1
0
Table 28
Number of data
bytes
Note: The maximum number of data bytes received/transmitted is 8, even if the data length code is set to a value
greater than 8. Thus, if the data length code is greater or
equal to eight bytes, the bits DLC2 to DLC0 are ignored.
Data Length Coding
Number of data
bytes
Data Length Coding
Storage of Standard Messages
During the processing of standard frames, the ExtendedIdentifier-bit (IDE) is set to “0”. The bits ID1[3:0], ID0[15:0]
are “don’t care” bits. A standard frame with eight data bytes
is shown in Table28.
IDE
Identifier Extension. IDE is set to “0” to indicate
that the message is a standard frame using 11
identifier bits. If IDE is set to “1”, the message
stored in the buffer is handled as an extended
frame.
Remote Transmission Request. RTR is set to
“1” to indicate that the message is a remote
frame. For a data frame, the RTR bit is set to
“0”.
RTR
ID[10:0]The ID buffer bits ID10 to ID0 are used for the 11
standard frame identifier bits.
Standard Frame with 8 Data Bytes
ADDR
BUFFER
register
15
14
13
12
11
10
9
8
7
6
5
4
3
xxxE16
ID1
ID10
ID9
ID8
ID7
ID6
ID5
ID4
ID3
ID2
ID1
ID0
RTR
IDE
xxxC 16
ID0
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1.0
2.7
2.6
2.5
2.4
2.3
2.2
2.1
2.0
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
xxxA 16
xxx8 16
xxx6 16
xxx4 16
DATA0
DATA1
DATA2
DATA3
2
1
0
don’t care
don’t care
Data
3.7
3.6
3.5
3.4
3.3
3.2
3.1
3.0
4.7
4.6
4.5
4.4
4.3
4.2
4.1
4.0
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
5.7
5.6
5.5
5.4
5.3
5.2
5.1
5.0
6.7
6.6
6.5
6.4
6.3
6.2
6.1
6.0
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
7.7
7.6
7.5
7.4
7.3
7.2
7.1
7.0
8.7
8.6
8.5
8.4
8.3
8.2
8.1
8.0
TSTP
14
TSTP
13
TSTP
12
TSTP
11
TSTP
10
TSTP
9
TSTP
8
TSTP
7
TSTP
6
TSTP
5
TSTP
4
TSTP
3
TSTP
2
TSTP
1
TSTP
0
DLC2
DLC1
DLC0
PRI3
PRI2
PRI1
PRI0
ST3
ST2
ST1
ST0
xxx2 16
TSTP
TSTP
15
xxx0 16
CNTSTAT
DLC3
Reserved
107
www.national.com
20.9.5
Storage of Messages with Less Than 8 Data
Bytes
20.9.6
Storage of Extended Messages
If the IDE bit is set to “1”, the buffer handles extended frames.
The storage of the extended ID follows the descriptions in
Table29. The SRR bit is at the bit position of the RTR bit for
standard frame and needs to be transmitted as “1”.
The data bytes that are not used for data transfer are “don’t
cares”. If the message is transmitted, the data within these
bytes will be ignored. If the message is received, the data
within these bytes will be overwritten with invalid data.
Table 29 Extended Messages with 8 Data Bytes
ADDR
BUFFER
register
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
xxxE16
ID1
ID28
ID27
ID26
ID25
ID24
ID23
ID22
ID21
ID20
ID19
ID18
SRR
IDE
ID17
ID16
ID15
xxxC 1 6
ID0
ID14
ID13
ID12
ID11
ID10
ID9
ID8
ID7
ID6
ID5
ID4
ID3
ID2
ID1
ID0
RTR
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
1.7
1.6
1.5
1.4
1.3
1.2
1.1
1.0
2.7
2.6
2.5
2.4
2.3
2.2
2.1
2.0
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
xxxA 1 6
xxx816
xxx616
xxx416
DATA0
DATA1
DATA2
DATA3
3.7
3.6
3.5
3.4
3.3
3.2
3.1
3.0
4.7
4.6
4.5
4.4
4.3
4.2
4.1
4.0
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
5.7
5.6
5.5
5.4
5.3
5.2
5.1
5.0
6.7
6.6
6.5
6.4
6.3
6.2
6.1
6.0
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
Data
7.7
7.6
7.5
7.4
7.3
7.2
7.1
7.0
8.7
8.6
8.5
8.4
8.3
8.2
8.1
8.0
TSTP
14
TSTP
13
TSTP
12
TSTP
11
TSTP
10
TSTP
9
TSTP
8
TSTP
7
TSTP
6
TSTP
5
TSTP
4
TSTP
3
TSTP
2
TSTP
1
TSTP
0
DLC2
DLC1
DLC0
PRI3
PRI2
PRI1
PRI0
ST3
ST2
ST1
ST0
xxx216
TSTP
TSTP
15
xxx016
CNTSTAT
DLC3
SRR
IDE
RTR
ID[28:0]
Reserved
Substitute Remote Request. SRR replaces the
RTR bit used in standard frames at this bit position. The SRR bit needs to be set to “1” by the
user if the buffer is configured to transmit a
message with an extended identifier. It will be
received as monitored on the CAN bus.
Identifier Extension. IDE is set to “0” to indicate
that the message is a standard frame using 11
identifier bits. If IDE is set to “1”, the message
stored in the buffer is handled as an extended
frame.
Remote Transmission Request. RTR is set to
“1” to indicate that the message is a remote
frame. For a data frame, the RTR bit is set to
“0”.
The ID bits 28 to 0 are used to build the 29-bit
identifier of an extended frame.
www.national.com
108
20.9.7
Storage of Remote Messages
ceived, the contents of these registers will be overwritten with
invalid data. The structure of a message buffer set up for a
remote frame with extended identifier is shown in Table30.
During remote frame transfer, the buffer registers DATA[3:0]
are “don’t cares”. If a remote frame is transmitted, the contents of these registers are ignored. If a remote frame is re-
Table 30 Extended Remote Frame
ADDR
BUFFER
register
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
xxxE16
ID1
ID28
ID27
ID26
ID25
ID24
ID23
ID22
ID21
ID20
ID19
ID18
SRR
IDE
ID17
ID16
ID15
xxxC 16
ID0
ID14
ID13
ID12
ID11
ID10
ID9
ID8
ID7
ID6
ID5
ID4
ID3
ID2
ID1
ID0
RTR
xxxA 16
DATA0
don’t care
xxx8 1 6
DATA1
don’t care
xxx6 1 6
DATA2
don’t care
xxx4 1 6
DATA3
don’t care
xxx2 1 6
TSTP
TSTP15
TSTP14
TSTP13
TSTP12
TSTP
7
TSTP
6
TSTP
5
TSTP
4
TSTP
3
TSTP
2
TSTP
1
TSTP
0
xxx0 1 6
CNTSTAT
DLC3
DLC2
DLC1
DLC0
PRI3
PRI2
PRI1
PRI0
ST3
ST2
ST1
ST0
SRR
IDE
RTR
ID[28:0]
TSTP11
TSTP10
TSTP
9
TSTP
8
Reserved
Substitute Remote Request. SRR replaces the
RTR bit used in standard frames at this bit position. The SRR bit needs to be set to “1” by the
user.
Identifier Extension. IDE is set to “0” to indicate
that the message is a standard frame using 11
identifier bits. If IDE is set to “1”, the message
stored in the buffer is handled as an extended
frame.
Remote Transmission Request. RTR is set to
“1” to indicate that the message is a remote
frame. For a data frame, the RTR bit is set to
“0”.
The ID bits 28 to 0 are used to build the 29-bit
identifier of an extended frame. The ID1 buffer
bits ID28 to ID18 are used for the 11 standard
frame identifier bits.
20.9.8
CAN Global Configuration Register (CGCR)
The CAN Global Configuration Register (CGCR) is a 16-bit
wide register used to:
• enable/disable the CR16CAN
• configure the BUFFLOCK function for the message buffer
0...14
• enable/disable the time stamp synchronization
• set the logic levels of the CAN Input/Output pins
CANRX/CANTX
• choose the data storage direction (DDIR)
• select the error interrupt type (EIT)
• enable/disable diagnostic functions
15
12
Reserved
7
6
5
11
10
9
8
EIT DIAGEN INTERNAL LOOPBACK
0
r/w
4
3
2
1
0
IGNACK LO DDIR TSTPEN BUFFLOCK CRX CTX CANEN
0
r/w
CANEN
CTX
109
CAN Enable. This bit enables/disables the
CR16CAN. When the CR16CAN is disabled,
all internal states and the TEC and REC
counter registers are cleared. In addition the
CR16CAN clock is disabled. All CR16CAN
control registers and the contents of the message memory are left unchanged.
The user needs to make sure that no message
is pending for transmission before the
CR16CAN is disabled.
“0” CR16CAN is disabled
“1” CR16CAN is enabled
Control Transmit. This bit configures the logic
level of the CAN transmit pin CANTX.
“0” dominate state is “0”; recessive state is “1”
www.national.com
CRX
BUFFLOC
TSTPEN
“1” dominate state is “1”; recessive state is “0”
Control Receive. This bit configures the logic
level of the CAN receive pin CANRX.
“0” dominate state is “0”; recessive state is “1”
“1” dominate state is “1”; recessive state is “0”
Buffer Lock. With this bit the user can configure
the buffer lock function. If this feature is enabled, a buffer will be locked upon a successful
frame reception. The buffer will be unlocked
again by writing RX_READY in the buffer status register, i.e., after reading data.
“0” lock function is disabled for all buffers
“1” lock function is enabled for all buffers
Time Sync Enable. The Time Sync bit enables
or disables the time stamp synchronization
function of the CR16CAN.
“0” Time synchronization disabled. The Time
Stamp counter value is not reset upon re-
ception or transmission of a message to/
from buffer 0.
“1” Time synchronization enabled. The Time
Stamp counter value is reset upon reception or transmission of a message to/from
buffer 0.
Data Direction. By setting or resetting the DDIR
bit, the user can select the direction the data
bytes are transmitted and received. The
CR16CAN transmits and receives the CAN
data byte Data1 first and the data byte Data8
last (Data1, Data2,...,Data7, Data8).
If DDIR is set to “0” the data contents of a received message is stored with the first byte at
the highest data address and the last data at
the lowest data address (see Figure69). The
same applies for transmitted data.
DDIR
Sequence of Data Bytes on the Bus
ID
Data1
Data2
Data3
Data4
Data5
Data6
Data7
Data8
CRC
t
Storage of Data Bytes
in the Buffer Memory
ADDR offset
Data Bytes
0A 16
Data1
Data2
08 16
Data3
Data4
06 16
Data5
Data6
04 16
Data7
Data8
Figure 69. Data Direction Bit set to ‘0’
www.national.com
110
Setting the DDIR bit to “1” will cause the direction of the data
storage to be reversed — the last byte received is stored at
the highest address and the first byte is stored at the lowest
address. See Figure70 for illustration.
Sequence of Data Bytes on the Bus
ID
Data1
Data2
Data3
Data4
Data5
Data6
Data7
Data8
CRC
t
Storage of Data Bytes
in the Buffer Memory
Figure 70.
LO
ADDRoffset
Data Bytes
0A 16
Data8
Data7
08 16
Data6
Data5
06 16
Data4
Data3
04 16
Data2
Data1
Data Direction Bit set to ‘1’
Listen Only — By setting the LO-bit to “1” the
CR16CAN interface is configured to behave
only as a receiver. This means:
•
it cannot transmit any message.
•
it cannot send a dominant ACK bit.
•
when errors are detected on the bus, the
CR16CAN will behave as in the error passive mode.
Using this listen only function, the CR16CAN
interface can be adjusted when it gets connected to an operating network with unknown bus
speed.
IGNACK
Ignore Acknowledge. If the ignore ACK function is enabled, then by setting the IGNACK bit
to “1”, CR16CAN does not expect to receive a
dominant ACK bit to indicate the validity of a
transmitted message. It will not send an error
frame when the transmitted frame in not acknowledged by any other CAN node.
This feature can be used in conjunction with
the LOOPBACK option for stand-alone tests
outside of a CAN network.
LOOPBACK Loopback. By setting the LOOPBACK bit, all
messages sent by CR16CAN can also be received by a CR16CAN buffer with a matching
buffer ID. However, CR16CAN does not acknowledge a message sent by itself. Therefore
CR16CAN will send an error frame when no
other device connected to the bus has acknowledged the message.
INTERNAL Internal. If the INTERNAL function is enabled,
the TX- and RX-pin of the CR16CAN are internally connected to each other. This feature can
be used in conjunction with the LOOPBACK
mode. This means that CR16CAN can receive
its own sent messages without connecting an
external transceiver chip to the RX- and TXpin; it allows the user to run real stand-alone
tests without any peripheral devices.
DIAGEN
EIT
111
Diagnostic Enable. The DIAGEN bit globally
enables or disables the special diagnostic features of CR16CAN. This includes the following
functions:
•
LO (Listen Only)
•
IGNACK (Ignore Acknowledge)
•
LOOPBACK (Loopback)
•
INTERNAL (Internal Loopback)
•
write access to hidden receive buffer
Error Interrupt Type. This bit configures when
the Error Interrupt Pending Bit (CIPND.EIPND)
is set and an error interrupt is generated if enabled by the Error Interrupt Enable
(CIEN.EIEN).
“0” The EIPND bit is set on every error on the
CAN bus.
“1” The EIPND bit is set only if the error state
(CSTPND.NS) changes as a result of incrementing either the receive or transmit
error counter.
www.national.com
20.9.9
CAN Timing Register (CTIM)
The Can Timing Register (CTIM) defines the configuration of
the Bit Time Logic (BTL).
15
9
PSC[6:0]
PSC[6:0]
Table 33
8
7 6
3 2
0
SJW[1:0] TSEG1[3:0] TSEG2[2:0]
0
r/w
Prescaler Configuration. These bits set the
CAN prescaler. The settings are shown in
Table31
Table 31
CAN Prescaler Settings
Time Segment 1 Settings
TSEG
13
TSEG
12
TSEG
11
TSEG
10
Length of Time
(TSEG1)
0
0
0
0
not recommended
0
0
0
1
2 tq
0
0
1
0
3 tq
0
0
1
1
4 tq
0
1
0
0
5 tq
0
1
0
1
6 tq
0
1
1
0
7 tq
0
1
1
1
8 tq
PS
C6
PS
C5
PS
C4
PS
C3
PS
C2
PS
C1
PS
C0
CAN
prescaler
(PSC)
0
0
0
0
0
0
0
2
1
0
0
0
9 tq
0
0
0
0
0
0
1
3
1
0
0
1
10 tq
0
0
0
0
0
1
0
4
1
0
1
0
11 tq
0
0
0
0
0
1
1
5
1
0
1
1
12 tq
1
1
0
0
13 tq
0
0
0
0
1
0
0
6
1
1
0
1
14 tq
:
:
:
:
:
:
:
1
1
1
0
15 tq
1
1
1
1
1
0
1
127
1
1
1
1
16 tq
1
1
1
1
1
1
0/1
128
SJW[1:0]
Synchronization Jump Width. These bits set
the Synchronization Jump Width which can be
programmed between 1 and 4 time quanta
(see Table32).
SJW0
Table 34
Time Segment 2 Settings
TSEG22
TSEG21
TSEG20
Length of
TSEG2
Synchronization Jump Width
(SJW)
0
0
0
1 tq
0
0
1
2 tq
Table 32
SJW1
TSEG2[2:0] Time Segment 2. The TSEG2[2:0] bits set the
number of time quanta (tq) for phase segment
2 (see Table34).
SJW Settings
0
0
1 tq
0
1
0
3 tq
0
1
2 tq
0
1
1
4 tq
1
0
3 tq
1
0
0
5 tq
1
1
4 tq
1
0
1
6 tq
1
1
0
7 tq
1
1
1
8 tq
Note: The settings of SJW has to be configured to be smaller or equal to TSEG1 and TSEG2
20.9.10 Global Mask Registers (GMSK — GMSKB and
GMSKX)
TSEG1[3:0] Time Segment 1. These bits configure the
length of the Time Segment 1 (TSEG1). It is not
recommended to configure the time segment 1
to be smaller than 2tq. (see Table33).
www.national.com
The GMSKB and GMSKX registers allow you to globally
mask, or “don’t care” the incoming extended/standard identifier bits, RTR/XRTR and IDE. Throughout this document, the
GMSKB and GMSKX 16-bit registers are referenced as a 32bit register GMSK.
112
GM[28:15]
The following are the bits for the GMSKB register.
15
5
4
RTR
0
r/w
GM[28:18]
GM[14:0]
3
IDE
2
BM[14:0]
0
15
1
GM[14:0]
0
r/w
For all BMSKB and BMSKX register bits the following applies:
0
XRTR
— “0” incoming identifier bit must match the corresponding bit in the message buffer identifier register.
— “1” accept “1” or “0” (“don’t care”) of the incoming ID bit
independent from the corresponding bit in the message buffer ID registers. The corresponding ID bit in
the message buffer will be overwritten by the incoming
identifier bits.
For all GMSKB and GMSKX register bits, the following applies:
— “0” is the incoming identifier bit must match the corresponding bit in the message buffer identifier register.
— “1” accept “1” or “0” (“don’t care”) of the incoming ID bit
independent from the corresponding bit in the message buffer ID registers. The corresponding ID bit in
the message buffer will be overwritten by the incoming
identifier bits.
When an extended frame is received from the CAN bus all
Basic Mask bits BM28 through BM0, IDE, RTR and XRTR
are used to mask the incoming message.
During the reception of standard frames only the Basic Mask
bits BM28 to BM18, RTR and IDE are utilized.
When an extended frame is received from the CAN bus, all
Global Mask bits GM28 through GM0, IDE, RTR and XRTR
are used to mask the incoming message.
ID[10:0]
a
RTR
IDE
IDE
GM[17:0]
XRTR
unused
extended frame
ID[28:18]
SRR IDE
ID[17:0]
RTR
a. the RTR bit has a different position in standard and
extended frames
— for standard frames the GMSK_RTR bit is used to
mask this bit
— for extended frames the GMSK_XRTR bit is used to
mask this bit
Basic Mask
BM[28:18]
standard frame
ID[10:0]
BM[28:15]
5
BM[28:18]
4
RTR
0
r/w
3
IDE
2
0
BM[17:15]
RTR
IDE
BM[17:0]
IDE
XRTR
unused
The CAN Interrupt Enable (CIEN) register enables the transmit/receive interrupts of the message buffers 0 through 14 as
well as the CAN Error Interrupt.
15
EIEN
14
0
IEN[14:0]
0
r/w
EIEN
The following are the bits for the BMSKB register.
15
a
20.9.12 CAN Interrupt Enable Register (CIEN)
20.9.11 Basic Mask Registers (BMSK — BMSKB and
BMSKX)
The two registers BMSKB and BMSKX allow to mask the
buffer 14, or “don’t care” the incoming extended/standard
identifier bits, RTR/XRTR and IDE. Throughout this document, the two 16-bit registers BMSKB and BMSKX are referenced to as a 32-bit register BMSK.
RTR
extended frame ID[28:18]
SRR
IDE
ID[17:0]
RTR
a. the RTR bit has a different position in standard and
extended frames
— for standard frames the BMSK_RTR bit is used to
mask this bit
— for extended frames the BMSK_XRTR bit is used to
mask this bit
During the reception of standard frames only the Global
Mask bits GM28 to GM18, RTR and IDE are utilized.
GM[28:18] RTR
0
XRTR
BM[14:0]
0
r/w
15
Global Mask
1
GM[17:15]
The following are the bits for the GMSKX register.
standard frame
The following are the bits for the BMSKX register.
IEN[14:0]
113
Error Interrupt Enable. This bit allows the
CR16CAN to interrupt the CPU if any kind of
CAN receive/transmit errors are detected. This
means any error status change in the error
counter registers REC/TEC is able to generate
an error interrupt if EIEN is enabled.
“0” The error interrupt is disabled and no error
interrupt will be generated.
“1” The error interrupt is enabled and a
change in REC/TEC will cause an interrupt to be generated.
Buffer Interrupt Enable. The IEN[14:0] allow
the user to enable/disable interrupt source for
each of the message buffers i.e., IEN14 configures buffer14 and IEN0 configures buffer0.
“0” buffer as interrupt source disabled
“1” buffer as interrupt source enabled
www.national.com
20.9.13 CAN Interrupt Pending Register (CIPND)
EICEN
The CIPND register indicates any CAN Receive/Transmit Interrupt Requests caused by the message buffers 0..14 and
CAN error occurrences.
15
14
EIPND
EIPND
IPND[14:0]
0
ICEN[14:0]
IPND[14:0]
0
r
Error Interrupt Pending — EIPND indicates the
status change of TEC/REC and will execute an
error interrupt if EIEN is set. The user has the
responsibility to reset EIPND by means of the
CICLR register.
“0” CAN status is not changed
“1” CAN status is changed
Buffer Interrupt Pending — IPND[14:0] bits are
set by CR16CAN following a successful transmission or reception of a message to or from
message buffer 0...14, IPND14 for buffer 14
and IPND0 for buffer 0.
“0” no interrupt pending for this message buffer
“1” message buffer has generated an interrupt
20.9.16 CAN Status Pending Register (CSTPND)
The CAN Status Pending Register (CSTPND) contains the
status of the CAN Node and the Interrupt Code.
15
NS[2:0]
ICLR[14:0]
7
5
NS[2:0]
0
r
0
ICLR[14:0]
0
w
4
IRQ
3
0
IST[3:0]
CAN Node Status. This bits indicate the status
of the CAN node as it is described in Table35.
Table 35
The bits in the CICLR register separately clear all CAN interrupt pending flags caused by the message buffers 0...14 and
from the Error Management Logic.
EICLR
8
Reserved
20.9.14 CAN Interrupt Clear Register (CICLR)
15
14
EICLR
Error Interrupt Code Enable:
“0” error interrupt pending is not indicated in
the interrupt code
“1” error interrupt pending is indicated in the
interrupt code
Buffer Interrupt Code Enable:
“0” buffer interrupt pending is not indicated in
the interrupt code
“1” buffer interrupt pending is indicated in the
interrupt code
CAN Node Status
NS2
NS1
NS0
Node Status
0
0
0
Not Active
0
0
1
1
0
1
Error active
Error Warning Level
1
0
X
Error passive
1
1
X
Bus off
IRQ,IST[3:0] Interrupt Code. This section of the Status
Pending Register represents the interrupt
source of the highest priority interrupt currently
pending and enabled in the CICEN register.
Table36 shows the several interrupt codes for
CICEN=FFFF.
Error Interrupt Clear. The EICLR bit can clear
the EIPND bit:
“0” the contents of the EIPND bit is unchanged
“1” the contents of the EIPND bit is reset
Buffer Interrupt Clear. The user is able to clear
the buffer interrupt pending bits by ICLR[14:0]:
“0” the contents of the respective IPND bit is
unchanged
“1” the contents of the respective IPND bit is
reset
Table 36
Highest Priority Interrupt Code
(CICEN = FFFF)
CAN interrupt
request
IRQ
IST3
IST2
IST1
IST0
no request
0
0
0
0
0
Error interrupt
1
0
0
0
0
20.9.15 CAN Interrupt Code Enable Register (CICEN)
Buffer 0
1
0
0
0
1
The CAN Interrupt Code Enable Register (CICEN) determines whether the interrupt pending flag in IPND should be
translated into the Interrupt Code field of the CSTPND register. All interrupt requests, CAN error and buffer 0...14 interrupts can be enabled/disabled separately for the interrupt
code indication field.
Buffer 1
1
0
0
1
0
Buffer 2
1
0
0
1
1
Buffer 3
1
0
1
0
0
Buffer 4
1
0
1
0
1
Buffer 5
1
0
1
1
0
Buffer 6
1
0
1
1
1
Buffer 7
1
1
0
0
0
Buffer 8
1
1
0
0
1
Buffer 9
1
1
0
1
0
15
EICEN
14
0
ICEN[14:0]
0
r/w
www.national.com
114
Table 36
Highest Priority Interrupt Code
(CICEN = FFFF)
CAN interrupt
request
IRQ
IST3
IST2
Table 37 Error Field Identifier
IST1
Field
EFID3
EFID2
EFID1
EFID0
IST0
IDE
1
0
1
0
1
0
1
1
Buffer 10
1
1
0
1
1
EXTENDED
ARBITRATION
Buffer 11
1
1
1
0
0
R1/R0
1
1
0
0
Buffer 12
1
1
1
0
1
DLC
1
1
0
1
Buffer 13
1
1
1
1
0
DATA
1
1
1
0
Buffer 14
1
1
1
1
1
CRC
1
1
1
1
20.9.17 CAN Error Counter Register (CANEC)
EBID[5:0]
The Can Error Counter Register contains the value of the
CAN Receive Error Counter and the CAN Transmit Error
Counter.
15
8
7
0
REC[7:0]
TEC[7:0]
0
r
REC[7:0]
CAN Receive Error Counter. The bits REC[7:0]
holds the value of the receive error counter.
CAN Transmit Error Counter. The bits TEC[7:0]
holds the value of the transmit error counter.
TEC[7:0]
Error Bit Identifier. The EDIAG[9:4] bits contain
the number (position) of the incorrect bit within
the erroneous frame field. The bit number
starts with the value equal to the respective
frame field length minus one at the beginning of
each field and is decremented with each CAN
bit. Figure71 shows an example on how the
EBID is calculated.
r
r
r
r
The CAN Error Diagnostic (CEDIAG) register provides information about the last detected error. CR16CAN is able to
identify the field within the CAN frame format, in which the error occurred, and it identifies the bit number of the erroneous
bit within the according frame field. The CPU has read only
access and all bits will be cleared upon reset.
15
14
13
12
11
10
9
4
EBID[5:0]
3
0
EFID[3:0]
0
r
EFID[3:0]
Error Field Identifier. The EDIAG bits 3...0 identify the frame field in which the last error occurred. How the various frame fields are coded
into the EFID bits is shown in Table37.
Table 37 Error Field Identifier
Field
EFID3
EFID2
EFID1
EFID0
ERROR
0
0
0
0
ERROR DEL
0
0
0
1
ERROR ECHO
0
0
1
0
BUS IDLE
0
0
1
1
ACK
0
1
0
0
EOF
0
1
0
1
INTERMISSION
0
1
1
0
SUSPEND
TRANSMISSION
0
1
1
1
SOF
1
0
0
0
ARBITRATION
1
0
0
1
r
incorrect
bit
20.9.18 CAN Error Diagnostic Register (CEDIAG)
Reserved DRIVE MON CRC STUFF TXE
r
data field
Figure 71. EBID Example
Assume the EFID resulted in 1110 2 and the EBID showed a
value of 111001 2 . This means that faulty field was the data
field. To calculate the bit position of the error, the DLC of the
message needs to be known. For example, for a DLC of 8
data bytes, the bit counter starts with the value: 8 x 8 - 1 = 63;
so when EBID[5:0]=111001 2 = 57, then the bit number was
63 - 57 = 6.
The following bits provide an information of the error type.
TXE
STUFF
CRC
MON
DRIVE
115
Transmit Error. If set, this bit indicates that the
CR16CAN was an active transmitter at the time
the error occurred. If reset, the CR16CAN was
a receiver.
Stuff Error. if set, this bit indicates that a the bit
stuffing rule was violated at the time the error
occurred. Note that certain bit fields do no use
bit stuffing and therefore this bit may be ignored in those.
CRC Error. if set, this bit indicates that the CRC
is invalid. This bit should only be used if the
EFID shows the code of the ACK field.
Monitor. This bit shows the bus value on the
CANRX pin as seen by the CR16CAN at the
time of the error.
Drive. This bit shows the output value on the
CANTX pin at the time of the error. Note that a
www.national.com
receiver will not drive the bus except during
ACK and during an active error flag.
Table 38
External CR16CAN Pins
20.9.19 CAN Timer Register (CTMR)
The current value of the Time Stamp counter as described in
section 20.8 can be monitored via the CAN Timer Register.
15
0
CTMR[15:0]
0
r
Description
CANTX
Output
Transmit data to the CAN bus
CANRX
Input
Receive data from the CAN bus
20.10.2 Transceiver Connection
An external Transceiver Chip needs to be connected between the CAN block and the bus. It is used to establish a bus
connection in differential mode and furthermore provides the
driver and protection requirements.
Figure72 shows a possible ISO-High-Speed configuration.
As described in Time Stamp Counter on page 105, the contents of CTMR are captured into the Time Stamp register of
the message buffer after successfully sending or receiving a
frame.
120 termination
CAN Bus Line
CORE BUS
SYSTEM START-UP AND MULTI-INPUT
WAKE-UP
to other modules
After system start-up, all CR16CAN related registers are in
their reset state. The CR16CAN module can be enabled after
all configuration registers are set to their desired value. The
following initial setting need to be made:
Transceiver Chip
VCC 3
5 REF BUS_H 7
CANRX 4 RX
BUS_L 6
1 TX
RS
GND
CANTX
8
2
CR16CAN
— configure the CAN Timing register (CTIM) See “Bit
Time Logic” on page94.
— configure every buffer to its function as receive/transmit Buffer Status/Control Register (CNSTAT) on page
106.
— set the acceptance filtering masks. See “Acceptance
Filtering” on page96.
— enable the CR16CAN interface. See “CAN Global
Configuration Register (CGCR)” on page109.
VCC
GND GND
120
Figure 72.
External Transceiver Connection
(ISO-High-Speed)
20.10.3 Timing Requirements
Processing messages and updating message buffers require
a certain number of clock cycles by CR16CAN as shown in
Table39. These requirements may lead to some restrictions
regarding the Bit Time Logic settings and the overall
CR16CAN performance which are described below in more
detail.
Table 39 CR16CAN Internal Timing
Before disabling the CR16CAN module, the user has to
make sure that no transmission is still pending.
Note: The device can be awaken from a power saving mode
by an activity on the CAN bus by selecting the CAN RX pin
as an input to the Multi-Input Wake-Up module. In this case
the CR16CAN module must not be disabled before entering
the power saving mode. Disabling the CR16CAN module
also disables the CAN RX pin.
As an alternative, the CAN RX pin can be connected to any
other input pin of the Multi-Input Wake-Up module. This input
channel must then be configured to trigger a wake-up event
on a falling edge (if a dominant bit is represented by a low level). In this case the CR16CAN module can be disabled before entering a power saving mode. After the device has
been waken up, the user has to manually enable the
CR16CAN again. All configuration and buffer registers still
contain the same data as prior to the power down phase.
task
# cycles a
occurrence/
frame b
copy hidden buffer to receive
17
0-1
message buffer
update status from TX_RTR
3
0-15
to TX_ONCE_RTR
schedule a message for trans2
0-1
mission
a. Wait cycles need to be added for CPU access to
the message memory as described in CPU
Access to CR16CAN Registers/Memory on
page 105.
20.10.1 External Connection
The CR16CAN uses two external pins, CANTX and CANRX
to connect to the physical layer of the CAN interface. They
provide the functionality as described in Table38.
www.national.com
Type
The logic levels are configurable by means of two control
flags CTX and CRX of the Global Configuration Register
CGCR (see “CAN Global Configuration Register (CGCR) ” on
page 109.
The CAN Time register is a free running 16-bit counter. It contains the number of CAN bits recognized by CR16CAN since
the register has been reset. The counter starts to increment
from the value 0000 16 after a hardware reset. If the Timer
Stamp enable flag (TSTPEN) in the CAN global configuration
register (CGCR) is set, the counter will also be reset upon a
message transfer of the message buffer 0.
20.10
Signal Name
b. Depends on the number of matching identifiers.
116
The critical path derives from receiving a remote frame which
triggers the transmission of one or more data frames. There
are a minimum of four bit times in-between two consecutive
frames. These bit times start at the validation point of received frame (reception of 6th EOF bit) and end at the earliest possible transmission start of the next frame, which is
after the third intermission bit at 100% burst bus load.
20.10.4 Bit Time Logic Calculation Examples
These four bit times have to be set in perspective with the
timing requirements of the CR16CAN.
TSEG1 = TSEG1[3:0] + 1 = 3 + 1 = 4
The minimum duration of the four CAN bit times is determined by the following Bit Time Logic settings:
SJW = TSEG2 = 3
PSC
The calculation of the CAN bus clocks using CKI = 16MHz is
shown in the following examples. The desired baud rate for
both examples is 1Mbit/s.
Example 1
PSC = PSC[5:0] + 2 = 0 + 2 = 2
TSEG2 = TSEG2[2:0] + 1 = 2 + 1 = 3
— sample point positioned at 62.5% of bit time
— bit time = 125ns x (1 + 4 + 3 ± 3) = (1 ± 0.375)µs
— busclock = 16MHz / (2 x (1 + 4 + 3)) = 1Mbit/s (nominal)
= PSCmin = 2
TSEG1 = TSEG1min = 2
Example 2
TSEG2 = TSEG2min = 1
PSC = PSC[5:0] + 1 = 2 + 2 = 4
bit time = Synch + Time Segment 1 + Time Segment 2
= (1 + 2 + 1) tq = 4 tq
= (4 tq x PSC) clock cycles
= (4 tq x 2) clock cycles = 8 clock cycles
TSEG1 = TSEG1[3:0] + 1 = 1 + 1 = 2
TSEG2 = TSEG2[2:0] + 1 = 0 + 1 = 1
SJW = TSEG2 = 1
For these minimum BTL settings, four CAN bit times take 32
clock cycles.
• sample point positioned at 75% of bit time
• bit time = 250ns x (1 + 2 + 1 ± 1) = (1 ± 0.25) µs
• busclock = 16MHz / (2 x (1 + 4 + 3)) = 1Mbit/s (nominal)
The following is an example that assumes typical case:
20.10.5 Acceptance Filter Considerations
—
—
—
—
minimum BTL settings
reception and copy of a remote frame
update of one buffer from TX_RTR
schedule of one buffer from transmit
As outlined in Table39 the copy process, update and scheduling the next transmission gives a total of 17+3+2=22 clock
cycles. Therefore under these conditions there is no timing
restriction.
The CR16CAN provides two acceptance filter masks GMSK
and BMSK as described in Acceptance Filtering on page 96,
Global Mask Registers (GMSK — GMSKB and GMSKX) on
page 112 and Basic Mask Registers (BMSK — BMSKB and
BMSKX) on page 113. These masks allow filtering of up to 32
bits of the message, which includes the standard identifier,
the extended identifier as well as the frame control bits RTR,
SRR and IDE.
The following example assumes the worst case:
20.10.6 Remote Frames
—
—
—
—
minimum BTL settings
reception and copy of a remote frame
update of the 14 remaining buffers from TX_RTR
schedule of one buffer for transmit
All these actions in total require 17 + 14 x 3 + 2 = 61 clock
cycles to be executed by CR16CAN. This leads to the limitation of the Bit Time Logic of 61 / 4 = 15.25 clock cycles per
CAN bit as a minimum, resulting in the minimum clock frequencies listed below (the frequency depends on the desired
baud rate and assumes the worst case scenario can occur in
the application).
Remote frames can be automatically processed by the
CR16CAN interface. However, to fully enable that feature,
the RTR/XRTR bits (for both standard and extended frames)
within the BMSK and/or GMSK register need to be set to
“don’t care”. This is because a remote frame with the RTR bit
being set to “1” should trigger the transmission of a data
frame with the RTR bit set to “0” and therefore the ID bits of
the received message need to pass through the acceptance
filter. The same applies to transmitting remote frames and
switching to receive the corresponding data frames.
Table40 gives examples for the minimum clock frequency in
order to ensure proper functionality at various CAN bus
speeds.
Table 40 Min. Clock Frequency Requirements
Baud Rate
min. clock frequency
1Mbit/sec
15.25MHz
500kbit/sec
7.625MHz
250kbit/sec
3,81MHz
117
www.national.com
21.0 Analog Comparators
The Dual Analog Comparator (ACMP2) module contains two
independent analog comparators with all necessary control
logic. Each comparator unit compares the analog input voltages applied to two input pins and determines which voltage
is higher. The comparison results can be placed on two output pins and/or read by the software from a register.
CMP1EN
Figure73 is a block diagram of the Dual Analog Comparator
module.
CMP2EN
The two comparators are designated Comparator 1 (CMP1)
and Comparator 2 (CMP2). Each comparator has a positive
and a negative input, called CMP1P and CMP1N for Comparator 1 and CMP2P and CMP2N for Comparator 2. An optional output, CMP1O for Comparator 1 or CMP2O for
Comparator 2, allows the external hardware to read the comparison results. If the positive input is greater than the negative input, the result is a logic 1. Otherwise, the result is a
logic 0. These same results are available to the software by
reading the CMPCTRL register. CMP1OP and CMP2OP are
the direct outputs of the analog comparator. These signals
are connected to the channels of the Multi-Wake-Up module.
21.1
CMP1OE
CMP2OE
ANALOG COMPARATOR CONTROL/
STATUS REGISTER (CMPCTRL)
21.2
The CMPCTRL register is a byte-wide, read/write register
that controls the comparator module and contains the comparison results. The control bits are read/write bits and the result bits are read-only bits. This register is cleared upon
reset. The register format is shown below.
7
6
5
4
3
2
1
CMP2RD
0
Using a comparator's output pin is optional. If it is to be used,
it must be configured to operate as an output in the alternate
function mode. The comparison result bits in the CMPCTRL
register are available to the CPU whether or not the output
pin is enabled.
Comparator 1 Read. This read-only bit contains the output of Comparator 1 when the
comparator
is
enabled
(CMP1EN=1).
CMP1RD is set to 1 when the voltage on
CMP1P is greater than the voltage on CMP1N.
This bit is always 0 when Comparator 1 is disabled.
Comparator 2 Read. This read-only bit contains the output of Comparator 2 when the
CMP1P
CMP1N
ANALOG COMPARATOR USAGE
The comparator I/O pins are alternate functions of the Port L
pins. In order for a comparator to operate, its two input pins
must be configured to operate as inputs in the alternate function mode.
Reserved CMP2OE CMP1OE CMP2EN CMP1EN CMP2RD CMP1RD
CMP1RD
comparator
is
enabled
(CMP2EN=1).
CMP2RD is set to 1 when the voltage on
CMP2P is greater than the voltage on CMP2N.
This bit is always 0 when Comparator 2 is disabled.
Comparator 1 Enable. This read/write bit enables (1) or disables (0) Comparator 1.
Comparator 2 Enable. This read/write bit enables (1) or disables (0) Comparator 2.
Comparator 1 Output Enable. This read/write
bit, when set to 1, enables the use of the
CMP1O pin as the output of Comparator 1
when Comparator 1 is enabled (CMP1EN=1).
If Comparator 1 is disabled (CMP1EN=0), setting the CMP1OE bit results in a logic 0 on the
CMP1O output pin.
Comparator 2 Output Enable. This read/write
bit, when set to 1, enables the use of the
CMP2O pin as the output of Comparator 2
when Comparator 2 is enabled (CMP2EN=1).
If Comparator 2 is disabled (CMP2EN=0), setting the CMP2OE bit results in a logic 0 on the
CMP2O output pin.
The comparators uses DC current whenever they are enabled. Therefore, in order to reduce power consumption, it is
recommended that the comparators be disabled when they
are not needed, especially before entering any of the Power
Save modes.
CMP1OP
+
CMP1
_
CMP1O
CMP1EN
CMP1OE
CONTROL + STATUS
CMP2EN
CMP2P
CMP2N
CMP2OE
+
CMP2
_
CMP2OP
Figure 73. Dual Analog Comparator Block Diagram
www.national.com
118
CMP2O
22.0 A/D Converter
The A/D Converter (ADC) module is a 12-channel, multiplexed-input, analog-to-digital converter. The A/D Converter
receives an analog voltage on an input pin and converts that
voltage into an 8-bit digital value using successive approximation. The CPU can then read the result from a memorymapped register. The module supports four automated operating modes, providing single-channel or 4-channel scanned
operation in single-conversion or continuous mode.
Table41. The A/D converter must be disabled when switching to a different mode.
Table 41
SCAN CONT
ADC Operation Modes
Mode
00
0
Single Channel, Single Conversion
00
1
Single Channel, Continuous Conversion
Figure74 is a block diagram of the A/D Converter module.
01
0
4 Channels Scan, Single Conversion
The analog input signal is selected from the analog inputs using a 12-channel analog multiplexer. The input pins are alternate functions of Port I.
01
1
4 Channel Scan, Continuous Conversion
A sample-and-hold circuit samples the analog voltage prior
to conversion and holds it stable throughout the conversion
process. A programmable initial delay period allows the sampled voltage to stabilize before the conversion process begins.
The input voltage range is from 0V to VREF (the A/D reference voltage). The device has a separate pin, VREF, for the
reference voltage.
A capacitor should be connected between the VREF and the
AVCC pin in order to minimize noise. The recommended value for this capacitor is about 0.47µF.
The internal analog-to-digital converter block is based on a
successive approximation algorithm, which compares the
sampled voltage against an internally generated sequence of
analog voltages. The result is a linear conversion of the analog voltage to an unsigned 8-bit value ranging from 00 hex for
0.0 volts to FF hex for VREF.
The clock used by the converter block is generated by a clock
divider that scales down the system clock by a programmable factor. The conversion algorithm requires ten A/D Converter clock cycles, or 10 microseconds at the maximum
allowed A/D Converter clock rate of 2 MHz.
Conversion can start after the power supply is stable and ADCEN set for 30 µs.
The conversion results are stored in a 4-level data buffer. Depending on the operating mode, the buffer can hold the results of four successive conversions from a single channel or
four conversions from adjacent channels scanned in sequence.
22.1
OPERATING MODES
The A/D Converter can be configured to operate in any one
of four modes:
—
—
—
—
Single channel, single conversion
Single channel, continuous conversion
4-channel scan, single conversion
4-channel scan, continuous conversion
The configuration is set by the SCAN and CONT fields in the
ADC Control 2 Register (ADCCNT2), as indicated in
22.1.1
Single Channel, Single Conversion Mode
In the single channel, single conversion mode, the A/D Converter performs a single conversion using a specified channel.
The software starts a conversion by setting the START bit in
the ADCCNT2 register. Upon completion of the conversion,
the A/D Converter places the result in register ADDATA0,
clears the START bit, and sets the EOC (end of conversion)
bit in the ADCST register. If the A/D Converter interrupt is enabled, an interrupt to the CPU is generated at this time.
22.1.2
Single Channel, Continuous Conversion Mode
In the single channel, continuous conversion mode, the A/D
Converter performs conversions repeatedly using the same
specified channel.
The software starts a conversion sequence by setting the
START bit. The A/D Converter performs four A/D conversions in sequence using the same channel, pausing only for
the programmable sampling delay time used in all conversion operations. It loads the four results into the A/D data registers in sequence, starting with ADDATA0 and ending with
ADDATA3. After it loads all four registers, it sets the EOC
(end of conversion) bit. If the A/D Converter interrupt is enabled, an interrupt to the CPU is generated at this time.
The START bit remains set until cleared by the software. If
the software does not clear the START bit, the A/D Converter
continues performing conversions using the same input
channel, storing the results in ADDATA0 following ADDATA3.
To prevent an overrun error, the software must read the results from the data registers before the A/D Converter writes
the next result into ADDATA0 following ADDATA3.
When the software clears the START bit, the A/D Converter
first completes the conversion currently in progress, then
stops and sets the EOC bit. A 2-bit buffer pointer in the
ADCST register points to the register containing the final result.
22.1.3
4-Channel Scan, Single Conversion Mode
In the 4-channel scan, single conversion mode, the A/D Converter performs four conversions using four adjacent input
channels.
The software starts the conversion sequence by setting the
START bit. The A/D Converter performs four A/D conversions in sequence using four adjacent channels, starting with
the specified channel and pausing only for the programmable
sampling delay time. It loads the four results into the A/D data
registers in sequence, starting with ADDATA0 and ending
119
www.national.com
4
CH0
CH1
CH2
CH3
CH4
CH5
CH6
CH7
CH8
CH9
CH10
CH11
CONFIGURATION
STATUS
&
CONTROL
PERIPHERAL
BUS
ANALOG
12:1
ANALOG
MUX
DATA
BUFFER
TO
SAMPLE
&
HOLD
DIGITAL
CONVERTER
VREF
CLK
CLOCK
DIVIDER
CLK
Figure 74. A/D Converter Block Diagram
with ADDATA3. After it loads all four registers, it clears the
START bit and sets the EOC (end of conversion) bit. If the A/
D Converter interrupt is enabled, an interrupt to the CPU is
generated at this time.
22.1.4
22.2.1
The ADCST register is a byte-wide register that indicates the
current status of the A/D Converter. One bit in this register,
the OVF flag bit, is cleared by writing a 1 to its bit position.
The other bits are read-only bits, so the values written to
them are ignored. Upon reset, the register is set to 30 hex.
The register format is shown below.
Channel Scan, Continuous Conversion Mode
In the 4-channel scan, continuous conversion mode, the A/D
Converter performs conversions repeatedly using four adjacent input channels.
7
6
Reserved
The software starts conversion operations by setting the
START bit. The A/D Converter performs four A/D conversions in sequence using four adjacent channels, starting with
the specified channel and pausing only for the programmable
sampling delay time. It loads the four results into the A/D data
registers in sequence, starting with ADDATA0 and ending
with ADDATA3. After it loads all four registers, it sets the EOC
(end of conversion) bit. If the A/D Converter interrupt is enabled, an interrupt to the CPU is generated at this time.
EOC
The START bit remains set until cleared by the software. If
the software does not clear the START bit, the A/D Converter
continues performing conversions, repeating the same sequence using the same four input channels and the same sequence of data registers. To prevent an overrun error, the
software must read the results from the data registers before
the A/D Converter writes the next result into ADDATA0.
BUSY
OVF
When the software clears the START bit, the A/D Converter
first completes the 4-channel conversion sequence currently
in progress, then stops and sets the EOC bit.
22.2
A/D CONVERTER REGISTERS
The software controls the A/D Converter and reads the A/D
results by accessing the ADC registers. There are eight such
registers:
—
—
—
—
—
BUFPTR
ADC Status Register (ADCST)
ADC Control 1 Register (ADCCNT1)
ADC Control 2 Register (ADCCNT2)
ADC Control 3 Register (ADCCNT3)
ADC Data Registers (ADDATA0 through ADDATA3)
www.national.com
ADC Status Register (ADCST)
120
5
4
BUFPTR
3
Reserved
2
OVF
1
BUSY
0
EOC
End of Conversion. This read-only bit reports
the status of the most recent A/D Converter operation. When cleared to 0, it indicates that the
conversion is not complete. When set to 1, it indicates that the conversion is complete. The
hardware sets this bit when it places the conversion results in the buffer and clears it when
any of the data registers are read.
ADC Busy. This read-only bit is set to 1 when
the A/D Converter is busy converting data and
is cleared to 0 when the A/D Converter is idle
or disabled.
Overflow. The hardware sets this bit to 1 when
the A/D Converter finishes a conversion and attempts to store the results in one of the data
registers (ADDATA0-ADDATA3) while the register is full. When this happens, the A/D Converter overwrites the data in the data register,
sets the OVF flag, and continues operating.
The OVF flag remains set until cleared by the
software. The software clears the flag by writing a 1 to it. Writing a 0 to this bit has no effect.
Buffer Pointer. This 2-bit, read-only field identifies the data register that was most recently
written with new data:
00 = ADDATA0
01 = ADDATA1
10 = ADDATA2
11 = ADDATA3
This register is initialized to 11 when a new
conversion is started (when ADCCNT2.START
is changed from 0 to 1) and is automatically incremented every time a result is written to buffers ADDATA0-ADDATA3. The result is a fourentry cyclic FIFO buffer, with BUFPTR pointing
to the last entry written by the A/D Converter.
22.2.2
CONT
ADC Control 1 Register (ADCCNT1)
The ADCCNT1 register is a byte-wide, read/write register
used to enable the A/D Converter and its interrupts, and also
to control the reference voltage source. When writing to this
register, all reserved bits must be written with 0 for the A/D
Converter to function properly. Changing any bits other than
ADCEN (bit 0) is not allowed while the A/D Converter is active (ADCST.BUSY or ADCCNT2.START set). Upon reset,
all non-reserved bits are cleared to 0. The register format is
shown below.
7
6 5 4
Reserved
ADCEN
3
2
INTE
1
Reserved
SCAN
START
0
ADCEN
A/D Converter Enable. Setting this bit enables
the A/D Converter and allows a conversion to
be started by setting the start bit
(ADCCNT2.START). Clearing the ADCEN bit
disables the A/D Converter, terminates any
conversion in progress, and clears the ADC
status flags (ADCST.EOC, ADCST.BUSY,
ADCST.OVF, and ADCCNT2.START).
Interrupt Enable. This bit enables (1) or disables (0) A/D Converter interrupts. If enabled,
and interrupt occurs at the end of a conversion
sequence or when the ADC data buffer is full,
depending on the operating mode.
INTE
22.2.4
ADC Control 2 Register (ADCCNT2)
7
Reserved
CDIV
The ADCCNT2 register is a byte-wide, read/write register
used to specify the A/D Converter operating mode and to
start conversion operations. All register fields other than the
START bit should be changed only while the A/D Converter
is inactive (START=0). Data written to the SCAN and CONT
fields is ignored if the START bit is already set. Upon reset,
the non-reserved bits of this register are cleared to 0. The
register format is shown below.
7
START
CHANNEL
6
5
SCAN
4
CONT
3
2
1
0
CHANNEL
Channel Select. This 4-bit field selects one of
the twelve analog input channels as follows:
0000 = ACH0
0001 = ACH1
0010 = ACH2
0011 = ACH3
0100 = ACH4
0101 = ACH5
0110 = ACH6
0111 = ACH7
1000 = ACH8
1001 = ACH9
ADC Control 3 Register (ADCCNT3)
The ADCCNT3 register is a byte-wide, read/write register
used to specify the analog sampling time delay and the divide-by factor for generating the ADC clock. This register
should be written only when the A/D Converter is disabled
(ADCCNT1.ADCEN=0). Upon reset, the non-reserved bits of
the ADCCNT3 register are cleared to 0. The register format
is shown below.
All reserved bits must be written with 0 for ADC to operate
properly.
22.2.3
1010 = ACH10
1011 = ACH11
11XX = reserved
Continuous Conversion. When cleared to 0,
the A/D Converter stops operating upon completion of the programmed conversion cycle (a
single conversion or a sequence of four conversions on four channels). When set to 1, the
A/D Converter operates continuously by repeating the programmed conversion cycle.
Scan Mode. This 2-bit field selects the singleconversion mode or 4-channel scan mode as
follows:
00 = single-conversion mode
01 = 4-channel scan mode
1X = reserved
Start Conversion. The software sets this bit to
1 to start a conversion or a 4-channel conversion cycle. In the “continuous” mode, this bit remains set until cleared by the software. In the
“single” (non-continuous) mode, the hardware
clears this bit upon completion of the programmed conversion cycle. The software
should not attempt to set this bit while the A/D
Converter is busy (ADCST.BUSY=1).
DELAY
121
6
PWREN
5
4
DELAY
3
2
1
CDIV
0
Clock Divide. This 3-bit field sets the divide-by
factor for generating the A/D Converter clock
from the system clock. The frequency of the A/
D Converter clock is equal to the system clock
divided by the programmed factor. The resulting A/D Converter clock frequency must be
less than or equal to 2 MHz. The divide-by factor is defined as follows:
000 = divide by 1
001 = divide by 2
010 = divide by 4
011 = divide by 8
100 = divide by 16
101 = divide by 32
Other = reserved
Sampling Time Delay. This 3-bit field defines
the number of A/D Converter clock cycles of
delay from the time that the input channel is selected until the analog voltage is sampled. The
programmed delay should be sufficient, dependent on the source impedance, to allow the
sampled signal to reach its final level before the
conversion begins. The delay is defined as follows:
www.national.com
PWREN
22.2.5
000 = 1 A/D Converter clock cycle
001 = 2 A/D Converter clock cycles
010 = 4 A/D Converter clock cycles
011 = 8 A/D Converter clock cycles
100 = 16 A/D Converter clock cycles
101 = 32 A/D Converter clock cycles
110 = 64 A/D Converter clock cycles
111 = reserved
Power Down Enable. controls the condition
when the ADC is powered down. When
PWREN is cleared (0), the ADC powers down
upon reset. When PWREN is set (1), the ADC
powers down when the ADCEN bit is low.
D conversion time is 16 microseconds (ten clock A/D Converter clock cycles).
The programmable sampling time delay should be made
small for faster operation, but large enough to allow the input
voltage to settle. The internal resistance and capacitance of
the A/D Converter, together with the source resistance of the
device that drives the A/D input determine the charge-up time
required for the voltage to settle. Figure75 shows a schematic of the charge-up circuit. For the values of RAIN and C AIN ,
see Section25.0.
Interrupts or polling can be used to read the A/D Converter
results. For interrupts, the A/D Converter interrupt must be
enabled by setting the ADCCNT1.INTE bit. The interrupt is
cleared automatically when any one of the data registers
(ADDATA0-ADDATA3) is read. For polling, the software
reads the ADCST.EOC bit to determine whether the conversion sequence is completed.
ADC Data Registers (ADDATA0-ADDATA3)
The four ADC Data Registers (ADDATA0 through ADDATA3)
are byte-wide, read/write registers that hold the conversion
results, which are stored sequentially starting with ADDATA0
and ending with ADDATA3. The results held in these registers are valid only after the ADCST.EOC flag is set. Upon reset, the contents of these registers are undefined.
Once the A/D Converter configuration has been set up, the
software can use the following procedure to perform an A/D
conversion sequence:
The value read from a data register is a linear mapping of the
analog input voltage to an 8-bit value. The value 00 hex represents 0.0 volts and the value FF hex represents the reference voltage, V REF.
22.3
1. Enable the A/D Converter by setting the ADCCNT1.ADCEN bit and wait 30 µs before performing any conversion.
2. Select the operating mode and channel by writing to the
SCAN, CONT, and CHANNEL fields of the ADCCNT2
register. At the same time, start the conversion by setting
the START bit in the same register.
3. Wait until the conversion is finished, either by polling or
using the A/D Converter interrupt.
4. Read the conversion results from the data registers,
ADDATA0 through ADDATA3 (or just ADDATA0 in the
single-channel, single-conversion mode).
5. In the continuous conversion modes, repeat Step 3 and
Step 4 for as long as samples are needed. Then stop the
A/D Converter by clearing either the START bit
(ADCCNT2.START) or the A/D Converter enable bit
(ADCCNT1.ADCEN).
A/D CONVERTER PROGRAMMING
The software should set the A/D Converter configuration before it enables the A/D Converter module. The configuration
consists of the following settings:
— ADC clock rate: ADCCNT3.CDIV
— Sampling delay: ADCCNT3.DELAY
— Interrupt enable (if required): ADCCNT1.INTE
The ADC clock is created by scaling down the system clock.
The fastest allowable clock for the A/D Converter is 2 MHz.
Therefore, for the fastest possible operation of the A/D Converter, use the smallest available divide-by factor that results
in a clock frequency of 1 MHz or lower. The available divideby factors are 1, 2, 4, 8, 16, and 32.
To minimize power consumption, the A/D Converter should
be disabled when it is not needed, especially before entering
a Power Save mode.
For example, if the system clock is 10 MHz, use a divide-by
factor of 16. In that case, the A/D Converter clock frequency
is 625 kHz, the clock period is 1.6 microseconds, and the A/
RAIN
RSOURCE
Input
signal
CAIN
Analog
Multiplexer
Sample &
Hold
A/D Converter
Figure 75. Sample-and-Hold Charge-Up Schematic
www.national.com
122
23.0 Memory Map
The CompactRISC architecture supports a uniform linear address space of 2 megabytes. The device implementation of
this architecture uses only the lowest 128K bytes of address
space, ranging from 0000 to 1FFFF hex. Table42 is a memory map showing the types of memory and peripherals that
occupy this memory space. Address ranges not listed in the
table are reserved and should not be read or written.
Table 42
Device Memory Map
Address
Range (hex)
Description
0000-7FFF
Flash Program Memorya
8000-BFFF
Flash Program Memory (48K bytes)
C000-CBFF
Static RAM (3K bytes)
E000-E5FF
ISP Memory(1.5K bytes)
E800-EFFF
Lower Endurance Flash EEPROM Data
Memory (2K bytes)
F000-F07F
High Endurance Flash EEPROM Data
Memory (128 bytes)
F400-F7FF
CAN buffers and registers (1K bytes)
F800-FAFF
BIU Peripherals (768 bytes)
FB00-FB06
Port B registers
FB00-FBFF
I/O Expansion + Ports PB & PC (256bytes)
FB10-FB16
Port C registers
FC00-FFFF
Peripherals and other I/O Ports (1K bytes)
FC40-FC8A
Clock, Power Management, and Wake-Up
registers
FCA0-FCA8
Port G registers
FCC0-FCC8
Port H registers
FF00-FF08
Port L registers
FD20-FD28
Port F registers
FE00-FE1E
Interrupt Control Unit registers
FE40-FE4E
USART 1 registers
FE60-FE66
MICROWIRE registers
FE80-FE8E
USART 2 registers
FEC0-FECA
ACCESS.bus registers
FEE0-FEE8
Port I registers
FF20-FF2A
Timer and WATCHDOG registers
FF40-FF50
Multi-function Timer1 registers
FF60-FF70
Multi-function Timer2 registers
FF80-FFA4
Versatile Timer Unit registers
FFC0-FFD0
A/D Converter registers
FFE0-FFE0
Analog Comparator register
1C000-1FFFF
Flash Program Memory (16K bytes)b
Table43 is a detailed memory map showing the specific
memory address of the memory, I/O ports, and registers. The
table shows the starting address, the size, and a brief description of each memory block and register. For detailed information on using these memory locations, see the
applicable sections in the data sheet.
All addresses not listed in the table are reserved and should
not be read or written. An attempt to access an unlisted address will have unpredictable results.
Each byte-wide register occupies a single address and can
be accessed only in a byte-wide transaction. Each word-wide
register occupies two consecutive memory addresses and
can be accessed only in a word-wide transaction. Both the
byte-wide and word-wide registers reside at word boundaries
(even addresses). Thus, each byte-wide register uses only
the lowest eight bits of the internal data bus.
Most device registers are read/write registers. However,
some registers are read-only or write-only, as indicated in the
table. An attempt to read a write-only register or to write a
read-only register will have unpredictable results.
When the software writes to a register in which one or more
bits are reserved, it must write a zero to each reserved bit unless indicated otherwise in the description of the register.
Reading a reserved bit returns an undefined value.
a
. 32K ROM or Flash, size depends on device specifications.
123
www.national.com
Table 43
Device Detailed Memory Map
Size
Register
Address
(hex)
32K/48K
3K
0000
C000
Flash EEPROM Program Memory
On-Chip RAM
2K
E800
Low Endurance Flash EEPROM Data Memory
1.5K
128
E000
F000
ISP Memory
High Endurance Flash EEPROM Data Memory
CMB0_CNTSTAT
word
F400
Read/Write
CAN message buffer 0 Status Register
CMB0_TSTP
CMB0_DATA3
word
word
F402
F404
Read/Write
Read/Write
CAN message buffer 0 Time stamp Register
CAN message buffer 0 Data 3 Register
CMB0_DATA2
word
F406
Read/Write
CAN message buffer 0 Data 2 Register
CMB0_DATA1
CMB0_DATA0
word
word
F408
F40A
Read/Write
Read/Write
CAN message buffer 0 Data 1 Register
CAN message buffer 0 Data 0 Register
CMB0_ID0
word
F40C
Read/Write
CAN message buffer 0 Identifier 0 Register
CMB0_ID1
CMB1
word
8-word
F40E
F410
Read/Write
Read/Write
CAN message buffer 0 Identifier 1 Register
CAN message buffer 1 Register
CMB2
8-word
F420
Read/Write
CAN message buffer 2 Register
CMB3
CMB4
8-word
8-word
F430
F440
Read/Write
Read/Write
CAN message buffer 3 Register
CAN message buffer 4 Register
CMB5
8-word
F450
Read/Write
CAN message buffer 5 Register
CMB6
CMB7
8-word
8-word
F460
F470
Read/Write
Read/Write
CAN message buffer 6 Register
CAN message buffer 7 Register
CMB8
8-word
F480
Read/Write
CAN message buffer 8 Register
CMB9
CMB10
8-word
8-word
F490
F4A0
Read/Write
Read/Write
CAN message buffer 9 Register
CAN message buffer 10 Register
CMB11
8-word
F4B0
Read/Write
CAN message buffer 11 Register
CMB12
CMB13
8-word
8-word
F4C0
F4D0
Read/Write
Read/Write
CAN message buffer 12 Register
CAN message buffer 13 Register
CMB14
8-word
F4E0
Read/Write
CAN message buffer 14 Register
CGCR
CTIM
word
word
F500
F502
Read/Write
Read/Write
CAN Global Configuration Register
CAN Timing Register
Register Name
Access Type
Contents
GMSKX and GMSK
word
F504
Read/Write
CAN Global Mask Registers
GMSKX and GMSKB
CIEN
word
word
F508
F50C
Read/Write
Read/Write
CAN Basic Mask Registers
CAN Interrupt Enabled Register
CIPND
word
F50E
Read/Write
CAN Interrupt Pending Register
CICLR
CICEN
word
word
F510
F512
Read/Write
Read/Write
CAN Interrupt Clear Register
CAN Interrupt Code Enable Register
CSTPND
word
F514
Read/Write
CAN Status Pending Register
CANEC
CEDIAG
word
word
F516
F518
Read/Write
Read/Write
CAN Error Counter Register
CAN Error Diagnostic Register
CTMR
word
F51A
Read/Write
CAN Timer Register
BCFG
IOCFG
byte
word
F900
F902
Read/Write
Read/Write
BIU Configuration Register
I/O Zone Configuration Register
SZCFG0
word
F904
Read/Write
Static Zone 0 Configuration Register
SZCFG1
MCFG
word
byte
F906
F910
Read/Write
Read/Write
Static Zone 1 Configuration Register
Module Configuration Register
MSTAT
byte
F914
Read Only
Module Status Register
www.national.com
124
Table 43
Device Detailed Memory Map
Register Name
Size
Register
Address
(hex)
Access Type
FLCTRL1
byte
F930
Read/Write
Flash EEPROM Program Memory Control Register 1
FLSEC
ISPKEY
byte
byte
F932
F934
Read/Write
Read/Write
Flash EEPROM Program Memory Security Register
ISP Memory Write Key Register
FLCTRL2
word
F936
Read/Write
Flash EEPROM Program Memory Control Register 2
DMCSR
DMPSLR
byte
byte
F940
F942
Read/Write
Read/Write
EEPROM Data Memory Control and Status Register
EEPROM Data Memory Prescaler Register
DMSTART
byte
F944
Read/Write
Data Memory Start Time Reload Register
DMTRAN
DMPROG
byte
byte
F946
F948
Read/Write
Read/Write
Data Memory Transition Time Reload Register
Data Memory Programming Time Reload Register
DMERASE
byte
F94A
Read/Write
Data Memory Erase Time Reload Register
DMEND
DMPCNT
byte
byte
F94C
F94E
Read/Write
Read/Write
Data Memory End Time Reload Register
Data Memory Prescaler Count Register
DMCNT
word
F950
Read/Write
Data Memory Timer Count Register
DMKEY
byte
F954
Read/Write
FLCSR
byte
F960
Read/Write
EEPROM Data Memory Write Key Register
Flash EEPROM Program Memory Control and Status
Register
FLPSLR
FLSTART
byte
byte
F962
F964
Read/Write
Read/Write
Flash EEPROM Program Memory Prescaler Register
Program Memory Start Time Reload Register
FLTRAN
byte
F966
Read/Write
Program Memory Transition Time Reload Register
FLPROG
FLERASE
byte
byte
F968
F96A
Read/Write
Read/Write
Program Memory Programming Time Reload Register
Program Memory Erase Time Reload Register
Contents
FLEND
byte
F96C
Read/Write
Program Memory End Time Reload Register
FLPCNT
FLCNT1
byte
byte
F96E
F970
Read/Write
Read/Write
Program Memory Prescaler Count Reload Register
Program Memory Timer Count Register 1
FLCNT2
byte
F972
Read/Write
Program Memory Timer Count Register 2
PGMKEY
PBDIR
byte
byte
F974
FB00
Read/Write
Read/Write
Flash EEPROM Program Memory Write Key Register
Port B Direction Register
PBDIN
byte
FB02
Read Only
Port B Data Input Register
PBDOUT
PBWKPU
byte
byte
FB04
FB06
Read/Write
Read/Write
Port B Data Output Register
Port B Weak Pull-Up Register
PCDIR
byte
FB10
Read/Write
Port C Direction Register
PCDIN
PCDOUT
byte
byte
FB12
FB14
Read Only
Read/Write
Port C Data Input Register
Port C Data Output Register
PCWKPU
byte
FB16
Read/Write
Port C Weak Pull-Up Register
CRCTRL
PRSSC
byte
byte
FC40
FC42
Read/Write
Read/Write
Clock and Reset Control Register
Slow Clock Prescaler Register
PRSSC1
byte
FC44
Read/Write
Prescaler Slow Clock 1 Register
PMCSR
WKEDG
byte
word
FC60
FC80
Read/Write
Read/Write
Power Management Control/Status Register
Wake-Up Edge Detection Register
WKENA
word
FC82
Read/Write
Wake-Up Enable Register
WKICT1
WKICTL2
word
word
FC84
FC86
Read/Write
Read Set
Wake-Up Interrupt Control Register 1
Wake-Up Interrupt Control Register 2
WKPND
word
FC88
Write Only
Wake-Up Pending Register
WKPCL
PGALT
word
byte
FC8A
FCA0
Read/Write
Read/Write
Wake-Up Pending Clear Register
Port G Alternate Function Register
125
www.national.com
Table 43
Device Detailed Memory Map
Register Name
Size
Register
Address
(hex)
Access Type
PGDIR
PGDIN
byte
byte
FCA2
FCA4
Read/Write
Read Only
Port G Direction Register
Port G Data Input Register
PGDOUT
byte
FCA6
Read/Write
Port G Data Output Register
PGWKPU
PHALT
byte
byte
FCA8
FCC0
Read/Write
Read/Write
Port G Weak Pull-Up Register
Port H Alternate Function Register
PHDIR
byte
FCC2
Read/Write
Port H Direction Register
PHDIN
PHDOUT
byte
byte
FCC4
FCC6
Read Only
Read/Write
Port H Data Input Register
Port H Data Output Register
PHWKUP
byte
FCC8
Read/Write
Port H Weak Pull-Up Register
PFALT
PFDIR
byte
byte
FD20
FD22
Read/Write
Read/Write
Port F Alternate Function Register
Port F Direction Register
Contents
PFDIN
byte
FD24
Read Only
Port F Data Input Register
PFDOUT
PFWKPU
byte
byte
FD26
FD28
Read/Write
Read/Write
Port F Data Output Register
Port F Weak Pull-Up Register
IVCT
byte
FE00
Read Only
Interrupt Vector Register
NMISTAT
EXNMI
byte
byte
FE02
FE04
Read Only
Read/Write
NMI Status Register
External NMI Control/Status Register
ISTAT0
word
FE0A
Read Only
Interrupt Status Register 0
ISTAT1
IENAM0
word
word
FE0C
FE0E
Read Only
Read/Write
Interrupt Status Register 1
Interrupt and Enable Mask Register 0
IENAM1
word
FE10
Read/Write
Interrupt and Enable Mask Register 1
IDBG
U1TBUF
word
byte
FE1A
FE40
Read/Write
Read/Write
Interrupt Debug Register
USART 1 Transmit Data Buffer
U1RBUF
byte
FE42
Read Only
USART 1 Receive Data Buffer
U1ICTRL
U1STAT
byte
byte
FE44
FE46
Read/Write
Read Only
USART 1 Interrupt Control Register
USART 1 Status Register
U1FRS
byte
FE48
Read/Write
USART 1 Frame Select Register
U1MDSL
U1BAUD
byte
byte
FE4A
FE4C
Read/Write
Read/Write
USART 1 Mode Select Register
USART 1 Baud Rate Divisor Register
U1PSR
byte
FE4E
Read/Write
USART 1 Baud Rate Prescaler
MWDAT
MWCTL
byte
byte
FE60
FE62
Read/Write
Read/Write
MICROWIRE Data Register
MICROWIRE Control Register
MWSTAT
word
FE64
Read/Write
MICROWIRE status Register
U2TBUF
U2RBUF
byte
byte
FE80
FE82
Read/Write
Read Only
USART 2 Transmit Data Buffer
USART 2 Receive Data Buffer
U2ICTRL
byte
FE84
Read/Write
USART 2 Interrupt Control Register
U2STAT
U2FRS
byte
byte
FE86
FE88
Read Only
Read/Write
USART 2 Status Register
USART 2 Frame Select Register
U2MDSL
byte
FE8A
Read/Write
USART 2 Mode Select Register
ACBSDA
U2BAUD
byte
byte
FEC0
FE8C
Read/Write
Read/Write
ACB Serial Data Register
USART 2 Baud Rate Divisor Register
U2PSR
byte
FE8E
Read/Write
USART 2 Baud Rate Prescaler
ACBST
byte
FEC2
Read/Write
ACB Status Register
www.national.com
126
Table 43
Device Detailed Memory Map
Register Name
Size
Register
Address
(hex)
Access Type
Contents
ACBCST
byte
FEC4
Read/Write
ACB Control Status Register
ACBCTL1
ACBADDR
byte
byte
FEC6
FEC8
Read/Write
Read/Write
ACB Control 1 Register
ACB Own Address Register
ACBCTL2
byte
FECA
Read/Write
ACB Control 2 Register
PIALT
PIDIR
byte
byte
FEE0
FEE2
Read/Write
Read/Write
Port I Alternate Function Register
Port I Direction Register
PIDIN
byte
FEE4
Read Only
Port I Data Input Register
PIDOUT
PIWKPU
byte
byte
FEE6
FEE8
Read/Write
Read/Write
Port I Data Output Register
Port I Weak Pull-Up Register
PLALT
byte
FF00
Read/Write
Port L Alternate Function Register
PLDIR
PLDIN
byte
byte
FF02
FF04
Read/Write
Read Only
Port L Direction Register
Port L Data Input Register
PLDOUT
byte
FF06
Read/Write
Port L Data Output Register
PLWKPU
TWCFG
byte
byte
FF08
FF20
Read/Write
Read/Write
Port L Weak Pull-Up Register
Timer and WATCHDOG Configuration Register
TWCP
byte
FF22
Read/Write
Timer and WATCHDOG Clock Prescaler Register
TWMT0
T0CSR
word
byte
FF24
FF26
Read/Write
Read/Write
TWM Timer 0 Register
TWMT0 Control and Status Register
WDCNT
byte
FF28
Write Only
WATCHDOG Count Register
WDSDM
T1CNT1
byte
word
FF2A
FF40
Write Only
Read/Write
WATCHDOG Service Data Match Register
T1 Timer/Counter I Register
T1CRA
word
FF42
Read/Write
T1 Reload/Capture A Register
T1CRB
T1CNT2
word
word
FF44
FF46
Read/Write
Read/Write
T1 Reload/Capture B Register
T1 Timer/Counter II Register
T1PRSC
byte
FF48
Read/Write
T1 Clock Prescaler Register
T1CKC
T1CTRL
byte
byte
FF4A
FF4C
Read/Write
Read/Write
T1 Clock Unit Control Register
T1 Timer Mode Control Register
T1ICTL
byte
FF4E
Read/Write
T1 Timer Interrupt Control Register
T1ICLR
T2CNT2
byte
word
FF50
FF60
Read/Write
Read/Write
T1 Timer Interrupt Clear Register
T2 Timer/Counter I Register
T2CRA
word
FF62
Read/Write
T2 Reload/Capture A Register
T2CRB
T2CNT2
word
word
FF64
FF66
Read/Write
Read/Write
T2 Reload/Capture B Register
T2 Timer/Counter II Register
T2PRSC
byte
FF68
Read/Write
T2 Clock Prescaler Register
T2CKC
T2CTRL
byte
byte
FF6A
FF6C
Read/Write
Read/Write
T2 Clock Unit Control Register
T2 Timer Mode Control Register
T2ICTL
byte
FF6E
Read/Write
T2 Timer Interrupt Control Register
T2ICLR
MODE
byte
word
FF70
FF80
Read/Write
Read/Write
T2 Timer Interrupt Clear Register
Mode Control Register
IO1CTL
word
FF82
Read/Write
I/O Control Register 1
IO2CTL
INTCTL
word
word
FF84
FF86
Read/Write
Read/Write
I/O Control Register 2
Interrupt Control Register
INTPND
word
FF88
Read/Write
Interrupt Pending Register
CLK1PS
COUNT1
word
word
FF8A
FF8C
Read/Write
Read/Write
Clock Prescaler Register 1
Counter Register 1
127
www.national.com
Table 43
Device Detailed Memory Map
Register Name
Size
Register
Address
(hex)
Access Type
PERCAP1
DTYCAP1
word
word
FF8E
FF90
Read/Write
Read/Write
Period/Capture Register 1
Duty Cycle/Capture Register 1
COUNT2
word
FF92
Read/Write
Count Register 2
PERCAP2
DTYCAP2
word
word
FF94
FF96
Read/Write
Read/Write
Period/Capture Register 2
Duty Cycle/Capture Register 2
Contents
CLK2PS
word
FF98
Read/Write
Clock Prescaler Register 2
COUNT3
PERCAP3
word
word
FF9A
FF9C
Read/Write
Read/Write
Count Register 3
Period/Capture Register 3
DTYCAP3
word
FF9E
Read/Write
Duty Cycle/Capture Register 3
COUNT4
PERCAP4
word
word
FFA0
FFA2
Read/Write
Read/Write
Count Register 4
Period/Capture Register 4
DTYCAP4
word
FFA4
Read/Write
Duty Cycle/Capture Register 4
ADCST
ADCCNT1
byte
byte
FFC0
FFC2
Read/Write
Read/Write
A/D Converter Status Register
A/D Converter Control 1 Register
ADCCNT2
byte
FFC4
Read/Write
A/D Converter Control 2 Register
ADCCNT3
ADDATA0
byte
byte
FFC6
FFCA
Read/Write
Read Only
A/D Converter Control 3 Register
A/D Converter Data 0 Register
ADDATA1
byte
FFCC
Read Only
A/D Converter Data 1 Register
ADDATA2
ADDATA3
byte
byte
FFCE
FFD0
Read Only
Read Only
A/D Converter Data 2 Register
A/D Converter Data 3 Register
CMPCTRL
byte
FFE0
Read/Write
Analog Comparator Control/Status Register
www.national.com
128
24.0 Register Layouts
The following tables show the functions of the bit fields of the device registers. For more information on using these registers,
see the detailed description of the applicable function elsewhere in this data sheet.
24.1
REGISTER LAYOUT
CAN Memory
Registers
15
14
13
12
11
10
9
8
7
6
5
4
3
2
1
0
CMBn.ID1
XI28
ID10
XI27
ID9
XI26
ID8
XI25
ID7
XI24
ID6
XI23
ID5
XI22
ID4
XI21
ID3
XI20
ID2
XI19
ID1
XI18
ID0
SRR
RTR
IDE
XI17
XI16
XI15
CMBn.ID0
XI14
XI13
XI12
XI11
XI10
XI9
XI8
XI7
XI6
XI5
XI4
XI3
XI2
XI1
XI0
RTR
CMBn.DATA0
Data
1.7
Data
1.6
Data
1.5
Data
1.4
Data
1.3
Data
1.2
Data
1.1
Data
1.0
Data
2.7
Data
2.6
Data
2.5
Data
2.4
Data
2.3
Data
2.2
Data
2.1
Data
2.0
CMBn.DATA1
Data
3.7
Data
3.6
Data
3.5
Data
3.4
Data
3.3
Data
3.2
Data
3.1
Data
3.0
Data
4.7
Data
4.6
Data
4.5
Data
4.4
Data
4.3
Data
4.2
Data
4.1
Data
4.0
CMBn.DATA2
Data
5.7
Data
5.6
Data
5.5
Data
5.4
Data
5.3
Data
5.2
Data
5.1
Data
5.0
Data
6.7
Data
6.6
Data
6.5
Data
6.4
Data
6.3
Data
6.2
Data
6.1
Data
6.0
CMBn.DATA3
Data
7.7
Data
7.6
Data
7.5
Data
7.4
Data
7.3
Data
7.2
Data
7.1
Data
7.0
Data
8.7
Data
8.6
Data
8.5
Data
8.4
Data
8.3
Data
8.2
Data
8.1
Data
8.0
CMBn.TSTP
TSTP15 TSTP14 TSTP13 TSTP12 TSTP11 TSTP10 TSTP9 TSTP8 TSTP7 TSTP6 TSTP5 TSTP4 TSTP3 TSTP2 TSTP1 TSTP0
CMBn.CNTSTAT
CAN
Control/
Status
DLC3
15
CGCR
DLC2
14
13
DLC1
12
Reserved
CTIM
DLC0
Reserved
PRI1
PRI0
ST3
ST2
10
9
8
7
6
5
4
3
EIT
DIAGEN
INTERNAL
LOOPBACK
IGNACK
LO
DDIR
TSTPEN
BUFFLOCK
PSC[6:0]
SJW[1:0]
RTR
ST1
2
TSEG1[3:0]
GM[28:18]
GMSKX
ST0
1
0
CRX CTX CANEN
TSEG2[2:0]
IDE
GM[17:15]
GM[14:0]
BMSKB
XRTR
BM[28:18]
BMSKX
RTR
IDE
BM[17:15]
BM[14:0]
XRTR
EIEN
IEN[14:0]
CIPND
EIPND
IPND[14:0]
CICLR
EICLR
ICLR[14:0]
CICEN
EICEN
ICEN[14:0]
CSTPND
Reserved
CANEC
REC[7:0]
CEDIAG
PRI2
11
GMSKB
CIEN
PRI3
Reserve
d
DRIVE
MON CRC
STUFF
NS[2:0]
IST[3:0]
TEC[7:0]
TXE
EFID[3:0]
EBID[5:0]
CTMR
System
Configuration
Registers
IRQ
CTMR[15:0]
7
6
5
4
129
3
2
1
0
www.national.com
MCFG
Reserved
CLK2OE
MSTAT
Reserved
Reserved
BIU Registers
15
12
11
10
PGMBUSY
9
8
BCFG
7
6
5
CLK1OE
CLKOE
Reserved
Reserved
OENV1
OENV0
4
3
2
1
0
Reserved
IOCFG
Reserved
EWR
IPST Res
BW
Reserved
HOLD
WAIT
SZCFG0
Reserved
FRE IPRE IPST Res
BW
Reserved
HOLD
WAIT
SZCFG1
Reserved
FRE IPRE IPST Res
BW
Reserved
HOLD
WAIT
ISP Registers
15 13
12
10
FLCTRL1
FLCTRL2
9
8
7
6
5
4
Reserved
EMPTY
Reserved
ISPKEY
Reserved
Flash Data
Memory
Registers
15
14
10
9
DMCSR
FROMWR
ISPKYVAL
8
7
6
5
4
3
Reserved
2
1
0
ERASE DMBUSY ZEROWS Reserved
FTDIV
DMSTART
Reserved
FTSTART
DMTRAN
Reserved
FTTRAN
DMPROG
Reserved
FTPROG
DMERASE
Reserved
FTER
DMEND
Reserved
FTEND
DMPCNT
Reserved
FTPCNT
Reserved
FTCNT
Reserved
Flash EEPROM Program
Memory Registers
7
FLCSR
MERASE
DMKEYVAL
6
5
4
Reserved
3
FTDIV
FLSTART
FTSTART
FLTRAN
FTTRAN
130
2
PMLFULL PMBUSY
FLPSLR
www.national.com
0
FROMRD
Reserved
DMKEY
1
CODEAREA
DMPSLR
DMCNT
2
BOOTAREA
Reserved
FLSEC
3
1
0
PMER
Reserved
FLPROG
FTPROG
FLERASE
FTER
FLEND
FTEND
FLPCNT
FTPCNT
FLCNT1
FTCNTL (0:7)
FLCNT2
Reserved
PGMKEY
FTCNTL (8:9)
PMKEYVAL
GPIO Registers
7
6
5
4
2
1
PxALT
Px Pins Alternate Function Enable
PxDIR
Px Port Direction
PxDIN
Px Port Output Data
PxDOUT
Px Port Input Data
PxWPU
Px Port Weak Pull-up Enable
ICU31L Registers
15
IVCT
12
11
8
Reserved
7
6
0
0
5
4
1
0
EXT
Reserved
ENLCK
ISTAT0
IST(15:0)
ISTAT1
IST(31:16)
IENAM0
IENA(15:0)
IENAM1
IENA(31:16)
Reserved
15
2
Reserved
EXNMI
IDBG
3
0
INTVECT
NMISTAT
MIWU16
Registers
3
14
13
IRQVECT
12
11
10
PIN
EN
INTVECT
9
8
WKEDG
WKED
WKENA
WKEN
7
6
5
4
3
2
1
0
WKICTL1
WKINTR7
WKINTR6
WKINTR5
WKINTR4
WKINTR3
WKINTR2
WKINTR1 WKINTR0
WKICTL2
WKINTR15
WKINTR14
WKINTR13
WKINTR12
WKINTR11
WKINTR10
WKINTR9 WKINTR8
WKPND
WKPD
WKPCL
WKCL
131
www.national.com
Dual Clock + Reset Registers
7
6
5
CRCTRL
4
3
Reserved
PRSSC
SCDIV2
Power Management Register
PMCSR
USART Registers
SCLK
6
5
4
3
2
1
0
OLFC
OHFC
WBPSM
Reserved
HALT
IDLE
DHF
PSM
3
2
7
6
5
4
UnRBUF
UnRBUF
UnICTRL
UnEEI
UnERI
UnETI
UnSTAT
Reserved
UnXMIP
UnRB9
UnFRS
Reserved
UnPEN
UnMDSL
Reserved
UnBKD
UnPSEL
Reserved
UnBAUD
UnERR
UnDOE
UnXB9
UnSTP
UnCKS
UnBRK
1
0
UnRBF
UnTBE
UnFE
UnPE
UnCHAR
UnATN
UnMOD
UnDIV[7]: UnDIV[0]
UnPSR
UnPSC
15
9
8
7
UnDIV[10]: UnDIV[8]
6
5
MWDAT
www.national.com
POR
7
UnTBUF
MWSTAT
0
SCDIV1
UnTBUF
MWCTL
1
SCDIV
PRSSC1
MWSPI16
Registers
2
4
3
2
1
0
MEIO
MECH
MMOD
MMNS
MEN
MOVR
MRBF
MBSY
MWDAT
MCDV
MIDL
MSKM
MEIW
Reserved
132
MEIR
ACB Registers
7
6
5
4
ACBSDA
ACBST
SLVSTP
SDAST
STASTRE
ACBADDR
SAEN
0
NEGACK
STASTR
NMATCH
MASTER
XMIT
TGSCL
TSDA
GMATCH
MATCH
BB
BUSY
NMINTE
GCMEN
ACK
Reserved
INTEN
STOP
START
ADDR
ACBCTL2
SCLFRQ
15
8
7
6
5
ENABLE
4
TnCNT1
TnCNT1
TnCRA
TnCRA
TnCRB
TnCRB
TnCNT2
TnCNT2
TnPRSC
3
Reserved
TnCKC
Reserved
TnCTRL
Reserved TnAOUT
TnICLR
1
BER
Reserved
ACBCTL1
TnICTL
2
DATA
ACBCST
TIMER Registers
3
TnDIEN
2
0
CLKPS
C2CSEL
TnCIEN
1
C1CSEL
TnBEN
TnAEN
TnBEDG
TnAEDG
TnBIEN
TnAIEN
TnDPND
TnCPND
TnBPND
TnAPND
TnDCLR
TnCCLR
TnBCLR
TnACLR
Reserved
133
MDSEL
www.national.com
VTU
Registers
15
MODE
14
TMOD4
13
12
11
T8RUN T7RUN
10
TMOD3
9
8
7
T6RUN T5RUN
6
5
TMOD2
4
3
T4RUN T3RUN
2
1
TMOD1
0
T2RUN T1RUN
IO1CTL
P4POL
C4EDG
P3POL
C3EDG
P2POL
C2EDG
P1POL
C1EDG
IO2CTL
P8POL
C8EDG
P7POL
C7EDG
P6POL
C6EDG
P5POL
C5EDG
INTCTL
I4DEN
I4CEN
I4BEN I4AEN
I3DEN
I3CEN
I3BEN I3AEN
I2DEN
I2CEN
I2BEN I2AEN
I1DEN
I1CEN
I1BEN I1AEN
INTPND
I4DPD
I4CPD
I4BPD I4APD
I3DPD
I3CPD
I3BPD I3APD
I2DPD
I2CPD
I2BPD I2APD
I1DPD
I1CPD
I1BPD I1APD
CLK1PS
C2PRSC
C1PRSC
COUNT1
CNT1
PERCAP1
PCAP1
DTYCAP1
DCAP1
COUNT2
CNT2
PERCAP2
PCAP2
DTYCAP2
DCAP2
CLK2PS
C4PRSC
C3PRSC
COUNT3
CNT3
PERCAP3
PCAP3
DTYCAP3
DCAP3
COUNT4
CNT4
PERCAP4
PCAP4
DTYCAP4
DCAP4
TWM Registers
TWCFG
TWCP
15
8
7
6
Reserved
5
4
3
2
WDSDME
WDCT0I
LWDCNT
LTWMT0
Reserved
TIMER0
T0CSR
0
LTWCP LTWCFG
MDIV
PRESET
Reserved
T0INTE
WDCNT
PRESET
WDSDM
RSTDATA
www.national.com
1
134
TC
RST
A/D Registers
7
ADCST
6
5
Reserved
ADCCNT1
4
BUFPTR
Reserved
ADCCNT2
START
ADCCNT3
Reserved
SCAN
3
2
1
0
Reserved
OVF
BUSY
EOC
Reserved
INTE
Reserved
ADCEN
CONT
PWREN
CHANNEL
DELAY
CDIV
ADDATA0
RESULT 1 DATA
ADDATA1
RESULT 2 DATA
ADDATA2
RESULT 3 DATA
ADDATA3
RESULT 4 DATA
Analog Comp. Registers
CMPCTRL
7
6
Reserved
5
4
3
2
1
0
CMP2OE
CMP1OE
CMP2EN
CMP1EN
CMP2RD
CMP1RD
135
www.national.com
25.0 ELECTRICAL AND THERMAL CHARACTERISTICS
Absolute Maximum Ratings
If Military/Aerospace specified devices are required, please
contact the National Semiconductor Sales Office/Distributors
for availability and specifications.
Supply Voltage (VCC )
Voltage at Any Pin *
ESD Protection Level
Total Current into VCC Pin (Source)
Total Current out of GND Pin (Sink)
Storage Temperature Range
200 mA
–65°C to +150°C
Note: Absolute maximum ratings indicate limits beyond
which damage to the device may occur. DC and AC electrical
specifications are not ensured when operating the device at
absolute maximum ratings. * The latch-up tolerance on Access Bus pins 14 and 15 exceeds 150mA.
7V
–0.6V to VCC +0.6V
2 kV
(Human Body Model)
200 mA
Thermal Characteristics
Characteristics
Symbol
Value
Unit
Average junction temperature
TJ
TA + (PD X ~JA)
°C
Ambient temperature
TA
User-determined
°C
Package thermal resistance (junction-to-ambient)
80-pin quad flat pack (QFP)
~JA
49.8
°C/W
PD
PINT + PI/O
or
K
T J + 273°C
W
Device internal power
dissipation
PINT
IDD X VDD
W
I/O pin power dissipation2
PI/O
User-determined
W
K
PD x (TA + 273°C) +
~ JA x PD 2
W, °C
Total power dissipation1
A constant3
1. This is an approximate value, neglecting P I/O .
2. For most applications PI/O
<< PINT and can be neglected.
3. K is a constant pertaining to the device. Solve for K with a known TA and a measured P D (at equilibrium). Use this value of K to solve for
P D and T J iterntively for any value of TA .
DC Electrical Characteristics: –40°C ≤ TA ≤ +85°C (also supports -40°C to +125°C)
Symbol
Parameter
Conditions
Operating Voltage
VIH
Logical 1 CMOS Input Voltage (except ACB & clocks)
VIL
Logical 0 CMOS Input Voltage (except ACB & clocks)
VIHACB
SDA, SCL Logical 1 CMOS Input Voltage
VILACB
SDA, SCL Logical 0 CMOS Input Voltage
Vxl
Low Level Input Voltage OSC
External X1 clock
Vxh
High Level Input Voltage OSC
External X1 clock
Vxl2
X2CKI Logical 0 Input Voltage
External X2 clock
Vxh2
X2CKI Logical 1 Input Voltage
External X2clock
Min
Max
Units
4.5
5.5
V
0.8Vcc
Vcc + 0.5
V
-0.5
0.2Vcc
V
0.7Vcc
Width a
V
0.3Vcc
V
0
0.2Vcc
V
0.5Vcc
Vcc
V
0.3
V
1.2
V
0.1Vcc
V
Vhys
Hysteresis Loop
I OH
Logical 1 CMOS Output Current
VOH = 3.8V, Vcc=4.5V
-1.6
mA
IO L
Logical 0 CMOS Output Current
VOL = 0.45V, Vcc=4.5V
1.6
mA
www.national.com
136
Symbol
Parameter
Conditions
Min
Max
Units
I OLACB
SDA, SCL Logical 0 CMOS Output Current
VOL = 0.4V, Vcc=4.5V
3.0
mA
I OHW
Weak Pull-up Current
VOH = 3.8V, Vcc=4.5V
-10
µA
I IL
RESET pin Weak Pull-down Current
VIL = 0.9V, Vcc=4.5V
IL
High Impedance Input Leakage Current
0V ≤ Vin ≤ Vcc
- 2.0
2.0
I O(Off)
Output Leakage Current
(I/O pins in input mode)
0V ≤ Vout ≤ Vcc
- 2.0
2.0j
µA
Icca1
Digital Supply Current Active Mode b
Vcc= 5.5V
95
mA
Digital Supply Current Active Mode
c
Vcc= 5.5V
115
mA
Icca2
Digital Supply Current Active Mode
d
Vcc = 5.5V
58
mA
Iccps
Digital Supply Current Power Save Mode e
Vcc= 5.5V
9
mA
Vcc = 5.5V
200
µA
Vcc = 5.5V
k
µA
Iccprog
Iccid
Iccq
Digital Supply Current Idle Mode
f
Digital Supply Current Halt Mode
f
0.4
µA
j
µA
20
g
Iacc
Analog Supply Current Active Mode
Vcc= 5.5V
a. Guaranteed by design
b. Run from internal memory, Iout=0mA, X1CKI=20MHz, not programming flash memory
c. Same conditions as Icca1 but programming or erasing one of the flash memory arrays
d. CPU executing an WAIT instruction, Iout=0mA, X1CKI=20MHz, peripherals not active
e. Running from internal memory, Iout=0mA, X1CKI=20MHz, X2CKI=32.768kHz
f. Iout=0mA, X1CKI=Vcc, X2CKI=32.768kHz
g. ADC and analog comparators enabled
j. I L adn I O are 2.0 µA at 85°C and 5.0 µA at 125°C
k. I acq is 20 µA at 85°C and 50 µA at 125°C
3
mA
A/D Converter Characteristics
VCC = 5V, TA = 25°C
Symbol
Parameter
Integral Errorb
NIL
Conditionsa
Min
Typ
Max
Units
VREF = VCC
±0.5
LSB
NDL
Differential Error
VREF = VCC
±1.0
LSB
VABSOLUTE
Absolute Error
VREF = VCC
±1.5
LSB
VIN
Input Voltage Range
VREF < VCC - 0.1
0
VREF
V
VREFEX
External Reference Voltage
3.0
VDD
V
IVREF
VREF input current
VREF = 5V
1.2
mA
IAL
Analog input leakage current
VREF = VCC
±1
µA
200
Ω
5
pF
500
ns
c
d
RAIN
Analog input resistance
CAIN
Analog input capacitancee
tADCCLK
Conversion Clock period
CREFEX
External Vref bypass capacitance
0.47
µF
tACT
First conversion after Vcc stable
30
µs
MMONOTON-
MONOTONICITYf
IC
GUARANTEED
a. All parameters specified for fOSC =2 MHz, V DD = 5.0V ± 10% unless otherwise noted.
b. Integral (Non-linearity) Error — The maximum difference between the best-fit straight line reference and the actual
conversion curves.
c. Differential (Non-linearity) Error — The maximum difference between the best-fit step size of 1 LSB and any actual
step size.
d. The resistance between the device input and the internal analog input capacitance.
137
www.national.com
e. The input signal is measured across the internal capacitance.
f. Conversion result never decreases with an increase in input voltage and has no missing codes.
www.national.com
138
Analog Comparator Characteristics
Symbol
Parameter
Conditions
VO S
Input Offset Voltage
VCM
I CS
Input Common Mode Voltage Range
DC Supply Current per Comparator (When
Enabled)
Response Time
Min
Typ
Vcc = 5V,
0.4V ≤ VIN ≤ VCC – 1.5V
0.4
VCC =5.5V
1V Step / 100mV Overdrive
Max
Units
±25
mV
VCC -1.5
V
250
µA
1
µs
Flash EEPROM Program Memory Programming
Symbol
t PWP
Parameter
Programming pulse width
Conditions
a
b
t EWP
Erase pulse width
t SDP
Charge pump power-up delayc
t TTP
Program/erase transition time
d
t PAH
Min
Max
Units
30
40
µs
1
-
ms
10
-
µs
5
-
µs
Programming address hold, new address setup
time
2
-
clock
cycles
t PEP
Charge pump enable hold time
1
-
clock
cycles
t EDP
Charge pump power hold timee
5
t CHVP
Cumulative program high voltage period for each
row after erase.f
-
25
100
-
Data retention
µs
ms
years
100K cycles
a. The programming pulse width is determined by the following equation:
tP W P = Tclk x (FTDIV+1) x (FTPROG+1), where Tclk is the system clock period, FTDIV is the contents of the
FLPSLR register and FTPROG is the contents of the FLPROG register.
b. The erase pulse width is determined by the following equation:
tE W P = Tclk x (FTDIV+1) x 4 x (FTER+1), where Tclk is the system clock period, FTDIV is the contents of the
FLPSLR register and FTER is the contents of the FLERASE register.
c. The program/erase start delay time is determined by the following equation:
tSDP = Tclk x (FTDIV+1) x (FTSTART+1), where Tclk is the system clock period, FTDIV is the contents of the
FLPSLR register and FTSTART is the contents of the FLSTART register.
d. The program/erase transition time is determined by the following equation:
tTTP = Tclk x (FTDIV+1) x (FTTRAN+1), where Tclk is the system clock period, FTDIV is the contents of the
FLPSLR register and FTTRAN is the contents of the FLTRAN register.
e. The program/erase end delay time is determined by the following equation:
tEDP = Tclk x (FTDIV+1) x (FTEND+1), where Tclk is the system clock period, FTDIV is the contents of the FLPSLR
register and FTEND is the contents of the FLEND register.
f. Cumulative program high voltage period for each row after erase t CHVP is the accumulated duration a flash cell is
exposed to the programming voltage after the last erase cycle. It is the sum of all tHV after the last erase.
139
www.national.com
Flash EEPROM Data Programming
Symbol
Parameter
Conditions
re-programming timea
t PWD
Programming pulse width
t EWD
Erase pulse widthc
b
Min
Max
Units
1.32
-
ms
30
40
µs
1
-
ms
10
-
µs
t TTD
Program/erase transition time
e
5
-
µs
t PED
Charge pump enable hold time
1
-
clock
cycles
t EDD
Charge pump power hold timef
5
-
µs
Write/erase endurance (high endurance)
100,000
-
cycles
Write/erase endurance (low endurance)
25,000
-
cycles
-
years
t SDD
d
Charge pump power-up time
Data retention
100
a. One re-programming cycle involves one erase pulse followed by programming of four bytes.
b. The programming pulse width is determined by the following equation:
tP W D = Tclk x (FTDIV+1) x (FTPROG+1), where Tclk is the system clock period, FTDIV is the contents of the
DMPSLR register and FTPROG is the contents of the DMPROG register.
c. The erase pulse width is determined by the following equation:
tE W D = Tclk x (FTDIV+1) x 4 x (FTER+1), where Tclk is the system clock period, FTDIV is the contents of the
DMPSLR register and FTER is the contents of the DMERASE register.
d. The program/erase start delay time is determined by the following equation:
tSDD = Tclk x (FTDIV+1) x (FTSTART+1), where Tclk is the system clock period, FTDIV is the contents of the
DMPSLR register and FTSTART is the contents of the DMSTART register.
e. The program/erase transition time is determined by the following equation:
tTTD = Tclk x (FTDIV+1) x (FTTRAN+1), where Tclk is the system clock period, FTDIV is the contents of the
DMPSLR register and FTTRAN is the contents of the DMTRAN register.
f. The program/erase end delay time is determined by the following equation:
tEDD = Tclk x (FTDIV+1) x (FTEND+1), where Tclk is the system clock period, FTDIV is the contents of the
DMPSLR register and FTEND is the contents of the DMEND register.
Flash EEPROM ISP-Memory Programming
Symbol
t PWI
t EWI
Parameter
Programming pulse
Erase pulse
Conditions
witha
widthb
Data retention
a. Programming timing is controlled by the flash EEPROM data memory interface
b. Erase timing is controlled by the flash EEPROM data memory interface
www.national.com
140
Min
Max
Units
30
40
µs
1
-
ms
100
-
years
-
100K
cycles
Row Select/
Start Charge
Pump
Select
Charge Pump/
Enable
Programming
Voltage
Programming
Pulse
tSD
tTT
tPP
tPW
t PW
tPE
t ED
Figure 76. Flash EEPROM Memory Programming Timing
(Sample Sequence for Programming two Words into Flash EEPROM Program Memory
Output Signal Levels
All output signals are powered by the digital supply (VCC).
Table44 summarizes the states of the output signals during
the reset state (when VCC power exists in the reset state)
and during the Power Save mode.
Table 44
Signals on a pin
The RESET and NMI input pins are active during the Power
Save mode. In order to guarantee that the Power Save current not exceed 1mA, these inputs must be driven to a voltage lower than 0.5V or higher than VCC-0.5V. An input
voltage between 0.5V and (VCC-0.5V) may result in power
consumption exceeding 1 mA.
Output Pins During Reset and Power-Save
Reset state
(with Vcc)
Power Save mode
PF[0:7]
TRI-STATE
Previous state
PG[0:7]
TRI-STATE
Previous state
PI[0:7]
TRI-STATE
Previous state
PL[0:7]
TRI-STATE
Previous state
PB[0:7]
TRI-STATE
Previous state
PC[0:7]
TRI-STATE
Previous state
141
Comments
I/O ports will maintain their values when
entering power-save mode
www.national.com
25.0.1
Timing Waveforms
tX1p
X1
t X1h
t X1l
tX2p
X2
t X2h
t X2l
tCLKp
t CLKr
t CLKf
CLK
t CLKh
tCLKl
Output Valid
Output Hold
Output
Signal
Input Hold
Input Setup
Input
Signal
Output Active/Inactive time
Control
Signal 1
Output Active/Inactive time
Control
Signal 2
ac-1
Figure 77. Clock Waveforms
www.national.com
142
CLK
t Is
tIh
t Iw
ISE
t Is
tIh
t Iw
NMI
Figure 78. ISE & NMI Signal Timing
CLK
t RST
RESET
Figure 79. Non-Power-On Reset
1
2
2
1
1
2
1
2
1
2
1
2
CLK
tCOv1
tCOv1
TXDn
tIS
RXDn
t IH
Figure 80.
USART Asynchronous Mode Timing
tCLKX
CKXn
t TXD
TXDn
tRXS
RXDn
tRXH
Figure 81. USART Synchronous Mode Timing
143
www.national.com
1
2
2
1
1
2
2
1
1
2
2
1
CLK
tI s
PORTS B, C (input)
tIh
t Of
PORTS B, C (output)
tCOv1
t COv2
tCOv1
tCOv2
BUZCLK
Figure 82. Port Signals Timing
tMSKp
MSKn
tMSKs
tMSKh
tMSKhd
tMSKl
msb
Data In
tMDIs
MDIDOn
(slave)
lsb
tMDIh
msb
tMDOf
lsb
tMDOf
tMDOv
tMDOh
MDODIn
(master)
msb
lsb
tMSKd
MCSn
(Slave)
tMCSs
tMCSh
Figure 83. MICROWIRE Transaction Timing, Normal Mode, MIDL Bit = 0
www.national.com
144
tMSKp
MSKn
tMSKs
tMSKl
tMSKhd
tMSKh
msb
Data In
tMDIs
MDIDOn
(slave)
lsb
tMDIh
msb
lsb
tMDOv
tMDOf
tMDOf
tMDOh
MDODIn
(master)
msb
lsb
MCSn
(Slave)
tMCSs
tMCSh
Figure 84. MICROWIRE Transaction Timing, Normal Mode, MIDL bit = 1
tMSKp
MSKn
tMSKs
tMSKh
tMSKhd
tMSKl
msb
Data In
tMDIs
MDIDOn
(Slave)
lsb
tMDIh
msb
lsb
tMDOv
tMDOf
tMDOf
tMDOh
MDODIn
(Master)
msb
lsb
MCSn
(Slave)
tMCSh
tMCSs
Figure 85.
MICROWIRE Transaction Timing, Alternate Mode, MIDL bit = 0
145
www.national.com
tMSKp
MSKn
tMSKs
tMSKl
Data In
tMSKhd
tMSKh
msb
tMDIs
MDIDOn
(Slave)
lsb
tMDIh
msb
lsb
tMDOf
tMDOv
tMDOf
tMDOh
MDODIn
(Master)
msb
lsb
tSKd
MCSn
(Slave only)
tMCSs
Figure 86.
tMCSh
MICROWIRE Transaction Timing, Alternate Mode, MIDL bit = 1
tMSKp
MSKn
tMSKs
tMSKh
tMSKhd
tMSKl
DI msb
MDODIn
(Slave)
tMDIs
DI lsb
tMDIh
tMITOp
MDIDOn
(Slave)
tMITOp
DO msb
DO lsb
tMDOnf
tMDOf
MCSn
tMCSs
tMCSh
Figure 87. MICROWIRE Transaction Timing, Data Echoed to Output, Normal Mode, MIDLBit= 0, MECHBit= 1, Slave
www.national.com
146
SDA
0.7VCC
0.7V CC
0.3VCC
0.3VCC
tSDAr
SCL
tSDAf
0.7VCC
0.7V CC
0.3VCC
0.3VCC
tSCLr
tSCLf
Note: In the timing tables the parameter name is added with an “o” for
output signal timing and “i” for input signal timing.
Figure 88. ACB signals (SDA and SCL) Rise Time and Fall Timing
Stop Condition
Start Condition
SDA
tDLCs
SCL
tCSTOs
tBUF
tCSTRh
Note: In the timing tables the parameter name is added with an “o” for
output signal timing and “i” for input signal timing.
Figure 89. ACB Start and Stop Condition Timing
Start Condition
SDA
SCL
tDHCs
tCSTRs
tCSTRh
Note: In the timing tables the parameter name is added with an “o” for
output signal timing and “i” for input signal timing.
Figure 90. ACB Start Conditioning Timing
147
www.national.com
Figure 91.
CLK
ACB Data Bits Timing
tTIOL tTIOH
TIOx
Figure 92. Versatile-Timer-Unit Input Timing
CLK
tTIOL tTIOH
TIOx
Figure 93. Versatile-Timer-Unit Input Timing
www.national.com
148
25.0.2
Timing Tables
Table 45 Output Signals
Symbol Figure
a
77
t CLKh
77
tCLKl
77
t CLKr
tCLKf
Tclk
Description
Reference
Min (ns) Max (ns)
64000a
CLK clock period
R.E. CLK to next R.E. CLK
43.4
CLK high time
At 2.0V
(Both Edges)
17.3
CLK low time
At 0.8V
(Both Edges)
17.3
77
CLK rise time on R.E. CLK
0.8V to 2.0V
3
77
CLK fall time on F.E. CLK
2.0V to 0.8V
3
CMOS output valid
All signals with prop. delay from CLK
R.E.
After R.E. CLK
tCOv1
35
USART Output Signals
tTXD
84
TXDn output valid
After R.E. CLKXn
35
MICROWIRE / SPI Output Signals
t MSKh
86
MICROWIRE Clock High
At 2.0V (both edges)
80
tMSKl
86
MICROWIRE Clock Low
At 0.8V (both edges)
80
MICROWIRE Clock Period
MnIDL bit = 0: R.E. MSK to next R.E.
MSKn
86
t MSKp
MnIDL bit = 1: F.E. MSK to next F.E.
MSKn
87
t MSKd
86
MSK Leading Edge Delayed (master
only)
Data Out Bit #7 Valid
tMDOf
86
MICROWIRE Data Float b
(slave only)
After R.E. MCSn
t MDOh
86
MICROWIRE Data Out Hold
Normal Mode: After F.E. MSK
tMDOnf
90
tMDOv
86
tMITOp
90
Alternate Mode: After R.E. MSK
MICROWIRE Data No Float (slave only) After F.E. MWCS
MICROWIRE Data Out Valid
200
0.5 t MSK 1.5 tMSK
56
0.0
0
56
Normal Mode: After F.E. MSK
56
Alternate Mode: After R.E. MSK
MDODI to MDIDO
(slave only)
Propagation Time
Value is the same in all clocking modes of
the MICROWIRE
56
CAN Output Signals
tCANTx
CANTx output valid
After R.E. CLKXn
13
ACCESS.bus Output Signals
tBUFo
89
Bus free time between Stop and Start
Condition
tSCLhigho
tCSTOso
89
SCL setup time
Before Stop Condition
tSCLhigho
tCSTRho
89
SCL hold time
After Start Condition
tSCLhigho
tCSTRso
90
SCL setup time
Before Start Condition
tSCLhigho
tDHCso
90
Data High setup time
Before SCL R.E.
tSCLhigho
-tSDAro
tDLCso
89
Data Low setup time
Before SCL R.E.
tSCLhigho
-tSDAfo
tSCLfo
88
SCL signal Fall time
300c
149
www.national.com
Table 45 Output Signals
Symbol Figure
Description
Reference
Min (ns) Max (ns)
-d
tSCLro
88
SCL signal Rise time
tSCLlowo
91
SCL low time
After SCL F.E.
K*t CLK 1e
tSCLhigho
91
SCL high time
After SCL R.E.
K*t CLK 1e
tSDAfo
88
SDA signal Fall time
tSDAro
88
SDA signal Rise time
tSDAho
91
SDA hold time
After SCL F.E.
tSDAvo
91
SDA valid time
After SCL F.E.
300
-
7*tCLK t SCLfo
7*tCLK+
tRD
a. Tclk is the actual clock period of the CPU clock used in the system.
The value of Tclk is system dependent.
The maximum cycle time of 64000ns is for Power Save mode; in active mode, the maximum cycle time is limited to 250ns by
the high frequency oscillator.
b. Guaranteed by design, but not fully tested.
c. Assuming signal’s capacitance up to 400pF.
d. Depends on the signal’s capacitance and the pull-up value. Must be less than 1ms.
e. K is as defined in ACBCTL2.SCLFRQ.
Table 46 Input Signal Requirements
Symbol Figure
Description
Reference
Min (ns)
t X1p
77
X1 period
R.E. X1 to next R.E. X1
40
t X1h
77
X1 high time, external clock
At 2V level (Both Edges)
0.5 Tclk - 4
tX1l
t X2p
77
77
X1 low time, external clock
X2 period a
At 0.8V level (Both Edges)
R.E. X2 to next R.E. X2
0.5 Tclk - 4
10,000
t X2h
77
X2 high time, external clock
At 2V level (both edges)
0.5 Tclk - 500
tX2l
77
81
At 0.8V level (both edges)
Before R.E. CLK
0.5 Tclk - 500
t Is
X2 low time, external clock
Input setup time
ISE
tIh
81
Input hold time
ISE, NMI, RXD1, RXD2
After R.E. CLK
t RST
82
Reset time
Reset active to reset end
12
0
4Tclk
Input Signals
Input Pulse Width
1*Tclk+13
USART Input Signals
t Is
80
Input setup time
RXDn (asynchronous mode)
Before R.E. CLK
tIh
80
Input hold time
RXDn (asynchronous mode)
After R.E. CLK
t CLKX
81
tRXS
81
tRXH
81
CKXn input period
(synchronous mode)
RDXn setup time
(synchronous mode)
RDXn hold time
(synchronous mode)
12
0
200
Before F.E. CKX in synchronous mode
After F.E. CKX in synchronous mode
4
2
MICROWIRE / SPI Input Signals
tMSKh
83
MICROWIRE Clock High
At 2.0V (both edges)
80
tMSKl
83
MICROWIRE Clock Low
At 0.8V (both edges)
80
www.national.com
150
Max (ns)
Table 46 Input Signal Requirements
Symbol Figure
Description
Reference
Min (ns)
200
tMSKp
83
84
MICROWIRE Clock Period
MnIDL bit = 0; R.E. MSK to next R.E. MSK
MIDL bit = 1; F.E. MSK to next F.E. MSK
tMSKh
83
MSK Hold (slave only)
After MCS becomes inactive
40
t MSKs
83
83
MSK Setup (slave only)
MCS Hold (slave only)
Before MCS becomes active
MIDL bit = 0: After F.E. MSK
80
tMCSh
tMCSs
84
83
84
83
tMDIh
MIDL bit = 1: After R.E. MSK
MCS Setup (slave only)
MIDL bit = 0: Before R.E. MSK
MIDL bit = 1: Before F.E. MSK
MICROWIRE Data In Hold (master) Normal Mode: After R.E. MSK
85
83
MICROWIRE Data In Hold (slave)
Alternate Mode: After F.E. MSK
85
tMDIs
83
Alternate Mode: After F.E. MSK
Normal Mode: After R.E. MSK
MICROWIRE Data In Setup
Normal Mode: Before R.E. MSK
Alternate Mode: Before F.E. MSK
85
Max (ns)
40
80
0
40
80
CAN Input Signals
t Is
CANRx Input setup time)
Before R.E. CLK
12
tIh
CANRx Input hold time
After R.E. CLK
0
ACCESS.bus Input Signals
Bus free time between Stop and
Start Condition
tBUFi
89
tSCLhigho
tCSTOsi
tCSTRhi
89
89
SCL setup time
SCL hold time
Before Stop Condition
After Start Condition
8*tCLK - tSCLri
8*tCLK - tSCLri
tCSTRsi
90
SCL setup time
Before Start Condition
8*tCLK - tSCLri
tDHCsi
tDLCsi
90
89
Data High setup time
Data Low setup time
Before SCL R.E.
Before SCL R.E.
tSCLfi
88
SCL signal Rise time
tSCLri
tSCLlowi
88
91
SCL signal Fall time
SCL low time
After SCL F.E.
16*tCLK
tSCLhighi
91
SCL high time
After SCL R.E.
16*tCLK
tSDAri
tSDAfi
88
88
SDA signal Rise time
SDA signal Fall time
tSDAhi
91
SDA hold time
After SCL F.E.
tSDAsi
91
SDA setup time
Before SCL R.E.
2*tCLK
2*tCLK
300
1000
300
1000
0
2*tCLK
Multi-Function Timer Input Signals
tTAH
92
TnA High Time
R.E. CLK
TCLK+5
tTAL
92
TnA Low Time
R.E. CLK
TCLK+5
92
92
TnB High Time
TnB Low Time
R.E. CLK
R.E. CLK
TCLK+5
TCLK+5
tTBH
tTBL
Versatile Timer Input Signals
tTIOH
96
TIOx Input High Time
RE CLK
1.5T CLK+5ns
tTIOL
96
TIOx Input Low Time
RE CLK
1.5T CLK+5ns
a. Only when operating with an external square wave on X2CKI; otherwise a 32kHz crystal network must be used
between X2CKI and X2CKO. If the slow clock is internally generated from the fast clock, it may not exceed this
given limit.
151
www.national.com
26.0 Appendix
The following document describes problems identified in the
CR16 modules.
26.1
CR16CAN
26.1.1
CR16CAN Problem Descriptions:
Applied to the CAN communication sequence described
above, this means that the transmitted message, currently
present in the hidden receive buffer will be copied into the
same receive buffer and the message received from the other CAN node will be overwritten.
Under certain conditions it occurs that the CR16CAN module
receives a frame, sent by itself even though the loopback feature is disabled.
Example
Buffer Settings
This condition consists of two parts, which both must be true
to cause this malfunction.
Filter Masks:
A) The first part is that a transmit buffer and at least one receive buffer are configured with the same identifier. Let's call
this identifier ID_RX_TX here. With regard to the receive
buffer, this means that the buffer identifier and the corresponding filter masks are setup in a way that the buffer is able
to receive frames with the identifier ID_RX_TX.
GMSKB = 0x0000
GMSKX = 0x000F
Buffer configuration:
CAN
Buffer
Number
B) The second part is that the CAN communication must take
place in the following sequence:
1. A message with the identifier ID_RX_TX from another
CAN node is received into the receive buffer.
2. A message with the identifier ID_RX_TX is sent by the
CR16CAN module immediately after the reception took
place (Note).
0
After this communication the frame sent by the CR16CAN
module will be copied into the next receive buffer available
for the identifier ID_RX_TX.
Identifier
Mask
TX_NOT_ACTIVE 0x15555550 0x1555555X
1
RX_READY
0x15555551 0x1555555X
2
RX_READY
0x15555552 0x1555555X
3
RX_READY
0x15555003 0x1555500X
CAN Communication Sequence A:
(BUFFLOCK disabled)
1. Message sent from another CAN node received into
buffer 1.
Buffer 1 and buffer 2 are tagged for reception of this
message.
2. CPU reads out data from CAN buffer 1 and resets the
buffer state from RX_FULL to RX_READY (Note 1).
3. CAN buffer 0 sends a frame (status set to TX_ONCE).
4. Status of CAN buffer 1 changes to RX_FULL, because it
has received the message sent by buffer 0 (Note 2).
Note: 1. Step 2 does not need to be done. In case the buffer
1 status is not updated to RX_READY, the buffer status will
change from RX_FULL to RX_OVERRUN in step 4.
CR16CAN Problem Cause
When a frame is received into the hidden receive buffer, the
CR16CAN module scans through all CAN message buffers.
During this sequence all receive buffers (RX-buffers) capable
of receiving this frame are tagged (RX-tag). If the message
was received correctly the frame is copied into first tagged
buffer (lowest buffer number).
Every CAN node also monitors frames being transmitted in
order to switch from transmitter to receiver after a lost arbitration. In order to do this the CR16CAN module also receives
transmitted frames into the hidden receive buffer.
Note: 2. As BUFFLOCK is disabled, all messages with the
identifier ID_RX_TX will be copied into buffer 1. Buffer 2 does
not receive any message.
The scanning sequence is also applied to transmitted
frames. This means, the identifier in the hidden receive buffer
is compared with the RX-buffer identifiers. As the identifier is
the same as with the last receive scanning sequence, the
RX-tags will not be changed.
CAN Communication Sequence B:
(BUFFLOCK enabled)
A CAN buffer receive tag is only updated in the following cases:
1. Message sent from another CAN node received into
buffer 1. Buffer 1 is locked now.
Buffer 1 and buffer 2 are tagged for reception of this
message.
2. CAN buffer 0 sends a frame.
3. Status of CAN buffer 2 changes to RX_FULL, because it
has received the message sent by buffer 0.
— A new scanning sequence has overwritten the old receive tags due to a different identifier mask under comparison.
— The CPU has changed the CAN buffer status in the
CNSTAT.ST-field to any value which disables the buffer to receive a message (e.g. RX_NOT_ACTIVE or
any TX-state)
— The CPU has changed the CAN buffer identifier.
www.national.com
Buffer
Identifier
X = don't care
Note: If a frame with an identifier different to ID_RX_TX is
sent or received in between the steps 1 and 2, the problem
does not occur.
26.1.2
CAN Buffer
Status
CAN Communication Sequence C
(CR16CAN does NOT receive a frame sent by itself.)
152
1. Message sent from another CAN node received into
buffer 1.
Buffer 1 and buffer 2 are tagged for reception of this
message.
2. Message sent from another CAN node received into
buffer 3 (ID=0x15555003).
Only buffer 3 is now tagged for reception.
3. CAN buffer 0 sends a frame (status set to TX_ONCE).
4. Status of CAN buffer 1 and 2 remains RX_READY, because they have not received the message sent by buffer 0.
26.1.3
should write the sequence RX_NOT_ACTIVE - RX_READY
to this receive buffer, which has received the latest message.
Modified CAN Communication Sequence:
In the CAN communication example described below, the
buffer 14 is set up as basic CAN path, which is able to receive
all standard frames. The buffers 1 to 13 cannot receive the
frame sent by buffer 0.
Filter Masks:
BMSKB = 0xFFF0
BMSKX = 0x0000
CR16CAN Problem Solutions
Reset receive buffer tags before transmitting a message
Buffer configuration:
The receive tag of a CAN receive buffer is reset when the
CPU updates the buffer status in the CNSTAT.ST-field to any
value which disables the receive buffer. Therefore the user
should write the sequence RX_NOT_ACTIVE - RX_READY
to all receive buffers which have an identifier filter matching
the identifier of the frame to be sent next before the message
is sent.
Modified CAN Communication Sequence:
The same CAN buffer settings as described in also apply to
this example.
Advantage:
No receive buffer is overwritten by a message sent by the
same CR16CAN node.
Disadvantage:
The corresponding receive buffers must be disabled for a
short period of time. During this time, when the receive buffers are in the RX_NOT_ACTIVE state, correct incoming
messages from other CAN nodes will get lost.
This method is more suitable compared to the method described in Section, if the number of transmit buffers with
identifier ID_RX_TX is lower than the number of receive buffers set up with the corresponding identifier mask.
Reset receive buffer tags after reception of a message
The receive tag of a CAN receive buffer is reset when the
CPU updates the buffer status in the CNSTAT.ST-field to any
value which disables the receive buffer. Therefore the user
CAN Buffer Status
Buffer Identifier
0
TX_NOT_ACTIVE
any standard frame
14
RX_READY
ID1.IDE bit = 1
1. Message sent from another CAN node received into
buffer 14.
Buffer 14 is tagged for reception of this message.
2. CPU reads out data from CAN buffer 14.
3. Write RX_NOT_ACTIVE to CNSTAT.ST-field of buffer
14.
Buffer 14 is NOT tagged for reception anymore.
4. Write RX_READY to CNSTAT.ST-field of buffer 14.
5. CAN buffer 0 sends a frame (status set to TX_ONCE).
6. Status of CAN buffer 14 remains RX_READY, because
it has NOT received the message sent by buffer 0.
(BUFFLOCK disabled)
1. Message sent from another CAN node received into
buffer 1.
Buffer 1 and buffer 2 are tagged for reception of this
message.
2. CPU reads out data from CAN buffer 1 and resets the
buffer state from RX_FULL to RX_READY.
3. Write RX_NOT_ACTIVE to CNSTAT.ST-field of buffer 1
and buffer 2.
Buffer 1 and buffer 2 are NOT tagged for reception anymore.
4. Write RX_READY to CNSTAT.ST-field of buffer 1 and
buffer 2.
5. CAN buffer 0 sends a frame (status set to TX_ONCE).
6. Status of CAN buffer 1 remains RX_READY, because it
has NOT received the message sent by buffer 0.
CAN Buffer
Number
Advantage:
No receive buffer is overwritten by a message sent by the
same CR16CAN node.
Disadvantage:
The corresponding receive buffers must be disabled for a
short period of time. During this time, when the receive buffers are in the RX_NOT_ACTIVE state, correctly incoming
messages from other CAN nodes will get lost.
This method is more suitable compared to the method described in Section, if the number of transmit buffers with
identifier ID_RX_TX is higher than the number of receive
buffers set up with the corresponding identifier mask. This is
the case if only the basic CAN path to buffer 14 is configured
to receive a range of identifiers, including the identifier
ID_RX_TX. All other buffers are configured with unique identifier filters.
Receive all frames and discard those, which were sent
by the same CR16CAN node.
Another approach to overcome this problem uses the Time
Stamp counter of the CR16CAN module to determine,
whether a message was sent and received at the same time.
This is the case when a transmitted frame is received by the
same CAN node.
153
www.national.com
When a frame was successfully sent by CR16CAN the contents of the Time Stamp counter are captured into the Time
Stamp register (TSTP) of the transmit buffer during the ACKslot of the frame currently being sent. Also, when a message
is received, the TSTP-register of the receiving buffer is loaded with the Time Stamp counter value during the ACK-slot of
the CAN frame currently being received.
26.2
8/16-BIT MICROWIRE/SPI (MWSPI16)
26.2.1
MWSPI16 Problem Description
According to the specification, the MSKn clock output in master mode should have the value of the MnIDL bit of the
MWnCTL register, even when the module is disabled. However, the MSKn pin is enabled and the module is disabled.
thus, even if the MnIDL bit is set, the MSKn clock will change
to a low level as soon as the module is disabled. If any slave
is selected at this time, i will interpret this unwanted transition
as a shift clock.
This means, in the case where a message is received in one
buffer, which was sent from another buffer of the same
CR16CAN node, the TSTP-register contents are equal after
this transaction.
A comparison of the two TSTP-register values can be inserted into a "read from CAN receive buffer" software routine, to
distinguish whether the data received are from another CAN
node (valid) or from the same CAN node (invalid).
26.2.2
The flowchart below shows a possible implementation. The
same CAN buffer settings as described in Section also apply
to this example.
26.2.3
Even if the module is disabled and the alternate function of
the MSKn pin is enabled, the module can still influence the
MSKn pin and drives the default value ‘0’.
MWSPI16 Problem Solutions
When the MSKn idle level of ‘1’ is to be used, the following
procedure should be followed when the module is disabled:
Modified CAN Receive Sequence:
1. Set the MSKn pin to high level in the corresponding port
data output register.
2. Configure the MSKn pin to an output in the corresponding port direction register.
3. Disable the alternate function of the MSKn pin in the corresponding port alternate function register.
4. Disable the MWSPI16 module.
Message received
into Buffer 1
Time Stamp of
Buffer 1
=
Time Stamp of
Buffer 0
26.3
TIMING AND WATCHDOG MODULE
26.3.1
Timing and WATCHDOG Module Problem
Description
The available window for a valid WATCHDOG service varies
with the TWM configuration and the operating mode of the
R16MCS9. Therefore it is not possible to generally provide
the limits for the maximum service window. However, the limits for the minimum service window is guaranteed and should
be used.
Read out Buffer 1 Data
26.3.2
Reset Buffer 1 Status to
RX_READY
Timing and WATCHDOG Module Problem
Cause
The timing and WATCHDOG module uses two different clock
signals for its operation, the slow system clock as well as the
fast system clock.
Exit
The slow system clock can either be generated by an external 32 kHz quartz or it can be derived from the fast system
clock by means of a prescaler counter in the CLK2RES modules. The TWM can operate off a maximum slow system
clock of 100 kHz. The WATCHDOG counter (down-counter)
is either clocked directly by the slow system (T0IN) or it is
decremented every time the counter T0 underflows
(T0OUT).
Advantage:
None of the CAN receive buffers must be disabled at any
time.
Disadvantage:
The receive buffer contents are overwritten by an invalid
message sent from the same CR16CAN node.
www.national.com
MWSPI16 Problem Cause
The fast system clock is used for accesses to TWM registers,
which build the user interface of the TWM. These user interface registers include all memory-mapped registers of the
TWM.
154
Every time the user (CR16B core) writes to a TWM configuration register or to the WATCHDOG Service Data Match
register, this “high speed operation” must be synchronized to
the internal TWM logic running at the slow clock rate. This
synchronization process takes a variable number of low
speed clock cycles, depending on the ratio between the lowspeed and the high-speed system clock and the phase shift
between the two clock signals. The more the two frequencies
differ from each other, the longer it takes the synchronization
process.
Figure 94.
In other words, write operations to the TWM registers take a
certain number of low-speed clock cycles to show the desired effects to the TWM logic.
This fact is especially critical for the write operation for the
WATCHDOG service, as it affects the allowed window for a
valid WATCHDOG service.
If the device runs in active mode, the synchronization process can take up to four WATCHDOG counter clock cycles.
This limits the available WATCHDOG service to the window
shown in figure 94:
WATCHDOG Services Windows in Active Mode
If the device runs in power save mode, the synchronization
process can take up to eight WATCHDOG counter clock cy-
cles. This limits the available WATCHDOG service to the window shown in figure 95:
Figure 95. WATCHDOG Services Windows in Power Save Mode
26.3.3
Timing and WATCHDOG Module Problem
Solutions
In order to guarantee a valid WATCHDOG service under all
circumstances, the WATCHDOG should only be serviced
within the guaranteed minimum valid window, as illustrated in
figure 94 and figure 95 in the previous section.
155
www.national.com
CR16HCS5/CR16HCS9/CR16MAR5/CR16MAS5 CR16MAS9/CR16MBR5/CR16MCS5/CR16MCS9 Family of 16-bit
CAN-enabled CompactRISC Microcontrollers
LIFE SUPPORT POLICY
NATIONAL’S PRODUCTS ARE NOT AUTHORIZED FOR USE AS CRITICAL COMPONENTS IN LIFE SUPPORT
DEVICES OR SYSTEMS WITHOUT THE EXPRESS WRITTEN APPROVAL OF THE PRESIDENT AND GENERAL
COUNSEL OF NATIONAL SEMICONDUCTOR CORPORATION. As used herein:
1. Life support devices or systems are devices or systems 2. A critical component is any component of a life support
which, (a) are intended for surgical implant into the body,
device or system whose failure to perform can be reaor (b) support or sustain life, and whose failure to personably expected to cause the failure of the life support
form, when properly used in accordance with instructions
device or system, or to affect its safety or effectiveness.
for use provided in the labeling, can be reasonably expected to result in a significant injury to the user.
National Semiconductor
Corporation
Tel: 1-800-272-9959
Fax: 1-800-737-7018
Email: [email protected]
National Semiconductor
Europe
Fax:
+49 (0) 180-530 85 86
Email:
[email protected]
Deutsch Tel:
+49 (0) 69 9508 6208
English Tel:
+44 (0) 870 24 0 2171
Francais Tel:
+33 (0) 1 41 91 8790
National Semiconductor
Asia Pacific
Customer Response Group
Tel: 65-254-4466
Fax: 65-250-4466
Email: [email protected]
National Semiconductor
Japan Ltd.
Tel: 81-3-5639-7560
Fax: 81-3-5639-7507
www.national.com
National does not assume any responsibility for use of any circuitry described, no circuit patent licenses are implied, and National reserves the right, at any time without notice, to change said circuitry or specifications.