Data Sheet

Features
• High-performance, Low-power AVR® 8-bit Microcontroller

• Advanced RISC Architecture
– 131 Powerful Instructions – Most Single-clock Cycle Execution
– 32 x 8 General Purpose Working Registers
– Fully Static Operation
– Up to 20 MIPS Throughput at 20 MHz
– On-chip 2-cycle Multiplier
• Nonvolatile Program and Data Memories
– 16/32/64K Bytes of In-System Self-Programmable Flash
Endurance: 10,000 Write/Erase Cycles
– Optional Boot Code Section with Independent Lock Bits
8-bit
In-System Programming by On-chip Boot Program
True Read-While-Write Operation
Microcontroller
– 512B/1K/2K Bytes EEPROM
Endurance: 100,000 Write/Erase Cycles with 16/32/64K
– 1/2/4K Bytes Internal SRAM
– Programming Lock for Software Security Bytes In-System
• JTAG (IEEE std. 1149.1 Compliant) Interface
– Boundary-scan Capabilities According to the JTAG Standard Programmable
– Extensive On-chip Debug Support
•
– Programming of Flash, EEPROM, Fuses, and Lock Bits through the JTAG Interface
Peripheral Features
Flash
– Two 8-bit Timer/Counters with Separate Prescalers and Compare Modes
– One 16-bit Timer/Counter with Separate Prescaler, Compare Mode, and Capture
Mode ATmega164P/V
– Real Time Counter with Separate Oscillator
– Six PWM Channels ATmega324P/V
– 8-channel, 10-bit ADC
Differential mode with selectable gain at 1x, 10x or 200x ATmega644P/V
– Byte-oriented Two-wire Serial Interface
– Two Programmable Serial USART
– Master/Slave SPI Serial Interface
– Programmable Watchdog Timer with Separate On-chip Oscillator Preliminary
– On-chip Analog Comparator
– Interrupt and Wake-up on Pin Change
• Special Microcontroller Features
– Power-on Reset and Programmable Brown-out Detection
– Internal Calibrated RC Oscillator
– External and Internal Interrupt Sources
– Six Sleep Modes: Idle, ADC Noise Reduction, Power-save, Power-down, Standby
and Extended Standby
• I/O and Packages
– 32 Programmable I/O Lines
– 40-pin PDIP, 44-lead TQFP, and 44-pad QFN/MLF
• Operating Voltages
– 1.8 - 5.5V for ATmega164P/324P/644PV
– 2.7 - 5.5V for ATmega164P/324P/644P
• Speed Grades
– ATmega164P/324P/644PV: 0 - 4MHz @ 1.8 - 5.5V, 0 - 10MHz @ 2.7 - 5.5V
– ATmega164P/324P/644P: 0 - 10MHz @ 2.7 - 5.5V, 0 - 20MHz @ 4.5 - 5.5V
• Power Consumption at 1 MHz, 1.8V, 25°C for ATmega164P/324P/644P
– Active: 338/398/TBD µA
– Power-down Mode:0.035 /0.027/TBD µA
– Power-save Mode:0.5 /0.5/TBD µA (Including 32 kHz RTC)
8011D–AVR–02/07
1. Pin Configurations
Figure 1-1. Pinout ATmega164P/324P/644P
PDIP
(PCINT8/XCK0/T0) PB0 PA0 (ADC0/PCINT0)

(PCINT9/CLKO/T1) PB1 PA1 (ADC1/PCINT1)
(PCINT10/INT2/AIN0) PB2 PA2 (ADC2/PCINT2)
(PCINT11/OC0A/AIN1) PB3 PA3 (ADC3/PCINT3)
(PCINT12/OC0B/SS) PB4 PA4 (ADC4/PCINT4)
(PCINT13/MOSI) PB5 PA5 (ADC5/PCINT5)
(PCINT14/MISO) PB6 PA6 (ADC6/PCINT6)
(PCINT15/SCK) PB7 PA7 (ADC7/PCINT7)
RESET AREF
VCC GND
GND AVCC
XTAL2 PC7 (TOSC2/PCINT23)
(PCINT24/RXD0) PD0 PC5 (TDI/PCINT21)
(PCINT25/TXD0) PD1 PC4 (TDO/PCINT20)
(PCINT26/RXD1/INT0) PD2 PC3 (TMS/PCINT19)
(PCINT27/TXD1/INT1) PD3 PC2 (TCK/PCINT18)
(PCINT28/XCK1/OC1B) PD4 PC1 (SDA/PCINT17)
(PCINT29/OC1A) PD5 PC0 (SCL/PCINT16)
(PCINT30/OC2B/ICP) PD6 PD7 (OC2A/PCINT31)
TQFP/QFN/MLF
PB3 (AIN1/OC0A/PCINT11)
PB2 (AIN0/INT2/PCINT10)
PB4 (SS/OC0B/PCINT12)
PB1 (T1/CLKO/PCINT9)
PB0 (XCK0/T0/PCINT8)
PA0 (ADC0/PCINT0)
PA1 (ADC1/PCINT1)
PA2 (ADC2/PCINT2)
PA3 (ADC3/PCINT3)
GND
VCC
(PCINT13/MOSI) PB5 PA4 (ADC4/PCINT4)

(PCINT14/MISO) PB6 PA5 (ADC5/PCINT5)
(PCINT15/SCK) PB7 PA6 (ADC6/PCINT6)
RESET PA7 (ADC7/PCINT7)
VCC AREF
GND GND
XTAL2 AVCC
(PCINT24/RXD0) PD0 PC6 (TOSC1/PCINT22)
(PCINT25/TXD0) PD1 PC5 (TDI/PCINT21)
(PCINT26/RXD1/INT0) PD2 PC4 (TDO/PCINT20)
PD3
PD4
PD5
PD6
PD7
VCC
GND
(PCINT16/SCL) PC0
(PCINT17/SDA) PC1
(PCINT18/TCK) PC2
(PCINT19/TMS) PC3
(PCINT27/TXD1/INT1)
(PCINT28/XCK1/OC1B)
(PCINT29/OC1A)
(PCINT30/OC2B/ICP)
(PCINT31/OC2A)
Note: The large center pad underneath the QFN/MLF package should be soldered to ground on the
board to ensure good mechanical stability.
2 ATmega164P/324P/644P
8011D–AVR–02/07
ATmega164P/324P/644P
2. Overview
The ATmega164P/324P/644P is a low-power CMOS 8-bit microcontroller based on the AVR
enhanced RISC architecture. By executing powerful instructions in a single clock cycle, the
ATmega164P/324P/644P achieves throughputs approaching 1 MIPS per MHz allowing the sys-
tem designer to optimize power consumption versus processing speed.
2.1 Block Diagram
Figure 2-1. Block Diagram
PA7..0 PB7..0
VCC
Power
Supervision
RESET PORT A (8)
POR / BOD & PORT B (8)
RESET
Watchdog
GND Timer
Watchdog A/D Analog

Comparator USART 0
Oscillator Converter
XTAL1
Oscillator Internal
Circuits / EEPROM Bandgap reference SPI
Clock
Generation
XTAL2
8bit T/C 0
CPU
JTAG/OCD 16bit T/C 1
USART 1
TWI FLASH SRAM 8bit T/C 2
PORT C (8) PORT D (8)
TOSC2/PC7 TOSC1/PC6 PC5..0 PD7..0
The AVR core combines a rich instruction set with 32 general purpose working registers. All the
32 registers are directly connected to the Arithmetic Logic Unit (ALU), allowing two independent
registers to be accessed in one single instruction executed in one clock cycle. The resulting
architecture is more code efficient while achieving throughputs up to ten times faster than con-
ventional CISC microcontrollers.
3
8011D–AVR–02/07
The ATmega164P/324P/644P provides the following features: 16/32/64K bytes of In-System
Programmable Flash with Read-While-Write capabilities, 512B/1K/2K bytes EEPROM, 1/2/4K
bytes SRAM, 32 general purpose I/O lines, 32 general purpose working registers, Real Time
Counter (RTC), three flexible Timer/Counters with compare modes and PWM, 2 USARTs, a byte
oriented 2-wire Serial Interface, a 8-channel, 10-bit ADC with optional differential input stage
with programmable gain, programmable Watchdog Timer with Internal Oscillator, an SPI serial
port, IEEE std. 1149.1 compliant JTAG test interface, also used for accessing the On-chip
Debug system and programming and six software selectable power saving modes. The Idle
mode stops the CPU while allowing the SRAM, Timer/Counters, SPI port, and interrupt system
to continue functioning. The Power-down mode saves the register contents but freezes the
Oscillator, disabling all other chip functions until the next interrupt or Hardware Reset. In Power-
save mode, the asynchronous timer continues to run, allowing the user to maintain a timer base
while the rest of the device is sleeping. The ADC Noise Reduction mode stops the CPU and all
I/O modules except Asynchronous Timer and ADC, to minimize switching noise during ADC
conversions. In Standby mode, the Crystal/Resonator Oscillator is running while the rest of the
device is sleeping. This allows very fast start-up combined with low power consumption. In
Extended Standby mode, both the main Oscillator and the Asynchronous Timer continue to run.
The device is manufactured using Atmel’s high-density nonvolatile memory technology. The On-
chip ISP Flash allows the program memory to be reprogrammed in-system through an SPI serial
interface, by a conventional nonvolatile memory programmer, or by an On-chip Boot program
running on the AVR core. The boot program can use any interface to download the application
program in the application Flash memory. Software in the Boot Flash section will continue to run
while the Application Flash section is updated, providing true Read-While-Write operation. By
combining an 8-bit RISC CPU with In-System Self-Programmable Flash on a monolithic chip,
the Atmel ATmega164P/324P/644P is a powerful microcontroller that provides a highly flexible
and cost effective solution to many embedded control applications.
The ATmega164P/324P/644P AVR is supported with a full suite of program and system devel-
opment tools including: C compilers, macro assemblers, program debugger/simulators, in-circuit
emulators, and evaluation kits.
2.2 Comparison Between ATmega164P, ATmega324P and ATmega644P
Table 2-1. Differences between ATmega164P and ATmega644P

Device Flash EEPROM RAM
ATmega164P 16 Kbyte 512 Bytes 1 Kbyte
ATmega324P 32 Kbyte 1 Kbyte 2 Kbyte
ATmega644P 64 Kbyte 2 Kbyte 4 Kbyte
2.3 Pin Descriptions

2.3.1 VCC
Digital supply voltage.

2.3.2 GND
Ground.
8011D–AVR–02/07
2.3.3 Port A (PA7:PA0)
Port A serves as analog inputs to the Analog-to-digital Converter.

Port A also serves as an 8-bit bi-directional I/O port with internal pull-up resistors (selected for
each bit). The Port A output buffers have symmetrical drive characteristics with both high sink
and source capability. As inputs, Port A pins that are externally pulled low will source current if
the pull-up resistors are activated. The Port A pins are tri-stated when a reset condition becomes
active, even if the clock is not running.
Port A also serves the functions of various special features of the ATmega164P/324P/644P as
listed on page 79.
2.3.4 Port B (PB7:PB0)
Port B is an 8-bit bi-directional I/O port with internal pull-up resistors (selected for each bit). The
Port B output buffers have symmetrical drive characteristics with both high sink and source
capability. As inputs, Port B pins that are externally pulled low will source current if the pull-up
resistors are activated. The Port B pins are tri-stated when a reset condition becomes active,
even if the clock is not running.
Port B also serves the functions of various special features of the ATmega164P/324P/644P as
listed on page 81.
2.3.5 Port C (PC7:PC0)
Port C is an 8-bit bi-directional I/O port with internal pull-up resistors (selected for each bit). The
Port C output buffers have symmetrical drive characteristics with both high sink and source
capability. As inputs, Port C pins that are externally pulled low will source current if the pull-up
resistors are activated. The Port C pins are tri-stated when a reset condition becomes active,
Port C also serves the functions of the JTAG interface, along with special features of the
ATmega164P/324P/644P as listed on page 84.
2.3.6 Port D (PD7:PD0)
Port D is an 8-bit bi-directional I/O port with internal pull-up resistors (selected for each bit). The
Port D output buffers have symmetrical drive characteristics with both high sink and source
capability. As inputs, Port D pins that are externally pulled low will source current if the pull-up
resistors are activated. The Port D pins are tri-stated when a reset condition becomes active,
Port D also serves the functions of various special features of the ATmega164P/324P/644P as
listed on page 86.
2.3.7 RESET
Reset input. A low level on this pin for longer than the minimum pulse length will generate a
reset, even if the clock is not running. The minimum pulse length is given in ”System and Reset
Characteristics” on page 329. Shorter pulses are not guaranteed to generate a reset.
2.3.8 XTAL1
Input to the inverting Oscillator amplifier and input to the internal clock operating circuit.
5
8011D–AVR–02/07
2.3.9 XTAL2
Output from the inverting Oscillator amplifier.

2.3.10 AVCC
AVCC is the supply voltage pin for Port F and the Analog-to-digital Converter. It should be exter-
nally connected to VCC, even if the ADC is not used. If the ADC is used, it should be connected
to VCC through a low-pass filter.
2.3.11 AREF
This is the analog reference pin for the Analog-to-digital Converter.
8011D–AVR–02/07
3. Resources
A comprehensive set of development tools, application notes and datasheetsare available for
download on http://www.atmel.com/avr.
7
8011D–AVR–02/07
4. About Code Examples
This documentation contains simple code examples that briefly show how to use various parts of
the device. Be aware that not all C compiler vendors include bit definitions in the header files
and interrupt handling in C is compiler dependent. Please confirm with the C compiler documen-
tation for more details.
The code examples assume that the part specific header file is included before compilation. For
I/O registers located in extended I/O map, "IN", "OUT", "SBIS", "SBIC", "CBI", and "SBI" instruc-
tions must be replaced with instructions that allow access to extended I/O. Typically "LDS" and
"STS" combined with "SBRS", "SBRC", "SBR", and "CBR".
8011D–AVR–02/07
5. AVR CPU Core

5.1 Overview
This section discusses the AVR core architecture in general. The main function of the CPU core
is to ensure correct program execution. The CPU must therefore be able to access memories,
perform calculations, control peripherals, and handle interrupts.
Figure 5-1. Block Diagram of the AVR Architecture
Data Bus 8-bit
Program Status
Flash
Counter and Control
Program
Memory
Interrupt
32 x 8 Unit
Instruction General
Register Purpose SPI
Registrers Unit
Instruction Watchdog
Decoder Timer
Indirect Addressing
Direct Addressing
ALU Analog
Control Lines Comparator
I/O Module1
Data I/O Module 2

SRAM
I/O Module n
EEPROM
I/O Lines
In order to maximize performance and parallelism, the AVR uses a Harvard architecture – with
separate memories and buses for program and data. Instructions in the program memory are
executed with a single level pipelining. While one instruction is being executed, the next instruc-
tion is pre-fetched from the program memory. This concept enables instructions to be executed
in every clock cycle. The program memory is In-System Reprogrammable Flash memory.
The fast-access Register File contains 32 x 8-bit general purpose working registers with a single
clock cycle access time. This allows single-cycle Arithmetic Logic Unit (ALU) operation. In a typ-
9
8011D–AVR–02/07
ical ALU operation, two operands are output from the Register File, the operation is executed,
and the result is stored back in the Register File – in one clock cycle.
Six of the 32 registers can be used as three 16-bit indirect address register pointers for Data
Space addressing – enabling efficient address calculations. One of the these address pointers
can also be used as an address pointer for look up tables in Flash program memory. These
added function registers are the 16-bit X-, Y-, and Z-register, described later in this section.
The ALU supports arithmetic and logic operations between registers or between a constant and
a register. Single register operations can also be executed in the ALU. After an arithmetic opera-
tion, the Status Register is updated to reflect information about the result of the operation.
Program flow is provided by conditional and unconditional jump and call instructions, able to
directly address the whole address space. Most AVR instructions have a single 16-bit word for-
mat. Every program memory address contains a 16- or 32-bit instruction.
Program Flash memory space is divided in two sections, the Boot Program section and the
Application Program section. Both sections have dedicated Lock bits for write and read/write
protection. The SPM instruction that writes into the Application Flash memory section must
reside in the Boot Program section.
During interrupts and subroutine calls, the return address Program Counter (PC) is stored on the
Stack. The Stack is effectively allocated in the general data SRAM, and consequently the Stack
size is only limited by the total SRAM size and the usage of the SRAM. All user programs must
initialize the SP in the Reset routine (before subroutines or interrupts are executed). The Stack
Pointer (SP) is read/write accessible in the I/O space. The data SRAM can easily be accessed
through the five different addressing modes supported in the AVR architecture.
The memory spaces in the AVR architecture are all linear and regular memory maps.
A flexible interrupt module has its control registers in the I/O space with an additional Global
Interrupt Enable bit in the Status Register. All interrupts have a separate Interrupt Vector in the
Interrupt Vector table. The interrupts have priority in accordance with their Interrupt Vector posi-
tion. The lower the Interrupt Vector address, the higher the priority.
The I/O memory space contains 64 addresses for CPU peripheral functions as Control Regis-
ters, SPI, and other I/O functions. The I/O Memory can be accessed directly, or as the Data
Space locations following those of the Register File, 0x20 - 0x5F. In addition, the
ATmega164P/324P/644P has Extended I/O space from 0x60 - 0xFF in SRAM where only the
ST/STS/STD and LD/LDS/LDD instructions can be used.
5.2 ALU – Arithmetic Logic Unit

The high-performance AVR ALU operates in direct connection with all the 32 general purpose
working registers. Within a single clock cycle, arithmetic operations between general purpose
registers or between a register and an immediate are executed. The ALU operations are divided
into three main categories – arithmetic, logical, and bit-functions. Some implementations of the
architecture also provide a powerful multiplier supporting both signed/unsigned multiplication
and fractional format. See the “Instruction Set” section for a detailed description.
5.3 Status Register

The Status Register contains information about the result of the most recently executed arith-
metic instruction. This information can be used for altering program flow in order to perform
conditional operations. Note that the Status Register is updated after all ALU operations, as
10 ATmega164P/324P/644P
8011D–AVR–02/07
specified in the Instruction Set Reference. This will in many cases remove the need for using the
dedicated compare instructions, resulting in faster and more compact code.
The Status Register is not automatically stored when entering an interrupt routine and restored
when returning from an interrupt. This must be handled by software.
5.3.1 SREG – Status Register
The AVR Status Register – SREG – is defined as:

Bit 7 6 5 4 3 2 1 0
0x3F (0x5F) I T H S V N Z C SREG
Read/Write R/W R/W R/W R/W R/W R/W R/W R/W
Initial Value 0 0 0 0 0 0 0 0
• Bit 7 – I: Global Interrupt Enable

The Global Interrupt Enable bit must be set for the interrupts to be enabled. The individual inter-
rupt enable control is then performed in separate control registers. If the Global Interrupt Enable
Register is cleared, none of the interrupts are enabled independent of the individual interrupt
enable settings. The I-bit is cleared by hardware after an interrupt has occurred, and is set by
the RETI instruction to enable subsequent interrupts. The I-bit can also be set and cleared by
the application with the SEI and CLI instructions, as described in the instruction set reference.
• Bit 6 – T: Bit Copy Storage

The Bit Copy instructions BLD (Bit LoaD) and BST (Bit STore) use the T-bit as source or desti-
nation for the operated bit. A bit from a register in the Register File can be copied into T by the
BST instruction, and a bit in T can be copied into a bit in a register in the Register File by the
BLD instruction.
• Bit 5 – H: Half Carry Flag

The Half Carry Flag H indicates a Half Carry in some arithmetic operations. Half Carry Is useful
in BCD arithmetic. See the “Instruction Set Description” for detailed information.
• Bit 4 – S: Sign Bit, S = N ⊕ V

The S-bit is always an exclusive or between the Negative Flag N and the Two’s Complement
Overflow Flag V. See the “Instruction Set Description” for detailed information.
• Bit 3 – V: Two’s Complement Overflow Flag

The Two’s Complement Overflow Flag V supports two’s complement arithmetics. See the
“Instruction Set Description” for detailed information.
• Bit 2 – N: Negative Flag

The Negative Flag N indicates a negative result in an arithmetic or logic operation. See the
“Instruction Set Description” for detailed information.
• Bit 1 – Z: Zero Flag

The Zero Flag Z indicates a zero result in an arithmetic or logic operation. See the “Instruction
Set Description” for detailed information.
• Bit 0 – C: Carry Flag

The Carry Flag C indicates a carry in an arithmetic or logic operation. See the “Instruction Set
Description” for detailed information.
11
8011D–AVR–02/07
5.4 General Purpose Register File
The Register File is optimized for the AVR Enhanced RISC instruction set. In order to achieve
the required performance and flexibility, the following input/output schemes are supported by the
Register File:
• One 8-bit output operand and one 8-bit result input
• Two 8-bit output operands and one 8-bit result input
• Two 8-bit output operands and one 16-bit result input
• One 16-bit output operand and one 16-bit result input
Figure 5-2 shows the structure of the 32 general purpose working registers in the CPU.
Figure 5-2. AVR CPU General Purpose Working Registers
7 0 Addr.
R0 0x00
R1 0x01
R2 0x02
…
R13 0x0D
General R14 0x0E
Purpose R15 0x0F
Working R16 0x10
Registers R17 0x11
…
R26 0x1A X-register Low Byte
R27 0x1B X-register High Byte
R28 0x1C Y-register Low Byte
R29 0x1D Y-register High Byte
R30 0x1E Z-register Low Byte
R31 0x1F Z-register High Byte
Most of the instructions operating on the Register File have direct access to all registers, and
most of them are single cycle instructions.
As shown in Figure 5-2, each register is also assigned a data memory address, mapping them
directly into the first 32 locations of the user Data Space. Although not being physically imple-
mented as SRAM locations, this memory organization provides great flexibility in access of the
registers, as the X-, Y- and Z-pointer registers can be set to index any register in the file.
12 ATmega164P/324P/644P
8011D–AVR–02/07
5.4.1 The X-register, Y-register, and Z-register
The registers R26..R31 have some added functions to their general purpose usage. These reg-
isters are 16-bit address pointers for indirect addressing of the data space. The three indirect
address registers X, Y, and Z are defined as described in Figure 5-3.
Figure 5-3. The X-, Y-, and Z-registers
15 XH XL 0
X-register 7 0 7 0
R27 (0x1B) R26 (0x1A)
15 YH YL 0
Y-register 7 0 7 0
R29 (0x1D) R28 (0x1C)
15 ZH ZL 0
Z-register 7 0 7 0
R31 (0x1F) R30 (0x1E)
In the different addressing modes these address registers have functions as fixed displacement,
automatic increment, and automatic decrement (see the instruction set reference for details).
5.5 Stack Pointer

The Stack is mainly used for storing temporary data, for storing local variables and for storing
return addresses after interrupts and subroutine calls. The Stack Pointer Register always points
to the top of the Stack. Note that the Stack is implemented as growing from higher memory loca-
tions to lower memory locations. This implies that a Stack PUSH command decreases the Stack
Pointer.
The Stack Pointer points to the data SRAM Stack area where the Subroutine and Interrupt
Stacks are located. This Stack space in the data SRAM must be defined by the program before
any subroutine calls are executed or interrupts are enabled. The Stack Pointer must be set to
point above 0x0100. The initial value of the stack pointer is the last address of the internal
SRAM. The Stack Pointer is decremented by one when data is pushed onto the Stack with the
PUSH instruction, and it is decremented by three when the return address is pushed onto the
Stack with subroutine call or interrupt. The Stack Pointer is incremented by one when data is
popped from the Stack with the POP instruction, and it is incremented by three when data is
popped from the Stack with return from subroutine RET or return from interrupt RETI.
The AVR Stack Pointer is implemented as two 8-bit registers in the I/O space. The number of
bits actually used is implementation dependent. Note that the data space in some implementa-
tions of the AVR architecture is so small that only SPL is needed. In this case, the SPH Register
will not be present.
13
8011D–AVR–02/07
5.5.1 SPH and SPL – Stack Pointer High and Stack pointer Low
Bit 15 14 13 12 11 10 9 8
0x3E (0x5E) – – – SP12 SP11 SP10 SP9 SP8 SPH
0x3D (0x5D) SP7 SP6 SP5 SP4 SP3 SP2 SP1 SP0 SPL
7 6 5 4 3 2 1 0
Read/Write R R R R/W R/W R/W R/W R/W
R/W R/W R/W R/W R/W R/W R/W R/W
1 1 1 1 1 1 1 1
5.5.2 RAMPZ – Extended Z-pointer Register for ELPM/SPM
Bit 7 6 5 4 3 2 1 0
0x3B (0x5B) RAMPZ7 RAMPZ6 RAMPZ5 RAMPZ4 RAMPZ3 RAMPZ2 RAMPZ1 RAMPZ0 RAMPZ
For ELPM/SPM instructions, the Z-pointer is a concatenation of RAMPZ, ZH, and ZL, as shown
in Figure 5-4. Note that LPM is not affected by the RAMPZ setting.
Figure 5-4. The Z-pointer used by ELPM and SPM
Bit ( 7 0 7 0 7 0
Individually)
RAMPZ ZH ZL
Bit (Z-pointer) 23 16 15 8 7 0
The actual number of bits is implementation dependent. Unused bits in an implementation will
always read as zero. For compatibility with future devices, be sure to write these bits to zero.
5.6 Instruction Execution Timing

This section describes the general access timing concepts for instruction execution. The AVR
CPU is driven by the CPU clock clkCPU, directly generated from the selected clock source for the
chip. No internal clock division is used.
Figure 5-5 on page 15 shows the parallel instruction fetches and instruction executions enabled
by the Harvard architecture and the fast-access Register File concept. This is the basic pipelin-
ing concept to obtain up to 1 MIPS per MHz with the corresponding unique results for functions
per cost, functions per clocks, and functions per power-unit.
14 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 5-5. The Parallel Instruction Fetches and Instruction Executions
T1 T2 T3 T4
clkCPU
1st Instruction Fetch
1st Instruction Execute
2nd Instruction Fetch
2nd Instruction Execute
3rd Instruction Fetch
3rd Instruction Execute
4th Instruction Fetch
Figure 5-6 shows the internal timing concept for the Register File. In a single clock cycle an ALU
operation using two register operands is executed, and the result is stored back to the destina-
tion register.
Figure 5-6. Single Cycle ALU Operation

T1 T2 T3 T4
clkCPU
Total Execution Time
Register Operands Fetch
ALU Operation Execute
Result Write Back
5.7 Reset and Interrupt Handling

The AVR provides several different interrupt sources. These interrupts and the separate Reset
Vector each have a separate program vector in the program memory space. All interrupts are
assigned individual enable bits which must be written logic one together with the Global Interrupt
Enable bit in the Status Register in order to enable the interrupt. Depending on the Program
Counter value, interrupts may be automatically disabled when Boot Lock bits BLB02 or BLB12
are programmed. This feature improves software security. See the section ”Memory Program-
ming” on page 291 for details.
The lowest addresses in the program memory space are by default defined as the Reset and
Interrupt Vectors. The complete list of vectors is shown in ”Interrupts” on page 61. The list also
determines the priority levels of the different interrupts. The lower the address the higher is the
priority level. RESET has the highest priority, and next is INT0 – the External Interrupt Request
0. The Interrupt Vectors can be moved to the start of the Boot Flash section by setting the IVSEL
bit in the MCU Control Register (MCUCR). Refer to ”Interrupts” on page 61 for more information.
The Reset Vector can also be moved to the start of the Boot Flash section by programming the
BOOTRST Fuse, see ”Memory Programming” on page 291.
When an interrupt occurs, the Global Interrupt Enable I-bit is cleared and all interrupts are dis-
abled. The user software can write logic one to the I-bit to enable nested interrupts. All enabled
15
8011D–AVR–02/07
interrupts can then interrupt the current interrupt routine. The I-bit is automatically set when a
Return from Interrupt instruction – RETI – is executed.
There are basically two types of interrupts. The first type is triggered by an event that sets the
Interrupt Flag. For these interrupts, the Program Counter is vectored to the actual Interrupt Vec-
tor in order to execute the interrupt handling routine, and hardware clears the corresponding
Interrupt Flag. Interrupt Flags can also be cleared by writing a logic one to the flag bit position(s)
to be cleared. If an interrupt condition occurs while the corresponding interrupt enable bit is
cleared, the Interrupt Flag will be set and remembered until the interrupt is enabled, or the flag is
cleared by software. Similarly, if one or more interrupt conditions occur while the Global Interrupt
Enable bit is cleared, the corresponding Interrupt Flag(s) will be set and remembered until the
Global Interrupt Enable bit is set, and will then be executed by order of priority.
The second type of interrupts will trigger as long as the interrupt condition is present. These
interrupts do not necessarily have Interrupt Flags. If the interrupt condition disappears before the
interrupt is enabled, the interrupt will not be triggered.
When the AVR exits from an interrupt, it will always return to the main program and execute one
more instruction before any pending interrupt is served.
Note that the Status Register is not automatically stored when entering an interrupt routine, nor
restored when returning from an interrupt routine. This must be handled by software.
When using the CLI instruction to disable interrupts, the interrupts will be immediately disabled.
No interrupt will be executed after the CLI instruction, even if it occurs simultaneously with the
CLI instruction. The following example shows how this can be used to avoid interrupts during the
timed EEPROM write sequence..
Assembly Code Example

in r16, SREG ; store SREG value
cli ; disable interrupts during timed sequence
sbi EECR, EEMPE ; start EEPROM write
sbi EECR, EEPE
out SREG, r16 ; restore SREG value (I-bit)
C Code Example
char cSREG;
cSREG = SREG; /* store SREG value */
/* disable interrupts during timed sequence */
__disable_interrupt();
EECR |= (1<<EEMPE); /* start EEPROM write */
EECR |= (1<<EEPE);
SREG = cSREG; /* restore SREG value (I-bit) */
When using the SEI instruction to enable interrupts, the instruction following SEI will be exe-
cuted before any pending interrupts, as shown in this example.
16 ATmega164P/324P/644P
8011D–AVR–02/07

sei ; set Global Interrupt Enable
sleep; enter sleep, waiting for interrupt
; note: will enter sleep before any pending
; interrupt(s)
C Code Example
__enable_interrupt(); /* set Global Interrupt Enable */
__sleep(); /* enter sleep, waiting for interrupt */
/* note: will enter sleep before any pending interrupt(s) */
5.7.1 Interrupt Response Time
The interrupt execution response for all the enabled AVR interrupts is five clock cycles minimum.
After five clock cycles the program vector address for the actual interrupt handling routine is exe-
cuted. During these five clock cycle period, the Program Counter is pushed onto the Stack. The
vector is normally a jump to the interrupt routine, and this jump takes three clock cycles. If an
interrupt occurs during execution of a multi-cycle instruction, this instruction is completed before
the interrupt is served. If an interrupt occurs when the MCU is in sleep mode, the interrupt exe-
cution response time is increased by five clock cycles. This increase comes in addition to the
start-up time from the selected sleep mode.
A return from an interrupt handling routine takes five clock cycles. During these five clock cycles,
the Program Counter (three bytes) is popped back from the Stack, the Stack Pointer is incre-
mented by three, and the I-bit in SREG is set.
17
8011D–AVR–02/07
6. AVR Memories
6.1 Overview
This section describes the different memories in the ATmega164P/324P/644P. The AVR archi-
tecture has two main memory spaces, the Data Memory and the Program Memory space. In
addition, the ATmega164P/324P/644P features an EEPROM Memory for data storage. All three
memory spaces are linear and regular.
6.2 In-System Reprogrammable Flash Program Memory

The ATmega164P/324P/644P contains 16/32/64K bytes On-chip In-System Reprogrammable
Flash memory for program storage. Since all AVR instructions are 16 or 32 bits wide, the Flash
is organized as 32/64 x 16. For software security, the Flash Program memory space is divided
into two sections, Boot Program section and Application Program section.
The Flash memory has an endurance of at least 10,000 write/erase cycles. The
ATmega164P/324P/644P Program Counter (PC) is 15/16 bits wide, thus addressing the 32/64K
program memory locations. The operation of Boot Program section and associated Boot Lock
bits for software protection are described in detail in ”Memory Programming” on page 291.
”Memory Programming” on page 291 contains a detailed description on Flash data serial down-
loading using the SPI pins or the JTAG interface.
Constant tables can be allocated within the entire program memory address space (see the LPM
– Load Program Memory instruction description.
Timing diagrams for instruction fetch and execution are presented in ”Instruction Execution Tim-
ing” on page 14.
18 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 6-1. Program Memory Map
Program Memory
0x0000
Application Flash Section
Boot Flash Section

0x1FFF
6.3 SRAM Data Memory

Figure 6-2 shows how the ATmega164P/324P/644P SRAM Memory is organized.
The ATmega164P/324P/644P is a complex microcontroller with more peripheral units than can
be supported within the 64 location reserved in the Opcode for the IN and OUT instructions. For
the Extended I/O space from $060 - $FF in SRAM, only the ST/STS/STD and LD/LDS/LDD
instructions can be used.
The first 4,352 Data Memory locations address both the Register File, the I/O Memory,
Extended I/O Memory, and the internal data SRAM. The first 32 locations address the Register
file, the next 64 location the standard I/O Memory, then 160 locations of Extended I/O memory
and the next 4,096 locations address the internal data SRAM.
The five different addressing modes for the data memory cover: Direct, Indirect with Displace-
ment, Indirect, Indirect with Pre-decrement, and Indirect with Post-increment. In the Register file,
registers R26 to R31 feature the indirect addressing pointer registers.
The direct addressing reaches the entire data space.
The Indirect with Displacement mode reaches 63 address locations from the base address given
by the Y- or Z-register.
When using register indirect addressing modes with automatic pre-decrement and post-incre-
ment, the address registers X, Y, and Z are decremented or incremented.
19
8011D–AVR–02/07
The 32 general purpose working registers, 64 I/O registers, 160 Extended I/O Registers and the
1024/2048/4096 bytes of internal data SRAM in the ATmega164P/324P/644P are all accessible
through all these addressing modes. The Register File is described in ”General Purpose Regis-
ter File” on page 12.
Figure 6-2. Data Memory Map
Data Memory
32 Registers $0000 - $001F
64 I/O Registers $0020 - $005F
160 Ext I/O Reg. $0060 - $00FF
$0100
Internal SRAM
(1024/2048/4096 x 8)
$10FF
6.3.1 Data Memory Access Times
This section describes the general access timing concepts for internal memory access. The
internal data SRAM access is performed in two clkCPU cycles as described in Figure 6-3.
Figure 6-3. On-chip Data SRAM Access Cycles

T1 T2 T3
clkCPU
Address Compute Address Address valid
Data
Write
WR
Data
RD Read
Memory Access Instruction Next Instruction
20 ATmega164P/324P/644P
8011D–AVR–02/07
6.4 EEPROM Data Memory
The ATmega164P/324P/644P contains 512B/1K/2K bytes of data EEPROM memory. It is orga-
nized as a separate data space, in which single bytes can be read and written. The EEPROM
has an endurance of at least 100,000 write/erase cycles. The access between the EEPROM and
the CPU is described in the following, specifying the EEPROM Address Registers, the EEPROM
Data Register, and the EEPROM Control Register.
For a detailed description of SPI, JTAG and Parallel data downloading to the EEPROM, see
page 306, page 310, and page 295 respectively.
6.4.1 EEPROM Read/Write Access
The EEPROM Access Registers are accessible in the I/O space. See ”Register Description” on
page 23 for details.
The write access time for the EEPROM is given in Table 6-2 on page 25. A self-timing function,
however, lets the user software detect when the next byte can be written. If the user code con-
tains instructions that write the EEPROM, some precautions must be taken. In heavily filtered
power supplies, VCC is likely to rise or fall slowly on power-up/down. This causes the device for
some period of time to run at a voltage lower than specified as minimum for the clock frequency
used. See Section “6.4.2” on page 21. for details on how to avoid problems in these situations.
In order to prevent unintentional EEPROM writes, a specific write procedure must be followed.
Refer to the description of the EEPROM Control Register for details on this.
When the EEPROM is read, the CPU is halted for four clock cycles before the next instruction is
executed. When the EEPROM is written, the CPU is halted for two clock cycles before the next
instruction is executed.
6.4.2 Preventing EEPROM Corruption
During periods of low VCC, the EEPROM data can be corrupted because the supply voltage is
too low for the CPU and the EEPROM to operate properly. These issues are the same as for
board level systems using EEPROM, and the same design solutions should be applied.
An EEPROM data corruption can be caused by two situations when the voltage is too low. First,
a regular write sequence to the EEPROM requires a minimum voltage to operate correctly. Sec-
ondly, the CPU itself can execute instructions incorrectly, if the supply voltage is too low.
EEPROM data corruption can easily be avoided by following this design recommendation:
Keep the AVR RESET active (low) during periods of insufficient power supply voltage. This can
be done by enabling the internal Brown-out Detector (BOD). If the detection level of the internal
BOD does not match the needed detection level, an external low VCC reset Protection circuit can
be used. If a reset occurs while a write operation is in progress, the write operation will be com-
pleted provided that the power supply voltage is sufficient.
21
8011D–AVR–02/07
6.5 I/O Memory
The I/O space definition of the ATmega164P/324P/644P is shown in ”Register Summary” on
page 386.
All ATmega164P/324P/644P I/Os and peripherals are placed in the I/O space. All I/O locations
may be accessed by the LD/LDS/LDD and ST/STS/STD instructions, transferring data between
the 32 general purpose working registers and the I/O space. I/O Registers within the address
range 0x00 - 0x1F are directly bit-accessible using the SBI and CBI instructions. In these regis-
ters, the value of single bits can be checked by using the SBIS and SBIC instructions. Refer to
the instruction set section for more details. When using the I/O specific commands IN and OUT,
the I/O addresses 0x00 - 0x3F must be used. When addressing I/O Registers as data space
using LD and ST instructions, 0x20 must be added to these addresses. The
ATmega164P/324P/644P is a complex microcontroller with more peripheral units than can be
supported within the 64 location reserved in Opcode for the IN and OUT instructions. For the
Extended I/O space from 0x60 - 0xFF in SRAM, only the ST/STS/STD and LD/LDS/LDD instruc-
tions can be used.
For compatibility with future devices, reserved bits should be written to zero if accessed.
Reserved I/O memory addresses should never be written.
Some of the Status Flags are cleared by writing a logical one to them. Note that, unlike most
other AVRs, the CBI and SBI instructions will only operate on the specified bit, and can therefore
be used on registers containing such Status Flags. The CBI and SBI instructions work with reg-
isters 0x00 to 0x1F only.
The I/O and peripherals control registers are explained in later sections.
The ATmega164P/324P/644P contains three General Purpose I/O Registers, see ”Register
Description” on page 23. These registers can be used for storing any information, and they are
particularly useful for storing global variables and Status Flags. General Purpose I/O Registers
within the address range 0x00 - 0x1F are directly bit-accessible using the SBI, CBI, SBIS, and
SBIC instructions.
22 ATmega164P/324P/644P
8011D–AVR–02/07
6.6 Register Description
6.6.1 EEARH and EEARL – The EEPROM Address Register
Bit 15 14 13 12 11 10 9 8
0x22 (0x42) – – – – EEAR11 EEAR10 EEAR9 EEAR8 EEARH
0x21 (0x41) EEAR7 EEAR6 EEAR5 EEAR4 EEAR3 EEAR2 EEAR1 EEAR0 EEARL
7 6 5 4 3 2 1 0
Read/Write R R R R R/W R/W R/W R/W
Initial Value 0 0 0 0 X X X X
X X X X X X X X
• Bits 15:12 – Res: Reserved Bits

These bits are reserved bits in the ATmega164P/324P/644P and will always read as zero.
• Bits 11:0 – EEAR8:0: EEPROM Address

The EEPROM Address Registers – EEARH and EEARL specify the EEPROM address in the 4K
bytes EEPROM space. The EEPROM data bytes are addressed linearly between 0 and 4096.
The initial value of EEAR is undefined. A proper value must be written before the EEPROM may
be accessed.
6.6.2 EEDR – The EEPROM Data Register
Bit 7 6 5 4 3 2 1 0
0x20 (0x40) MSB LSB EEDR
• Bits 7:0 – EEDR7:0: EEPROM Data

For the EEPROM write operation, the EEDR Register contains the data to be written to the
EEPROM in the address given by the EEAR Register. For the EEPROM read operation, the
EEDR contains the data read out from the EEPROM at the address given by EEAR.
6.6.3 EECR – The EEPROM Control Register
Bit 7 6 5 4 3 2 1 0
0x1F (0x3F) – – EEPM1 EEPM0 EERIE EEMPE EEPE EERE EECR
Read/Write R R R/W R/W R/W R/W R/W R/W
Initial Value 0 0 X X 0 0 X 0

• Bits 5:4 – EEPM1 and EEPM0: EEPROM Programming Mode Bits

The EEPROM Programming mode bit setting defines which programming action that will be trig-
gered when writing EEPE. It is possible to program data in one atomic operation (erase the old
value and program the new value) or to split the Erase and Write operations in two different
operations. The Programming times for the different modes are shown in Table 6-1 on page 24.
23
8011D–AVR–02/07
While EEPE is set, any write to EEPMn will be ignored. During reset, the EEPMn bits will be
reset to 0b00 unless the EEPROM is busy programming.
Table 6-1. EEPROM Mode Bits

Programming
EEPM1 EEPM0 Time Operation
0 0 3.4 ms Erase and Write in one operation (Atomic Operation)
0 1 1.8 ms Erase Only
1 0 1.8 ms Write Only
1 1 – Reserved for future use
• Bit 3 – EERIE: EEPROM Ready Interrupt Enable

Writing EERIE to one enables the EEPROM Ready Interrupt if the I bit in SREG is set. Writing
EERIE to zero disables the interrupt. The EEPROM Ready interrupt generates a constant inter-
rupt when EEPE is cleared.
• Bit 2 – EEMPE: EEPROM Master Programming Enable

The EEMPE bit determines whether setting EEPE to one causes the EEPROM to be written.
When EEMPE is set, setting EEPE within four clock cycles will write data to the EEPROM at the
selected address If EEMPE is zero, setting EEPE will have no effect. When EEMPE has been
written to one by software, hardware clears the bit to zero after four clock cycles. See the
description of the EEPE bit for an EEPROM write procedure.
• Bit 1 – EEPE: EEPROM Programming Enable

The EEPROM Write Enable Signal EEPE is the write strobe to the EEPROM. When address
and data are correctly set up, the EEPE bit must be written to one to write the value into the
EEPROM. The EEMPE bit must be written to one before a logical one is written to EEPE, other-
wise no EEPROM write takes place. The following procedure should be followed when writing
the EEPROM (the order of steps 3 and 4 is not essential):
1. Wait until EEPE becomes zero.
2. Wait until SELFPRGEN in SPMCSR becomes zero.
3. Write new EEPROM address to EEAR (optional).
4. Write new EEPROM data to EEDR (optional).
5. Write a logical one to the EEMPE bit while writing a zero to EEPE in EECR.
6. Within four clock cycles after setting EEMPE, write a logical one to EEPE.
The EEPROM can not be programmed during a CPU write to the Flash memory. The software
must check that the Flash programming is completed before initiating a new EEPROM write.
Step 2 is only relevant if the software contains a Boot Loader allowing the CPU to program the
Flash. If the Flash is never being updated by the CPU, step 2 can be omitted. See ”Memory Pro-
gramming” on page 291 for details about Boot programming.
Caution: An interrupt between step 5 and step 6 will make the write cycle fail, since the
EEPROM Master Write Enable will time-out. If an interrupt routine accessing the EEPROM is
interrupting another EEPROM access, the EEAR or EEDR Register will be modified, causing the
interrupted EEPROM access to fail. It is recommended to have the Global Interrupt Flag cleared
during all the steps to avoid these problems.
24 ATmega164P/324P/644P
8011D–AVR–02/07
When the write access time has elapsed, the EEPE bit is cleared by hardware. The user soft-
ware can poll this bit and wait for a zero before writing the next byte. When EEPE has been set,
the CPU is halted for two cycles before the next instruction is executed.
• Bit 0 – EERE: EEPROM Read Enable

The EEPROM Read Enable Signal EERE is the read strobe to the EEPROM. When the correct
address is set up in the EEAR Register, the EERE bit must be written to a logic one to trigger the
EEPROM read. The EEPROM read access takes one instruction, and the requested data is
available immediately. When the EEPROM is read, the CPU is halted for four cycles before the
next instruction is executed.
The user should poll the EEPE bit before starting the read operation. If a write operation is in
progress, it is neither possible to read the EEPROM, nor to change the EEAR Register.
The calibrated Oscillator is used to time the EEPROM accesses. Table 6-2 on page 25 lists the
typical programming time for EEPROM access from the CPU.
Table 6-2. EEPROM Programming Time

Symbol Number of Calibrated RC Oscillator Cycles Typ Programming Time
EEPROM write
26,368 3.3 ms
(from CPU)
25
8011D–AVR–02/07
The following code examples show one assembly and one C function for writing to the
EEPROM. The examples assume that interrupts are controlled (e.g. by disabling interrupts glo-
bally) so that no interrupts will occur during execution of these functions. The examples also
assume that no Flash Boot Loader is present in the software. If such code is present, the
EEPROM write function must also wait for any ongoing SPM command to finish.
Assembly Code Example()

EEPROM_write:
; Wait for completion of previous write
sbic EECR,EEPE
rjmp EEPROM_write
; Set up address (r18:r17) in address register
out EEARH, r18
out EEARL, r17
; Write data (r16) to Data Register
out EEDR,r16
; Write logical one to EEMPE
sbi EECR,EEMPE
; Start eeprom write by setting EEPE
sbi EECR,EEPE
ret
C Code Example(1)
void EEPROM_write(unsigned int uiAddress, unsigned char ucData)
{
/* Wait for completion of previous write */
while(EECR & (1<<EEPE))
;
/* Set up address and Data Registers */
EEAR = uiAddress;
EEDR = ucData;
/* Write logical one to EEMPE */
EECR |= (1<<EEMPE);
/* Start eeprom write by setting EEPE */
EECR |= (1<<EEPE);
}
Note: 1. See “About Code Examples” on page 8.
26 ATmega164P/324P/644P
8011D–AVR–02/07
The next code examples show assembly and C functions for reading the EEPROM. The exam-
ples assume that interrupts are controlled so that no interrupts will occur during execution of
these functions.
Assembly Code Example(1)

EEPROM_read:
; Wait for completion of previous write
sbic EECR,EEPE
rjmp EEPROM_read
; Set up address (r18:r17) in address register
out EEARH, r18
out EEARL, r17
; Start eeprom read by writing EERE
sbi EECR,EERE
; Read data from Data Register
in r16,EEDR
ret
C Code Example(1)
unsigned char EEPROM_read(unsigned int uiAddress)
{
/* Wait for completion of previous write */
while(EECR & (1<<EEPE))
;
/* Set up address register */
EEAR = uiAddress;
/* Start eeprom read by writing EERE */
EECR |= (1<<EERE);
/* Return data from Data Register */
return EEDR;
}
27
8011D–AVR–02/07
6.6.4 GPIOR2 – General Purpose I/O Register 2
Bit 7 6 5 4 3 2 1 0
0x2B (0x4B) MSB LSB GPIOR2
Bit 7 6 5 4 3 2 1 0
0x2A (0x4A) MSB LSB GPIOR1
Bit 7 6 5 4 3 2 1 0
0x1E (0x3E) MSB LSB GPIOR0
Note: 1. SRWn1 = SRW11 (upper sector) or SRW01 (lower sector), SRWn0 = SRW10 (upper sector) or
SRW00 (lower sector). The ALE pulse in period T4 is only present if the next instruction
accesses the RAM (internal or external).
28 ATmega164P/324P/644P
8011D–AVR–02/07
7. System Clock and Clock Options

7.1 Clock Systems and their Distribution
Figure 7-1 presents the principal clock systems in the AVR and their distribution. All of the clocks
need not be active at a given time. In order to reduce power consumption, the clocks to modules
not being used can be halted by using different sleep modes, as described in ”Power Manage-
ment and Sleep Modes” on page 42. The clock systems are detailed below.
Figure 7-1. Clock Distribution

Asynchronous General I/O Flash and
ADC CPU Core RAM
Timer/Counter Modules EEPROM
clkADC
clkI/O AVR Clock clkCPU

Control Unit
clkASY clkFLASH
Reset Logic Watchdog Timer
Source clock Watchdog clock
System Clock Watchdog

Prescaler Oscillator
Clock
Multiplexer
Timer/Counter Crystal Low-frequency Calibrated RC

External Clock
Oscillator Oscillator Crystal Oscillator Oscillator
7.1.1 CPU Clock – clkCPU
The CPU clock is routed to parts of the system concerned with operation of the AVR core.
Examples of such modules are the General Purpose Register File, the Status Register and the
data memory holding the Stack Pointer. Halting the CPU clock inhibits the core from performing
general operations and calculations.
7.1.2 I/O Clock – clkI/O
The I/O clock is used by the majority of the I/O modules, like Timer/Counters, SPI, and USART.
The I/O clock is also used by the External Interrupt module, but note that some external inter-
rupts are detected by asynchronous logic, allowing such interrupts to be detected even if the I/O
clock is halted. Also note that start condition detection in the USI module is carried out asynchro-
nously when clkI/O is halted, TWI address recognition in all sleep modes.
29
8011D–AVR–02/07
7.1.3 Flash Clock – clkFLASH
The Flash clock controls operation of the Flash interface. The Flash clock is usually active simul-
taneously with the CPU clock.
7.1.4 Asynchronous Timer Clock – clkASY
The Asynchronous Timer clock allows the Asynchronous Timer/Counter to be clocked directly
from an external clock or an external 32 kHz clock crystal. The dedicated clock domain allows
using this Timer/Counter as a real-time counter even when the device is in sleep mode.
7.1.5 ADC Clock – clkADC
The ADC is provided with a dedicated clock domain. This allows halting the CPU and I/O clocks
in order to reduce noise generated by digital circuitry. This gives more accurate ADC conversion
results.
7.2 Clock Sources

The device has the following clock source options, selectable by Flash Fuse bits as shown
below. The clock from the selected source is input to the AVR clock generator, and routed to the
appropriate modules.
Table 7-1. Device Clocking Options Select(1)

Device Clocking Option CKSEL3..0
Low Power Crystal Oscillator 1111 - 1000
Full Swing Crystal Oscillator 0111 - 0110
Low Frequency Crystal Oscillator 0101 - 0100
Internal 128 kHz RC Oscillator 0011
Calibrated Internal RC Oscillator 0010
External Clock 0000
Reserved 0001
Note: 1. For all fuses “1” means unprogrammed while “0” means programmed.
7.2.1 Default Clock Source
The device is shipped with internal RC oscillator at 8.0MHz and with the fuse CKDIV8 pro-
grammed, resulting in 1.0MHz system clock. The startup time is set to maximum and time-out
period enabled. (CKSEL = "0010", SUT = "10", CKDIV8 = "0"). The default setting ensures that
all users can make their desired clock source setting using any available programming interface.
7.2.2 Clock Startup Sequence
Any clock source needs a sufficient VCC to start oscillating and a minimum number of oscillating
cycles before it can be considered stable.
To ensure sufficient VCC, the device issues an internal reset with a time-out delay (tTOUT) after
the device reset is released by all other reset sources. ”On-chip Debug System” on page 46
describes the start conditions for the internal reset. The delay (tTOUT) is timed from the Watchdog
Oscillator and the number of cycles in the delay is set by the SUTx and CKSELx fuse bits. The
30 ATmega164P/324P/644P
8011D–AVR–02/07
selectable delays are shown in Table 7-2. The frequency of the Watchdog Oscillator is voltage
dependent as shown in ”Typical Characteristics” on page 336.
Table 7-2. Number of Watchdog Oscillator Cycles

Typ Time-out (VCC = 5.0V) Typ Time-out (VCC = 3.0V) Number of Cycles
0 ms 0 ms 0
4.1 ms 4.3 ms 512
65 ms 69 ms 8K (8,192)
Main purpose of the delay is to keep the AVR in reset until it is supplied with minimum Vcc. The
delay will not monitor the actual voltage and it will be required to select a delay longer than the
Vcc rise time. If this is not possible, an internal or external Brown-Out Detection circuit should be
used. A BOD circuit will ensure sufficient Vcc before it releases the reset, and the time-out delay
can be disabled. Disabling the time-out delay without utilizing a Brown-Out Detection circuit is
not recommended.
The oscillator is required to oscillate for a minimum number of cycles before the clock is consid-
ered stable. An internal ripple counter monitors the oscillator output clock, and keeps the internal
reset active for a given number of clock cycles. The reset is then released and the device will
start to execute. The recommended oscillator start-up time is dependent on the clock type, and
varies from 6 cycles for an externally applied clock to 32K cycles for a low frequency crystal.
The start-up sequence for the clock includes both the time-out delay and the start-up time when
the device starts up from reset. When starting up from Power-save or Power-down mode, Vcc is
assumed to be at a sufficient level and only the start-up time is included.
7.2.3 Clock Source Connections
The pins XTAL1 and XTAL2 are input and output, respectively, of an inverting amplifier which
can be configured for use as an On-chip Oscillator, as shown in Figure 7-2 on page 31. Either a
quartz crystal or a ceramic resonator may be used.
C1 and C2 should always be equal for both crystals and resonators. The optimal value of the
capacitors depends on the crystal or resonator in use, the amount of stray capacitance, and the
electromagnetic noise of the environment. For ceramic resonators, the capacitor values given by
the manufacturer should be used.
Figure 7-2. Crystal Oscillator Connections
C2
XTAL2
C1
XTAL1
GND
31
8011D–AVR–02/07
7.3 Low Power Crystal Oscillator
This Crystal Oscillator is a low power oscillator, with reduced voltage swing on the XTAL2 out-
put. It gives the lowest power consumption, but is not capable of driving other clock inputs, and
may be more susceptible to noise in noisy environments. In these cases, refer to the ”Full Swing
Crystal Oscillator” on page 33.
Some initial guidelines for choosing capacitors for use with crystals are given in Table 7-3. The
crystal should be connected as described in ”Clock Source Connections” on page 31.
The Low Power Oscillator can operate in three different modes, each optimized for a specific fre-
quency range. The operating mode is selected by the fuses CKSEL3..1 as shown in Table 7-3.
Table 7-3. Low Power Crystal Oscillator Operating Modes(3)

Recommended Range for Capacitors C1
Frequency Range(1) (MHz) CKSEL3..1 and C2 (pF)
0.4 - 0.9 100(2) –
0.9 - 3.0 101 12 - 22
3.0 - 8.0 110 12 - 22
8.0 - 16.0 111 12 - 22
Notes: 1. The frequency ranges are preliminary values. Actual values are TBD.
2. This option should not be used with crystals, only with ceramic resonators.
3. If 8 MHz frequency exceeds the specification of the device (depends on VCC), the CKDIV8
Fuse can be programmed in order to divide the internal frequency by 8. It must be ensured
that the resulting divided clock meets the frequency specification of the device.
The CKSEL0 Fuse together with the SUT1..0 Fuses select the start-up times as shown in Table
7-4.
Table 7-4. Start-up Times for the Low Power Crystal Oscillator Clock Selection
Start-up Time from Additional Delay
Oscillator Source / Power-down and from Reset
Power Conditions Power-save (VCC = 5.0V) CKSEL0 SUT1..0
Ceramic resonator, fast
258 CK 14CK + 4.1 ms(1) 0 00
rising power
Ceramic resonator, slowly
258 CK 14CK + 65 ms(1) 0 01
rising power
Ceramic resonator, BOD
1K CK 14CK(2) 0 10
enabled
1K CK 14CK + 4.1 ms(2) 0 11
rising power
1K CK 14CK + 65 ms(2) 1 00
rising power
32 ATmega164P/324P/644P
8011D–AVR–02/07
Table 7-4. Start-up Times for the Low Power Crystal Oscillator Clock Selection (Continued)
Crystal Oscillator, BOD
16K CK 14CK 1 01
enabled
Crystal Oscillator, fast
16K CK 14CK + 4.1 ms 1 10
rising power
Crystal Oscillator, slowly
16K CK 14CK + 65 ms 1 11
rising power
Notes: 1. These options should only be used when not operating close to the maximum frequency of the
device, and only if frequency stability at start-up is not important for the application. These
options are not suitable for crystals.
2. These options are intended for use with ceramic resonators and will ensure frequency stability
at start-up. They can also be used with crystals when not operating close to the maximum fre-
quency of the device, and if frequency stability at start-up is not important for the application.
7.4 Full Swing Crystal Oscillator

This Crystal Oscillator is a full swing oscillator, with rail-to-rail swing on the XTAL2 output. This is
useful for driving other clock inputs and in noisy environments. The current consumption is
higher than the ”Low Power Crystal Oscillator” on page 32. Note that the Full Swing Crystal
Oscillator will only operate for Vcc = 2.7 - 5.5 volts.
Some initial guidelines for choosing capacitors for use with crystals are given in Table 7-6. The
crystal should be connected as described in ”Clock Source Connections” on page 31.
The operating mode is selected by the fuses CKSEL3..1 as shown in Table 7-5.
Table 7-5. Full Swing Crystal Oscillator operating modes(2)

Recommended Range for Capacitors C1
Frequency Range(1) (MHz) CKSEL3..1 and C2 (pF)
0.4 - 20 011 12 - 22
Notes: 1. The frequency ranges are preliminary values. Actual values are TBD.
Fuse can be programmed in order to divide the internal frequency by 8. It must be ensured
that the resulting divided clock meets the frequency specification of the device.
Table 7-6. Start-up Times for the Full Swing Crystal Oscillator Clock Selection
258 CK 14CK + 4.1 ms(1) 0 00
rising power
258 CK 14CK + 65 ms(1) 0 01
rising power
Ceramic resonator, BOD
1K CK 14CK(2) 0 10
enabled
33
8011D–AVR–02/07
Table 7-6. Start-up Times for the Full Swing Crystal Oscillator Clock Selection
1K CK 14CK + 4.1 ms(2) 0 11
rising power
1K CK 14CK + 65 ms(2) 1 00
rising power
Crystal Oscillator, BOD
16K CK 14CK 1 01
enabled
Crystal Oscillator, fast
16K CK 14CK + 4.1 ms 1 10
rising power
Crystal Oscillator, slowly
16K CK 14CK + 65 ms 1 11
rising power
Notes: 1. These options should only be used when not operating close to the maximum frequency of the
device, and only if frequency stability at start-up is not important for the application. These
options are not suitable for crystals.
2. These options are intended for use with ceramic resonators and will ensure frequency stability
at start-up. They can also be used with crystals when not operating close to the maximum fre-
quency of the device, and if frequency stability at start-up is not important for the application.
7.5 Low Frequency Crystal Oscillator

The Low-frequency Crystal Oscillator is optimized for use with a 32.768 kHz watch crystal.
When selecting crystals, load capasitance and crystal’s Equivalent Series Resistance, ESR
must be taken into consideration. Both values are specified by the crystal vendor.
ATmega164P/324P/644P oscillator is optimized for very low power consumption, and thus when
selecting crystals, see Table 7-7 on page 34 for maximum ESR recommendations on 9 pF and
12.5 pF crystals
Table 7-7. Maximum ESR Recommendation for 32.768 kHz Watch Crystal
Crystal CL (pF) Max ESR [kΩ](1)

9.0 65
12.5 30
Note: 1. Maximum ESR is typical value based on characterization

The Low-frequency Crystal Oscillator provides an internal load capacitance of typical 8.0 pF.
Crystals with recommended 8.0 pF load capacitance can be without external capacitors as
shown in Figure 7-3 on page 35.
34 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 7-3. Crystal Oscillator Connections
TOSC2
TOSC1
Table 7-8. Low-frequency Crystal Oscillator Internal load Capacitance
Min. (pF) Typ. (pF) Max. (pF)

TBD 8.0 TBD
Crystals specifying load capacitance (CL) higher than 8.0 pF, require external capacitors applied
as described in Figure 7-2 on page 31.
To find suitable load capacitance for a 32.768 kHz crysal, please consult the crystal datasheet.
When this oscillator is selected, start-up times are determined by the SUT Fuses and CKSEL0
as shown in Table 7-9.
Table 7-9. Start-up Times for the Low Frequency Crystal Oscillator Clock Selection
Power-down and from Reset
(1)
BOD enabled 1K CK 14CK 0 00
(1)
Fast rising power 1K CK 14CK + 4.1 ms 0 01
Slowly rising power 1K CK 14CK + 65 ms(1) 0 10
Reserved 0 11
BOD enabled 32K CK 14CK 1 00
Fast rising power 32K CK 14CK + 4.1 ms 1 01
Slowly rising power 32K CK 14CK + 65 ms 1 10
Reserved 1 11
Note: 1. These options should only be used if frequency stability at start-up is not important for the
application.
35
8011D–AVR–02/07
7.6 Calibrated Internal RC Oscillator
By default, the Internal RC Oscillator provides an approximate 8 MHz clock. Though voltage and
temperature dependent, this clock can be very accurately calibrated by the the user. See Table
26-3 on page 328 and ”Internal Oscillator Speed” on page 354 and page 378 for more details.
The device is shipped with the CKDIV8 Fuse programmed. See ”System Clock Prescaler” on
page 38 for more details.
This clock may be selected as the system clock by programming the CKSEL Fuses as shown in
Table 7-10. If selected, it will operate with no external components. During reset, hardware loads
the pre-programmed calibration value into the OSCCAL Register and thereby automatically cal-
ibrates the RC Oscillator. The accuracy of this calibration is shown as Factory calibration in
Table 26-3 on page 328.
By changing the OSCCAL register from SW, see ”OSCCAL – Oscillator Calibration Register” on
page 40, it is possible to get a higher calibration accuracy than by using the factory calibration.
The accuracy of this calibration is shown as User calibration in Table 26-3 on page 328.
When this Oscillator is used as the chip clock, the Watchdog Oscillator will still be used for the
Watchdog Timer and for the Reset Time-out. For more information on the pre-programmed cali-
bration value, see the section ”Calibration Byte” on page 294.
Table 7-10. Internal Calibrated RC Oscillator Operating Modes(1)(3)

Frequency Range(2) (MHz) CKSEL3..0
7.3 - 8.1 0010
Notes: 1. The device is shipped with this option selected.

2. The frequency ranges are preliminary values. Actual values are TBD.
Fuse can be programmed in order to divide the internal frequency by 8.
When this Oscillator is selected, start-up times are determined by the SUT Fuses as shown in
Table 7-11. Start-up times for the Internal Calibrated RC Oscillator clock selection
Start-up Time from Power- Additional Delay from
Power Conditions down and Power-save Reset (VCC = 5.0V) SUT1..0
BOD enabled 6 CK 14CK 00
Fast rising power 6 CK 14CK + 4.1 ms 01
Slowly rising power 6 CK 14CK + 65 ms(1) 10
Reserved 11
Note: 1. The device is shipped with this option selected.
36 ATmega164P/324P/644P
8011D–AVR–02/07
7.7 128 kHz Internal Oscillator
The 128 kHz internal Oscillator is a low power Oscillator providing a clock of 128 kHz. The fre-
quency is nominal at 3V and 25°C. This clock may be select as the system clock by
programming the CKSEL Fuses to “0011” as shown in Table 7-12.
Table 7-12. 128 kHz Internal Oscillator Operating Modes

Nominal Frequency CKSEL3..0
128 kHz 0011
Note: 1. The frequency is preliminary value. Actual value is TBD.

When this clock source is selected, start-up times are determined by the SUT Fuses as shown in
Table 7-13.
Table 7-13. Start-up Times for the 128 kHz Internal Oscillator
Power Conditions down and Power-save Reset SUT1..0
Fast rising power 6 CK 14CK + 4 ms 01
Slowly rising power 6 CK 14CK + 64 ms 10
Reserved 11
7.8 External Clock

To drive the device from an external clock source, XTAL1 should be driven as shown in Figure
7-4. To run the device on an external clock, the CKSEL Fuses must be programmed to “0000”.
Figure 7-4. External Clock Drive Configuration
NC XTAL2
EXTERNAL
CLOCK XTAL1
SIGNAL
GND
When this clock source is selected, start-up times are determined by the SUT Fuses as shown in
Table 7-15.
Table 7-14. Crystal Oscillator Clock Frequency

Nominal Frequency CKSEL3..0
0 - 20 kHz 0000
37
8011D–AVR–02/07
Table 7-15. Start-up Times for the External Clock Selection
Power Conditions down and Power-save Reset (VCC = 5.0V) SUT1..0
Fast rising power 6 CK 14CK + 4.1 ms 01
Slowly rising power 6 CK 14CK + 65 ms 10
Reserved 11
When applying an external clock, it is required to avoid sudden changes in the applied clock fre-
quency to ensure stable operation of the MCU. A variation in frequency of more than 2% from
one clock cycle to the next can lead to unpredictable behavior. If changes of more than 2% is
required, ensure that the MCU is kept in Reset during the changes.
Note that the System Clock Prescaler can be used to implement run-time changes of the internal
clock frequency while still ensuring stable operation. Refer to ”System Clock Prescaler” on page
38 for details.
7.9 Timer/Counter Oscillator

ATmega164P/324P/644P uses the same type of crystal oscillator for Low-frequency Crystal
Oscillator and Timer/Counter Oscillator. See ”Low Frequency Crystal Oscillator” on page 34 for
details on the oscillator and crystal requirements.
The device can operate its Timer/Counter2 from an external 32.768 kHz watch crystal or a exter-
nal clock source. See ”Clock Source Connections” on page 31 for details.
Applying an external clock source to TOSC1 can be done if EXTCLK in the ASSR Register is
written to logic one. See ”The Output Compare Register B contains an 8-bit value that is contin-
uously compared with the counter value (TCNT2). A match can be used to generate an Output
Compare interrupt, or to generate a waveform output on the OC2B pin.” on page 156 for further
description on selecting external clock as input instead of a 32.768 kHz watch crystal.
7.10 Clock Output Buffer

The device can output the system clock on the CLKO pin. To enable the output, the CKOUT
Fuse has to be programmed. This mode is suitable when the chip clock is used to drive other cir-
cuits on the system. The clock also will be output during reset, and the normal operation of I/O
pin will be overridden when the fuse is programmed. Any clock source, including the internal RC
Oscillator, can be selected when the clock is output on CLKO. If the System Clock Prescaler is
used, it is the divided system clock that is output.
7.11 System Clock Prescaler

The ATmega164P/324P/644P has a system clock prescaler, and the system clock can be
divided by setting the ”CLKPR – Clock Prescale Register” on page 40. This feature can be used
to decrease the system clock frequency and the power consumption when the requirement for
processing power is low. This can be used with all clock source options, and it will affect the
clock frequency of the CPU and all synchronous peripherals. clkI/O, clkADC, clkCPU, and clkFLASH
are divided by a factor as shown in Table 7-16 on page 41.
When switching between prescaler settings, the System Clock Prescaler ensures that no
glitches occurs in the clock system. It also ensures that no intermediate frequency is higher than
38 ATmega164P/324P/644P
8011D–AVR–02/07
neither the clock frequency corresponding to the previous setting, nor the clock frequency corre-
sponding to the new setting.
The ripple counter that implements the prescaler runs at the frequency of the undivided clock,
which may be faster than the CPU's clock frequency. Hence, it is not possible to determine the
state of the prescaler - even if it were readable, and the exact time it takes to switch from one
clock division to the other cannot be exactly predicted. From the time the CLKPS values are writ-
ten, it takes between T1 + T2 and T1 + 2 * T2 before the new clock frequency is active. In this
interval, 2 active clock edges are produced. Here, T1 is the previous clock period, and T2 is the
period corresponding to the new prescaler setting.
To avoid unintentional changes of clock frequency, a special write procedure must be followed
to change the CLKPS bits:
1. Write the Clock Prescaler Change Enable (CLKPCE) bit to one and all other bits in
CLKPR to zero.
2. Within four cycles, write the desired value to CLKPS while writing a zero to CLKPCE.
Interrupts must be disabled when changing prescaler setting to make sure the write procedure is
not interrupted.
39
8011D–AVR–02/07
7.12.1 OSCCAL – Oscillator Calibration Register
Bit 7 6 5 4 3 2 1 0
(0x66) CAL7 CAL6 CAL5 CAL4 CAL3 CAL2 CAL1 CAL0 OSCCAL
Initial Value Device Specific Calibration Value
• Bits 7:0 – CAL7:0: Oscillator Calibration Value

The Oscillator Calibration Register is used to trim the Calibrated Internal RC Oscillator to
remove process variations from the oscillator frequency. A pre-programmed calibration value is
automatically written to this register during chip reset, giving the Factory calibrated frequency as
specified in Table 26-3 on page 328. The application software can write this register to change
the oscillator frequency. The oscillator can be calibrated to frequencies as specified in Table 26-
3 on page 328. Calibration outside that range is not guaranteed.
Note that this oscillator is used to time EEPROM and Flash write accesses, and these write
times will be affected accordingly. If the EEPROM or Flash are written, do not calibrate to more
than 8.8 MHz. Otherwise, the EEPROM or Flash write may fail.
The CAL7 bit determines the range of operation for the oscillator. Setting this bit to 0 gives the
lowest frequency range, setting this bit to 1 gives the highest frequency range. The two fre-
quency ranges are overlapping, in other words a setting of OSCCAL = 0x7F gives a higher
frequency than OSCCAL = 0x80.
The CAL6..0 bits are used to tune the frequency within the selected range. A setting of 0x00
gives the lowest frequency in that range, and a setting of 0x7F gives the highest frequency in the
range.
7.12.2 CLKPR – Clock Prescale Register
Bit 7 6 5 4 3 2 1 0
(0x61) CLKPCE – – – CLKPS3 CLKPS2 CLKPS1 CLKPS0 CLKPR
Read/Write R/W R R R R/W R/W R/W R/W
Initial Value 0 0 0 0 See Bit Description
• Bit 7 – CLKPCE: Clock Prescaler Change Enable

The CLKPCE bit must be written to logic one to enable change of the CLKPS bits. The CLKPCE
bit is only updated when the other bits in CLKPR are simultaneously written to zero. CLKPCE is
cleared by hardware four cycles after it is written or when CLKPS bits are written. Rewriting the
CLKPCE bit within this time-out period does neither extend the time-out period, nor clear the
CLKPCE bit.
• Bits 3:0 – CLKPS3:0: Clock Prescaler Select Bits 3 - 0

These bits define the division factor between the selected clock source and the internal system
clock. These bits can be written run-time to vary the clock frequency to suit the application
requirements. As the divider divides the master clock input to the MCU, the speed of all synchro-
nous peripherals is reduced when a division factor is used. The division factors are given in
40 ATmega164P/324P/644P
8011D–AVR–02/07
The CKDIV8 Fuse determines the initial value of the CLKPS bits. If CKDIV8 is unprogrammed,
the CLKPS bits will be reset to “0000”. If CKDIV8 is programmed, CLKPS bits are reset to
“0011”, giving a division factor of 8 at start up. This feature should be used if the selected clock
source has a higher frequency than the maximum frequency of the device at the present operat-
ing conditions. Note that any value can be written to the CLKPS bits regardless of the CKDIV8
Fuse setting. The Application software must ensure that a sufficient division factor is chosen if
the selected clock source has a higher frequency than the maximum frequency of the device at
the present operating conditions. The device is shipped with the CKDIV8 Fuse programmed.
Table 7-16. Clock Prescaler Select

CLKPS3 CLKPS2 CLKPS1 CLKPS0 Clock Division Factor
0 0 0 0 1
0 0 0 1 2
0 0 1 0 4
0 0 1 1 8
0 1 0 0 16
0 1 0 1 32
0 1 1 0 64
0 1 1 1 128
1 0 0 0 256
1 0 0 1 Reserved
1 0 1 0 Reserved
1 0 1 1 Reserved
1 1 0 0 Reserved
1 1 0 1 Reserved
1 1 1 0 Reserved
1 1 1 1 Reserved
41
8011D–AVR–02/07
8. Power Management and Sleep Modes
8.1 Overview
Sleep modes enable the application to shut down unused modules in the MCU, thereby saving-
power. The AVR provides various sleep modes allowing the user to tailor the power
consumption to the application’s requirements.
When enabled, the Brown-out Detector (BOD) actively monitors the power supply voltage during
the sleep periods. To further save power, it is possible to disable the BOD in some sleep modes.
See ”BOD Disable” on page 43 for more details.
8.2 Sleep Modes

Figure 7-1 on page 29 presents the different clock systems in the ATmega164P/324P/644P, and
their distribution. The figure is helpful in selecting an appropriate sleep mode. Table 8-1 shows
the different sleep modes, their wake up sources and BOD disable ability.
Table 8-1. Active Clock Domains and Wake-up Sources in the Different Sleep Modes.
Active Clock Domains Oscillators Wake-up Sources
EEPROM Ready
WDT Interrupt
BOD Disdable
TWI Address
Pin Change
Main Clock
INT2:0 and
Timer Osc
Other I/O
Software
Enabled
Enabled
clkFLASH
Source
Timer2
clkADC
Match
clkCPU
clkASY
SPM/
ADC
clkIO
Sleep Mode
Idle X X X X X(2) X X X X X X X
(2) (3) (2)
ADCNRM X X X X X X X X X X
(3)
Power-down X X X X
Power-save X X(2) X(3) X X X X
(1) (3)
Standby X X X X X
Extended
X(2) X X(2) X(3) X X X X
Standby
Notes: 1. Only recommended with external crystal or resonator selected as clock source.
2. If either LCD controller or Timer/Counter2 is running in asynchronous mode.
3. For INT0, only level interrupt.
To enter any of the sleep modes, the SE bit in SMCR must be written to logic one and a SLEEP
instruction must be executed. The SM2, SM1, and SM0 bits in the SMCR Register select which
sleep mode will be activated by the SLEEP instruction. See Table 8-2 on page 47 for a
summary.
If an enabled interrupt occurs while the MCU is in a sleep mode, the MCU wakes up. The MCU
is then halted for four cycles in addition to the start-up time, executes the interrupt routine, and
resumes execution from the instruction following SLEEP. The contents of the Register File and
SRAM are unaltered when the device wakes up from sleep. If a reset occurs during sleep mode,
the MCU wakes up and executes from the Reset Vector.
42 ATmega164P/324P/644P
8011D–AVR–02/07
8.3 BOD Disable
When the Brown-out Detector (BOD) is enabled by BODLEVEL fuses, Table 25-3 on page 292,
the BOD is actively monitoring the power supply voltage during a sleep period. To save power, it
is possible to disable the BOD by software for some of the sleep modes, see Table 8-1 on page
42. The sleep mode power consumption will then be at the same level as when BOD is globally
disabled by fuses. If BOD is disabled in software, the BOD function is turned off immediately
after entering the sleep mode. Upon wake-up from sleep, BOD is automatically enabled again.
This ensures safe operation in case the VCC level has dropped during the sleep period.
When the BOD has been disabled, the wake-up time from sleep mode will be approximately 60
µs to ensure that the BOD is working correctly before the MCU continues executing code.
BOD disable is controlled by bit 6, BODS (BOD Sleep) in the control register MCUCR, see
”MCUCR – MCU Control Register” on page 48. Writing this bit to one turns off the BOD in rele-
vant sleep modes, while a zero in this bit keeps BOD active. Default setting keeps BOD active,
i.e. BODS set to zero.
Writing to the BODS bit is controlled by a timed sequence and an enable bit, see ”MCUCR –
MCU Control Register” on page 48.
8.4 Idle Mode

When the SM2..0 bits are written to 000, the SLEEP instruction makes the MCU enter Idle
mode, stopping the CPU but allowing the SPI, USART, Analog Comparator, ADC, 2-wire Serial
Interface, Timer/Counters, Watchdog, and the interrupt system to continue operating. This sleep
mode basically halts clkCPU and clkFLASH, while allowing the other clocks to run.
Idle mode enables the MCU to wake up from external triggered interrupts as well as internal
ones like the Timer Overflow and USART Transmit Complete interrupts. If wake-up from the
Analog Comparator interrupt is not required, the Analog Comparator can be powered down by
setting the ACD bit in the Analog Comparator Control and Status Register – ACSR. This will
reduce power consumption in Idle mode. If the ADC is enabled, a conversion starts automati-
cally when this mode is entered.
8.5 ADC Noise Reduction Mode

When the SM2..0 bits are written to 001, the SLEEP instruction makes the MCU enter ADC
Noise Reduction mode, stopping the CPU but allowing the ADC, the external interrupts, 2-wire
Serial Interface address match, Timer/Counter2 and the Watchdog to continue operating (if
enabled). This sleep mode basically halts clkI/O, clkCPU, and clkFLASH, while allowing the
other clocks to run.
This improves the noise environment for the ADC, enabling higher resolution measurements. If
the ADC is enabled, a conversion starts automatically when this mode is entered. Apart form the
ADC Conversion Complete interrupt, only an External Reset, a Watchdog System Reset, a
Watchdog interrupt, a Brown-out Reset, a 2-wire serial interface interrupt, a Timer/Counter2
interrupt, an SPM/EEPROM ready interrupt, an external level interrupt on INT7:4 or a pin
change interrupt can wakeup the MCU from ADC Noise Reduction mode.
43
8011D–AVR–02/07
8.6 Power-down Mode
When the SM2..0 bits are written to 010, the SLEEP instruction makes the MCU enter Power-
down mode. In this mode, the external Oscillator is stopped, while the external interrupts, the 2-
wire Serial Interface, and the Watchdog continue operating (if enabled). Only an External Reset,
a Watchdog Reset, a Brown-out Reset, 2-wire Serial Interface address match, an external level
interrupt on PCINT7:4, an external interrupt on INT2:0, or a pin change interrupt can wake up
the MCU. This sleep mode basically halts all generated clocks, allowing operation of asynchro-
nous modules only.
Note that if a level triggered interrupt is used for wake-up from Power-down mode, the changed
level must be held for some time to wake up the MCU. Refer to ”External Interrupts” on page 66
for details.
When waking up from Power-down mode, there is a delay from the wake-up condition occurs
until the wake-up becomes effective. This allows the clock to restart and become stable after
having been stopped. The wake-up period is defined by the same CKSEL Fuses that define the
Reset Time-out period, as described in ”Clock Sources” on page 30.
8.7 Power-save Mode

When the SM2:0 bits are written to 011, the SLEEP instruction makes the MCU enter Power-
save mode. This mode is identical to Power-down, with one exception:
If Timer/Counter2 is enabled, it will keep running during sleep. The device can wake up from
either Timer Overflow or Output Compare event from Timer/Counter2 if the corresponding
Timer/Counter2 interrupt enable bits are set in TIMSK2, and the Global Interrupt Enable bit in
SREG is set.
If Timer/Counter2 is not running, Power-down mode is recommended instead of Power-save
mode.
The Timer/Counter2 can be clocked both synchronously and asynchronously in Power-save
mode. If the Timer/Counter2 is not using the asynchronous clock, the Timer/Counter Oscillator is
stopped during sleep. If the Timer/Counter2 is not using the synchronous clock, the clock source
is stopped during sleep. Note that even if the synchronous clock is running in Power-save, this
clock is only available for the Timer/Counter2.
8.8 Standby Mode

When the SM2..0 bits are 110 and an external crystal/resonator clock option is selected, the
SLEEP instruction makes the MCU enter Standby mode. This mode is identical to Power-down
with the exception that the Oscillator is kept running. From Standby mode, the device wakes up
in six clock cycles.
8.9 Extended Standby Mode

When the SM2..0 bits are 111 and an external crystal/resonator clock option is selected, the
SLEEP instruction makes the MCU enter Extended Standby mode. This mode is identical to
Power-save mode with the exception that the Oscillator is kept running. From Extended Standby
mode, the device wakes up in six clock cycles.
44 ATmega164P/324P/644P
8011D–AVR–02/07
8.10 Power Reduction Register
The Power Reduction Register(PRR), see ”PRR – Power Reduction Register” on page 48, pro-
vides a method to stop the clock to individual peripherals to reduce power consumption. The
current state of the peripheral is frozen and the I/O registers can not be read or written.
Resources used by the peripheral when stopping the clock will remain occupied, hence the
peripheral should in most cases be disabled before stopping the clock. Waking up a module,
which is done by clearing the bit in PRR, puts the module in the same state as before shutdown.
Module shutdown can be used in Idle mode and Active mode to significantly reduce the overall
power consumption. In all other sleep modes, the clock is already stopped.
8.11 Minimizing Power Consumption

There are several issues to consider when trying to minimize the power consumption in an AVR
controlled system. In general, sleep modes should be used as much as possible, and the sleep
mode should be selected so that as few as possible of the device’s functions are operating. All
functions not needed should be disabled. In particular, the following modules may need special
consideration when trying to achieve the lowest possible power consumption.
8.11.1 Analog to Digital Converter
If enabled, the ADC will be enabled in all sleep modes. To save power, the ADC should be dis-
abled before entering any sleep mode. When the ADC is turned off and on again, the next
conversion will be an extended conversion. Refer to ”ADC - Analog-to-digital Converter” on page
238 for details on ADC operation.
8.11.2 Analog Comparator
When entering Idle mode, the Analog Comparator should be disabled if not used. When entering
ADC Noise Reduction mode, the Analog Comparator should be disabled. In other sleep modes,
the Analog Comparator is automatically disabled. However, if the Analog Comparator is set up
to use the Internal Voltage Reference as input, the Analog Comparator should be disabled in all
sleep modes. Otherwise, the Internal Voltage Reference will be enabled, independent of sleep
mode. Refer to ”AC - Analog Comparator” on page 235 for details on how to configure the Ana-
log Comparator.
8.11.3 Brown-out Detector
If the Brown-out Detector is not needed by the application, this module should be turned off. If
the Brown-out Detector is enabled by the BODLEVEL Fuses, it will be enabled in all sleep
modes, and hence, always consume power. In the deeper sleep modes, this will contribute sig-
nificantly to the total current consumption. Refer to ”Brown-out Detection” on page 53 for details
on how to configure the Brown-out Detector.
8.11.4 Internal Voltage Reference
The Internal Voltage Reference will be enabled when needed by the Brown-out Detection, the
Analog Comparator or the ADC. If these modules are disabled as described in the sections
above, the internal voltage reference will be disabled and it will not be consuming power. When
turned on again, the user must allow the reference to start up before the output is used. If the
reference is kept on in sleep mode, the output can be used immediately. Refer to ”Internal Volt-
age Reference” on page 54 for details on the start-up time.
45
8011D–AVR–02/07
8.11.5 Watchdog Timer
If the Watchdog Timer is not needed in the application, the module should be turned off. If the
Watchdog Timer is enabled, it will be enabled in all sleep modes, and hence, always consume
power. In the deeper sleep modes, this will contribute significantly to the total current consump-
tion. Refer to ”Interrupts” on page 61 for details on how to configure the Watchdog Timer.
8.11.6 Port Pins
When entering a sleep mode, all port pins should be configured to use minimum power. The
most important is then to ensure that no pins drive resistive loads. In sleep modes where both
the I/O clock (clkI/O) and the ADC clock (clkADC) are stopped, the input buffers of the device will
be disabled. This ensures that no power is consumed by the input logic when not needed. In
some cases, the input logic is needed for detecting wake-up conditions, and it will then be
enabled. Refer to the section ”Digital Input Enable and Sleep Modes” on page 75 for details on
which pins are enabled. If the input buffer is enabled and the input signal is left floating or have
an analog signal level close to VCC/2, the input buffer will use excessive power.
For analog input pins, the digital input buffer should be disabled at all times. An analog signal
level close to VCC/2 on an input pin can cause significant current even in active mode. Digital
input buffers can be disabled by writing to the Digital Input Disable Registers (DIDR1 and
DIDR0). Refer to ”DIDR1 – Digital Input Disable Register 1” on page 237 and ”DIDR0 – Digital
Input Disable Register 0” on page 257 for details.
8.11.7 On-chip Debug System
If the On-chip debug system is enabled by the OCDEN Fuse and the chip enters sleep mode,
the main clock source is enabled, and hence, always consumes power. In the deeper sleep
modes, this will contribute significantly to the total current consumption.
There are three alternative ways to disable the OCD system:
• Disable the OCDEN Fuse.
• Disable the JTAGEN Fuse.
• Write one to the JTD bit in MCUCR.
46 ATmega164P/324P/644P
8011D–AVR–02/07
8.12.1 SMCR – Sleep Mode Control Register
The Sleep Mode Control Register contains control bits for power management.
Bit 7 6 5 4 3 2 1 0
0x33 (0x53) – – – – SM2 SM1 SM0 SE SMCR
• Bits 3, 2, 1 – SM2:0: Sleep Mode Select Bits 2, 1, and 0

These bits select between the five available sleep modes as shown in Table 8-2.
Table 8-2. Sleep Mode Select

SM2 SM1 SM0 Sleep Mode
0 0 0 Idle
0 0 1 ADC Noise Reduction
0 1 0 Power-down
0 1 1 Power-save
1 0 0 Reserved
1 0 1 Reserved
1 1 0 Standby(1)
1 1 1 Extended Standby(1)
Note: 1. Standby modes are only recommended for use with external crystals or resonators.
• Bit 0 – SE: Sleep Enable

The SE bit must be written to logic one to make the MCU enter the sleep mode when the SLEEP
instruction is executed. To avoid the MCU entering the sleep mode unless it is the programmer’s
purpose, it is recommended to write the Sleep Enable (SE) bit to one just before the execution of
the SLEEP instruction and to clear it immediately after waking up.
47
8011D–AVR–02/07
8.12.2 MCUCR – MCU Control Register
Bit 7 6 5 4 3 2 1 0
0x35 (0x55) JTD BODS BODSE PUD – – IVSEL IVCE MCUCR
Read/Write R/W R R R/W R R R/W R/W
• Bit 6 – BODS: BOD Sleep

The BODS bit must be written to logic one in order to turn off BOD during sleep, see Table 8-1
on page 42. Writing to the BODS bit is controlled by a timed sequence and an enable bit,
BODSE in MCUCR. To disable BOD in relevant sleep modes, both BODS and BODSE must first
be set to one. Then, to set the BODS bit, BODS must be set to one and BODSE must be set to
zero within four clock cycles.
The BODS bit is active three clock cycles after it is set. A sleep instruction must be executed
while BODS is active in order to turn off the BOD for the actual sleep mode. The BODS bit is
automatically cleared after three clock cycles.
• Bit 5 – BODSE: BOD Sleep Enable

BODSE enables setting of BODS control bit, as explained in BODS bit description. BOD disable
is controlled by a timed sequence.
8.12.3 PRR – Power Reduction Register
Bit 7 6 5 4 3 2 1 0
(0x64) PRTWI PRTIM2 PRTIM0 PRUSART1 PRTIM1 PRSPI PRUSART0 PRADC PRR
• Bit 7 - PRTWI: Power Reduction TWI

Writing a logic one to this bit shuts down the TWI by stopping the clock to the module. When
waking up the TWI again, the TWI should be re initialized to ensure proper operation.
• Bit 6 - PRTIM2: Power Reduction Timer/Counter2

Writing a logic one to this bit shuts down the Timer/Counter2 module in synchronous mode (AS2
is 0). When the Timer/Counter2 is enabled, operation will continue like before the shutdown.

Writing a logic one to this bit shuts down the Timer/Counter0 module. When the Timer/Counter0
is enabled, operation will continue like before the shutdown.
• Bit 4 - PRUSART1: Power Reduction USART1

Writing a logic one to this bit shuts down the USART1 by stopping the clock to the module.
When waking up the USART1 again, the USART1 should be reinitialized to ensure proper
operation.

Writing a logic one to this bit shuts down the Timer/Counter1 module. When the Timer/Counter1
is enabled, operation will continue like before the shutdown.
48 ATmega164P/324P/644P
8011D–AVR–02/07
• Bit 2 - PRSPI: Power Reduction Serial Peripheral Interface
Writing a logic one to this bit shuts down the Serial Peripheral Interface by stopping the clock to
the module. When waking up the SPI again, the SPI should be re initialized to ensure proper
operation.
• Bit 1 - PRUSART0: Power Reduction USART0

Writing a logic one to this bit shuts down the USART0 by stopping the clock to the module.
When waking up the USART0 again, the USART0 should be reinitialized to ensure proper
operation.
• Bit 0 - PRADC: Power Reduction ADC

Writing a logic one to this bit shuts down the ADC. The ADC must be disabled before shut down.
The analog comparator cannot use the ADC input MUX when the ADC is shut down.
49
8011D–AVR–02/07
9. System Control and Reset
9.0.1 Resetting the AVR
During reset, all I/O Registers are set to their initial values, and the program starts execution
from the Reset Vector. The instruction placed at the Reset Vector must be a JMP – Absolute
Jump – instruction to the reset handling routine. If the program never enables an interrupt
source, the Interrupt Vectors are not used, and regular program code can be placed at these
locations. This is also the case if the Reset Vector is in the Application section while the Interrupt
Vectors are in the Boot section or vice versa. The circuit diagram in Figure 9-1 on page 51
shows the reset logic. ”System and Reset Characteristics” on page 329 defines the electrical
parameters of the reset circuitry.
The I/O ports of the AVR are immediately reset to their initial state when a reset source goes
active. This does not require any clock source to be running.
After all reset sources have gone inactive, a delay counter is invoked, stretching the internal
reset. This allows the power to reach a stable level before normal operation starts. The time-out
period of the delay counter is defined by the user through the SUT and CKSEL Fuses. The dif-
ferent selections for the delay period are presented in ”Clock Sources” on page 30.
9.0.2 Reset Sources
The ATmega164P/324P/644P has five sources of reset:

• Power-on Reset. The MCU is reset when the supply voltage is below the Power-on Reset
threshold (VPOT).
• External Reset. The MCU is reset when a low level is present on the RESET pin for longer than
the minimum pulse length.
• Watchdog Reset. The MCU is reset when the Watchdog Timer period expires and the
Watchdog is enabled.
• Brown-out Reset. The MCU is reset when the supply voltage VCC is below the Brown-out Reset
threshold (VBOT) and the Brown-out Detector is enabled.
• JTAG AVR Reset. The MCU is reset as long as there is a logic one in the Reset Register, one
of the scan chains of the JTAG system. Refer to the section ”IEEE 1149.1 (JTAG) Boundary-
scan” on page 264 for details.
50 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 9-1. Reset Logic
DATA BUS
MCU Status
Register (MCUSR)
PORF
BORF
EXTRF
WDRF
JTRF
Power-on Reset
Circuit
Brown-out
BODLEVEL [2..0] Reset Circuit
Pull-up Resistor
SPIKE
FILTER
JTAG Reset
Register
Watchdog
Oscillator
Clock CK Delay Counters

Generator TIMEOUT
CKSEL[3:0]
SUT[1:0]
9.0.3 Power-on Reset
A Power-on Reset (POR) pulse is generated by an On-chip detection circuit. The detection level
is defined in ”System and Reset Characteristics” on page 329. The POR is activated whenever
VCC is below the detection level. The POR circuit can be used to trigger the start-up Reset, as
well as to detect a failure in supply voltage.
A Power-on Reset (POR) circuit ensures that the device is reset from Power-on. Reaching the
Power-on Reset threshold voltage invokes the delay counter, which determines how long the
device is kept in RESET after VCC rise. The RESET signal is activated again, without any delay,
when VCC decreases below the detection level.
51
8011D–AVR–02/07
Figure 9-2. MCU Start-up, RESET Tied to VCC
VPOT
VCC
VRST
RESET
tTOUT
TIME-OUT
INTERNAL
RESET
Figure 9-3. MCU Start-up, RESET Extended Externally

VPOT
VCC
VRST
RESET
tTOUT
TIME-OUT
INTERNAL
RESET
9.0.4 External Reset
An External Reset is generated by a low level on the RESET pin. Reset pulses longer than the
minimum pulse width (see ”System and Reset Characteristics” on page 329) will generate a
reset, even if the clock is not running. Shorter pulses are not guaranteed to generate a reset.
When the applied signal reaches the Reset Threshold Voltage – VRST – on its positive edge, the
delay counter starts the MCU after the Time-out period – tTOUT – has expired.
Figure 9-4. External Reset During Operation

CC
52 ATmega164P/324P/644P
8011D–AVR–02/07
9.0.5 Brown-out Detection
ATmega164P/324P/644P has an On-chip Brown-out Detection (BOD) circuit for monitoring the
VCC level during operation by comparing it to a fixed trigger level. The trigger level for the BOD
can be selected by the BODLEVEL Fuses. The trigger level has a hysteresis to ensure spike
free Brown-out Detection. The hysteresis on the detection level should be interpreted as VBOT+ =
VBOT + VHYST/2 and VBOT- = VBOT - VHYST/2.
When the BOD is enabled, and VCC decreases to a value below the trigger level (VBOT- in Figure
9-5 on page 53), the Brown-out Reset is immediately activated. When VCC increases above the
trigger level (VBOT+ in Figure 9-5 on page 53), the delay counter starts the MCU after the Time-
out period tTOUT has expired.
The BOD circuit will only detect a drop in VCC if the voltage stays below the trigger level for
longer than tBOD given in ”System and Reset Characteristics” on page 329.
Figure 9-5. Brown-out Reset During Operation
VCC VBOT+
VBOT-
RESET
TIME-OUT tTOUT
INTERNAL
RESET
9.0.6 Watchdog Reset
When the Watchdog times out, it will generate a short reset pulse of one CK cycle duration. On
the falling edge of this pulse, the delay timer starts counting the Time-out period tTOUT. Refer to
page 61 for details on operation of the Watchdog Timer.
Figure 9-6. Watchdog Reset During Operation

CC
CK
53
8011D–AVR–02/07
9.1 Internal Voltage Reference
ATmega164P/324P/644P features an internal bandgap reference. This reference is used for
Brown-out Detection, and it can be used as an input to the Analog Comparator or the ADC.
9.1.1 Voltage Reference Enable Signals and Start-up Time
The voltage reference has a start-up time that may influence the way it should be used. The
start-up time is given in ”System and Reset Characteristics” on page 329. To save power, the
reference is not always turned on. The reference is on during the following situations:
1. When the BOD is enabled (by programming the BODLEVEL [2:0] Fuse).
2. When the bandgap reference is connected to the Analog Comparator (by setting the
ACBG bit in ACSR).
3. When the ADC is enabled.
Thus, when the BOD is not enabled, after setting the ACBG bit or enabling the ADC, the user
must always allow the reference to start up before the output from the Analog Comparator or
ADC is used. To reduce power consumption in Power-down mode, the user can avoid the three
conditions above to ensure that the reference is turned off before entering Power-down mode.
54 ATmega164P/324P/644P
8011D–AVR–02/07
9.2 Watchdog Timer
9.2.1 Features
• Clocked from separate On-chip Oscillator

• 3 Operating modes
– Interrupt
– System Reset
– Interrupt and System Reset
• Selectable Time-out period from 16ms to 8s
• Possible Hardware fuse Watchdog always on (WDTON) for fail-safe mode
9.2.2 Overview
ATmega164P/324P/644P has an Enhanced Watchdog Timer (WDT). The WDT is a timer count-
ing cycles of a separate on-chip 128 kHz oscillator. The WDT gives an interrupt or a system
reset when the counter reaches a given time-out value. In normal operation mode, it is required
that the system uses the WDR - Watchdog Timer Reset - instruction to restart the counter before
the time-out value is reached. If the system doesn't restart the counter, an interrupt or system
reset will be issued.
Figure 9-7. Watchdog Timer
128kHz
OSCILLATOR
OSC/2K
OSC/4K
OSC/8K
OSC/16K
OSC/32K
OSC/64K
OSC/128K
OSC/256K
OSC/512K
OSC/1024K
WDP0
WDP1
WATCHDOG WDP2
RESET WDP3
WDE
MCU RESET
WDIF
INTERRUPT
WDIE
In Interrupt mode, the WDT gives an interrupt when the timer expires. This interrupt can be used
to wake the device from sleep-modes, and also as a general system timer. One example is to
limit the maximum time allowed for certain operations, giving an interrupt when the operation
has run longer than expected. In System Reset mode, the WDT gives a reset when the timer
expires. This is typically used to prevent system hang-up in case of runaway code. The third
mode, Interrupt and System Reset mode, combines the other two modes by first giving an inter-
rupt and then switch to System Reset mode. This mode will for instance allow a safe shutdown
by saving critical parameters before a system reset.
The Watchdog always on (WDTON) fuse, if programmed, will force the Watchdog Timer to Sys-
tem Reset mode. With the fuse programmed the System Reset mode bit (WDE) and Interrupt
mode bit (WDIE) are locked to 1 and 0 respectively. To further ensure program security, alter-
ations to the Watchdog set-up must follow timed sequences. The sequence for clearing WDE
and changing time-out configuration is as follows:
55
8011D–AVR–02/07
1. In the same operation, write a logic one to the Watchdog change enable bit (WDCE) and
WDE. A logic one must be written to WDE regardless of the previous value of the WDE
bit.
2. Within the next four clock cycles, write the WDE and Watchdog prescaler bits (WDP) as
desired, but with the WDCE bit cleared. This must be done in one operation.
The following code example shows one assembly and one C function for turning off the Watch-
dog Timer. The example assumes that interrupts are controlled (e.g. by disabling interrupts
globally) so that no interrupts will occur during the execution of these functions.

WDT_off:
; Turn off global interrupt
cli
; Reset Watchdog Timer
wdr
; Clear WDRF in MCUSR
in r16, MCUSR
andi r16, (0xff & (0<<WDRF))
out MCUSR, r16
; Write logical one to WDCE and WDE
; Keep old prescaler setting to prevent unintentional time-out
in r16, WDTCSR
ori r16, (1<<WDCE) | (1<<WDE)
out WDTCSR, r16
; Turn off WDT
ldi r16, (0<<WDE)
out WDTCSR, r16
; Turn on global interrupt
sei
ret
C Code Example(1)
void WDT_off(void)
{
__watchdog_reset();
/* Clear WDRF in MCUSR */
MCUSR &= ~(1<<WDRF);
/* Write logical one to WDCE and WDE */
/* Keep old prescaler setting to prevent unintentional time-out */
WDTCSR |= (1<<WDCE) | (1<<WDE);
/* Turn off WDT */
WDTCSR = 0x00;
__enable_interrupt();
}
Note: 1. The example code assumes that the part specific header file is included.
56 ATmega164P/324P/644P
8011D–AVR–02/07
Note: If the Watchdog is accidentally enabled, for example by a runaway pointer or brown-out
condition, the device will be reset and the Watchdog Timer will stay enabled. If the code is not
set up to handle the Watchdog, this might lead to an eternal loop of time-out resets. To avoid this
situation, the application software should always clear the Watchdog System Reset Flag
(WDRF) and the WDE control bit in the initialisation routine, even if the Watchdog is not in use.
The following code example shows one assembly and one C function for changing the time-out
value of the Watchdog Timer.

WDT_Prescaler_Change:
; Turn off global interrupt
cli
; Reset Watchdog Timer
wdr
; Start timed sequence
in r16, WDTCSR
ori r16, (1<<WDCE) | (1<<WDE)
out WDTCSR, r16
; -- Got four cycles to set the new values from here -
; Set new prescaler(time-out) value = 64K cycles (~0.5 s)
ldi r16, (1<<WDE) | (1<<WDP2) | (1<<WDP0)
out WDTCSR, r16
; -- Finished setting new values, used 2 cycles -
; Turn on global interrupt
sei
ret
C Code Example(1)
void WDT_Prescaler_Change(void)
{
__watchdog_reset();
/* Start timed equence */
WDTCSR |= (1<<WDCE) | (1<<WDE);
/* Set new prescaler(time-out) value = 64K cycles (~0.5 s) */
WDTCSR = (1<<WDE) | (1<<WDP2) | (1<<WDP0);
__enable_interrupt();
}
Note: The Watchdog Timer should be reset before any change of the WDP bits, since a change
in the WDP bits can result in a time-out when switching to a shorter time-out period.
57
8011D–AVR–02/07
9.3.1 MCUSR – MCU Status Register
The MCU Status Register provides information on which reset source caused an MCU reset.
Bit 7 6 5 4 3 2 1 0
0x34 (0x54) – – – JTRF WDRF BORF EXTRF PORF MCUSR
Initial Value 0 0 0 See Bit Description
• Bit 4 – JTRF: JTAG Reset Flag

This bit is set if a reset is being caused by a logic one in the JTAG Reset Register selected by
the JTAG instruction AVR_RESET. This bit is reset by a Power-on Reset, or by writing a logic
zero to the flag.
• Bit 3 – WDRF: Watchdog Reset Flag

This bit is set if a Watchdog Reset occurs. The bit is reset by a Power-on Reset, or by writing a
logic zero to the flag.
• Bit 2 – BORF: Brown-out Reset Flag

This bit is set if a Brown-out Reset occurs. The bit is reset by a Power-on Reset, or by writing a
• Bit 1 – EXTRF: External Reset Flag

This bit is set if an External Reset occurs. The bit is reset by a Power-on Reset, or by writing a
• Bit 0 – PORF: Power-on Reset Flag

This bit is set if a Power-on Reset occurs. The bit is reset only by writing a logic zero to the flag.
To make use of the Reset Flags to identify a reset condition, the user should read and then
Reset the MCUSR as early as possible in the program. If the register is cleared before another
reset occurs, the source of the reset can be found by examining the Reset Flags.
58 ATmega164P/324P/644P
8011D–AVR–02/07
9.3.2 WDTCSR – Watchdog Timer Control Register
Bit 7 6 5 4 3 2 1 0
(0x60) WDIF WDIE WDP3 WDCE WDE WDP2 WDP1 WDP0 WDTCSR
Initial Value 0 0 0 0 X 0 0 0
• Bit 7 - WDIF: Watchdog Interrupt Flag

This bit is set when a time-out occurs in the Watchdog Timer and the Watchdog Timer is config-
ured for interrupt. WDIF is cleared by hardware when executing the corresponding interrupt
handling vector. Alternatively, WDIF is cleared by writing a logic one to the flag. When the I-bit in
SREG and WDIE are set, the Watchdog Time-out Interrupt is executed.
• Bit 6 - WDIE: Watchdog Interrupt Enable

When this bit is written to one and the I-bit in the Status Register is set, the Watchdog Interrupt is
enabled. If WDE is cleared in combination with this setting, the Watchdog Timer is in Interrupt
Mode, and the corresponding interrupt is executed if time-out in the Watchdog Timer occurs.
If WDE is set, the Watchdog Timer is in Interrupt and System Reset Mode. The first time-out in
the Watchdog Timer will set WDIF. Executing the corresponding interrupt vector will clear WDIE
and WDIF automatically by hardware (the Watchdog goes to System Reset Mode). This is use-
ful for keeping the Watchdog Timer security while using the interrupt. To stay in Interrupt and
System Reset Mode, WDIE must be set after each interrupt. This should however not be done
within the interrupt service routine itself, as this might compromise the safety-function of the
Watchdog System Reset mode. If the interrupt is not executed before the next time-out, a Sys-
tem Reset will be applied.
Table 9-1. Watchdog Timer Configuration

WDTON WDE WDIE Mode Action on Time-out
0 0 0 Stopped None
0 0 1 Interrupt Mode Interrupt
0 1 0 System Reset Mode Reset
Interrupt and System Reset Interrupt, then go to System
0 1 1
Mode Reset Mode
1 x x System Reset Mode Reset
• Bit 4 - WDCE: Watchdog Change Enable

This bit is used in timed sequences for changing WDE and prescaler bits. To clear the WDE bit,
and/or change the prescaler bits, WDCE must be set.
Once written to one, hardware will clear WDCE after four clock cycles.
• Bit 3 - WDE: Watchdog System Reset Enable

WDE is overridden by WDRF in MCUSR. This means that WDE is always set when WDRF is
set. To clear WDE, WDRF must be cleared first. This feature ensures multiple resets during con-
ditions causing failure, and a safe start-up after the failure.
59
8011D–AVR–02/07
• Bit 5, 2:0 - WDP3:0: Watchdog Timer Prescaler 3, 2, 1 and 0
The WDP3:0 bits determine the Watchdog Timer prescaling when the Watchdog Timer is run-
ning. The different prescaling values and their corresponding time-out periods are shown in
.
Table 9-2. Watchdog Timer Prescale Select

Number of WDT Oscillator Typical Time-out at
WDP3 WDP2 WDP1 WDP0 Cycles VCC = 5.0V
0 0 0 0 2K (2048) cycles 16 ms
0 0 0 1 4K (4096) cycles 32 ms
0 0 1 0 8K (8192) cycles 64 ms
0 0 1 1 16K (16384) cycles 0.125 s
0 1 0 0 32K (32768) cycles 0.25 s
0 1 0 1 64K (65536) cycles 0.5 s
0 1 1 0 128K (131072) cycles 1.0 s
0 1 1 1 256K (262144) cycles 2.0 s
1 0 0 0 512K (524288) cycles 4.0 s
1 0 0 1 1024K (1048576) cycles 8.0 s
1 0 1 0
1 0 1 1
1 1 0 0
Reserved
1 1 0 1
1 1 1 0
1 1 1 1
60 ATmega164P/324P/644P
8011D–AVR–02/07
10. Interrupts
10.1 Overview
This section describes the specifics of the interrupt handling as performed in
ATmega164P/324P/644P. For a general explanation of the AVR interrupt handling, refer to
”Reset and Interrupt Handling” on page 15.
10.2 Interrupt Vectors in ATmega164P/324P/644P
Table 10-1. Reset and Interrupt Vectors

Vector Program
No. Address(2) Source Interrupt Definition
External Pin, Power-on Reset, Brown-out Reset,
1 $0000(1) RESET
Watchdog Reset, and JTAG AVR Reset
2 $0002 INT0 External Interrupt Request 0
5 $0008 PCINT0 Pin Change Interrupt Request 0
6 $000A PCINT1 Pin Change Interrupt Request 1
7 $000C PCINT2 Pin Change Interrupt Request 2
8 $000E PCINT3 Pin Change Interrupt Request 3
9 $0010 WDT Watchdog Time-out Interrupt
10 $0012 TIMER2_COMPA Timer/Counter2 Compare Match A
11 $0014 TIMER2_COMPB Timer/Counter2 Compare Match B
12 $0016 TIMER2_OVF Timer/Counter2 Overflow
13 $0018 TIMER1_CAPT Timer/Counter1 Capture Event
14 $001A TIMER1_COMPA Timer/Counter1 Compare Match A
15 $001C TIMER1_COMPB Timer/Counter1 Compare Match B
16 $001E TIMER1_OVF Timer/Counter1 Overflow
17 $0020 TIMER0_COMPA Timer/Counter0 Compare Match A
18 $0022 TIMER0_COMPB Timer/Counter0 Compare match B
19 $0024 TIMER0_OVF Timer/Counter0 Overflow
20 $0026 SPI_STC SPI Serial Transfer Complete
21 $0028 USART0_RX USART0 Rx Complete
22 $002A USART0_UDRE USART0 Data Register Empty
23 $002C USART0_TX USART0 Tx Complete
24 $002E ANALOG_COMP Analog Comparator
25 $0030 ADC ADC Conversion Complete
26 $0032 EE_READY EEPROM Ready
27 $0034 TWI 2-wire Serial Interface
61
8011D–AVR–02/07
Table 10-1. Reset and Interrupt Vectors (Continued)
Vector Program
No. Address(2) Source Interrupt Definition
28 $0036 SPM_READY Store Program Memory Ready
29 $0038 USART1_RX USART1 Rx Complete
30 $003A USART1_UDRE USART1 Data Register Empty
31 $003C USART1_TX USART1 Tx Complete
Notes: 1. When the BOOTRST Fuse is programmed, the device will jump to the Boot Loader address at
reset, see ”Memory Programming” on page 291.
2. When the IVSEL bit in MCUCR is set, Interrupt Vectors will be moved to the start of the Boot
Flash Section. The address of each Interrupt Vector will then be the address in this table
added to the start address of the Boot Flash Section.
Table 10-2 shows reset and Interrupt Vectors placement for the various combinations of
BOOTRST and IVSEL settings. If the program never enables an interrupt source, the Interrupt
Vectors are not used, and regular program code can be placed at these locations. This is also
the case if the Reset Vector is in the Application section while the Interrupt Vectors are in the
Boot section or vice versa.
Table 10-2. Reset and Interrupt Vectors Placement(1)
BOOTRST IVSEL Reset Address Interrupt Vectors Start Address
1 0 0x0000 0x0002
1 1 0x0000 Boot Reset Address + 0x0002
0 0 Boot Reset Address 0x0002
0 1 Boot Reset Address Boot Reset Address + 0x0002
Note: 1. The Boot Reset Address is shown in Table 24-7 on page 286. For the BOOTRST Fuse “1”
means unprogrammed while “0” means programmed.
The most typical and general program setup for the Reset and Interrupt Vector Addresses in
ATmega164P/324P/644P is:
Address Labels Code Comments

0x0000 jmp RESET ; Reset
0x0002 jmp INT0 ; IRQ0
0x0008 jmp PCINT0 ; PCINT0
0x000A jmp PCINT1 ; PCINT1
0x000C jmp PCINT2 ; PCINT2
0x000E jmp PCINT3 ; PCINT3
0x0010 jmp WDT ; Watchdog Timeout
0x0012 jmp TIM2_COMPA ; Timer2 CompareA
0x0014 jmp TIM2_COMPB ; Timer2 CompareB
0x0016 jmp TIM2_OVF ; Timer2 Overflow
0x0018 jmp TIM1_CAPT ; Timer1 Capture
0x001A jmp TIM1_COMPA ; Timer1 CompareA
0x001C jmp TIM1_COMPB ; Timer1 CompareB
0x001E jmp TIM1_OVF ; Timer1 Overflow
0x0020 jmp TIM0_COMPA ; Timer0 CompareA
62 ATmega164P/324P/644P
8011D–AVR–02/07
0x0022 jmp TIM0_COMPB ; Timer0 CompareB
0x0024 jmp TIM0_OVF ; Timer0 Overflow
0x0026 jmp SPI_STC ; SPI Transfer Complete
0x0028 jmp USART0_RXC ; USART0 RX Complete
0x002A jmp USART0_UDRE ; USART0,UDR Empty
0x002C jmp USART0_TXC ; USART0 TX Complete
0x002E jmp ANA_COMP ; Analog Comparator
0x0030 jmp ADC ; ADC Conversion Complete
0x0032 jmp EE_RDY ; EEPROM Ready
0x0034 jmp TWI ; 2-wire Serial
0x0036 jmp SPM_RDY ; SPM Ready
0x0038 jmp USART1_RXC ; USART1 RX Complete
0x003A jmp USART1_UDRE ; USART1,UDR Empty
0x003C jmp USART1_TXC ; USART1 TX Complete
;
0x003E RESET: ldi r16, ; Main program start
high(RAMEND)
0x003F out SPH,r16 ; Set Stack Pointer to
top of RAM
0x0040 ldi r16,
low(RAMEND)
0x0041 out SPL,r16
0x0042 sei ; Enable interrupts
0x0043 <instr> xxx
... ... ... ...
When the BOOTRST Fuse is unprogrammed, the Boot section size set to 8K bytes and the
IVSEL bit in the MCUCR Register is set before any interrupts are enabled, the most typical and
general program setup for the Reset and Interrupt Vector Addresses is:

0x00000 RESET: ldi r16,high(RAMEND); Main program start
0x00001 out SPH,r16 ; Set Stack Pointer to top of RAM
0x00002 ldi r16,low(RAMEND)
0x00003 out SPL,r16
0x00004 sei ; Enable interrupts
0x00005 <instr> xxx
;
.org 0x1F002
0x1F002 jmp EXT_INT0 ; IRQ0 Handler
... ... ... ;
0x1FO36 jmp SPM_RDY ; SPM Ready Handler
When the BOOTRST Fuse is programmed and the Boot section size set to 8K bytes, the most
typical and general program setup for the Reset and Interrupt Vector Addresses is:

.org 0x0002
0x00002 jmp EXT_INT0 ; IRQ0 Handler
63
8011D–AVR–02/07
0x00004 jmp EXT_INT1 ; IRQ1 Handler
... ... ... ;
0x00036 jmp SPM_RDY ; SPM Ready Handler
;
.org 0x1F000
0x1F000 RESET: ldi r16,high(RAMEND); Main program start
0x1F001 out SPH,r16 ; Set Stack Pointer to top of RAM
0x1F002 ldi r16,low(RAMEND)
0x1F003 out SPL,r16
0x1F004 sei ; Enable interrupts
0x1F005 <instr> xxx
When the BOOTRST Fuse is programmed, the Boot section size set to 8K bytes and the IVSEL
bit in the MCUCR Register is set before any interrupts are enabled, the most typical and general
program setup for the Reset and Interrupt Vector Addresses is:
;
.org 0x1F000
0x1F000 jmp RESET ; Reset handler
... ... ... ;
0x1F036 jmp SPM_RDY ; SPM Ready Handler
;
0x1F03E RESET: ldi r16,high(RAMEND); Main program start
0x1F03F out SPH,r16 ; Set Stack Pointer to top of RAM
0x1F040 ldi r16,low(RAMEND)
0x1F041 out SPL,r16
0x1F042 sei ; Enable interrupts
0x1FO43 <instr> xxx
10.2.1 Moving Interrupts Between Application and Boot Space
The General Interrupt Control Register controls the placement of the Interrupt Vector table.

Bit 7 6 5 4 3 2 1 0
64 ATmega164P/324P/644P
8011D–AVR–02/07
• Bit 1 – IVSEL: Interrupt Vector Select

When the IVSEL bit is cleared (zero), the Interrupt Vectors are placed at the start of the Flash
memory. When this bit is set (one), the Interrupt Vectors are moved to the beginning of the Boot
Loader section of the Flash. The actual address of the start of the Boot Flash Section is deter-
mined by the BOOTSZ Fuses. Refer to the section ”Memory Programming” on page 291 for
details. To avoid unintentional changes of Interrupt Vector tables, a special write procedure must
be followed to change the IVSEL bit:
a. Write the Interrupt Vector Change Enable (IVCE) bit to one.
b. Within four cycles, write the desired value to IVSEL while writing a zero to IVCE.
Interrupts will automatically be disabled while this sequence is executed. Interrupts are disabled
in the cycle IVCE is set, and they remain disabled until after the instruction following the write to
IVSEL. If IVSEL is not written, interrupts remain disabled for four cycles. The I-bit in the Status
Register is unaffected by the automatic disabling.
Note: If Interrupt Vectors are placed in the Boot Loader section and Boot Lock bit BLB02 is programmed,
interrupts are disabled while executing from the Application section. If Interrupt Vectors are placed
in the Application section and Boot Lock bit BLB12 is programed, interrupts are disabled while
executing from the Boot Loader section. Refer to the section ”Memory Programming” on page 291
for details on Boot Lock bits.
• Bit 0 – IVCE: Interrupt Vector Change Enable

The IVCE bit must be written to logic one to enable change of the IVSEL bit. IVCE is cleared by
hardware four cycles after it is written or when IVSEL is written. Setting the IVCE bit will disable
interrupts, as explained in the IVSEL description above. See the following Code Example.

Move_interrupts:
; Enable change of Interrupt Vectors
ldi r16, (1<<IVCE)
out MCUCR, r16
; Move interrupts to Boot Flash section
ldi r16, (1<<IVSEL)
out MCUCR, r16
ret
C Code Example
void Move_interrupts(void)
{
/* Enable change of Interrupt Vectors */
MCUCR = (1<<IVCE);
/* Move interrupts to Boot Flash section */
MCUCR = (1<<IVSEL);
}
65
8011D–AVR–02/07
11. External Interrupts
11.1 Overview
The External Interrupts are triggered by the INT2:0 pin or any of the PCINT31:0 pins. Observe
that, if enabled, the interrupts will trigger even if the INT2:0 or PCINT31:0 pins are configured as
outputs. This feature provides a way of generating a software interrupt.
The Pin change interrupt PCI3 will trigger if any enabled PCINT31:24 pin toggle, Pin change
interrupt PCI2 will trigger if any enabled PCINT23:16 pin toggles, Pin change interrupt PCI1 if
any enabled PCINT15:8 toggles and Pin change interrupts PCI0 will trigger if any enabled
PCINT7:0 pin toggles. PCMSK3, PCMSK2, PCMSK1 and PCMSK0 Registers control which pins
contribute to the pin change interrupts. Pin change interrupts on PCINT31:0 are detected asyn-
chronously. This implies that these interrupts can be used for waking the part also from sleep
modes other than Idle mode.
The External Interrupts can be triggered by a falling or rising edge or a low level. This is set up
as indicated in the specification for the External Interrupt Control Registers – EICRA (INT2:0).
When the external interrupt is enabled and is configured as level triggered, the interrupt will trig-
ger as long as the pin is held low. Low level interrupts and the edge interrupt on INT2:0 are
detected asynchronously. This implies that these interrupts can be used for waking the part also
from sleep modes other than Idle mode. The I/O clock is halted in all sleep modes except Idle
mode.
Note that if a level triggered interrupt is used for wake-up from Power-down, the required level
must be held long enough for the MCU to complete the wake-up to trigger the level interrupt. If
the level disappears before the end of the Start-up Time, the MCU will still wake up, but no inter-
rupt will be generated. The start-up time is defined by the SUT and CKSEL Fuses as described
in ”System Clock and Clock Options” on page 29.

11.2.1 EICRA – External Interrupt Control Register A
The External Interrupt Control Register A contains control bits for interrupt sense control.
Bit 7 6 5 4 3 2 1 0
(0x69) – – ISC21 ISC20 ISC11 ISC10 ISC01 ISC00 EICRA
Read/Write R R R/W R/W R/W R/W R/W R/W
• Bits 7:6 – Reserved

These bits are reserved in the ATmega164P/324P/644P, and will always read as zero.
• Bits 5:0 – ISC21, ISC20 – ISC00, ISC00: External Interrupt 2 - 0 Sense Control Bits
The External Interrupts 2 - 0 are activated by the external pins INT2:0 if the SREG I-flag and the
corresponding interrupt mask in the EIMSK is set. The level and edges on the external pins that
activate the interrupts are defined in Table 11-1. Edges on INT2..INT0 are registered asynchro-
nously. Pulses on INT2:0 pins wider than the minimum pulse width given in ”External Interrupts
Characteristics” on page 329 will generate an interrupt. Shorter pulses are not guaranteed to
generate an interrupt. If low level interrupt is selected, the low level must be held until the com-
66 ATmega164P/324P/644P
8011D–AVR–02/07
pletion of the currently executing instruction to generate an interrupt. If enabled, a level triggered
interrupt will generate an interrupt request as long as the pin is held low. When changing the
ISCn bit, an interrupt can occur. Therefore, it is recommended to first disable INTn by clearing its
Interrupt Enable bit in the EIMSK Register. Then, the ISCn bit can be changed. Finally, the INTn
interrupt flag should be cleared by writing a logical one to its Interrupt Flag bit (INTFn) in the
EIFR Register before the interrupt is re-enabled.
Table 11-1. Interrupt Sense Control(1)

ISCn1 ISCn0 Description
0 0 The low level of INTn generates an interrupt request.
0 1 Any edge of INTn generates asynchronously an interrupt request.
1 0 The falling edge of INTn generates asynchronously an interrupt request.
1 1 The rising edge of INTn generates asynchronously an interrupt request.
Note: 1. n = 2, 1 or 0.
When changing the ISCn1/ISCn0 bits, the interrupt must be disabled by clearing its Interrupt
Enable bit in the EIMSK Register. Otherwise an interrupt can occur when the bits are changed.
11.2.2 EIMSK – External Interrupt Mask Register
Bit 7 6 5 4 3 2 1 0
0x1D (0x3D) – – – – – INT2 INT1 IINT0 EIMSK
Read/Write R R R R R R/W R/W R/W
• Bits 2:0 – INT2:0: External Interrupt Request 2 - 0 Enable

When an INT2:0 bit is written to one and the I-bit in the Status Register (SREG) is set (one), the
corresponding external pin interrupt is enabled. The Interrupt Sense Control bits in the External
Interrupt Control Register, EICRA, defines whether the external interrupt is activated on rising or
falling edge or level sensed. Activity on any of these pins will trigger an interrupt request even if
the pin is enabled as an output. This provides a way of generating a software interrupt.
11.2.3 EIFR –External Interrupt Flag Register
Bit 7 6 5 4 3 2 1 0
0x1C (0x3C) – – – – – INTF2 INTF1 IINTF0 EIFR
Read/Write R/W R R R R R/W R/W R/W
• Bits 2:0 – INTF2:0: External Interrupt Flags 2 - 0

When an edge or logic change on the INT2:0 pin triggers an interrupt request, INTF2:0 becomes
set (one). If the I-bit in SREG and the corresponding interrupt enable bit, INT2:0 in EIMSK, are
set (one), the MCU will jump to the interrupt vector. The flag is cleared when the interrupt routine
is executed. Alternatively, the flag can be cleared by writing a logical one to it. These flags are
always cleared when INT2:0 are configured as level interrupt. Note that when entering sleep
mode with the INT2:0 interrupts disabled, the input buffers on these pins will be disabled. This
may cause a logic change in internal signals which will set the INTF2:0 flags. See ”Digital Input
Enable and Sleep Modes” on page 75 for more information.
67
8011D–AVR–02/07
11.2.4 PCICR – Pin Change Interrupt Control Register
Bit 7 6 5 4 3 2 1 0
(0x68) – – – – PCIE3 PCIE2 PCIE1 PCIE0 PCICR
• Bit 3 – PCIE3: Pin Change Interrupt Enable 3

When the PCIE3 bit is set (one) and the I-bit in the Status Register (SREG) is set (one), pin
change interrupt 3 is enabled. Any change on any enabled PCINT31..24 pin will cause an inter-
rupt. The corresponding interrupt of Pin Change Interrupt Request is executed from the PCI3
Interrupt Vector. PCINT31..24 pins are enabled individually by the PCMSK3 Register.



change interrupt 0 is enabled. Any change on any enabled PCINT7..0 pin will cause an interrupt.
The corresponding interrupt of Pin Change Interrupt Request is executed from the PCI0 Interrupt
Vector. PCINT7..0 pins are enabled individually by the PCMSK0 Register.
11.2.5 PCIFR – Pin Change Interrupt Flag Register
Bit 7 6 5 4 3 2 1 0
0x1B (0x3B) – – PCIF3 PCIF2 PCIF1 PCIF0 PCIFR
• Bit 3– PCIF3: Pin Change Interrupt Flag 3

When a logic change on any PCINT31..24 pin triggers an interrupt request, PCIF3 becomes set
(one). If the I-bit in SREG and the PCIE3 bit in EIMSK are set (one), the MCU will jump to the
corresponding Interrupt Vector. The flag is cleared when the interrupt routine is executed. Alter-
natively, the flag can be cleared by writing a logical one to it.
• Bit 2 – PCIF2: Pin Change Interrupt Flag 2

68 ATmega164P/324P/644P
8011D–AVR–02/07


11.2.6 PCMSK3 – Pin Change Mask Register 3
Bit 7 6 5 4 3 2 1 0
(0x73) PCINT31 PCINT30 PCINT29 PCINT28 PCINT27 PCINT26 PCINT25 PCINT24 PCMSK2
• Bit 7:0 – PCINT31:24: Pin Change Enable Mask 31:24

Each PCINT31:24-bit selects whether pin change interrupt is enabled on the corresponding I/O
pin. If PCINT31:24 is set and the PCIE2 bit in PCICR is set, pin change interrupt is enabled on
the corresponding I/O pin. If PCINT31..24 is cleared, pin change interrupt on the corresponding
I/O pin is disabled.
Bit 7 6 5 4 3 2 1 0
(0x6D) PCINT23 PCINT22 PCINT21 PCINT20 PCINT19 PCINT18 PCINT17 PCINT16 PCMSK2
• Bit 7:0 – PCINT23:16: Pin Change Enable Mask 23..16

pin. If PCINT23:16 is set and the PCIE2 bit in PCICR is set, pin change interrupt is enabled on
the corresponding I/O pin. If PCINT23..16 is cleared, pin change interrupt on the corresponding
I/O pin is disabled.
Bit 7 6 5 4 3 2 1 0
(0x6C) PCINT15 PCINT14 PCINT13 PCINT12 PCINT11 PCINT10 PCINT9 PCINT8 PCMSK1

pin. If PCINT15:8 is set and the PCIE1 bit in EIMSK is set, pin change interrupt is enabled on the
corresponding I/O pin. If PCINT15:8 is cleared, pin change interrupt on the corresponding I/O
pin is disabled.
69
8011D–AVR–02/07
Bit 7 6 5 4 3 2 1 0
(0x6B) PCINT7 PCINT6 PCINT5 PCINT4 PCINT3 PCINT2 PCINT1 PCINT0 PCMSK0

Each PCINT7:0 bit selects whether pin change interrupt is enabled on the corresponding I/O pin.
If PCINT7:0 is set and the PCIE0 bit in PCICR is set, pin change interrupt is enabled on the cor-
responding I/O pin. If PCINT7..0 is cleared, pin change interrupt on the corresponding I/O pin is
disabled.
70 ATmega164P/324P/644P
8011D–AVR–02/07
12. I/O-Ports
12.1 Overview
All AVR ports have true Read-Modify-Write functionality when used as general digital I/O ports.
This means that the direction of one port pin can be changed without unintentionally changing
the direction of any other pin with the SBI and CBI instructions. The same applies when chang-
ing drive value (if configured as output) or enabling/disabling of pull-up resistors (if configured as
input). Each output buffer has symmetrical drive characteristics with both high sink and source
capability. The pin driver is strong enough to drive LED displays directly. All port pins have indi-
vidually selectable pull-up resistors with a supply-voltage invariant resistance. All I/O pins have
protection diodes to both VCC and Ground as indicated in Figure 12-1. Refer to ”Electrical Char-
acteristics” on page 323 for a complete list of parameters.
Figure 12-1. I/O Pin Equivalent Schematic
Rpu
Pxn Logic
Cpin
See Figure
"General Digital I/O" for
Details
All registers and bit references in this section are written in general form. A lower case “x” repre-
sents the numbering letter for the port, and a lower case “n” represents the bit number. However,
when using the register or bit defines in a program, the precise form must be used. For example,
PORTB3 for bit no. 3 in Port B, here documented generally as PORTxn. The physical I/O Regis-
ters and bit locations are listed in ”Register Description” on page 90.
Three I/O memory address locations are allocated for each port, one each for the Data Register
– PORTx, Data Direction Register – DDRx, and the Port Input Pins – PINx. The Port Input Pins
I/O location is read only, while the Data Register and the Data Direction Register are read/write.
However, writing a logic one to a bit in the PINx Register, will result in a toggle in the correspond-
ing bit in the Data Register. In addition, the Pull-up Disable – PUD bit in MCUCR disables the
pull-up function for all pins in all ports when set.
Using the I/O port as General Digital I/O is described in ”Ports as General Digital I/O” on page
72. Most port pins are multiplexed with alternate functions for the peripheral features on the
device. How each alternate function interferes with the port pin is described in ”Alternate Port
Functions” on page 77. Refer to the individual module sections for a full description of the alter-
nate functions.
71
8011D–AVR–02/07
Note that enabling the alternate function of some of the port pins does not affect the use of the
other pins in the port as general digital I/O.
12.2 Ports as General Digital I/O

The ports are bi-directional I/O ports with optional internal pull-ups. Figure 12-2 shows a func-
tional description of one I/O-port pin, here generically called Pxn.
Figure 12-2. General Digital I/O(1)
PUD
Q D
DDxn
Q CLR
WDx
RESET
RDx
DATA BUS
1
Pxn Q D
PORTxn 0
Q CLR
RESET
WRx WPx
SLEEP RRx
SYNCHRONIZER
RPx
D Q D Q
PINxn
L Q Q
clk I/O
WDx: WRITE DDRx

PUD: PULLUP DISABLE RDx: READ DDRx
SLEEP: SLEEP CONTROL WRx: WRITE PORTx
clkI/O: I/O CLOCK RRx: READ PORTx REGISTER
RPx: READ PORTx PIN
WPx: WRITE PINx REGISTER
Note: 1. WRx, WPx, WDx, RRx, RPx, and RDx are common to all pins within the same port. clkI/O,
SLEEP, and PUD are common to all ports.
12.2.1 Configuring the Pin
Each port pin consists of three register bits: DDxn, PORTxn, and PINxn. As shown in ”Register
Description” on page 90, the DDxn bits are accessed at the DDRx I/O address, the PORTxn bits
at the PORTx I/O address, and the PINxn bits at the PINx I/O address.
The DDxn bit in the DDRx Register selects the direction of this pin. If DDxn is written logic one,
Pxn is configured as an output pin. If DDxn is written logic zero, Pxn is configured as an input
pin.
If PORTxn is written logic one when the pin is configured as an input pin, the pull-up resistor is
activated. To switch the pull-up resistor off, PORTxn has to be written logic zero or the pin has to
be configured as an output pin. The port pins are tri-stated when reset condition becomes active,
even if no clocks are running.
72 ATmega164P/324P/644P
8011D–AVR–02/07
If PORTxn is written logic one when the pin is configured as an output pin, the port pin is driven
high (one). If PORTxn is written logic zero when the pin is configured as an output pin, the port
pin is driven low (zero).
12.2.2 Toggling the Pin
Writing a logic one to PINxn toggles the value of PORTxn, independent on the value of DDRxn.
Note that the SBI instruction can be used to toggle one single bit in a port.
12.2.3 Switching Between Input and Output
When switching between tri-state ({DDxn, PORTxn} = 0b00) and output high ({DDxn, PORTxn}
= 0b11), an intermediate state with either pull-up enabled {DDxn, PORTxn} = 0b01) or output
low ({DDxn, PORTxn} = 0b10) must occur. Normally, the pull-up enabled state is fully accept-
able, as a high-impedant environment will not notice the difference between a strong high driver
and a pull-up. If this is not the case, the PUD bit in the MCUCR Register can be set to disable all
pull-ups in all ports.
Switching between input with pull-up and output low generates the same problem. The user
must use either the tri-state ({DDxn, PORTxn} = 0b00) or the output high state ({DDxn, PORTxn}
= 0b11) as an intermediate step.
Table 12-1 summarizes the control signals for the pin value.
Table 12-1. Port Pin Configurations

PUD
DDxn PORTxn (in MCUCR) I/O Pull-up Comment
0 0 X Input No Tri-state (Hi-Z)
0 1 0 Input Yes Pxn will source current if ext. pulled low.
0 1 1 Input No Tri-state (Hi-Z)
1 0 X Output No Output Low (Sink)
1 1 X Output No Output High (Source)
12.2.4 Reading the Pin Value
Independent of the setting of Data Direction bit DDxn, the port pin can be read through the
PINxn Register bit. As shown in Figure 12-2, the PINxn Register bit and the preceding latch con-
stitute a synchronizer. This is needed to avoid metastability if the physical pin changes value
near the edge of the internal clock, but it also introduces a delay. Figure 12-3 shows a timing dia-
gram of the synchronization when reading an externally applied pin value. The maximum and
minimum propagation delays are denoted tpd,max and tpd,min respectively.
73
8011D–AVR–02/07
Figure 12-3. Synchronization when Reading an Externally Applied Pin value
SYSTEM CLK
INSTRUCTIONS XXX XXX in r17, PINx
SYNC LATCH
PINxn
r17 0x00 0xFF
t pd, max
t pd, min
Consider the clock period starting shortly after the first falling edge of the system clock. The latch
is closed when the clock is low, and goes transparent when the clock is high, as indicated by the
shaded region of the “SYNC LATCH” signal. The signal value is latched when the system clock
goes low. It is clocked into the PINxn Register at the succeeding positive clock edge. As indi-
cated by the two arrows tpd,max and tpd,min, a single signal transition on the pin will be delayed
between ½ and 1½ system clock period depending upon the time of assertion.
When reading back a software assigned pin value, a nop instruction must be inserted as indi-
cated in Figure 12-4. The out instruction sets the “SYNC LATCH” signal at the positive edge of
the clock. In this case, the delay tpd through the synchronizer is 1 system clock period.
Figure 12-4. Synchronization when Reading a Software Assigned Pin Value
SYSTEM CLK
r16 0xFF
INSTRUCTIONS out PORTx, r16 nop in r17, PINx
SYNC LATCH
PINxn
r17 0x00 0xFF
t pd
The following code example shows how to set port B pins 0 and 1 high, 2 and 3 low, and define
the port pins from 4 to 7 as input with pull-ups assigned to port pins 6 and 7. The resulting pin
values are read back again, but as previously discussed, a nop instruction is included to be able
to read back the value recently assigned to some of the pins.
74 ATmega164P/324P/644P
8011D–AVR–02/07

...
; Define pull-ups and set outputs high
; Define directions for port pins
ldi r16,(1<<PB7)|(1<<PB6)|(1<<PB1)|(1<<PB0)
ldi r17,(1<<DDB3)|(1<<DDB2)|(1<<DDB1)|(1<<DDB0)
out PORTB,r16
out DDRB,r17
; Insert nop for synchronization
nop
; Read port pins
in r16,PINB
...
C Code Example
unsigned char i;
...
/* Define pull-ups and set outputs high */
/* Define directions for port pins */
PORTB = (1<<PB7)|(1<<PB6)|(1<<PB1)|(1<<PB0);
DDRB = (1<<DDB3)|(1<<DDB2)|(1<<DDB1)|(1<<DDB0);
/* Insert nop for synchronization*/
__no_operation();
/* Read port pins */
i = PINB;
...
Note: 1. For the assembly program, two temporary registers are used to minimize the time from pull-
ups are set on pins 0, 1, 6, and 7, until the direction bits are correctly set, defining bit 2 and 3
as low and redefining bits 0 and 1 as strong high drivers.
12.2.5 Digital Input Enable and Sleep Modes
As shown in Figure 12-2, the digital input signal can be clamped to ground at the input of the
schmitt-trigger. The signal denoted SLEEP in the figure, is set by the MCU Sleep Controller in
Power-down mode, Power-save mode, and Standby mode to avoid high power consumption if
some input signals are left floating, or have an analog signal level close to VCC/2.
SLEEP is overridden for port pins enabled as external interrupt pins. If the external interrupt
request is not enabled, SLEEP is active also for these pins. SLEEP is also overridden by various
other alternate functions as described in ”Alternate Port Functions” on page 77.
If a logic high level (“one”) is present on an asynchronous external interrupt pin configured as
“Interrupt on Rising Edge, Falling Edge, or Any Logic Change on Pin” while the external interrupt
is not enabled, the corresponding External Interrupt Flag will be set when resuming from the
above mentioned Sleep mode, as the clamping in these sleep mode produces the requested
logic change.
75
8011D–AVR–02/07
12.2.6 Unconnected Pins
If some pins are unused, it is recommended to ensure that these pins have a defined level. Even
though most of the digital inputs are disabled in the deep sleep modes as described above, float-
ing inputs should be avoided to reduce current consumption in all other modes where the digital
inputs are enabled (Reset, Active mode and Idle mode).
The simplest method to ensure a defined level of an unused pin, is to enable the internal pull-up.
In this case, the pull-up will be disabled during reset. If low power consumption during reset is
important, it is recommended to use an external pull-up or pull-down. Connecting unused pins
directly to VCC or GND is not recommended, since this may cause excessive currents if the pin is
accidentally configured as an output.
76 ATmega164P/324P/644P
8011D–AVR–02/07
12.3 Alternate Port Functions
Most port pins have alternate functions in addition to being general digital I/Os. Figure 12-5
shows how the port pin control signals from the simplified Figure 12-2 on page 72 can be over-
ridden by alternate functions. The overriding signals may not be present in all port pins, but the
figure serves as a generic description applicable to all port pins in the AVR microcontroller
family.
Figure 12-5. Alternate Port Functions(1)

PUOExn
PUOVxn
1
0
PUD
DDOExn
DDOVxn
1
0 Q D
DDxn
Q CLR
WDx
PVOExn RESET
RDx
PVOVxn
DATA BUS
1 1
Pxn
0 Q D 0
PORTxn
PTOExn
Q CLR
DIEOExn
WPx
DIEOVxn RESET
1 WRx
RRx
0 SLEEP
SYNCHRONIZER
RPx
SET
D Q D Q
PINxn
L CLR Q CLR Q
clk I/O
DIxn
AIOxn
PUOExn: Pxn PULL-UP OVERRIDE ENABLE PUD: PULLUP DISABLE

PUOVxn: Pxn PULL-UP OVERRIDE VALUE WDx: WRITE DDRx
DDOExn: Pxn DATA DIRECTION OVERRIDE ENABLE RDx: READ DDRx
DDOVxn: Pxn DATA DIRECTION OVERRIDE VALUE RRx: READ PORTx REGISTER
PVOExn: Pxn PORT VALUE OVERRIDE ENABLE WRx: WRITE PORTx
PVOVxn: Pxn PORT VALUE OVERRIDE VALUE RPx: READ PORTx PIN
DIEOExn: Pxn DIGITAL INPUT-ENABLE OVERRIDE ENABLE WPx: WRITE PINx
DIEOVxn: Pxn DIGITAL INPUT-ENABLE OVERRIDE VALUE clkI/O: I/O CLOCK
SLEEP: SLEEP CONTROL DIxn: DIGITAL INPUT PIN n ON PORTx
PTOExn: Pxn, PORT TOGGLE OVERRIDE ENABLE AIOxn: ANALOG INPUT/OUTPUT PIN n ON PORTx
Note: 1. WRx, WPx, WDx, RRx, RPx, and RDx are common to all pins within the same port. clkI/O,
SLEEP, and PUD are common to all ports. All other signals are unique for each pin.
77
8011D–AVR–02/07
Table 12-2 summarizes the function of the overriding signals. The pin and port indexes from Fig-
ure 12-5 are not shown in the succeeding tables. The overriding signals are generated internally
in the modules having the alternate function.
Table 12-2. Generic Description of Overriding Signals for Alternate Functions

Signal Name Full Name Description
If this signal is set, the pull-up enable is controlled by the PUOV
Pull-up Override
PUOE signal. If this signal is cleared, the pull-up is enabled when
Enable
{DDxn, PORTxn, PUD} = 0b010.
If PUOE is set, the pull-up is enabled/disabled when PUOV is
Pull-up Override
PUOV set/cleared, regardless of the setting of the DDxn, PORTxn,
Value
and PUD Register bits.
If this signal is set, the Output Driver Enable is controlled by the
Data Direction
DDOE DDOV signal. If this signal is cleared, the Output driver is
Override Enable
enabled by the DDxn Register bit.
If DDOE is set, the Output Driver is enabled/disabled when
Data Direction
DDOV DDOV is set/cleared, regardless of the setting of the DDxn
Override Value
Register bit.
If this signal is set and the Output Driver is enabled, the port
Port Value value is controlled by the PVOV signal. If PVOE is cleared, and
PVOE
Override Enable the Output Driver is enabled, the port Value is controlled by the
PORTxn Register bit.
Port Value If PVOE is set, the port value is set to PVOV, regardless of the
PVOV
Override Value setting of the PORTxn Register bit.
Port Toggle
PTOE If PTOE is set, the PORTxn Register bit is inverted.
Override Enable
Digital Input If this bit is set, the Digital Input Enable is controlled by the
DIEOE Enable Override DIEOV signal. If this signal is cleared, the Digital Input Enable
Enable is determined by MCU state (Normal mode, sleep mode).
Digital Input If DIEOE is set, the Digital Input is enabled/disabled when
DIEOV Enable Override DIEOV is set/cleared, regardless of the MCU state (Normal
Value mode, sleep mode).
This is the Digital Input to alternate functions. In the figure, the
signal is connected to the output of the schmitt trigger but
DI Digital Input before the synchronizer. Unless the Digital Input is used as a
clock source, the module with the alternate function will use its
own synchronizer.
This is the Analog Input/output to/from alternate functions. The
Analog
AIO signal is connected directly to the pad, and can be used bi-
Input/Output
directionally.
The following subsections shortly describe the alternate functions for each port, and relate the
overriding signals to the alternate function. Refer to the alternate function description for further
details.
78 ATmega164P/324P/644P
8011D–AVR–02/07
12.3.1 Alternate Functions of Port A
The Port A has an alternate function as the address low byte and data lines for the External
Memory Interface.
Table 12-3. Port A Pins Alternate Functions

Port Pin Alternate Function
ADC7 (ADC input channel 7)
PA7
PCINT7 (Pin Change Interrupt 7)
PA6
PA5
PA4
PA3
PA2
PA1
PA0
• ADC7:0/PCINT7:0 – Port A, Bit 7:0

ADC7:0, Analog to Digital Converter, Channels 7:0.
PCINT7:0, Pin Change Interrupt source 7:0: The PA7:0 pins can serve as external interrupt
sources.
79
8011D–AVR–02/07
Table 12-4 on page 80 and Table 12-5 on page 80 relates the alternate functions of Port A to the
overriding signals shown in Figure 12-5 on page 77.
Table 12-4. Overriding Signals for Alternate Functions in PA7:PA4

Signal PA7/ADC7/ PA6/ADC6/ PA5/ADC5/ PA4/ADC4/
Name PCINT7 PCINT6 PCINT5 PCINT4
PUOE 0 0 0 0
PUOV 0 0 0 0
DDOE 0 0 0 0
DDOV 0 0 0 0
PVOE 0 0 0 0
PVOV 0 0 0 0
PCINT7 • PCIE0 + PCINT6 • PCIE0 + PCINT5 • PCIE0 + PCINT4 • PCIE0 +
DIEOE
ADC7D ADC6D ADC5D ADC4D
DIEOV PCINT7 • PCIE0 PCINT6 • PCIE0 PCINT5 • PCIE0 PCINT4 • PCIE0
DI PCINT7 INPUT PCINT6 INPUT PCINT5 INPUT PCINT4 INPUT
AIO ADC7 INPUT ADC6 INPUT ADC5 INPUT ADC4 INPUT
Table 12-5. Overriding Signals for Alternate Functions in PA3:PA0

Signal PA3/ADC3/ PA2/ADC2/ PA1/ADC1/ PA0/ADC0/
PUOE 0 0 0 0
PUOV 0 0 0 0
DDOE 0 0 0 0
DDOV 0 0 0 0
PVOE 0 0 0 0
PVOV 0 0 0 0
PCINT3 • PCIE0 + PCINT2 • PCIE0 + PCINT1 • PCIE0 + PCINT0 • PCIE0 +
DIEOE
ADC3D ADC2D ADC1D ADC0D
DIEOV PCINT3 • PCIE0 PCINT2 • PCIE0 PCINT1 • PCIE0 PCINT0 • PCIE0
AIO ADC3 INPUT ADC2 INPUT ADC1 INPUT ADC0 INPUT
80 ATmega164P/324P/644P
8011D–AVR–02/07
12.3.2 Alternate Functions of Port B
The Port B pins with alternate functions are shown in Table 12-6.
Table 12-6. Port B Pins Alternate Functions

Port Pin Alternate Functions
SCK (SPI Bus Master clock input)
PB7
MISO (SPI Bus Master Input/Slave Output)
PB6
MOSI (SPI Bus Master Output/Slave Input)
PB5
SS (SPI Slave Select input)
PB4 OC0B (Timer/Conter 0 Output Compare Match B Output)
AIN1 (Analog Comparator Negative Input)
PB3 OC0A (Timer/Conter 0 Output Compare Match A Output)
AIN0 (Analog Comparator Positive Input)
PB2 INT2 (External Interrupt 2 Input)
T1 (Timer/Counter 1 External Counter Input)
PB1 CLKO (Divided System Clock Output)
T0 (Timer/Counter 0 External Counter Input)
PB0 XCK0 (USART0 External Clock Input/Output)
The alternate pin configuration is as follows:
• SCK/PCINT15 – Port B, Bit 7

SCK: Master Clock output, Slave Clock input pin for SPI channel. When the SPI is enabled as a
slave, this pin is configured as an input regardless of the setting of DDB7. When the SPI0 is
enabled as a master, the data direction of this pin is controlled by DDB7. When the pin is forced
to be an input, the pull-up can still be controlled by the PORTB7 bit.
PCINT15, Pin Change Interrupt source 15: The PB7 pin can serve as an external interrupt
source.
• MISO/PCINT14 – Port B, Bit 6

MISO: Master Data input, Slave Data output pin for SPI channel. When the SPI is enabled as a
master, this pin is configured as an input regardless of the setting of DDB6. When the SPI is
enabled as a slave, the data direction of this pin is controlled by DDB6. When the pin is forced to
be an input, the pull-up can still be controlled by the PORTB6 bit.
source.
81
8011D–AVR–02/07
• MOSI/PCINT13 – Port B, Bit 5
MOSI: SPI Master Data output, Slave Data input for SPI channel. When the SPI is enabled as a
slave, this pin is configured as an input regardless of the setting of DDB5. When the SPI is
enabled as a master, the data direction of this pin is controlled by DDB5. When the pin is forced
to be an input, the pull-up can still be controlled by the PORTB5 bit.
source.
• SS/OC0B/PCINT12 – Port B, Bit 4

SS: Slave Port Select input. When the SPI is enabled as a slave, this pin is configured as an
input regardless of the setting of DDB4. As a slave, the SPI is activated when this pin is driven
low. When the SPI is enabled as a master, the data direction of this pin is controlled by DDB4.
When the pin is forced to be an input, the pull-up can still be controlled by the PORTB4 bit.
OC0B, Output Compare Match B output: The PB4 pin can serve as an external output for the
Timer/Counter0 Output Compare. The pin has to be configured as an output (DDB4 set “one”) to
serve this function. The OC0B pin is also the output pin for the PWM mode timer function.
source.
• AIN1/OC0A/PCINT11, Bit 3
AIN1, Analog Comparator Negative input. This pin is directly connected to the negative input of
the Analog Comparator.
OC0A, Output Compare Match A output: The PB3 pin can serve as an external output for the
Timer/Counter0 Output Compare. The pin has to be configured as an output (DDB3 set “one”) to
serve this function. The OC0A pin is also the output pin for the PWM mode timer function.
source.
• AIN0/INT2/PCINT10, Bit 2
AIN0, Analog Comparator Positive input. This pin is directly connected to the positive input of
the Analog Comparator.
INT2, External Interrupt source 2. The PB2 pin can serve as an External Interrupt source to the
MCU.
source.
• T1/CLKO/PCINT9, Bit 1
T1, Timer/Counter1 counter source.
CLKO, Divided System Clock: The divided system clock can be output on the PB1 pin. The
divided system clock will be output if the CKOUT Fuse is programmed, regardless of the
PORTB1 and DDB1 settings. It will also be output during reset.
PCINT9, Pin Change Interrupt source 9: The PB1 pin can serve as an external interrupt source.
82 ATmega164P/324P/644P
8011D–AVR–02/07
• T0/XCK0/PCINT8, Bit 0
T0, Timer/Counter0 counter source.
XCK0, USART0 External clock. The Data Direction Register (DDB0) controls whether the clock
is output (DDD0 set “one”) or input (DDD0 cleared). The XCK0 pin is active only when the
USART0 operates in Synchronous mode.
PCINT8, Pin Change Interrupt source 8: The PB0 pin can serve as an external interrupt source.
Table 12-7 and Table 12-8 relate the alternate functions of Port B to the overriding signals
shown in Figure 12-5 on page 77. SPI MSTR INPUT and SPI SLAVE OUTPUT constitute the
MISO signal, while MOSI is divided into SPI MSTR OUTPUT and SPI SLAVE INPUT. .
Table 12-7. Overriding Signals for Alternate Functions in PB7:PB4

Signal PB7/SCK/ PB6/MISO/ PB5/MOSI/ PB4/SS/OC0B/
PUOE SPE • MSTR SPE • MSTR SPE • MSTR SPE • MSTR
PUOV PORTB7 • PUD PORTB14 • PUD PORTB13 • PUD PORTB12 • PUD
DDOE SPE • MSTR SPE • MSTR SPE • MSTR SPE • MSTR
DDOV 0 0 0 0
PVOE SPE • MSTR SPE • MSTR SPE • MSTR OC0A ENABLE
SPI SLAVE
PVOV SCK OUTPUT SPI MSTR OUTPUT OC0A
OUTPUT
DIEOE PCINT15 • PCIE1 PCINT14 • PCIE1 PCINT13 • PCIE1 PCINT4 • PCIE1
DIEOV 1 1 1 1
SCK INPUT SPI MSTR INPUT SPI SLAVE INPUT SPI SS
DI
PCINT17 INPUT PCINT14 INPUT PCINT13 INPUT PCINT12 INPUT
AIO – – – –
Table 12-8. Overriding Signals for Alternate Functions in PB3:PB0

Signal PB3/AIN1/OC0B/ PB2/AIN0/INT2/ PB1/T1/CLKO/PCIN PB0/T0/XCK/
Name PCINT11 PCINT10 T9 PCINT8
PUOE 0 0 0 0
PUOV 0 0 0 0
DDOE 0 0 0 0
DDOV 0 0 0 0
PVOE OC0B ENABLE 0 0 0
PVOV OC0B 0 0 0
INT2 ENABLE
DIEOE PCINT11 • PCIE1 PCINT9 • PCIE1 PCINT8 • PCIE1
PCINT10 • PCIE1
DIEOV 1 1 1 1
INT2 INPUT T1 INPUT T0 INPUT
DI PCINT11 INPUT
PCINT10 INPUT PCINT9 INPUT PCINT8 INPUT
AIO – – – –
83
8011D–AVR–02/07
12.3.3 Alternate Functions of Port C
The Port C alternate function is as follows:
Table 12-9. Port C Pins Alternate Functions

TOSC2 (Timer Oscillator pin 2)
PC7
TOSC1 (Timer Oscillator pin 1)
PC6
TDI (JTAG Test Data Input)
PC5
TDO (JTAG Test Data Output)
PC4
TMS (JTAG Test Mode Select)
PC3
TCK (JTAG Test Clock)
PC2
SDA (2-wire Serial Bus Data Input/Output Line)
PC1
SCL (2-wire Serial Bus Clock Line)
PC0
• TOSC2/PCINT23 – Port C, Bit7

TOSC2, Timer Oscillator pin 2. The PC7 pin can serve as an external interrupt source to the
MCU.
PCINT23, Pin Change Interrupt source 23: The PC7 pin can serve as an external interrupt
source.
• TOSC1/PCINT22 – Port C, Bit 6

TOSC1, Timer Oscillator pin 1. The PC6 pin can serve as an external interrupt source to the
MCU.
source.
• TDI/PCINT21 – Port C, Bit 5

TDI, JTAG Test Data Input.
source.
• TDO/PCINT20 – Port C, Bit 4

TDO, JTAG Test Data Output.
source.
84 ATmega164P/324P/644P
8011D–AVR–02/07
• TMS/PCINT19 – Port C, Bit 3

TMS, JTAG Test Mode Select.
source.
• TCK/PCINT18 – Port C, Bit 2

TCK, JTAG Test Clock.
source.
• SDA/PCINT17 – Port C, Bit 1

SDA, 2-wire Serial Bus Data Input/Output Line.
source.
• SCL/PCINT16 – Port C, Bit 0

SCL, 2-wire Serial Busk Clock Line.
source.
Table 12-10 and Table 12-11 relate the alternate functions of Port C to the overriding signals
shown in Figure 12-5 on page 77.
Table 12-10. Overriding Signals for Alternate Functions in PC7:PC4

Signal PC7/TOSC2/ PC6/TOSC1/ PC5/TDI/ PC4/TDO/
PUOE AS2 • EXCLK AS2 JTAGEN JTAGEN
PUOV 0 0 1 1
DDOE AS2 • EXCLK AS2 JTAGEN JTAGEN
SHIFT_IR +
DDOV 0 0 0
SHIFT_DR
PVOE 0 0 0 JTAGEN
PVOV 0 0 0 TDO
AS2 • EXCLK +
DIEOE AS2 JTAGEN JTAGEN
PCINT23 • PCIE2
DIEOV 0 EXCLK 0 0
T/C2 OSC
AIO T/C2 OSC OUTPUT TDI INPUT –
INPUT
85
8011D–AVR–02/07
Table 12-11. Overriding Signals for Alternate Functions in PC3:PC0
Signal PC3/TMS/ PC2/TCK/ PC1/SDA/ PC0/SCL/
PUOE JTAGEN JTAGEN TWEN TWEN
PUOV 1 1 PORTC1 • PUD PORTC0 • PUD
DDOE JTAGEN JTAGEN TWEN TWEN
DDOV 0 0 0 0
PVOE 0 0 TWEN TWEN
PVOV 0 0 SDA OUT SCL OUT
DIEOE JTAGEN JTAGEN PCINT17 • PCIE2 PCINT16 • PCIE2
DIEOV 1 1 1 1
AIO TMS INPUT TCK INPUT SDA INPUT SCL INPUT
12.3.4 Alternate Functions of Port D
The Port D pins with alternate functions are shown in Table 12-12.
Table 12-12. Port D Pins Alternate Functions

OC2A (Timer/Counter2 Output Compare Match A Output)
PD7
ICP1 (Timer/Counter1 Input Capture Trigger)
PD6 OC2B (Timer/Counter2 Output Compare Match B Output)
OC1A (Timer/Counter1 Output Compare Match A Output)
PD5
OC1B (Timer/Counter1 Output Compare Match B Output)
PD4 XCK1 (USART1 External Clock Input/Output)
INT1 (External Interrupt1 Input)
PD3 TXD1 (USART1 Transmit Pin)
INT0 (External Interrupt0 Input)
PD2 RXD1 (USART1 Receive Pin)
TXD0 (USART0 Transmit Pin)
PD1
RXD0 (USART0 Receive Pin)
PD0
86 ATmega164P/324P/644P
8011D–AVR–02/07
The alternate pin configuration is as follows:
• OC2A/PCINT31 – Port D, Bit 7

OC2A, Output Compare Match A output: The PD7 pin can serve as an external output for the
Timer/Counter2 Output Compare A. The pin has to be configured as an output (DDD7 set (one))
to serve this function. The OC2A pin is also the output pin for the PWM mode timer function.
PCINT31, Pin Change Interrupt Source 31:The PD7 pin can serve as an external interrupt
source.
• ICP1/OC2B/PCINT30 – Port D, Bit 6

ICP1, Input Capture Pin 1: The PD6 pin can act as an input capture pin for Timer/Counter1.
OC2B, Output Compare Match B output: The PD6 pin can serve as an external output for the
Timer/Counter2 Output Compare B. The pin has to be configured as an output (DDD6 set (one))
to serve this function. The OC2B pin is also the output pin for the PWM mode timer function.
PCINT30, Pin Change Interrupt Source 30: The PD6 pin can serve as an external interrupt
source.
• OC1A/PCINT29 – Port D, Bit 5

OC1A, Output Compare Match A output: The PD5 pin can serve as an external output for the
Timer/Counter1 Output Compare A. The pin has to be configured as an output (DDD5 set (one))
to serve this function. The OC1A pin is also the output pin for the PWM mode timer function.
source.
• OC1B/XCK1/PCINT28 – Port D, Bit 4

OC1B, Output Compare Match B output: The PB4 pin can serve as an external output for the
Timer/Counter1 Output Compare B. The pin has to be configured as an output (DDD4 set (one))
to serve this function. The OC1B pin is also the output pin for the PWM mode timer function.
XCK1, USART1 External clock. The Data Direction Register (DDB4) controls whether the clock
is output (DDD4 set “one”) or input (DDD4 cleared). The XCK4 pin is active only when the
USART1 operates in Synchronous mode.
source.
• INT1/TXD1/PCINT27 – Port D, Bit 3

INT1, External Interrupt source 1. The PD3 pin can serve as an external interrupt source to the
MCU.
TXD1, Transmit Data (Data output pin for the USART1). When the USART1 Transmitter is
enabled, this pin is configured as an output regardless of the value of DDD3.
source.
87
8011D–AVR–02/07
• INT0/RXD1/PCINT26 – Port D, Bit 2
INT0, External Interrupt source 0. The PD2 pin can serve as an external interrupt source to the
MCU.
RXD1, RXD0, Receive Data (Data input pin for the USART1). When the USART1 receiver is
enabled this pin is configured as an input regardless of the value of DDD2. When the USART
forces this pin to be an input, the pull-up can still be controlled by the PORTD2 bit.
source.
• TXD0/PCINT25 – Port D, Bit 1

TXD0, Transmit Data (Data output pin for the USART0). When the USART0 Transmitter is
enabled, this pin is configured as an output regardless of the value of DDD1.
source.
• RXD0/PCINT24 – Port D, Bit 0

RXD0, Receive Data (Data input pin for the USART0). When the USART0 receiver is enabled
this pin is configured as an input regardless of the value of DDD0. When the USART forces this
pin to be an input, the pull-up can still be controlled by the PORTD0 bit.
source.
Table 12-13 on page 88 and Table 12-14 on page 89 relates the alternate functions of Port D to
the overriding signals shown in Figure 12-5 on page 77.
Table 12-13. Overriding Signals for Alternate Functions PD7:PD4

PD6/ICP1/
PD7/OC2A/ OC2B/ PD5/OC1A/ PD4/OC1B/XCK1/
Signal Name PCINT31 PCINT30 PCINT29 PCINT28
PUOE 0 0 0 0
PUOV 0 0 0 0
DDOE 0 0 0 0
DDOV 0 0 0 0
PVOE OC2A ENABLE OC2B ENABLE OC1A ENABLE OC1B ENABLE
PVOV OCA2A OC2B OC1A OC1B
DIEOE PCINT31 • PCIE3 PCINT30 • PCIE3 PCINT29 • PCIE3 PCINT28 • PCIE3
DIEOV 1 1 1 1
ICP1 INPUT
DI PCINT31 INPUT PCINT29 INPUT PCINT28 INPUT
PCINT30 INPUT
AIO – – – –
88 ATmega164P/324P/644P
8011D–AVR–02/07
Table 12-14. Overriding Signals for Alternate Functions in PD3:PD0(1)

PD3/INT1/TXD1/ PD2/INT0/RXD1/ PD1/TXD0/ PD0/RXD0/
Signal Name PCINT27 PCINT26 PCINT25 PCINT27
PUOE 0 TXEN RXEN
PUOV 0 PORTD2 • PUD PORTD1 • PUD PORTD0 • PUD
DDOE 0 RXEN1 TXEN RXEN
DDOV 0 0 SDA_OUT SCL_OUT
PVOE 0 0 TWEN TWEN
PVOV 0 0 0 0
INT1 ENABLE INT2 ENABLE INT1 ENABLE INT0 ENABLE
DIEOE
PCINT27 • PCIE3 PCINT26 • PCIE3 PCINT25 • PCIE3 PCINT24 • PCIE3
DIEOV 1 1 1 1
INT1 INPUT INT0 INPUT TXD RXD
DI
PCINT27 INPUT PCINT27 INPUT PCINT25 INPUT PCINT24 INPUT
AIO – – – –
Note: 1. When enabled, the 2-wire Serial Interface enables Slew-Rate controls on the output pins PD0
and PD1. This is not shown in this table. In addition, spike filters are connected between the
AIO outputs shown in the port figure and the digital logic of the TWI module.
89
8011D–AVR–02/07
Register Description
Bit 7 6 5 4 3 2 1 0
• Bit 4 – PUD: Pull-up Disable

When this bit is written to one, the pull-ups in the I/O ports are disabled even if the DDxn and
PORTxn Registers are configured to enable the pull-ups ({DDxn, PORTxn} = 0b01). See ”Con-
figuring the Pin” on page 72 for more details about this feature.
12.3.6 PORTA – Port A Data Register
Bit 7 6 5 4 3 2 1 0
0x02 (0x22) PORTA7 PORTA6 PORTA5 PORTA4 PORTA3 PORTA2 PORTA1 PORTA0 PORTA
12.3.7 DDRA – Port A Data Direction Register
Bit 7 6 5 4 3 2 1 0
0x01 (0x21) DDA7 DDA6 DDA5 DDA4 DDA3 DDA2 DDA1 DDA0 DDRA
12.3.8 PINA – Port A Input Pins Address
Bit 7 6 5 4 3 2 1 0
0x00 (0x20) PINA7 PINA6 PINA5 PINA4 PINA3 PINA2 PINA1 PINA0 PINA
Initial Value N/A N/A N/A N/A N/A N/A N/A N/A
12.3.9 PORTB – Port B Data Register
Bit 7 6 5 4 3 2 1 0
0x05 (0x25) PORTB7 PORTB6 PORTB5 PORTB4 PORTB3 PORTB2 PORTB1 PORTB0 PORTB
12.3.10 DDRB – Port B Data Direction Register
Bit 7 6 5 4 3 2 1 0
0x04 (0x24) DDB7 DDB6 DDB5 DDB4 DDB3 DDB2 DDB1 DDB0 DDRB
12.3.11 PINB – Port B Input Pins Address
Bit 7 6 5 4 3 2 1 0
0x03 (0x23) PINB7 PINB6 PINB5 PINB4 PINB3 PINB2 PINB1 PINB0 PINB
90 ATmega164P/324P/644P
8011D–AVR–02/07
12.3.12 PORTC – Port C Data Register
Bit 7 6 5 4 3 2 1 0
0x08 (0x28) PORTC7 PORTC6 PORTC5 PORTC4 PORTC3 PORTC2 PORTC1 PORTC0 PORTC
12.3.13 DDRC – Port C Data Direction Register
Bit 7 6 5 4 3 2 1 0
0x07 (0x27) DDC7 DDC6 DDC5 DDC4 DDC3 DDC2 DDC1 DDC0 DDRC
12.3.14 PINC – Port C Input Pins Address
Bit 7 6 5 4 3 2 1 0
0x06 (0x26) PINC7 PINC6 PINC5 PINC4 PINC3 PINC2 PINC1 PINC0 PINC
12.3.15 PORTD – Port D Data Register
Bit 7 6 5 4 3 2 1 0
0x0B (0x2B) PORTD7 PORTD6 PORTD5 PORTD4 PORTD3 PORTD2 PORTD1 PORTD0 PORTD
12.3.16 DDRD – Port D Data Direction Register
Bit 7 6 5 4 3 2 1 0
0x0A (0x2A) DDD7 DDD6 DDD5 DDD4 DDD3 DDD2 DDD1 DDD0 DDRD
12.3.17 PIND – Port D Input Pins Address
Bit 7 6 5 4 3 2 1 0
0x09 (0x29) PIND7 PIND6 PIND5 PIND4 PIND3 PIND2 PIND1 PIND0 PIND
91
8011D–AVR–02/07
13. 8-bit Timer/Counter0 with PWM
13.1 Features
• Two Independent Output Compare Units
• Double Buffered Output Compare Registers
• Clear Timer on Compare Match (Auto Reload)
• Glitch Free, Phase Correct Pulse Width Modulator (PWM)
• Variable PWM Period
• Frequency Generator
• Three Independent Interrupt Sources (TOV0, OCF0A, and OCF0B)
13.2 Overview
Timer/Counter0 is a general purpose 8-bit Timer/Counter module, with two independent Output
Compare Units, and with PWM support. It allows accurate program execution timing (event man-
agement) and wave generation.
A simplified block diagram of the 8-bit Timer/Counter is shown in Figure 13-1. For the actual
placement of I/O pins, see ”Pin Configurations” on page 2. CPU accessible I/O Registers, includ-
ing I/O bits and I/O pins, are shown in bold. The device-specific I/O Register and bit locations
are listed in the ”Register Description” on page 103.
Figure 13-1. 8-bit Timer/Counter Block Diagram

Count TOVn
Clear (Int.Req.)
Control Logic
Direction clkTn Clock Select
Edge
Tn
Detector
TOP BOTTOM
( From Prescaler )
Timer/Counter
TCNTn
= =0
OCnA
(Int.Req.)
Waveform
= Generation
OCnA
OCRnA
Fixed
OCnB
TOP
(Int.Req.)
Value
DATA BUS
Waveform
= Generation
OCnB
OCRnB
TCCRnA TCCRnB
13.2.1 Registers
The Timer/Counter (TCNT0) and Output Compare Registers (OCR0A and OCR0B) are 8-bit
registers. Interrupt request (abbreviated to Int.Req. in the figure) signals are all visible in the
Timer Interrupt Flag Register (TIFR0). All interrupts are individually masked with the Timer Inter-
rupt Mask Register (TIMSK0). TIFR0 and TIMSK0 are not shown in the figure.
The Timer/Counter can be clocked internally, via the prescaler, or by an external clock source on
the T0 pin. The Clock Select logic block controls which clock source and edge the Timer/Counter
92 ATmega164P/324P/644P
8011D–AVR–02/07
uses to increment (or decrement) its value. The Timer/Counter is inactive when no clock source
is selected. The output from the Clock Select logic is referred to as the timer clock (clkT0).
The double buffered Output Compare Registers (OCR0A and OCR0B) are compared with the
Timer/Counter value at all times. The result of the compare can be used by the Waveform Gen-
erator to generate a PWM or variable frequency output on the Output Compare pins (OC0A and
OC0B). See Section “13.5” on page 94. for details. The Compare Match event will also set the
Compare Flag (OCF0A or OCF0B) which can be used to generate an Output Compare interrupt
request.
13.2.2 Definitions
Many register and bit references in this section are written in general form. A lower case “n”
replaces the Timer/Counter number, in this case 0. A lower case “x” replaces the Output Com-
pare Unit, in this case Compare Unit A or Compare Unit B. However, when using the register or
bit defines in a program, the precise form must be used, i.e., TCNT0 for accessing
Timer/Counter0 counter value and so on.
The definitions in Table 13-1 are also used extensively throughout the document.
Table 13-1. Definitions

BOTTOM The counter reaches the BOTTOM when it becomes 0x00.
MAX The counter reaches its MAXimum when it becomes 0xFF (decimal 255).
TOP The counter reaches the TOP when it becomes equal to the highest value in the
count sequence. The TOP value can be assigned to be the fixed value 0xFF
(MAX) or the value stored in the OCR0A Register. The assignment is depen-
dent on the mode of operation.
13.3 Timer/Counter Clock Sources

The Timer/Counter can be clocked by an internal or an external clock source. The clock source
is selected by the Clock Select logic which is controlled by the Clock Select (CS02:0) bits
located in the Timer/Counter Control Register (TCCR0B). For details on clock sources and pres-
caler, see ”Timer/Counter Prescaler” on page 151.
13.4 Counter Unit

The main part of the 8-bit Timer/Counter is the programmable bi-directional counter unit. Figure
13-2 shows a block diagram of the counter and its surroundings.
Figure 13-2. Counter Unit Block Diagram

TOVn
DATA BUS (Int.Req.)
Clock Select
count Edge
Tn
clear clkTn Detector
TCNTn Control Logic
direction
( From Prescaler )
bottom top
Signal description (internal signals):
93
8011D–AVR–02/07
count Increment or decrement TCNT0 by 1.
direction Select between increment and decrement.
clear Clear TCNT0 (set all bits to zero).
clkTn Timer/Counter clock, referred to as clkT0 in the following.
top Signalize that TCNT0 has reached maximum value.
bottom Signalize that TCNT0 has reached minimum value (zero).
Depending of the mode of operation used, the counter is cleared, incremented, or decremented
at each timer clock (clkT0). clkT0 can be generated from an external or internal clock source,
selected by the Clock Select bits (CS02:0). When no clock source is selected (CS02:0 = 0) the
timer is stopped. However, the TCNT0 value can be accessed by the CPU, regardless of
whether clkT0 is present or not. A CPU write overrides (has priority over) all counter clear or
count operations.
The counting sequence is determined by the setting of the WGM01 and WGM00 bits located in
the Timer/Counter Control Register (TCCR0A) and the WGM02 bit located in the Timer/Counter
Control Register B (TCCR0B). There are close connections between how the counter behaves
(counts) and how waveforms are generated on the Output Compare outputs OC0A and OC0B.
For more details about advanced counting sequences and waveform generation, see ”Modes of
Operation” on page 97.
The Timer/Counter Overflow Flag (TOV0) is set according to the mode of operation selected by
the WGM02:0 bits. TOV0 can be used for generating a CPU interrupt.
13.5 Output Compare Unit

The 8-bit comparator continuously compares TCNT0 with the Output Compare Registers
(OCR0A and OCR0B). Whenever TCNT0 equals OCR0A or OCR0B, the comparator signals a
match. A match will set the Output Compare Flag (OCF0A or OCF0B) at the next timer clock
cycle. If the corresponding interrupt is enabled, the Output Compare Flag generates an Output
Compare interrupt. The Output Compare Flag is automatically cleared when the interrupt is exe-
cuted. Alternatively, the flag can be cleared by software by writing a logical one to its I/O bit
location. The Waveform Generator uses the match signal to generate an output according to
operating mode set by the WGM02:0 bits and Compare Output mode (COM0x1:0) bits. The max
and bottom signals are used by the Waveform Generator for handling the special cases of the
extreme values in some modes of operation (”Modes of Operation” on page 97).
Figure 13-3 shows a block diagram of the Output Compare unit.
94 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 13-3. Output Compare Unit, Block Diagram
DATA BUS
OCRnx TCNTn
= (8-bit Comparator )
OCFnx (Int.Req.)
top
bottom
Waveform Generator OCnx
FOCn
WGMn1:0 COMnX1:0
The OCR0x Registers are double buffered when using any of the Pulse Width Modulation
(PWM) modes. For the normal and Clear Timer on Compare (CTC) modes of operation, the dou-
ble buffering is disabled. The double buffering synchronizes the update of the OCR0x Compare
Registers to either top or bottom of the counting sequence. The synchronization prevents the
occurrence of odd-length, non-symmetrical PWM pulses, thereby making the output glitch-free.
The OCR0x Register access may seem complex, but this is not case. When the double buffering
is enabled, the CPU has access to the OCR0x Buffer Register, and if double buffering is dis-
abled the CPU will access the OCR0x directly.
13.5.1 Force Output Compare
In non-PWM waveform generation modes, the match output of the comparator can be forced by
writing a one to the Force Output Compare (FOC0x) bit. Forcing Compare Match will not set the
OCF0x Flag or reload/clear the timer, but the OC0x pin will be updated as if a real Compare
Match had occurred (the COM0x1:0 bits settings define whether the OC0x pin is set, cleared or
toggled).
13.5.2 Compare Match Blocking by TCNT0 Write
All CPU write operations to the TCNT0 Register will block any Compare Match that occur in the
next timer clock cycle, even when the timer is stopped. This feature allows OCR0x to be initial-
ized to the same value as TCNT0 without triggering an interrupt when the Timer/Counter clock is
enabled.
13.5.3 Using the Output Compare Unit
Since writing TCNT0 in any mode of operation will block all Compare Matches for one timer
clock cycle, there are risks involved when changing TCNT0 when using the Output Compare
Unit, independently of whether the Timer/Counter is running or not. If the value written to TCNT0
equals the OCR0x value, the Compare Match will be missed, resulting in incorrect waveform
95
8011D–AVR–02/07
generation. Similarly, do not write the TCNT0 value equal to BOTTOM when the counter is
down-counting.
The setup of the OC0x should be performed before setting the Data Direction Register for the
port pin to output. The easiest way of setting the OC0x value is to use the Force Output Com-
pare (FOC0x) strobe bits in Normal mode. The OC0x Registers keep their values even when
changing between Waveform Generation modes.
Be aware that the COM0x1:0 bits are not double buffered together with the compare value.
Changing the COM0x1:0 bits will take effect immediately.
13.6 Compare Match Output Unit

The Compare Output mode (COM0x1:0) bits have two functions. The Waveform Generator uses
the COM0x1:0 bits for defining the Output Compare (OC0x) state at the next Compare Match.
Also, the COM0x1:0 bits control the OC0x pin output source. Figure 13-4 shows a simplified
schematic of the logic affected by the COM0x1:0 bit setting. The I/O Registers, I/O bits, and I/O
pins in the figure are shown in bold. Only the parts of the general I/O Port Control Registers
(DDR and PORT) that are affected by the COM0x1:0 bits are shown. When referring to the
OC0x state, the reference is for the internal OC0x Register, not the OC0x pin. If a system reset
occur, the OC0x Register is reset to “0”.
Figure 13-4. Compare Match Output Unit, Schematic
COMnx1
COMnx0 Waveform
D Q
FOCn Generator
1
OCnx
OCnx Pin
0
D Q
DATA BUS
PORT
D Q
DDR
clk I/O
The general I/O port function is overridden by the Output Compare (OC0x) from the Waveform
Generator if either of the COM0x1:0 bits are set. However, the OC0x pin direction (input or out-
put) is still controlled by the Data Direction Register (DDR) for the port pin. The Data Direction
Register bit for the OC0x pin (DDR_OC0x) must be set as output before the OC0x value is visi-
ble on the pin. The port override function is independent of the Waveform Generation mode.
The design of the Output Compare pin logic allows initialization of the OC0x state before the out-
put is enabled. Note that some COM0x1:0 bit settings are reserved for certain modes of
operation. See Section “13.9” on page 103.
96 ATmega164P/324P/644P
8011D–AVR–02/07
13.6.1 Compare Output Mode and Waveform Generation
The Waveform Generator uses the COM0x1:0 bits differently in Normal, CTC, and PWM modes.
For all modes, setting the COM0x1:0 = 0 tells the Waveform Generator that no action on the
OC0x Register is to be performed on the next Compare Match. For compare output actions in
the non-PWM modes refer to Table 13-2 on page 103. For fast PWM mode, refer to Table 13-3
on page 103, and for phase correct PWM refer to Table 13-4 on page 104.
A change of the COM0x1:0 bits state will have effect at the first Compare Match after the bits are
written. For non-PWM modes, the action can be forced to have immediate effect by using the
FOC0x strobe bits.
13.7 Modes of Operation

The mode of operation, i.e., the behavior of the Timer/Counter and the Output Compare pins, is
defined by the combination of the Waveform Generation mode (WGM02:0) and Compare Output
mode (COM0x1:0) bits. The Compare Output mode bits do not affect the counting sequence,
while the Waveform Generation mode bits do. The COM0x1:0 bits control whether the PWM out-
put generated should be inverted or not (inverted or non-inverted PWM). For non-PWM modes
the COM0x1:0 bits control whether the output should be set, cleared, or toggled at a Compare
Match (See Section “14.8” on page 121.).
For detailed timing information see ”Timer/Counter Timing Diagrams” on page 101.
13.7.1 Normal Mode
The simplest mode of operation is the Normal mode (WGM02:0 = 0). In this mode the counting
direction is always up (incrementing), and no counter clear is performed. The counter simply
overruns when it passes its maximum 8-bit value (TOP = 0xFF) and then restarts from the bot-
tom (0x00). In normal operation the Timer/Counter Overflow Flag (TOV0) will be set in the same
timer clock cycle as the TCNT0 becomes zero. The TOV0 Flag in this case behaves like a ninth
bit, except that it is only set, not cleared. However, combined with the timer overflow interrupt
that automatically clears the TOV0 Flag, the timer resolution can be increased by software.
There are no special cases to consider in the Normal mode, a new counter value can be written
anytime.
The Output Compare Unit can be used to generate interrupts at some given time. Using the Out-
put Compare to generate waveforms in Normal mode is not recommended, since this will
occupy too much of the CPU time.
13.7.2 Clear Timer on Compare Match (CTC) Mode
In Clear Timer on Compare or CTC mode (WGM02:0 = 2), the OCR0A Register is used to
manipulate the counter resolution. In CTC mode the counter is cleared to zero when the counter
value (TCNT0) matches the OCR0A. The OCR0A defines the top value for the counter, hence
also its resolution. This mode allows greater control of the Compare Match output frequency. It
also simplifies the operation of counting external events.
The timing diagram for the CTC mode is shown in Figure 13-5. The counter value (TCNT0)
increases until a Compare Match occurs between TCNT0 and OCR0A, and then counter
(TCNT0) is cleared.
97
8011D–AVR–02/07
Figure 13-5. CTC Mode, Timing Diagram
OCnx Interrupt Flag Set
TCNTn
OCn
(COMnx1:0 = 1)
(Toggle)
Period 1 2 3 4
An interrupt can be generated each time the counter value reaches the TOP value by using the
OCF0A Flag. If the interrupt is enabled, the interrupt handler routine can be used for updating
the TOP value. However, changing TOP to a value close to BOTTOM when the counter is run-
ning with none or a low prescaler value must be done with care since the CTC mode does not
have the double buffering feature. If the new value written to OCR0A is lower than the current
value of TCNT0, the counter will miss the Compare Match. The counter will then have to count to
its maximum value (0xFF) and wrap around starting at 0x00 before the Compare Match can
occur.
For generating a waveform output in CTC mode, the OC0A output can be set to toggle its logical
level on each Compare Match by setting the Compare Output mode bits to toggle mode
(COM0A1:0 = 1). The OC0A value will not be visible on the port pin unless the data direction for
the pin is set to output. The waveform generated will have a maximum frequency of fOC0 =
fclk_I/O/2 when OCR0A is set to zero (0x00). The waveform frequency is defined by the following
equation:
f clk_I/O
f OCnx = -------------------------------------------------
-
2 ⋅ N ⋅ ( 1 + OCRnx )
The N variable represents the prescale factor (1, 8, 64, 256, or 1024).
As for the Normal mode of operation, the TOV0 Flag is set in the same timer clock cycle that the
counter counts from MAX to 0x00.
13.7.3 Fast PWM Mode
The fast Pulse Width Modulation or fast PWM mode (WGM02:0 = 3 or 7) provides a high fre-
quency PWM waveform generation option. The fast PWM differs from the other PWM option by
its single-slope operation. The counter counts from BOTTOM to TOP then restarts from BOT-
TOM. TOP is defined as 0xFF when WGM2:0 = 3, and OCR0A when WGM2:0 = 7. In non-
inverting Compare Output mode, the Output Compare (OC0x) is cleared on the Compare Match
between TCNT0 and OCR0x, and set at BOTTOM. In inverting Compare Output mode, the out-
put is set on Compare Match and cleared at BOTTOM. Due to the single-slope operation, the
operating frequency of the fast PWM mode can be twice as high as the phase correct PWM
mode that use dual-slope operation. This high frequency makes the fast PWM mode well suited
for power regulation, rectification, and DAC applications. High frequency allows physically small
sized external components (coils, capacitors), and therefore reduces total system cost.
In fast PWM mode, the counter is incremented until the counter value matches the TOP value.
The counter is then cleared at the following timer clock cycle. The timing diagram for the fast
98 ATmega164P/324P/644P
8011D–AVR–02/07
PWM mode is shown in Figure 13-6. The TCNT0 value is in the timing diagram shown as a his-
togram for illustrating the single-slope operation. The diagram includes non-inverted and
inverted PWM outputs. The small horizontal line marks on the TCNT0 slopes represent Com-
pare Matches between OCR0x and TCNT0.
Figure 13-6. Fast PWM Mode, Timing Diagram
OCRnx Interrupt Flag Set
OCRnx Update and

TOVn Interrupt Flag Set
TCNTn
OCnx (COMnx1:0 = 2)
OCnx (COMnx1:0 = 3)
Period 1 2 3 4 5 6 7
The Timer/Counter Overflow Flag (TOV0) is set each time the counter reaches TOP. If the inter-
rupt is enabled, the interrupt handler routine can be used for updating the compare value.
In fast PWM mode, the compare unit allows generation of PWM waveforms on the OC0x pins.
Setting the COM0x1:0 bits to two will produce a non-inverted PWM and an inverted PWM output
can be generated by setting the COM0x1:0 to three: Setting the COM0A1:0 bits to one allows
the OC0A pin to toggle on Compare Matches if the WGM02 bit is set. This option is not available
for the OC0B pin (See Table 13-3 on page 103). The actual OC0x value will only be visible on
the port pin if the data direction for the port pin is set as output. The PWM waveform is gener-
ated by setting (or clearing) the OC0x Register at the Compare Match between OCR0x and
TCNT0, and clearing (or setting) the OC0x Register at the timer clock cycle the counter is
cleared (changes from TOP to BOTTOM).
The PWM frequency for the output can be calculated by the following equation:
f clk_I/O
f OCnxPWM = -----------------
-
N ⋅ 256
The extreme values for the OCR0A Register represents special cases when generating a PWM
waveform output in the fast PWM mode. If the OCR0A is set equal to BOTTOM, the output will
be a narrow spike for each MAX+1 timer clock cycle. Setting the OCR0A equal to MAX will result
in a constantly high or low output (depending on the polarity of the output set by the COM0A1:0
bits.)
A frequency (with 50% duty cycle) waveform output in fast PWM mode can be achieved by set-
ting OC0x to toggle its logical level on each Compare Match (COM0x1:0 = 1). The waveform
generated will have a maximum frequency of fOC0 = fclk_I/O/2 when OCR0A is set to zero. This
99
8011D–AVR–02/07
feature is similar to the OC0A toggle in CTC mode, except the double buffer feature of the Out-
put Compare unit is enabled in the fast PWM mode.
13.7.4 Phase Correct PWM Mode
The phase correct PWM mode (WGM02:0 = 1 or 5) provides a high resolution phase correct
PWM waveform generation option. The phase correct PWM mode is based on a dual-slope
operation. The counter counts repeatedly from BOTTOM to TOP and then from TOP to BOT-
TOM. TOP is defined as 0xFF when WGM2:0 = 1, and OCR0A when WGM2:0 = 5. In non-
inverting Compare Output mode, the Output Compare (OC0x) is cleared on the Compare Match
between TCNT0 and OCR0x while upcounting, and set on the Compare Match while down-
counting. In inverting Output Compare mode, the operation is inverted. The dual-slope operation
has lower maximum operation frequency than single slope operation. However, due to the sym-
metric feature of the dual-slope PWM modes, these modes are preferred for motor control
applications.
In phase correct PWM mode the counter is incremented until the counter value matches TOP.
When the counter reaches TOP, it changes the count direction. The TCNT0 value will be equal
to TOP for one timer clock cycle. The timing diagram for the phase correct PWM mode is shown
on Figure 13-7. The TCNT0 value is in the timing diagram shown as a histogram for illustrating
the dual-slope operation. The diagram includes non-inverted and inverted PWM outputs. The
small horizontal line marks on the TCNT0 slopes represent Compare Matches between OCR0x
and TCNT0.
Figure 13-7. Phase Correct PWM Mode, Timing Diagram
OCRnx Update
TCNTn
OCnx (COMnx1:0 = 2)
OCnx (COMnx1:0 = 3)
Period 1 2 3
The Timer/Counter Overflow Flag (TOV0) is set each time the counter reaches BOTTOM. The
Interrupt Flag can be used to generate an interrupt each time the counter reaches the BOTTOM
value.
In phase correct PWM mode, the compare unit allows generation of PWM waveforms on the
OC0x pins. Setting the COM0x1:0 bits to two will produce a non-inverted PWM. An inverted
PWM output can be generated by setting the COM0x1:0 to three: Setting the COM0A0 bits to
100 ATmega164P/324P/644P
8011D–AVR–02/07
one allows the OC0A pin to toggle on Compare Matches if the WGM02 bit is set. This option is
not available for the OC0B pin (See Table 13-4 on page 104). The actual OC0x value will only
be visible on the port pin if the data direction for the port pin is set as output. The PWM wave-
form is generated by clearing (or setting) the OC0x Register at the Compare Match between
OCR0x and TCNT0 when the counter increments, and setting (or clearing) the OC0x Register at
Compare Match between OCR0x and TCNT0 when the counter decrements. The PWM fre-
quency for the output when using phase correct PWM can be calculated by the following
equation:
f clk_I/O
f OCnxPCPWM = -----------------
-
N ⋅ 510
The extreme values for the OCR0A Register represent special cases when generating a PWM
waveform output in the phase correct PWM mode. If the OCR0A is set equal to BOTTOM, the
output will be continuously low and if set equal to MAX the output will be continuously high for
non-inverted PWM mode. For inverted PWM the output will have the opposite logic values.
At the very start of period 2 in Figure 13-7 OCnx has a transition from high to low even though
there is no Compare Match. The point of this transition is to guarantee symmetry around BOT-
TOM. There are two cases that give a transition without Compare Match.
• OCR0A changes its value from MAX, like in Figure 13-7. When the OCR0A value is MAX the
OCn pin value is the same as the result of a down-counting Compare Match. To ensure
symmetry around BOTTOM the OCn value at MAX must correspond to the result of an up-
counting Compare Match.
• The timer starts counting from a value higher than the one in OCR0A, and for that reason
misses the Compare Match and hence the OCn change that would have happened on the way
up.
13.8 Timer/Counter Timing Diagrams
The Timer/Counter is a synchronous design and the timer clock (clkT0) is therefore shown as a
clock enable signal in the following figures. The figures include information on when Interrupt
Flags are set. Figure 13-8 contains timing data for basic Timer/Counter operation. The figure
shows the count sequence close to the MAX value in all modes other than phase correct PWM
mode.
Figure 13-8. Timer/Counter Timing Diagram, no Prescaling
clkI/O
clkTn
(clkI/O /1)
TCNTn MAX - 1 MAX BOTTOM BOTTOM + 1
TOVn
Figure 13-9 shows the same timing data, but with the prescaler enabled.
101
8011D–AVR–02/07
Figure 13-9. Timer/Counter Timing Diagram, with Prescaler (fclk_I/O/8)
clkI/O
clkTn
(clkI/O /8)
TOVn
Figure 13-10 shows the setting of OCF0B in all modes and OCF0A in all modes except CTC
mode and PWM mode, where OCR0A is TOP.
Figure 13-10. Timer/Counter Timing Diagram, Setting of OCF0x, with Prescaler (fclk_I/O/8)
clkI/O
clkTn
(clkI/O /8)
TCNTn OCRnx - 1 OCRnx OCRnx + 1 OCRnx + 2
OCRnx OCRnx Value
OCFnx
Figure 13-11 shows the setting of OCF0A and the clearing of TCNT0 in CTC mode and fast
PWM mode where OCR0A is TOP.
Figure 13-11. Timer/Counter Timing Diagram, Clear Timer on Compare Match mode, with Pres-
caler (fclk_I/O/8)
clkI/O
clkTn
(clkI/O /8)
TCNTn
TOP - 1 TOP BOTTOM BOTTOM + 1
(CTC)
OCRnx TOP
OCFnx
102 ATmega164P/324P/644P
8011D–AVR–02/07
13.9.1 TCCR0A – Timer/Counter Control Register A
Bit 7 6 5 4 3 2 1 0
0x24 (0x44) COM0A1 COM0A0 COM0B1 COM0B0 – – WGM01 WGM00 TCCR0A
Read/Write R/W R/W R/W R/W R R R/W R/W
• Bits 7:6 – COM0A1:0: Compare Match Output A Mode

These bits control the Output Compare pin (OC0A) behavior. If one or both of the COM0A1:0
bits are set, the OC0A output overrides the normal port functionality of the I/O pin it is connected
to. However, note that the Data Direction Register (DDR) bit corresponding to the OC0A pin
must be set in order to enable the output driver.
When OC0A is connected to the pin, the function of the COM0A1:0 bits depends on the
WGM02:0 bit setting. Table 13-2 shows the COM0A1:0 bit functionality when the WGM02:0 bits
are set to a normal or CTC mode (non-PWM).
Table 13-2. Compare Output Mode, non-PWM Mode

COM0A1 COM0A0 Description
0 0 Normal port operation, OC0A disconnected.
0 1 Toggle OC0A on Compare Match
1 0 Clear OC0A on Compare Match
1 1 Set OC0A on Compare Match
Table 13-3 shows the COM0A1:0 bit functionality when the WGM01:0 bits are set to fast PWM
mode.
Table 13-3. Compare Output Mode, Fast PWM Mode(1)

WGM02 = 0: Normal Port Operation, OC0A Disconnected.
0 1
WGM02 = 1: Toggle OC0A on Compare Match.
Clear OC0A on Compare Match, set OC0A at BOTTOM,
1 0
(non-inverting mode).
Set OC0A on Compare Match, clear OC0A at BOTTOM,
1 1
(inverting mode).
Note: 1. A special case occurs when OCR0A equals TOP and COM0A1 is set. In this case, the Com-
pare Match is ignored, but the set or clear is done at BOTTOM. See ”Fast PWM Mode” on
Table 13-4 on page 104 shows the COM0A1:0 bit functionality when the WGM02:0 bits are set
to phase correct PWM mode.
103
8011D–AVR–02/07
Table 13-4. Compare Output Mode, Phase Correct PWM Mode(1)
0 1
Clear OC0A on Compare Match when up-counting. Set OC0A on
1 0
Compare Match when down-counting.
Set OC0A on Compare Match when up-counting. Clear OC0A on
1 1
pare Match is ignored, but the set or clear is done at TOP. See ”Phase Correct PWM Mode” on
• Bits 5:4 – COM0B1:0: Compare Match Output B Mode

These bits control the Output Compare pin (OC0B) behavior. If one or both of the COM0B1:0
bits are set, the OC0B output overrides the normal port functionality of the I/O pin it is connected
to. However, note that the Data Direction Register (DDR) bit corresponding to the OC0B pin
When OC0B is connected to the pin, the function of the COM0B1:0 bits depends on the
WGM02:0 bit setting. Table 13-2 on page 103 shows the COM0A1:0 bit functionality when the
WGM02:0 bits are set to a normal or CTC mode (non-PWM).

COM0B1 COM0B0 Description
0 0 Normal port operation, OC0B disconnected.
0 1 Toggle OC0B on Compare Match
1 0 Clear OC0B on Compare Match
1 1 Set OC0B on Compare Match
Table 13-6 shows the COM0B1:0 bit functionality when the WGM02:0 bits are set to fast PWM
mode.

0 1 Reserved
Clear OC0B on Compare Match, set OC0B at BOTTOM,
1 0
Set OC0B on Compare Match, clear OC0B at BOTTOM,
1 1
(inverting mode).
Note: 1. A special case occurs when OCR0B equals TOP and COM0B1 is set. In this case, the Com-
pare Match is ignored, but the set or clear is done atBOTTOM. See ”Fast PWM Mode” on page
98 for more details.
104 ATmega164P/324P/644P
8011D–AVR–02/07
Table 13-7 on page 105 shows the COM0B1:0 bit functionality when the WGM02:0 bits are set
to phase correct PWM mode.

0 1 Reserved
Clear OC0B on Compare Match when up-counting. Set OC0B on
1 0
Set OC0B on Compare Match when up-counting. Clear OC0B on
1 1

• Bits 1:0 – WGM01:0: Waveform Generation Mode

Combined with the WGM02 bit found in the TCCR0B Register, these bits control the counting
sequence of the counter, the source for maximum (TOP) counter value, and what type of wave-
form generation to be used, see Table 13-8 on page 105. Modes of operation supported by the
Timer/Counter unit are: Normal mode (counter), Clear Timer on Compare Match (CTC) mode,
and two types of Pulse Width Modulation (PWM) modes (see ”Modes of Operation” on page
122).
Table 13-8. Waveform Generation Mode Bit Description

Timer/Counter
Mode of Update of TOV Flag
Mode WGM2 WGM1 WGM0 Operation TOP OCRx at Set on(1)(2)
0 0 0 0 Normal 0xFF Immediate MAX
PWM, Phase
1 0 0 1 0xFF TOP BOTTOM
Correct
2 0 1 0 CTC OCRA Immediate MAX
3 0 1 1 Fast PWM 0xFF BOTTOM MAX
4 1 0 0 Reserved – – –
PWM, Phase
5 1 0 1 OCRA TOP BOTTOM
Correct
6 1 1 0 Reserved – – –
7 1 1 1 Fast PWM OCRA BOTTOM TOP
Notes: 1. MAX = 0xFF

2. BOTTOM = 0x00
105
8011D–AVR–02/07
13.9.2 TCCR0B – Timer/Counter Control Register B
Bit 7 6 5 4 3 2 1 0
0x25 (0x45) FOC0A FOC0B – – WGM02 CS02 CS01 CS00 TCCR0B
Read/Write W W R R R/W R/W R/W R/W
• Bit 7 – FOC0A: Force Output Compare A

The FOC0A bit is only active when the WGM bits specify a non-PWM mode.
However, for ensuring compatibility with future devices, this bit must be set to zero when
TCCR0B is written when operating in PWM mode. When writing a logical one to the FOC0A bit,
an immediate Compare Match is forced on the Waveform Generation unit. The OC0A output is
changed according to its COM0A1:0 bits setting. Note that the FOC0A bit is implemented as a
strobe. Therefore it is the value present in the COM0A1:0 bits that determines the effect of the
forced compare.
A FOC0A strobe will not generate any interrupt, nor will it clear the timer in CTC mode using
OCR0A as TOP.
The FOC0A bit is always read as zero.
• Bit 6 – FOC0B: Force Output Compare B

The FOC0B bit is only active when the WGM bits specify a non-PWM mode.
TCCR0B is written when operating in PWM mode. When writing a logical one to the FOC0B bit,
an immediate Compare Match is forced on the Waveform Generation unit. The OC0B output is
changed according to its COM0B1:0 bits setting. Note that the FOC0B bit is implemented as a
strobe. Therefore it is the value present in the COM0B1:0 bits that determines the effect of the
forced compare.
A FOC0B strobe will not generate any interrupt, nor will it clear the timer in CTC mode using
OCR0B as TOP.
The FOC0B bit is always read as zero.

These bits are reserved bits and will always read as zero.
• Bit 3 – WGM02: Waveform Generation Mode

See the description in the ”TCCR0A – Timer/Counter Control Register A” on page 103.
106 ATmega164P/324P/644P
8011D–AVR–02/07
• Bits 2:0 – CS02:0: Clock Select

The three Clock Select bits select the clock source to be used by the Timer/Counter.
Table 13-9. Clock Select Bit Description

CS02 CS01 CS00 Description
0 0 0 No clock source (Timer/Counter stopped)
0 0 1 clkI/O/(No prescaling)
0 1 0 clkI/O/8 (From prescaler)
1 1 0 External clock source on T0 pin. Clock on falling edge.
1 1 1 External clock source on T0 pin. Clock on rising edge.
If external pin modes are used for the Timer/Counter0, transitions on the T0 pin will clock the
counter even if the pin is configured as an output. This feature allows software control of the
counting.
13.9.3 TCNT0 – Timer/Counter Register
Bit 7 6 5 4 3 2 1 0
0x26 (0x46) TCNT0[7:0] TCNT0
The Timer/Counter Register gives direct access, both for read and write operations, to the
Timer/Counter unit 8-bit counter. Writing to the TCNT0 Register blocks (removes) the Compare
Match on the following timer clock. Modifying the counter (TCNT0) while the counter is running,
introduces a risk of missing a Compare Match between TCNT0 and the OCR0x Registers.
13.9.4 OCR0A – Output Compare Register A
Bit 7 6 5 4 3 2 1 0
0x27 (0x47) OCR0A[7:0] OCR0A
The Output Compare Register A contains an 8-bit value that is continuously compared with the
counter value (TCNT0). A match can be used to generate an Output Compare interrupt, or to
generate a waveform output on the OC0A pin.
107
8011D–AVR–02/07
13.9.5 OCR0B – Output Compare Register B
Bit 7 6 5 4 3 2 1 0
0x28 (0x48) OCR0B[7:0] OCR0B
The Output Compare Register B contains an 8-bit value that is continuously compared with the
generate a waveform output on the OC0B pin.
13.9.6 TIMSK0 – Timer/Counter Interrupt Mask Register
Bit 7 6 5 4 3 2 1 0
(0x6E) – – – – – OCIE0B OCIE0A TOIE0 TIMSK0

These bits are reserved bits and will always read as zero.
• Bit 2 – OCIE0B: Timer/Counter Output Compare Match B Interrupt Enable

When the OCIE0B bit is written to one, and the I-bit in the Status Register is set, the
Timer/Counter Compare Match B interrupt is enabled. The corresponding interrupt is executed if
a Compare Match in Timer/Counter occurs, i.e., when the OCF0B bit is set in the Timer/Counter
Interrupt Flag Register – TIFR0.
• Bit 1 – OCIE0A: Timer/Counter0 Output Compare Match A Interrupt Enable

When the OCIE0A bit is written to one, and the I-bit in the Status Register is set, the
Timer/Counter0 Compare Match A interrupt is enabled. The corresponding interrupt is executed
if a Compare Match in Timer/Counter0 occurs, i.e., when the OCF0A bit is set in the
Timer/Counter 0 Interrupt Flag Register – TIFR0.
• Bit 0 – TOIE0: Timer/Counter0 Overflow Interrupt Enable

When the TOIE0 bit is written to one, and the I-bit in the Status Register is set, the
Timer/Counter0 Overflow interrupt is enabled. The corresponding interrupt is executed if an
overflow in Timer/Counter0 occurs, i.e., when the TOV0 bit is set in the Timer/Counter 0 Inter-
rupt Flag Register – TIFR0.
13.9.7 TIFR0 – Timer/Counter 0 Interrupt Flag Register
Bit 7 6 5 4 3 2 1 0
0x15 (0x35) – – – – – OCF0B OCF0A TOV0 TIFR0

108 ATmega164P/324P/644P
8011D–AVR–02/07
• Bit 2 – OCF0B: Timer/Counter 0 Output Compare B Match Flag

The OCF0B bit is set when a Compare Match occurs between the Timer/Counter and the data in
OCR0B – Output Compare Register0 B. OCF0B is cleared by hardware when executing the cor-
responding interrupt handling vector. Alternatively, OCF0B is cleared by writing a logic one to
the flag. When the I-bit in SREG, OCIE0B (Timer/Counter Compare B Match Interrupt Enable),
and OCF0B are set, the Timer/Counter Compare Match Interrupt is executed.
• Bit 1 – OCF0A: Timer/Counter 0 Output Compare A Match Flag

The OCF0A bit is set when a Compare Match occurs between the Timer/Counter0 and the data
in OCR0A – Output Compare Register0. OCF0A is cleared by hardware when executing the cor-
responding interrupt handling vector. Alternatively, OCF0A is cleared by writing a logic one to
the flag. When the I-bit in SREG, OCIE0A (Timer/Counter0 Compare Match Interrupt Enable),
and OCF0A are set, the Timer/Counter0 Compare Match Interrupt is executed.
• Bit 0 – TOV0: Timer/Counter0 Overflow Flag

The bit TOV0 is set when an overflow occurs in Timer/Counter0. TOV0 is cleared by hardware
when executing the corresponding interrupt handling vector. Alternatively, TOV0 is cleared by
writing a logic one to the flag. When the SREG I-bit, TOIE0 (Timer/Counter0 Overflow Interrupt
Enable), and TOV0 are set, the Timer/Counter0 Overflow interrupt is executed.
The setting of this flag is dependent of the WGM02:0 bit setting. Refer to Table 13-8, ”Waveform
Generation Mode Bit Description” on page 105.
109
8011D–AVR–02/07
14. 16-bit Timer/Counter1 with PWM
14.1 Features
• True 16-bit Design (i.e., Allows 16-bit PWM)
• Two independent Output Compare Units
• Double Buffered Output Compare Registers
• One Input Capture Unit
• Input Capture Noise Canceler
• Glitch-free, Phase Correct Pulse Width Modulator (PWM)
• Variable PWM Period
• External Event Counter
• Four independent interrupt Sources (TOV1, OCF1A, OCF1B, and ICF1)
14.2 Overview
The 16-bit Timer/Counter unit allows accurate program execution timing (event management),
wave generation, and signal timing measurement.
Most register and bit references in this section are written in general form. A lower case “n”
replaces the Timer/Counter number, and a lower case “x” replaces the Output Compare unit
channel. However, when using the register or bit defines in a program, the precise form must be
used, i.e., TCNT1 for accessing Timer/Counter1 counter value and so on.
A simplified block diagram of the 16-bit Timer/Counter is shown in Figure 14-1. For the actual
The PRTIM1 bit in ”PRR – Power Reduction Register” on page 48 must be written to zero to
enable Timer/Counter1 module.
110 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 14-1. 16-bit Timer/Counter Block Diagram(Note:)
Count TOVn
Clear (Int.Req.)
Control Logic
Direction clkTn Clock Select
Edge
Tn
Detector
TOP BOTTOM
( From Prescaler )
Timer/Counter
TCNTn
= =0
OCnA
(Int.Req.)
Waveform
= Generation
OCnA
OCRnA
Fixed OCnB
TOP (Int.Req.)
DATA BUS
Values
Waveform
= Generation
OCnB
OCRnB ( From Analog

Comparator Ouput )
ICFn (Int.Req.)
Edge Noise
ICRn
Detector Canceler
ICPn
TCCRnA TCCRnB
Note: Refer to Figure 1-1 on page 2 and ”Alternate Port Functions” on page 77 for Timer/Counter1 pin
placement and description.
14.2.1 Registers
The Timer/Counter (TCNTn), Output Compare Registers (OCRnA/B/C), and Input Capture Reg-
ister (ICRn) are all 16-bit registers. Special procedures must be followed when accessing the 16-
bit registers. These procedures are described in the section ”Accessing 16-bit Registers” on
page 112. The Timer/Counter Control Registers (TCCRnA/B/C) are 8-bit registers and have no
CPU access restrictions. Interrupt requests (abbreviated to Int.Req. in the figure) signals are all
visible in the Timer Interrupt Flag Register (TIFRn). All interrupts are individually masked with
the Timer Interrupt Mask Register (TIMSKn). TIFRn and TIMSKn are not shown in the figure.
The Timer/Counter can be clocked internally, via the prescaler, or by an external clock source on
the Tn pin. The Clock Select logic block controls which clock source and edge the Timer/Counter
uses to increment (or decrement) its value. The Timer/Counter is inactive when no clock source
is selected. The output from the Clock Select logic is referred to as the timer clock (clkTn).
The double buffered Output Compare Registers (OCRnA/B/C) are compared with the
Timer/Counter value at all time. The result of the compare can be used by the Waveform Gener-
ator to generate a PWM or variable frequency output on the Output Compare pin (OCnA/B/C).
111
8011D–AVR–02/07
See Section “14.7” on page 119.. The compare match event will also set the Compare Match
Flag (OCFnA/B/C) which can be used to generate an Output Compare interrupt request.
The Input Capture Register can capture the Timer/Counter value at a given external (edge trig-
gered) event on either the Input Capture pin (ICPn) or on the Analog Comparator pins (See
Section “20.” on page 235.) The Input Capture unit includes a digital filtering unit (Noise Can-
celer) for reducing the chance of capturing noise spikes.
The TOP value, or maximum Timer/Counter value, can in some modes of operation be defined
by either the OCRnA Register, the ICRn Register, or by a set of fixed values. When using
OCRnA as TOP value in a PWM mode, the OCRnA Register can not be used for generating a
PWM output. However, the TOP value will in this case be double buffered allowing the TOP
value to be changed in run time. If a fixed TOP value is required, the ICRn Register can be used
as an alternative, freeing the OCRnA to be used as PWM output.
14.2.2 Definitions
The following definitions are used extensively throughout the section:

BOTTOM The counter reaches the BOTTOM when it becomes 0x0000.
MAX The counter reaches its MAXimum when it becomes 0xFFFF (decimal 65535).
The counter reaches the TOP when it becomes equal to the highest value in the count
sequence. The TOP value can be assigned to be one of the fixed values: 0x00FF, 0x01FF,
TOP
or 0x03FF, or to the value stored in the OCRnA or ICRn Register. The assignment is
dependent of the mode of operation.
14.3 Accessing 16-bit Registers

The TCNTn, OCRnA/B/C, and ICRn are 16-bit registers that can be accessed by the AVR CPU
via the 8-bit data bus. The 16-bit register must be byte accessed using two read or write opera-
tions. Each 16-bit timer has a single 8-bit register for temporary storing of the high byte of the 16-
bit access. The same temporary register is shared between all 16-bit registers within each 16-bit
timer. Accessing the low byte triggers the 16-bit read or write operation. When the low byte of a
16-bit register is written by the CPU, the high byte stored in the temporary register, and the low
byte written are both copied into the 16-bit register in the same clock cycle. When the low byte of
a 16-bit register is read by the CPU, the high byte of the 16-bit register is copied into the tempo-
rary register in the same clock cycle as the low byte is read.
Not all 16-bit accesses uses the temporary register for the high byte. Reading the OCRnA/B/C
16-bit registers does not involve using the temporary register.
To do a 16-bit write, the high byte must be written before the low byte. For a 16-bit read, the low
byte must be read before the high byte.
The following code examples show how to access the 16-bit Timer Registers assuming that no
interrupts updates the temporary register. The same principle can be used directly for accessing
the OCRnA/B/C and ICRn Registers. Note that when using “C”, the compiler handles the 16-bit
access.
112 ATmega164P/324P/644P
8011D–AVR–02/07
Assembly Code Examples(1)

...
; Set TCNTn to 0x01FF
ldi r17,0x01
ldi r16,0xFF
out TCNTnH,r17
out TCNTnL,r16
; Read TCNTn into r17:r16
in r16,TCNTnL
in r17,TCNTnH
...
C Code Examples(1)
unsigned int i;
...
/* Set TCNTn to 0x01FF */
TCNTn = 0x1FF;
/* Read TCNTn into i */
i = TCNTn;
...
For I/O Registers located in extended I/O map, “IN”, “OUT”, “SBIS”, “SBIC”, “CBI”, and “SBI”
instructions must be replaced with instructions that allow access to extended I/O. Typically
“LDS” and “STS” combined with “SBRS”, “SBRC”, “SBR”, and “CBR”.
The assembly code example returns the TCNTn value in the r17:r16 register pair.
It is important to notice that accessing 16-bit registers are atomic operations. If an interrupt
occurs between the two instructions accessing the 16-bit register, and the interrupt code
updates the temporary register by accessing the same or any other of the 16-bit Timer Regis-
ters, then the result of the access outside the interrupt will be corrupted. Therefore, when both
the main code and the interrupt code update the temporary register, the main code must disable
the interrupts during the 16-bit access.
113
8011D–AVR–02/07
The following code examples show how to do an atomic read of the TCNTn Register contents.
Reading any of the OCRnA/B/C or ICRn Registers can be done by using the same principle.

TIM16_ReadTCNTn:
; Save global interrupt flag
in r18,SREG
; Disable interrupts
cli
; Read TCNTn into r17:r16
in r16,TCNTnL
in r17,TCNTnH
; Restore global interrupt flag
out SREG,r18
ret
C Code Example(1)
unsigned int TIM16_ReadTCNTn( void )
{
unsigned char sreg;
unsigned int i;
/* Save global interrupt flag */
sreg = SREG;
/* Disable interrupts */
_CLI();
/* Read TCNTn into i */
i = TCNTn;
/* Restore global interrupt flag */
SREG = sreg;
return i;
}
The assembly code example returns the TCNTn value in the r17:r16 register pair.
114 ATmega164P/324P/644P
8011D–AVR–02/07
The following code examples show how to do an atomic write of the TCNTn Register contents.
Writing any of the OCRnA/B/C or ICRn Registers can be done by using the same principle.

TIM16_WriteTCNTn:
; Save global interrupt flag
in r18,SREG
; Disable interrupts
cli
; Set TCNTn to r17:r16
out TCNTnH,r17
out TCNTnL,r16
; Restore global interrupt flag
out SREG,r18
ret
C Code Example(1)
void TIM16_WriteTCNTn( unsigned int i )
{
unsigned char sreg;
unsigned int i;
/* Save global interrupt flag */
sreg = SREG;
/* Disable interrupts */
_CLI();
/* Set TCNTn to i */
TCNTn = i;
/* Restore global interrupt flag */
SREG = sreg;
}
The assembly code example requires that the r17:r16 register pair contains the value to be writ-
ten to TCNTn.
14.3.1 Reusing the Temporary High Byte Register
If writing to more than one 16-bit register where the high byte is the same for all registers written,
then the high byte only needs to be written once. However, note that the same rule of atomic
operation described previously also applies in this case.

The Timer/Counter can be clocked by an internal or an external clock source. The clock source
is selected by the Clock Select logic which is controlled by the Clock Select (CSn2:0) bits
located in the Timer/Counter control Register B (TCCRnB). For details on clock sources and
prescaler, see ”Timer/Counter Prescaler” on page 151.
115
8011D–AVR–02/07
14.5 Counter Unit
The main part of the 16-bit Timer/Counter is the programmable 16-bit bi-directional counter unit.
Figure 14-2 shows a block diagram of the counter and its surroundings.

DATA BUS (8-bit)
TOVn
(Int.Req.)
TEMP (8-bit)
Clock Select
Count Edge
Tn
TCNTnH (8-bit) TCNTnL (8-bit) Clear clkTn Detector
Control Logic
Direction
TCNTn (16-bit Counter)
( From Prescaler )
TOP BOTTOM

Count Increment or decrement TCNTn by 1.
Direction Select between increment and decrement.
Clear Clear TCNTn (set all bits to zero).
clkTn Timer/Counter clock.
TOP Signalize that TCNTn has reached maximum value.
BOTTOM Signalize that TCNTn has reached minimum value (zero).
The 16-bit counter is mapped into two 8-bit I/O memory locations: Counter High (TCNTnH) con-
taining the upper eight bits of the counter, and Counter Low (TCNTnL) containing the lower eight
bits. The TCNTnH Register can only be indirectly accessed by the CPU. When the CPU does an
access to the TCNTnH I/O location, the CPU accesses the high byte temporary register (TEMP).
The temporary register is updated with the TCNTnH value when the TCNTnL is read, and
TCNTnH is updated with the temporary register value when TCNTnL is written. This allows the
CPU to read or write the entire 16-bit counter value within one clock cycle via the 8-bit data bus.
It is important to notice that there are special cases of writing to the TCNTn Register when the
counter is counting that will give unpredictable results. The special cases are described in the
sections where they are of importance.
Depending on the mode of operation used, the counter is cleared, incremented, or decremented
at each timer clock (clkTn). The clkTn can be generated from an external or internal clock source,
selected by the Clock Select bits (CSn2:0). When no clock source is selected (CSn2:0 = 0) the
timer is stopped. However, the TCNTn value can be accessed by the CPU, independent of
whether clkTn is present or not. A CPU write overrides (has priority over) all counter clear or
count operations.
The counting sequence is determined by the setting of the Waveform Generation mode bits
(WGMn3:0) located in the Timer/Counter Control Registers A and B (TCCRnA and TCCRnB).
There are close connections between how the counter behaves (counts) and how waveforms
are generated on the Output Compare outputs OCnx. For more details about advanced counting
sequences and waveform generation, see ”Modes of Operation” on page 122.
116 ATmega164P/324P/644P
8011D–AVR–02/07
The Timer/Counter Overflow Flag (TOVn) is set according to the mode of operation selected by
the WGMn3:0 bits. TOVn can be used for generating a CPU interrupt.
14.6 Input Capture Unit

The Timer/Counter incorporates an Input Capture unit that can capture external events and give
them a time-stamp indicating time of occurrence. The external signal indicating an event, or mul-
tiple events, can be applied via the ICPn pin or alternatively, via the analog-comparator unit. The
time-stamps can then be used to calculate frequency, duty-cycle, and other features of the sig-
nal applied. Alternatively the time-stamps can be used for creating a log of the events.
The Input Capture unit is illustrated by the block diagram shown in Figure 14-3. The elements of
the block diagram that are not directly a part of the Input Capture unit are gray shaded. The
small “n” in register and bit names indicates the Timer/Counter number.
Figure 14-3. Input Capture Unit Block Diagram

DATA BUS (8-bit)
TEMP (8-bit)
ICRnH (8-bit) ICRnL (8-bit) TCNTnH (8-bit) TCNTnL (8-bit)
WRITE ICRn (16-bit Register) TCNTn (16-bit Counter)
ACO* ACIC* ICNC ICES
Analog
Comparator Noise Edge
ICFn (Int.Req.)
Canceler Detector
ICPn
When a change of the logic level (an event) occurs on the Input Capture pin (ICPn), alternatively
on the Analog Comparator output (ACO), and this change confirms to the setting of the edge
detector, a capture will be triggered. When a capture is triggered, the 16-bit value of the counter
(TCNTn) is written to the Input Capture Register (ICRn). The Input Capture Flag (ICFn) is set at
the same system clock as the TCNTn value is copied into ICRn Register. If enabled (ICIEn = 1),
the Input Capture Flag generates an Input Capture interrupt. The ICFn Flag is automatically
cleared when the interrupt is executed. Alternatively the ICFn Flag can be cleared by software
by writing a logical one to its I/O bit location.
Reading the 16-bit value in the Input Capture Register (ICRn) is done by first reading the low
byte (ICRnL) and then the high byte (ICRnH). When the low byte is read the high byte is copied
into the high byte temporary register (TEMP). When the CPU reads the ICRnH I/O location it will
access the TEMP Register.
The ICRn Register can only be written when using a Waveform Generation mode that utilizes
the ICRn Register for defining the counter’s TOP value. In these cases the Waveform Genera-
117
8011D–AVR–02/07
tion mode (WGMn3:0) bits must be set before the TOP value can be written to the ICRn
Register. When writing the ICRn Register the high byte must be written to the ICRnH I/O location
before the low byte is written to ICRnL.
For more information on how to access the 16-bit registers refer to ”Accessing 16-bit Registers”
on page 112.
14.6.1 Input Capture Trigger Source
The main trigger source for the Input Capture unit is the Input Capture pin (ICPn).
Timer/Counter1 can alternatively use the Analog Comparator output as trigger source for the
Input Capture unit. The Analog Comparator is selected as trigger source by setting the Analog
Comparator Input Capture (ACIC) bit in the Analog Comparator Control and Status Register
(ACSR). Be aware that changing trigger source can trigger a capture. The Input Capture Flag
must therefore be cleared after the change.
Both the Input Capture pin (ICPn) and the Analog Comparator output (ACO) inputs are sampled
using the same technique as for the Tn pin (Figure 14-1 on page 111). The edge detector is also
identical. However, when the noise canceler is enabled, additional logic is inserted before the
edge detector, which increases the delay by four system clock cycles. Note that the input of the
noise canceler and edge detector is always enabled unless the Timer/Counter is set in a Wave-
form Generation mode that uses ICRn to define TOP.
An Input Capture can be triggered by software by controlling the port of the ICPn pin.
14.6.2 Noise Canceler
The noise canceler improves noise immunity by using a simple digital filtering scheme. The
noise canceler input is monitored over four samples, and all four must be equal for changing the
output that in turn is used by the edge detector.
The noise canceler is enabled by setting the Input Capture Noise Canceler (ICNCn) bit in
Timer/Counter Control Register B (TCCRnB). When enabled the noise canceler introduces addi-
tional four system clock cycles of delay from a change applied to the input, to the update of the
ICRn Register. The noise canceler uses the system clock and is therefore not affected by the
prescaler.
14.6.3 Using the Input Capture Unit
The main challenge when using the Input Capture unit is to assign enough processor capacity
for handling the incoming events. The time between two events is critical. If the processor has
not read the captured value in the ICRn Register before the next event occurs, the ICRn will be
overwritten with a new value. In this case the result of the capture will be incorrect.
When using the Input Capture interrupt, the ICRn Register should be read as early in the inter-
rupt handler routine as possible. Even though the Input Capture interrupt has relatively high
priority, the maximum interrupt response time is dependent on the maximum number of clock
cycles it takes to handle any of the other interrupt requests.
Using the Input Capture unit in any mode of operation when the TOP value (resolution) is
actively changed during operation, is not recommended.
Measurement of an external signal’s duty cycle requires that the trigger edge is changed after
each capture. Changing the edge sensing must be done as early as possible after the ICRn
Register has been read. After a change of the edge, the Input Capture Flag (ICFn) must be
118 ATmega164P/324P/644P
8011D–AVR–02/07
cleared by software (writing a logical one to the I/O bit location). For measuring frequency only,
the clearing of the ICFn Flag is not required (if an interrupt handler is used).
14.7 Output Compare Units

The 16-bit comparator continuously compares TCNTn with the Output Compare Register
(OCRnx). If TCNT equals OCRnx the comparator signals a match. A match will set the Output
Compare Flag (OCFnx) at the next timer clock cycle. If enabled (OCIEnx = 1), the Output Com-
pare Flag generates an Output Compare interrupt. The OCFnx Flag is automatically cleared
when the interrupt is executed. Alternatively the OCFnx Flag can be cleared by software by writ-
ing a logical one to its I/O bit location. The Waveform Generator uses the match signal to
generate an output according to operating mode set by the Waveform Generation mode
(WGMn3:0) bits and Compare Output mode (COMnx1:0) bits. The TOP and BOTTOM signals
are used by the Waveform Generator for handling the special cases of the extreme values in
some modes of operation (See Section “14.9” on page 122.)
A special feature of Output Compare unit A allows it to define the Timer/Counter TOP value (i.e.,
counter resolution). In addition to the counter resolution, the TOP value defines the period time
for waveforms generated by the Waveform Generator.
Figure 14-4 shows a block diagram of the Output Compare unit. The small “n” in the register and
bit names indicates the device number (n = n for Timer/Counter n), and the “x” indicates Output
Compare unit (A/B/C). The elements of the block diagram that are not directly a part of the Out-
put Compare unit are gray shaded.

DATA BUS (8-bit)
TEMP (8-bit)
OCRnxH Buf. (8-bit) OCRnxL Buf. (8-bit) TCNTnH (8-bit) TCNTnL (8-bit)
OCRnx Buffer (16-bit Register) TCNTn (16-bit Counter)
OCRnxH (8-bit) OCRnxL (8-bit)
OCRnx (16-bit Register)
OCFnx (Int.Req.)
TOP
BOTTOM
WGMn3:0 COMnx1:0
The OCRnx Register is double buffered when using any of the twelve Pulse Width Modulation
(PWM) modes. For the Normal and Clear Timer on Compare (CTC) modes of operation, the
double buffering is disabled. The double buffering synchronizes the update of the OCRnx Com-
pare Register to either TOP or BOTTOM of the counting sequence. The synchronization
119
8011D–AVR–02/07
prevents the occurrence of odd-length, non-symmetrical PWM pulses, thereby making the out-
put glitch-free.
The OCRnx Register access may seem complex, but this is not case. When the double buffering
is enabled, the CPU has access to the OCRnx Buffer Register, and if double buffering is dis-
abled the CPU will access the OCRnx directly. The content of the OCR1x (Buffer or Compare)
Register is only changed by a write operation (the Timer/Counter does not update this register
automatically as the TCNT1 and ICR1 Register). Therefore OCR1x is not read via the high byte
temporary register (TEMP). However, it is a good practice to read the low byte first as when
accessing other 16-bit registers. Writing the OCRnx Registers must be done via the TEMP Reg-
ister since the compare of all 16 bits is done continuously. The high byte (OCRnxH) has to be
written first. When the high byte I/O location is written by the CPU, the TEMP Register will be
updated by the value written. Then when the low byte (OCRnxL) is written to the lower eight bits,
the high byte will be copied into the upper 8-bits of either the OCRnx buffer or OCRnx Compare
Register in the same system clock cycle.
For more information of how to access the 16-bit registers refer to ”Accessing 16-bit Registers”
on page 112.
In non-PWM Waveform Generation modes, the match output of the comparator can be forced by
writing a one to the Force Output Compare (FOCnx) bit. Forcing compare match will not set the
OCFnx Flag or reload/clear the timer, but the OCnx pin will be updated as if a real compare
match had occurred (the COMn1:0 bits settings define whether the OCnx pin is set, cleared or
toggled).
14.7.2 Compare Match Blocking by TCNTn Write
All CPU writes to the TCNTn Register will block any compare match that occurs in the next timer
clock cycle, even when the timer is stopped. This feature allows OCRnx to be initialized to the
same value as TCNTn without triggering an interrupt when the Timer/Counter clock is enabled.
Since writing TCNTn in any mode of operation will block all compare matches for one timer clock
cycle, there are risks involved when changing TCNTn when using any of the Output Compare
channels, independent of whether the Timer/Counter is running or not. If the value written to
TCNTn equals the OCRnx value, the compare match will be missed, resulting in incorrect wave-
form generation. Do not write the TCNTn equal to TOP in PWM modes with variable TOP
values. The compare match for the TOP will be ignored and the counter will continue to 0xFFFF.
Similarly, do not write the TCNTn value equal to BOTTOM when the counter is downcounting.
The setup of the OCnx should be performed before setting the Data Direction Register for the
port pin to output. The easiest way of setting the OCnx value is to use the Force Output Com-
pare (FOCnx) strobe bits in Normal mode. The OCnx Register keeps its value even when
Be aware that the COMnx1:0 bits are not double buffered together with the compare value.
Changing the COMnx1:0 bits will take effect immediately.
120 ATmega164P/324P/644P
8011D–AVR–02/07
The Compare Output mode (COMnx1:0) bits have two functions. The Waveform Generator uses
the COMnx1:0 bits for defining the Output Compare (OCnx) state at the next compare match.
Secondly the COMnx1:0 bits control the OCnx pin output source. Figure 14-5 shows a simplified
schematic of the logic affected by the COMnx1:0 bit setting. The I/O Registers, I/O bits, and I/O
(DDR and PORT) that are affected by the COMnx1:0 bits are shown. When referring to the
OCnx state, the reference is for the internal OCnx Register, not the OCnx pin. If a system reset
occur, the OCnx Register is reset to “0”.
COMnx1
COMnx0 Waveform
D Q
FOCnx Generator
1
OCnx
OCnx Pin
0
D Q
DATA BUS
PORT
D Q
DDR
clk I/O
The general I/O port function is overridden by the Output Compare (OCnx) from the Waveform
Generator if either of the COMnx1:0 bits are set. However, the OCnx pin direction (input or out-
Register bit for the OCnx pin (DDR_OCnx) must be set as output before the OCnx value is visi-
ble on the pin. The port override function is generally independent of the Waveform Generation
mode, but there are some exceptions. Refer to Table 14-2, Table 14-3 and Table 14-4 for
details.
The design of the Output Compare pin logic allows initialization of the OCnx state before the out-
put is enabled. Note that some COMnx1:0 bit settings are reserved for certain modes of
operation. See Section “14.11” on page 131.
The COMnx1:0 bits have no effect on the Input Capture unit.
121
8011D–AVR–02/07
The Waveform Generator uses the COMnx1:0 bits differently in normal, CTC, and PWM modes.
For all modes, setting the COMnx1:0 = 0 tells the Waveform Generator that no action on the
OCnx Register is to be performed on the next compare match. For compare output actions in the
non-PWM modes refer to Table 14-2 on page 131. For fast PWM mode refer to Table 14-3 on
page 132, and for phase correct and phase and frequency correct PWM refer to Table 14-4 on
page 132.
A change of the COMnx1:0 bits state will have effect at the first compare match after the bits are
FOCnx strobe bits.

defined by the combination of the Waveform Generation mode (WGMn3:0) and Compare Output
mode (COMnx1:0) bits. The Compare Output mode bits do not affect the counting sequence,
while the Waveform Generation mode bits do. The COMnx1:0 bits control whether the PWM out-
the COMnx1:0 bits control whether the output should be set, cleared or toggle at a compare
match (See Section “14.8” on page 121.)
For detailed timing information refer to ”Timer/Counter Timing Diagrams” on page 129.
14.9.1 Normal Mode
The simplest mode of operation is the Normal mode (WGMn3:0 = 0). In this mode the counting
overruns when it passes its maximum 16-bit value (MAX = 0xFFFF) and then restarts from the
BOTTOM (0x0000). In normal operation the Timer/Counter Overflow Flag (TOVn) will be set in
the same timer clock cycle as the TCNTn becomes zero. The TOVn Flag in this case behaves
like a 17th bit, except that it is only set, not cleared. However, combined with the timer overflow
interrupt that automatically clears the TOVn Flag, the timer resolution can be increased by soft-
ware. There are no special cases to consider in the Normal mode, a new counter value can be
written anytime.
The Input Capture unit is easy to use in Normal mode. However, observe that the maximum
interval between the external events must not exceed the resolution of the counter. If the interval
between events are too long, the timer overflow interrupt or the prescaler must be used to
extend the resolution for the capture unit.
The Output Compare units can be used to generate interrupts at some given time. Using the
Output Compare to generate waveforms in Normal mode is not recommended, since this will
In Clear Timer on Compare or CTC mode (WGMn3:0 = 4 or 12), the OCRnA or ICRn Register
are used to manipulate the counter resolution. In CTC mode the counter is cleared to zero when
the counter value (TCNTn) matches either the OCRnA (WGMn3:0 = 4) or the ICRn (WGMn3:0 =
12). The OCRnA or ICRn define the top value for the counter, hence also its resolution. This
mode allows greater control of the compare match output frequency. It also simplifies the opera-
tion of counting external events.
122 ATmega164P/324P/644P
8011D–AVR–02/07
The timing diagram for the CTC mode is shown in Figure 14-6. The counter value (TCNTn)
increases until a compare match occurs with either OCRnA or ICRn, and then counter (TCNTn)
is cleared.

OCnA Interrupt Flag Set
or ICFn Interrupt Flag Set
(Interrupt on TOP)
TCNTn
OCnA
(COMnA1:0 = 1)
(Toggle)
Period 1 2 3 4
An interrupt can be generated at each time the counter value reaches the TOP value by either
using the OCFnA or ICFn Flag according to the register used to define the TOP value. If the
interrupt is enabled, the interrupt handler routine can be used for updating the TOP value. How-
ever, changing the TOP to a value close to BOTTOM when the counter is running with none or a
low prescaler value must be done with care since the CTC mode does not have the double buff-
ering feature. If the new value written to OCRnA or ICRn is lower than the current value of
TCNTn, the counter will miss the compare match. The counter will then have to count to its max-
imum value (0xFFFF) and wrap around starting at 0x0000 before the compare match can occur.
In many cases this feature is not desirable. An alternative will then be to use the fast PWM mode
using OCRnA for defining TOP (WGMn3:0 = 15) since the OCRnA then will be double buffered.
For generating a waveform output in CTC mode, the OCnA output can be set to toggle its logical
level on each compare match by setting the Compare Output mode bits to toggle mode
(COMnA1:0 = 1). The OCnA value will not be visible on the port pin unless the data direction for
the pin is set to output (DDR_OCnA = 1). The waveform generated will have a maximum fre-
quency of fOCnA = fclk_I/O/2 when OCRnA is set to zero (0x0000). The waveform frequency is
defined by the following equation:
f clk_I/O
f OCnA = --------------------------------------------------
-
2 ⋅ N ⋅ ( 1 + OCRnA )
The N variable represents the prescaler factor (1, 8, 64, 256, or 1024).
As for the Normal mode of operation, the TOVn Flag is set in the same timer clock cycle that the
The fast Pulse Width Modulation or fast PWM mode (WGMn3:0 = 5, 6, 7, 14, or 15) provides a
high frequency PWM waveform generation option. The fast PWM differs from the other PWM
options by its single-slope operation. The counter counts from BOTTOM to TOP then restarts
from BOTTOM. In non-inverting Compare Output mode, the Output Compare (OCnx) is cleared
on the compare match between TCNTn and OCRnx, and set at BOTTOM. In inverting Compare
Output mode output is set on compare match and cleared at BOTTOM. Due to the single-slope
operation, the operating frequency of the fast PWM mode can be twice as high as the phase cor-
123
8011D–AVR–02/07
rect and phase and frequency correct PWM modes that use dual-slope operation. This high
frequency makes the fast PWM mode well suited for power regulation, rectification, and DAC
applications. High frequency allows physically small sized external components (coils, capaci-
tors), hence reduces total system cost.
The PWM resolution for fast PWM can be fixed to 8-, 9-, or 10-bit, or defined by either ICRn or
OCRnA. The minimum resolution allowed is 2-bit (ICRn or OCRnA set to 0x0003), and the max-
imum resolution is 16-bit (ICRn or OCRnA set to MAX). The PWM resolution in bits can be
calculated by using the following equation:
log ( TOP + 1 )
R FPWM = ----------------------------------
-
log ( 2 )
In fast PWM mode the counter is incremented until the counter value matches either one of the
fixed values 0x00FF, 0x01FF, or 0x03FF (WGMn3:0 = 5, 6, or 7), the value in ICRn (WGMn3:0 =
14), or the value in OCRnA (WGMn3:0 = 15). The counter is then cleared at the following timer
clock cycle. The timing diagram for the fast PWM mode is shown in Figure 14-7. The figure
shows fast PWM mode when OCRnA or ICRn is used to define TOP. The TCNTn value is in the
timing diagram shown as a histogram for illustrating the single-slope operation. The diagram
includes non-inverted and inverted PWM outputs. The small horizontal line marks on the TCNTn
slopes represent compare matches between OCRnx and TCNTn. The OCnx Interrupt Flag will
be set when a compare match occurs.

OCRnx/TOP Update and
TOVn Interrupt Flag Set and
(Interrupt on TOP)
TCNTn
OCnx (COMnx1:0 = 2)
OCnx (COMnx1:0 = 3)
Period 1 2 3 4 5 6 7 8
The Timer/Counter Overflow Flag (TOVn) is set each time the counter reaches TOP. In addition
the OCnA or ICFn Flag is set at the same timer clock cycle as TOVn is set when either OCRnA
or ICRn is used for defining the TOP value. If one of the interrupts are enabled, the interrupt han-
dler routine can be used for updating the TOP and compare values.
When changing the TOP value the program must ensure that the new TOP value is higher or
equal to the value of all of the Compare Registers. If the TOP value is lower than any of the
Compare Registers, a compare match will never occur between the TCNTn and the OCRnx.
Note that when using fixed TOP values the unused bits are masked to zero when any of the
OCRnx Registers are written.
The procedure for updating ICRn differs from updating OCRnA when used for defining the TOP
value. The ICRn Register is not double buffered. This means that if ICRn is changed to a low
124 ATmega164P/324P/644P
8011D–AVR–02/07
value when the counter is running with none or a low prescaler value, there is a risk that the new
ICRn value written is lower than the current value of TCNTn. The result will then be that the
counter will miss the compare match at the TOP value. The counter will then have to count to the
MAX value (0xFFFF) and wrap around starting at 0x0000 before the compare match can occur.
The OCRnA Register however, is double buffered. This feature allows the OCRnA I/O location
to be written anytime. When the OCRnA I/O location is written the value written will be put into
the OCRnA Buffer Register. The OCRnA Compare Register will then be updated with the value
in the Buffer Register at the next timer clock cycle the TCNTn matches TOP. The update is done
at the same timer clock cycle as the TCNTn is cleared and the TOVn Flag is set.
Using the ICRn Register for defining TOP works well when using fixed TOP values. By using
ICRn, the OCRnA Register is free to be used for generating a PWM output on OCnA. However,
if the base PWM frequency is actively changed (by changing the TOP value), using the OCRnA
as TOP is clearly a better choice due to its double buffer feature.
In fast PWM mode, the compare units allow generation of PWM waveforms on the OCnx pins.
Setting the COMnx1:0 bits to two will produce a non-inverted PWM and an inverted PWM output
can be generated by setting the COMnx1:0 to three (see Table on page 132). The actual OCnx
value will only be visible on the port pin if the data direction for the port pin is set as output
(DDR_OCnx). The PWM waveform is generated by setting (or clearing) the OCnx Register at
the compare match between OCRnx and TCNTn, and clearing (or setting) the OCnx Register at
the timer clock cycle the counter is cleared (changes from TOP to BOTTOM).
f clk_I/O
f OCnxPWM = ----------------------------------
-
N ⋅ ( 1 + TOP )
The N variable represents the prescaler divider (1, 8, 64, 256, or 1024).
The extreme values for the OCRnx Register represents special cases when generating a PWM
waveform output in the fast PWM mode. If the OCRnx is set equal to BOTTOM (0x0000) the out-
put will be a narrow spike for each TOP+1 timer clock cycle. Setting the OCRnx equal to TOP
will result in a constant high or low output (depending on the polarity of the output set by the
COMnx1:0 bits.)
ting OCnA to toggle its logical level on each compare match (COMnA1:0 = 1). This applies only
if OCR1A is used to define the TOP value (WGM13:0 = 15). The waveform generated will have
a maximum frequency of fOCnA = fclk_I/O/2 when OCRnA is set to zero (0x0000). This feature is
similar to the OCnA toggle in CTC mode, except the double buffer feature of the Output Com-
pare unit is enabled in the fast PWM mode.
The phase correct Pulse Width Modulation or phase correct PWM mode (WGMn3:0 = 1, 2, 3,
10, or 11) provides a high resolution phase correct PWM waveform generation option. The
phase correct PWM mode is, like the phase and frequency correct PWM mode, based on a dual-
slope operation. The counter counts repeatedly from BOTTOM (0x0000) to TOP and then from
TOP to BOTTOM. In non-inverting Compare Output mode, the Output Compare (OCnx) is
cleared on the compare match between TCNTn and OCRnx while upcounting, and set on the
compare match while downcounting. In inverting Output Compare mode, the operation is
inverted. The dual-slope operation has lower maximum operation frequency than single slope
125
8011D–AVR–02/07
operation. However, due to the symmetric feature of the dual-slope PWM modes, these modes
are preferred for motor control applications.
The PWM resolution for the phase correct PWM mode can be fixed to 8-, 9-, or 10-bit, or defined
by either ICRn or OCRnA. The minimum resolution allowed is 2-bit (ICRn or OCRnA set to
0x0003), and the maximum resolution is 16-bit (ICRn or OCRnA set to MAX). The PWM resolu-
tion in bits can be calculated by using the following equation:
R PCPWM = log ( TOP + 1 )

-----------------------------------
log ( 2 )
In phase correct PWM mode the counter is incremented until the counter value matches either
one of the fixed values 0x00FF, 0x01FF, or 0x03FF (WGMn3:0 = 1, 2, or 3), the value in ICRn
(WGMn3:0 = 10), or the value in OCRnA (WGMn3:0 = 11). The counter has then reached the
TOP and changes the count direction. The TCNTn value will be equal to TOP for one timer clock
cycle. The timing diagram for the phase correct PWM mode is shown on Figure 14-8. The figure
shows phase correct PWM mode when OCRnA or ICRn is used to define TOP. The TCNTn
value is in the timing diagram shown as a histogram for illustrating the dual-slope operation. The
diagram includes non-inverted and inverted PWM outputs. The small horizontal line marks on
the TCNTn slopes represent compare matches between OCRnx and TCNTn. The OCnx Inter-
rupt Flag will be set when a compare match occurs.

OCRnx/TOP Update and
(Interrupt on TOP)

(Interrupt on Bottom)
TCNTn
OCnx (COMnx1:0 = 2)
OCnx (COMnx1:0 = 3)
Period 1 2 3 4
The Timer/Counter Overflow Flag (TOVn) is set each time the counter reaches BOTTOM. When
either OCRnA or ICRn is used for defining the TOP value, the OCnA or ICFn Flag is set accord-
ingly at the same timer clock cycle as the OCRnx Registers are updated with the double buffer
value (at TOP). The Interrupt Flags can be used to generate an interrupt each time the counter
reaches the TOP or BOTTOM value.
Note that when using fixed TOP values, the unused bits are masked to zero when any of the
126 ATmega164P/324P/644P
8011D–AVR–02/07
OCRnx Registers are written. As the third period shown in Figure 14-8 illustrates, changing the
TOP actively while the Timer/Counter is running in the phase correct mode can result in an
unsymmetrical output. The reason for this can be found in the time of update of the OCRnx Reg-
ister. Since the OCRnx update occurs at TOP, the PWM period starts and ends at TOP. This
implies that the length of the falling slope is determined by the previous TOP value, while the
length of the rising slope is determined by the new TOP value. When these two values differ the
two slopes of the period will differ in length. The difference in length gives the unsymmetrical
result on the output.
It is recommended to use the phase and frequency correct mode instead of the phase correct
mode when changing the TOP value while the Timer/Counter is running. When using a static
TOP value there are practically no differences between the two modes of operation.
In phase correct PWM mode, the compare units allow generation of PWM waveforms on the
OCnx pins. Setting the COMnx1:0 bits to two will produce a non-inverted PWM and an inverted
PWM output can be generated by setting the COMnx1:0 to three (See Table on page 132). The
actual OCnx value will only be visible on the port pin if the data direction for the port pin is set as
output (DDR_OCnx). The PWM waveform is generated by setting (or clearing) the OCnx Regis-
ter at the compare match between OCRnx and TCNTn when the counter increments, and
clearing (or setting) the OCnx Register at compare match between OCRnx and TCNTn when
the counter decrements. The PWM frequency for the output when using phase correct PWM can
be calculated by the following equation:
f clk_I/O
f OCnxPCPWM = ---------------------------
-
2 ⋅ N ⋅ TOP
The extreme values for the OCRnx Register represent special cases when generating a PWM
waveform output in the phase correct PWM mode. If the OCRnx is set equal to BOTTOM the
output will be continuously low and if set equal to TOP the output will be continuously high for
non-inverted PWM mode. For inverted PWM the output will have the opposite logic values. If
OCR1A is used to define the TOP value (WGM13:0 = 11) and COM1A1:0 = 1, the OC1A output
will toggle with a 50% duty cycle.
14.9.5 Phase and Frequency Correct PWM Mode
The phase and frequency correct Pulse Width Modulation, or phase and frequency correct PWM
mode (WGMn3:0 = 8 or 9) provides a high resolution phase and frequency correct PWM wave-
form generation option. The phase and frequency correct PWM mode is, like the phase correct
PWM mode, based on a dual-slope operation. The counter counts repeatedly from BOTTOM
(0x0000) to TOP and then from TOP to BOTTOM. In non-inverting Compare Output mode, the
Output Compare (OCnx) is cleared on the compare match between TCNTn and OCRnx while
upcounting, and set on the compare match while downcounting. In inverting Compare Output
mode, the operation is inverted. The dual-slope operation gives a lower maximum operation fre-
quency compared to the single-slope operation. However, due to the symmetric feature of the
dual-slope PWM modes, these modes are preferred for motor control applications.
The main difference between the phase correct, and the phase and frequency correct PWM
mode is the time the OCRnx Register is updated by the OCRnx Buffer Register, (see Figure 14-
8 and Figure 14-9).
The PWM resolution for the phase and frequency correct PWM mode can be defined by either
ICRn or OCRnA. The minimum resolution allowed is 2-bit (ICRn or OCRnA set to 0x0003), and
127
8011D–AVR–02/07
the maximum resolution is 16-bit (ICRn or OCRnA set to MAX). The PWM resolution in bits can
be calculated using the following equation:
log ( TOP + 1 )
R PFCPWM = ----------------------------------
-
log ( 2 )
In phase and frequency correct PWM mode the counter is incremented until the counter value
matches either the value in ICRn (WGMn3:0 = 8), or the value in OCRnA (WGMn3:0 = 9). The
counter has then reached the TOP and changes the count direction. The TCNTn value will be
equal to TOP for one timer clock cycle. The timing diagram for the phase correct and frequency
correct PWM mode is shown on Figure 14-9. The figure shows phase and frequency correct
PWM mode when OCRnA or ICRn is used to define TOP. The TCNTn value is in the timing dia-
gram shown as a histogram for illustrating the dual-slope operation. The diagram includes non-
inverted and inverted PWM outputs. The small horizontal line marks on the TCNTn slopes repre-
sent compare matches between OCRnx and TCNTn. The OCnx Interrupt Flag will be set when a
compare match occurs.
Figure 14-9. Phase and Frequency Correct PWM Mode, Timing Diagram
(Interrupt on TOP)
OCRnx/TOP Updateand
(Interrupt on Bottom)
TCNTn
OCnx (COMnx1:0 = 2)
OCnx (COMnx1:0 = 3)
Period 1 2 3 4
The Timer/Counter Overflow Flag (TOVn) is set at the same timer clock cycle as the OCRnx
Registers are updated with the double buffer value (at BOTTOM). When either OCRnA or ICRn
is used for defining the TOP value, the OCnA or ICFn Flag set when TCNTn has reached TOP.
The Interrupt Flags can then be used to generate an interrupt each time the counter reaches the
TOP or BOTTOM value.
As Figure 14-9 shows the output generated is, in contrast to the phase correct mode, symmetri-
cal in all periods. Since the OCRnx Registers are updated at BOTTOM, the length of the rising
and the falling slopes will always be equal. This gives symmetrical output pulses and is therefore
frequency correct.
128 ATmega164P/324P/644P
8011D–AVR–02/07
Using the ICRn Register for defining TOP works well when using fixed TOP values. By using
ICRn, the OCRnA Register is free to be used for generating a PWM output on OCnA. However,
if the base PWM frequency is actively changed by changing the TOP value, using the OCRnA as
TOP is clearly a better choice due to its double buffer feature.
In phase and frequency correct PWM mode, the compare units allow generation of PWM wave-
forms on the OCnx pins. Setting the COMnx1:0 bits to two will produce a non-inverted PWM and
an inverted PWM output can be generated by setting the COMnx1:0 to three (See Table on
page 132). The actual OCnx value will only be visible on the port pin if the data direction for the
port pin is set as output (DDR_OCnx). The PWM waveform is generated by setting (or clearing)
the OCnx Register at the compare match between OCRnx and TCNTn when the counter incre-
ments, and clearing (or setting) the OCnx Register at compare match between OCRnx and
TCNTn when the counter decrements. The PWM frequency for the output when using phase
and frequency correct PWM can be calculated by the following equation:
f clk_I/O
f OCnxPFCPWM = ---------------------------
-
2 ⋅ N ⋅ TOP
The extreme values for the OCRnx Register represents special cases when generating a PWM
waveform output in the phase correct PWM mode. If the OCRnx is set equal to BOTTOM the
output will be continuously low and if set equal to TOP the output will be set to high for non-
inverted PWM mode. For inverted PWM the output will have the opposite logic values. If OCR1A
is used to define the TOP value (WGM13:0 = 9) and COM1A1:0 = 1, the OC1A output will toggle
with a 50% duty cycle.

The Timer/Counter is a synchronous design and the timer clock (clkTn) is therefore shown as a
clock enable signal in the following figures. The figures include information on when Interrupt
Flags are set, and when the OCRnx Register is updated with the OCRnx buffer value (only for
modes utilizing double buffering). Figure 14-10 shows a timing diagram for the setting of OCFnx.
Figure 14-10. Timer/Counter Timing Diagram, Setting of OCFnx, no Prescaling
clkI/O
clkTn
(clkI/O /1)
OCRnx OCRnx Value
OCFnx
129
8011D–AVR–02/07
Figure 14-11. Timer/Counter Timing Diagram, Setting of OCFnx, with Prescaler (fclk_I/O/8)
clkI/O
clkTn
(clkI/O /8)
OCRnx OCRnx Value
OCFnx
Figure 14-12 shows the count sequence close to TOP in various modes. When using phase and
frequency correct PWM mode the OCRnx Register is updated at BOTTOM. The timing diagrams
will be the same, but TOP should be replaced by BOTTOM, TOP-1 by BOTTOM+1 and so on.
The same renaming applies for modes that set the TOVn Flag at BOTTOM.
clkI/O
clkTn
(clkI/O /1)
TCNTn
(CTC and FPWM)
TCNTn
TOP - 1 TOP TOP - 1 TOP - 2
(PC and PFC PWM)
TOVn (FPWM)
and ICFn (if used
as TOP)
OCRnx
Old OCRnx Value New OCRnx Value
(Update at TOP)
130 ATmega164P/324P/644P
8011D–AVR–02/07
clkI/O
clkTn
(clkI/O/8)
TCNTn
(CTC and FPWM)
TCNTn TOP - 1 TOP TOP - 1 TOP - 2

(PC and PFC PWM)
TOVn (FPWM)
and ICF n (if used
as TOP)
OCRnx
Old OCRnx Value New OCRnx Value
(Update at TOP)

14.11.1 TCCR1A – Timer/Counter1 Control Register A
Bit 7 6 5 4 3 2 1 0
(0x80) COM1A1 COM1A0 COM1B1 COM1B0 – – WGM11 WGM10 TCCR1A
• Bit 7:6 – COMnA1:0: Compare Output Mode for Channel A
• Bit 5:4 – COMnB1:0: Compare Output Mode for Channel B

The COMnA1:0 and COMnB1:0 control the Output Compare pins (OCnA and OCnB respec-
tively) behavior. If one or both of the COMnA1:0 bits are written to one, the OCnA output
overrides the normal port functionality of the I/O pin it is connected to. If one or both of the
COMnB1:0 bit are written to one, the OCnB output overrides the normal port functionality of the
I/O pin it is connected to. However, note that the Data Direction Register (DDR) bit correspond-
ing to the OCnA or OCnB pin must be set in order to enable the output driver.
When the OCnA or OCnB is connected to the pin, the function of the COMnx1:0 bits is depen-
dent of the WGMn3:0 bits setting. Table 14-2 on page 131 shows the COMnx1:0 bit functionality
when the WGMn3:0 bits are set to a Normal or a CTC mode (non-PWM).
Table 14-2. Compare Output Mode, non-PWM
COMnA1/COMnB1 COMnA0/COMnB0 Description
0 0 Normal port operation, OCnA/OCnB disconnected.
0 1 Toggle OCnA/OCnB on Compare Match.
Clear OCnA/OCnB on Compare Match (Set output to
1 0
low level).
Set OCnA/OCnB on Compare Match (Set output to
1 1
high level).
131
8011D–AVR–02/07
Table 14-3 on page 132 shows the COMnx1:0 bit functionality when the WGMn3:0 bits are set to
the fast PWM mode.
Table 14-3. Compare Output Mode, Fast PWM(1)
WGMn3:0 = 14 or 15: Toggle OC1A on Compare
Match, OC1B disconnected (normal port operation).
0 1
For all other WGM1 settings, normal port operation,
OC1A/OC1B disconnected.
Clear OCnA/OCnB on Compare Match, set
1 0
OCnA/OCnB at BOTTOM (non-inverting mode)
Set OCnA/OCnB on Compare Match, clear
1 1
OCnA/OCnB at BOTTOM (inverting mode)
Note: 1. A special case occurs when OCRnA/OCRnB equals TOP and COMnA1/COMnB1 is set. In
this case the compare match is ignored, but the set or clear is done at BOTTOM. See Section
“14.9.3” on page 123. for more details.
Table 14-4 on page 132 shows the COMnx1:0 bit functionality when the WGMn3:0 bits are set to
the phase correct or the phase and frequency correct, PWM mode.
Table 14-4. Compare Output Mode, Phase Correct and Phase and Frequency Correct
PWM(1)
WGMn3:0 = 9 or 11: Toggle OCnA on Compare
Match, OCnB disconnected (normal port operation).
0 1
For all other WGM1 settings, normal port operation,
OC1A/OC1B disconnected.
Clear OCnA/OCnB on Compare Match when up-
1 0 counting. Set OCnA/OCnB on Compare Match when
downcounting.
Set OCnA/OCnB on Compare Match when up-
1 1 counting. Clear OCnA/OCnB on Compare Match
when downcounting.
Note: 1. A special case occurs when OCRnA/OCRnB equals TOP and COMnA1/COMnB1 is set. See
Section “14.9.4” on page 125. for more details.
• Bit 1:0 – WGMn1:0: Waveform Generation Mode

Combined with the WGMn3:2 bits found in the TCCRnB Register, these bits control the counting
form generation to be used, see Table 14-5 on page 133. Modes of operation supported by the
Timer/Counter unit are: Normal mode (counter), Clear Timer on Compare match (CTC) mode,
and three types of Pulse Width Modulation (PWM) modes. (See Section “14.9” on page 122.).
132 ATmega164P/324P/644P
8011D–AVR–02/07
Table 14-5. Waveform Generation Mode Bit Description(1)
WGMn2 WGMn1 WGMn0 Timer/Counter Mode of Update of TOVn Flag
Mode WGMn3 (CTCn) (PWMn1) (PWMn0) Operation TOP OCRnx at Set on
0 0 0 0 0 Normal 0xFFFF Immediate MAX
1 0 0 0 1 PWM, Phase Correct, 8-bit 0x00FF TOP BOTTOM
4 0 1 0 0 CTC OCRnA Immediate MAX
5 0 1 0 1 Fast PWM, 8-bit 0x00FF BOTTOM TOP
PWM, Phase and Frequency
8 1 0 0 0 ICRn BOTTOM BOTTOM
Correct
PWM, Phase and Frequency
9 1 0 0 1 OCRnA BOTTOM BOTTOM
Correct
10 1 0 1 0 PWM, Phase Correct ICRn TOP BOTTOM
11 1 0 1 1 PWM, Phase Correct OCRnA TOP BOTTOM
12 1 1 0 0 CTC ICRn Immediate MAX
13 1 1 0 1 (Reserved) – – –
14 1 1 1 0 Fast PWM ICRn BOTTOM TOP
15 1 1 1 1 Fast PWM OCRnA BOTTOM TOP
Note: 1. The CTCn and PWMn1:0 bit definition names are obsolete. Use the WGMn2:0 definitions. However, the functionality and
location of these bits are compatible with previous versions of the timer.
14.11.2 TCCR1B – Timer/Counter1 Control Register B
Bit 7 6 5 4 3 2 1 0
(0x81) ICNC1 ICES1 – WGM13 WGM12 CS12 CS11 CS10 TCCR1B
Read/Write R/W R/W R R/W R/W R/W R/W R/W
• Bit 7 – ICNCn: Input Capture Noise Canceler

Setting this bit (to one) activates the Input Capture Noise Canceler. When the noise canceler is
activated, the input from the Input Capture pin (ICPn) is filtered. The filter function requires four
successive equal valued samples of the ICPn pin for changing its output. The Input Capture is
therefore delayed by four Oscillator cycles when the noise canceler is enabled.
• Bit 6 – ICESn: Input Capture Edge Select

This bit selects which edge on the Input Capture pin (ICPn) that is used to trigger a capture
event. When the ICESn bit is written to zero, a falling (negative) edge is used as trigger, and
when the ICESn bit is written to one, a rising (positive) edge will trigger the capture.
When a capture is triggered according to the ICESn setting, the counter value is copied into the
Input Capture Register (ICRn). The event will also set the Input Capture Flag (ICFn), and this
can be used to cause an Input Capture Interrupt, if this interrupt is enabled.
133
8011D–AVR–02/07
When the ICRn is used as TOP value (see description of the WGMn3:0 bits located in the
TCCRnA and the TCCRnB Register), the ICPn is disconnected and consequently the Input Cap-
ture function is disabled.
• Bit 5 – Reserved Bit

This bit is reserved for future use. For ensuring compatibility with future devices, this bit must be
written to zero when TCCRnB is written.
• Bit 4:3 – WGMn3:2: Waveform Generation Mode

See TCCRnA Register description.
• Bit 2:0 – CSn2:0: Clock Select

The three Clock Select bits select the clock source to be used by the Timer/Counter, see Figure
14-10 and Figure 14-11.

CSn2 CSn1 CSn0 Description
0 0 0 No clock source (Timer/Counter stopped).
0 0 1 clkI/O/1 (No prescaling)
1 1 0 External clock source on Tn pin. Clock on falling edge.
1 1 1 External clock source on Tn pin. Clock on rising edge.
If external pin modes are used for the Timer/Countern, transitions on the Tn pin will clock the
counting.
14.11.3 TCCR1C – Timer/Counter1 Control Register C
Bit 7 6 5 4 3 2 1 0
(0x82) FOC1A FOC1B – – – – – – TCCR1C
Read/Write R/W R/W R R R R R R
• Bit 7 – FOCnA: Force Output Compare for Channel A
• Bit 6 – FOCnB: Force Output Compare for Channel B

The FOCnA/FOCnB bits are only active when the WGMn3:0 bits specifies a non-PWM mode.
However, for ensuring compatibility with future devices, these bits must be set to zero when
TCCRnA is written when operating in a PWM mode. When writing a logical one to the
FOCnA/FOCnB bit, an immediate compare match is forced on the Waveform Generation unit.
The OCnA/OCnB output is changed according to its COMnx1:0 bits setting. Note that the
FOCnA/FOCnB bits are implemented as strobes. Therefore it is the value present in the
COMnx1:0 bits that determine the effect of the forced compare.
134 ATmega164P/324P/644P
8011D–AVR–02/07
A FOCnA/FOCnB strobe will not generate any interrupt nor will it clear the timer in Clear Timer
on Compare match (CTC) mode using OCRnA as TOP.
The FOCnA/FOCnB bits are always read as zero.
14.11.4 TCNT1H and TCNT1L –Timer/Counter1
Bit 7 6 5 4 3 2 1 0
(0x85) TCNT1[15:8] TCNT1H
(0x84) TCNT1[7:0] TCNT1L
The two Timer/Counter I/O locations (TCNTnH and TCNTnL, combined TCNTn) give direct
access, both for read and for write operations, to the Timer/Counter unit 16-bit counter. To
ensure that both the high and low bytes are read and written simultaneously when the CPU
accesses these registers, the access is performed using an 8-bit temporary High Byte Register
(TEMP). This temporary register is shared by all the other 16-bit registers. See Section “14.3” on
page 112.
Modifying the counter (TCNTn) while the counter is running introduces a risk of missing a com-
pare match between TCNTn and one of the OCRnx Registers.
Writing to the TCNTn Register blocks (removes) the compare match on the following timer clock
for all compare units.
14.11.5 OCR1AH and OCR1AL – Output Compare Register 1 A
Bit 7 6 5 4 3 2 1 0
(0x89) OCR1A[15:8] OCR1AH
(0x88) OCR1A[7:0] OCR1AL
14.11.6 OCR1BH and OCR1BL – Output Compare Register 1 B
Bit 7 6 5 4 3 2 1 0
(0x8B) OCR1B[15:8] OCR1BH
(0x8A) OCR1B[7:0] OCR1BL
The Output Compare Registers contain a 16-bit value that is continuously compared with the
counter value (TCNTn). A match can be used to generate an Output Compare interrupt, or to
generate a waveform output on the OCnx pin.
The Output Compare Registers are 16-bit in size. To ensure that both the high and low bytes are
written simultaneously when the CPU writes to these registers, the access is performed using an
8-bit temporary High Byte Register (TEMP). This temporary register is shared by all the other
16-bit registers. See Section “14.3” on page 112.
135
8011D–AVR–02/07
14.11.7 ICR1H and ICR1L – Input Capture Register 1
Bit 7 6 5 4 3 2 1 0
(0x87) ICR1[15:8] ICR1H
(0x86) ICR1[7:0] ICR1L
The Input Capture is updated with the counter (TCNTn) value each time an event occurs on the
ICPn pin (or optionally on the Analog Comparator output for Timer/Counter1). The Input Capture
can be used for defining the counter TOP value.
The Input Capture Register is 16-bit in size. To ensure that both the high and low bytes are read
simultaneously when the CPU accesses these registers, the access is performed using an 8-bit
temporary High Byte Register (TEMP). This temporary register is shared by all the other 16-bit
registers. See Section “14.3” on page 112.
14.11.8 TIMSK1 – Timer/Counter1 Interrupt Mask Register
Bit 7 6 5 4 3 2 1 0
(0x6F) – – ICIE1 – – OCIE1B OCIE1A TOIE1 TIMSK1
Read/Write R R R/W R R R/W R/W R/W
• Bit 7:6 – Res: Reserved Bits

These bits are unused bits in the ATmega164P/324P/644P, and will always read as zero.
• Bit 5 – ICIE1: Timer/Counter1, Input Capture Interrupt Enable

When this bit is written to one, and the I-flag in the Status Register is set (interrupts globally
enabled), the Timer/Counter1 Input Capture interrupt is enabled. The corresponding Interrupt
Vector (see “Interrupts” on page 61) is executed when the ICF1 Flag, located in TIFR1, is set.

• Bit 2 – OCIE1B: Timer/Counter1, Output Compare B Match Interrupt Enable

enabled), the Timer/Counter1 Output Compare B Match interrupt is enabled. The corresponding
Interrupt Vector (see “Interrupts” on page 61) is executed when the OCF1B Flag, located in
TIFR1, is set.
• Bit 1 – OCIE1A: Timer/Counter1, Output Compare A Match Interrupt Enable

enabled), the Timer/Counter1 Output Compare A Match interrupt is enabled. The corresponding
Interrupt Vector (see “Interrupts” on page 61) is executed when the OCF1A Flag, located in
TIFR1, is set.
• Bit 0 – TOIE1: Timer/Counter1, Overflow Interrupt Enable

enabled), the Timer/Counter1 Overflow interrupt is enabled. The corresponding Interrupt Vector
(See Section “9.2” on page 55.) is executed when the TOV1 Flag, located in TIFR1, is set.
136 ATmega164P/324P/644P
8011D–AVR–02/07
14.11.9 TIFR1 – Timer/Counter1 Interrupt Flag Register
Bit 7 6 5 4 3 2 1 0
0x16 (0x36) – – ICF1 – – OCF1B OCF1A TOV1 TIFR1
Read/Write R R R/W R R R/W R/W R/W

• Bit 5 – ICF1: Timer/Counter1, Input Capture Flag

This flag is set when a capture event occurs on the ICP1 pin. When the Input Capture Register
(ICR1) is set by the WGMn3:0 to be used as the TOP value, the ICF1 Flag is set when the
counter reaches the TOP value.
ICF1 is automatically cleared when the Input Capture Interrupt Vector is executed. Alternatively,
ICF1 can be cleared by writing a logic one to its bit location.

• Bit 2 – OCF1B: Timer/Counter1, Output Compare B Match Flag

This flag is set in the timer clock cycle after the counter (TCNT1) value matches the Output
Compare Register B (OCR1B).
Note that a Forced Output Compare (FOC1B) strobe will not set the OCF1B Flag.
OCF1B is automatically cleared when the Output Compare Match B Interrupt Vector is exe-
cuted. Alternatively, OCF1B can be cleared by writing a logic one to its bit location.
• Bit 1 – OCF1A: Timer/Counter1, Output Compare A Match Flag

This flag is set in the timer clock cycle after the counter (TCNT1) value matches the Output
Compare Register A (OCR1A).
Note that a Forced Output Compare (FOC1A) strobe will not set the OCF1A Flag.
OCF1A is automatically cleared when the Output Compare Match A Interrupt Vector is exe-
cuted. Alternatively, OCF1A can be cleared by writing a logic one to its bit location.
• Bit 0 – TOV1: Timer/Counter1, Overflow Flag

The setting of this flag is dependent of the WGMn3:0 bits setting. In Normal and CTC modes,
the TOV1 Flag is set when the timer overflows. Refer to Table 14-5 on page 133 for the TOV1
Flag behavior when using another WGMn3:0 bit setting.
TOV1 is automatically cleared when the Timer/Counter1 Overflow Interrupt Vector is executed.
Alternatively, TOV1 can be cleared by writing a logic one to its bit location.
137
8011D–AVR–02/07
15. 8-bit Timer/Counter2 with PWM and Asynchronous Operation
15.1 Features
• Single Channel Counter
• Glitch-free, Phase Correct Pulse Width Modulator (PWM)
• 10-bit Clock Prescaler
• Overflow and Compare Match Interrupt Sources (TOV2, OCF2A and OCF2B)
• Allows Clocking from External 32 kHz Watch Crystal Independent of the I/O Clock
15.2 Overview
Timer/Counter2 is a general purpose, single channel, 8-bit Timer/Counter module.
A simplified block diagram of the 8-bit Timer/Counter is shown in Figure 14-12.. For the actual
The Power Reduction Timer/Counter2 bit, PRTIM2, in ”PRR – Power Reduction Register” on
page 48 must be written to zero to enable Timer/Counter2 module.
Figure 15-1. 8-bit Timer/Counter Block Diagram
Count TOVn
Clear (Int.Req.)
Control Logic
Direction clkTn TOSC1
T/C
Oscillator
Prescaler TOSC2
TOP BOTTOM
clkI/O
Timer/Counter
TCNTn
= =0
OCnA
(Int.Req.)
Waveform
= Generation
OCnA
OCRnA
Fixed
OCnB
TOP
(Int.Req.)
Value
DATA BUS
Waveform
= Generation
OCnB
OCRnB clkI/O
Synchronized Status flags
Synchronization Unit
clkASY
asynchronous mode
Status flags select (ASn)
ASSRn
TCCRnA TCCRnB
138 ATmega164P/324P/644P
8011D–AVR–02/07
15.2.1 Registers
The Timer/Counter (TCNT2) and Output Compare Register (OCR2A and OCR2B) are 8-bit reg-
isters. Interrupt request (abbreviated to Int.Req.) signals are all visible in the Timer Interrupt Flag
Register (TIFR2). All interrupts are individually masked with the Timer Interrupt Mask Register
(TIMSK2). TIFR2 and TIMSK2 are not shown in the figure.
The Timer/Counter can be clocked internally, via the prescaler, or asynchronously clocked from
the TOSC1/2 pins, as detailed later in this section. The asynchronous operation is controlled by
the Asynchronous Status Register (ASSR). The Clock Select logic block controls which clock
source the Timer/Counter uses to increment (or decrement) its value. The Timer/Counter is inac-
tive when no clock source is selected. The output from the Clock Select logic is referred to as the
timer clock (clkT2).
The double buffered Output Compare Register (OCR2A and OCR2B) are compared with the
Timer/Counter value at all times. The result of the compare can be used by the Waveform Gen-
erator to generate a PWM or variable frequency output on the Output Compare pins (OC2A and
OC2B). See Section “15.5” on page 140. for details. The compare match event will also set the
Compare Flag (OCF2A or OCF2B) which can be used to generate an Output Compare interrupt
request.
15.2.2 Definitions
Many register and bit references in this document are written in general form. A lower case “n”
replaces the Timer/Counter number, in this case 2. However, when using the register or bit
defines in a program, the precise form must be used, i.e., TCNT2 for accessing Timer/Counter2
counter value and so on.
The definitions in Table 15-1 are also used extensively throughout the section.

BOTTOM The counter reaches the BOTTOM when it becomes zero (0x00).
MAX The counter reaches its MAXimum when it becomes 0xFF (decimal 255).
TOP The counter reaches the TOP when it becomes equal to the highest value in the
count sequence. The TOP value can be assigned to be the fixed value 0xFF
(MAX) or the value stored in the OCR2A Register. The assignment is depen-
dent on the mode of operation.

The Timer/Counter can be clocked by an internal synchronous or an external asynchronous
clock source. The clock source clkT2 is by default equal to the MCU clock, clkI/O. When the AS2
bit in the ASSR Register is written to logic one, the clock source is taken from the Timer/Counter
Oscillator connected to TOSC1 and TOSC2. For details on asynchronous operation, see ”ASSR
– Asynchronous Status Register” on page 156. For details on clock sources and prescaler, see
”Timer/Counter Prescaler” on page 151.
139
8011D–AVR–02/07
15.4 Counter Unit
The main part of the 8-bit Timer/Counter is the programmable bi-directional counter unit. Figure
15-2 shows a block diagram of the counter and its surrounding environment.

TOVn
DATA BUS (Int.Req.)
TOSC1
count
T/C
clear clk Tn
TCNTn Control Logic Prescaler Oscillator
direction
TOSC2
bottom top clkI/O

count Increment or decrement TCNT2 by 1.
direction Selects between increment and decrement.
clear Clear TCNT2 (set all bits to zero).
clkTn Timer/Counter clock, referred to as clkT2 in the following.
top Signalizes that TCNT2 has reached maximum value.
bottom Signalizes that TCNT2 has reached minimum value (zero).
Depending on the mode of operation used, the counter is cleared, incremented, or decremented
at each timer clock (clkT2). clkT2 can be generated from an external or internal clock source,
selected by the Clock Select bits (CS22:0). When no clock source is selected (CS22:0 = 0) the
timer is stopped. However, the TCNT2 value can be accessed by the CPU, regardless of
whether clkT2 is present or not. A CPU write overrides (has priority over) all counter clear or
count operations.
The counting sequence is determined by the setting of the WGM21 and WGM20 bits located in
the Timer/Counter Control Register (TCCR2A) and the WGM22 located in the Timer/Counter
Control Register B (TCCR2B). There are close connections between how the counter behaves
(counts) and how waveforms are generated on the Output Compare outputs OC2A and OC2B.
For more details about advanced counting sequences and waveform generation, see ”Modes of
Operation” on page 143.
The Timer/Counter Overflow Flag (TOV2) is set according to the mode of operation selected by
the WGM22:0 bits. TOV2 can be used for generating a CPU interrupt.
15.5 Output Compare Unit

The 8-bit comparator continuously compares TCNT2 with the Output Compare Register
(OCR2A and OCR2B). Whenever TCNT2 equals OCR2A or OCR2B, the comparator signals a
match. A match will set the Output Compare Flag (OCF2A or OCF2B) at the next timer clock
cycle. If the corresponding interrupt is enabled, the Output Compare Flag generates an Output
Compare interrupt. The Output Compare Flag is automatically cleared when the interrupt is exe-
cuted. Alternatively, the Output Compare Flag can be cleared by software by writing a logical
one to its I/O bit location. The Waveform Generator uses the match signal to generate an output
140 ATmega164P/324P/644P
8011D–AVR–02/07
according to operating mode set by the WGM22:0 bits and Compare Output mode (COM2x1:0)
bits. The max and bottom signals are used by the Waveform Generator for handling the special
cases of the extreme values in some modes of operation (”Modes of Operation” on page 143).
Figure 14-10 on page 129 shows a block diagram of the Output Compare unit.

DATA BUS
OCRnx TCNTn
OCFnx (Int.Req.)
top
bottom
FOCn
WGMn1:0 COMnX1:0
The OCR2x Register is double buffered when using any of the Pulse Width Modulation (PWM)
modes. For the Normal and Clear Timer on Compare (CTC) modes of operation, the double
buffering is disabled. The double buffering synchronizes the update of the OCR2x Compare
Register to either top or bottom of the counting sequence. The synchronization prevents the
occurrence of odd-length, non-symmetrical PWM pulses, thereby making the output glitch-free.
The OCR2x Register access may seem complex, but this is not case. When the double buffering
is enabled, the CPU has access to the OCR2x Buffer Register, and if double buffering is dis-
abled the CPU will access the OCR2x directly.
In non-PWM waveform generation modes, the match output of the comparator can be forced by
writing a one to the Force Output Compare (FOC2x) bit. Forcing compare match will not set the
OCF2x Flag or reload/clear the timer, but the OC2x pin will be updated as if a real compare
match had occurred (the COM2x1:0 bits settings define whether the OC2x pin is set, cleared or
toggled).
15.5.2 Compare Match Blocking by TCNT2 Write
All CPU write operations to the TCNT2 Register will block any compare match that occurs in the
next timer clock cycle, even when the timer is stopped. This feature allows OCR2x to be initial-
ized to the same value as TCNT2 without triggering an interrupt when the Timer/Counter clock is
enabled.
141
8011D–AVR–02/07
Since writing TCNT2 in any mode of operation will block all compare matches for one timer clock
cycle, there are risks involved when changing TCNT2 when using the Output Compare channel,
independently of whether the Timer/Counter is running or not. If the value written to TCNT2
equals the OCR2x value, the compare match will be missed, resulting in incorrect waveform
generation. Similarly, do not write the TCNT2 value equal to BOTTOM when the counter is
downcounting.
The setup of the OC2x should be performed before setting the Data Direction Register for the
port pin to output. The easiest way of setting the OC2x value is to use the Force Output Com-
pare (FOC2x) strobe bit in Normal mode. The OC2x Register keeps its value even when
Be aware that the COM2x1:0 bits are not double buffered together with the compare value.
Changing the COM2x1:0 bits will take effect immediately.

The Compare Output mode (COM2x1:0) bits have two functions. The Waveform Generator uses
the COM2x1:0 bits for defining the Output Compare (OC2x) state at the next compare match.
Also, the COM2x1:0 bits control the OC2x pin output source. Figure 15-4 shows a simplified
schematic of the logic affected by the COM2x1:0 bit setting. The I/O Registers, I/O bits, and I/O
(DDR and PORT) that are affected by the COM2x1:0 bits are shown. When referring to the
OC2x state, the reference is for the internal OC2x Register, not the OC2x pin.
COMnx1
COMnx0 Waveform
D Q
FOCnx Generator
1
OCnx
OCnx Pin
0
D Q
DATA BUS
PORT
D Q
DDR
clk I/O
The general I/O port function is overridden by the Output Compare (OC2x) from the Waveform
Generator if either of the COM2x1:0 bits are set. However, the OC2x pin direction (input or out-
142 ATmega164P/324P/644P
8011D–AVR–02/07
Register bit for the OC2x pin (DDR_OC2x) must be set as output before the OC2x value is visi-
ble on the pin. The port override function is independent of the Waveform Generation mode.
The design of the Output Compare pin logic allows initialization of the OC2x state before the out-
put is enabled. Note that some COM2x1:0 bit settings are reserved for certain modes of
operation. See ”Register Description” on page 151.
The Waveform Generator uses the COM2x1:0 bits differently in normal, CTC, and PWM modes.
For all modes, setting the COM2x1:0 = 0 tells the Waveform Generator that no action on the
OC2x Register is to be performed on the next compare match. For compare output actions in the
non-PWM modes refer to Table 15-5 on page 153. For fast PWM mode, refer to Table 15-6 on
page 153, and for phase correct PWM refer to Table 15-7 on page 153.
A change of the COM2x1:0 bits state will have effect at the first compare match after the bits are
FOC2x strobe bits.

defined by the combination of the Waveform Generation mode (WGM22:0) and Compare Output
mode (COM2x1:0) bits. The Compare Output mode bits do not affect the counting sequence,
while the Waveform Generation mode bits do. The COM2x1:0 bits control whether the PWM out-
the COM2x1:0 bits control whether the output should be set, cleared, or toggled at a compare
match (See Section “15.6” on page 142.).
For detailed timing information refer to ”Timer/Counter Timing Diagrams” on page 147.
15.7.1 Normal Mode
The simplest mode of operation is the Normal mode (WGM22:0 = 0). In this mode the counting
overruns when it passes its maximum 8-bit value (TOP = 0xFF) and then restarts from the bot-
tom (0x00). In normal operation the Timer/Counter Overflow Flag (TOV2) will be set in the same
timer clock cycle as the TCNT2 becomes zero. The TOV2 Flag in this case behaves like a ninth
bit, except that it is only set, not cleared. However, combined with the timer overflow interrupt
that automatically clears the TOV2 Flag, the timer resolution can be increased by software.
There are no special cases to consider in the Normal mode, a new counter value can be written
anytime.
The Output Compare unit can be used to generate interrupts at some given time. Using the Out-
put Compare to generate waveforms in Normal mode is not recommended, since this will
In Clear Timer on Compare or CTC mode (WGM22:0 = 2), the OCR2A Register is used to
manipulate the counter resolution. In CTC mode the counter is cleared to zero when the counter
value (TCNT2) matches the OCR2A. The OCR2A defines the top value for the counter, hence
also its resolution. This mode allows greater control of the compare match output frequency. It
also simplifies the operation of counting external events.
143
8011D–AVR–02/07
The timing diagram for the CTC mode is shown in Table 15-5 on page 144. The counter value
(TCNT2) increases until a compare match occurs between TCNT2 and OCR2A, and then
counter (TCNT2) is cleared.
TCNTn
OCnx
(COMnx1:0 = 1)
(Toggle)
Period 1 2 3 4
An interrupt can be generated each time the counter value reaches the TOP value by using the
OCF2A Flag. If the interrupt is enabled, the interrupt handler routine can be used for updating
the TOP value. However, changing TOP to a value close to BOTTOM when the counter is run-
ning with none or a low prescaler value must be done with care since the CTC mode does not
have the double buffering feature. If the new value written to OCR2A is lower than the current
value of TCNT2, the counter will miss the compare match. The counter will then have to count to
its maximum value (0xFF) and wrap around starting at 0x00 before the compare match can
occur.
For generating a waveform output in CTC mode, the OC2A output can be set to toggle its logical
level on each compare match by setting the Compare Output mode bits to toggle mode
(COM2A1:0 = 1). The OC2A value will not be visible on the port pin unless the data direction for
the pin is set to output. The waveform generated will have a maximum frequency of fOC2A =
fclk_I/O/2 when OCR2A is set to zero (0x00). The waveform frequency is defined by the following
equation:
f clk_I/O
f OCnx = -------------------------------------------------
-
2 ⋅ N ⋅ ( 1 + OCRnx )
The N variable represents the prescale factor (1, 8, 32, 64, 128, 256, or 1024).
As for the Normal mode of operation, the TOV2 Flag is set in the same timer clock cycle that the
The fast Pulse Width Modulation or fast PWM mode (WGM22:0 = 3 or 7) provides a high fre-
quency PWM waveform generation option. The fast PWM differs from the other PWM option by
its single-slope operation. The counter counts from BOTTOM to TOP then restarts from BOT-
TOM. TOP is defined as 0xFF when WGM22:0 = 3, and OCR2A when MGM22:0 = 7. In non-
inverting Compare Output mode, the Output Compare (OC2x) is cleared on the compare match
between TCNT2 and OCR2x, and set at BOTTOM. In inverting Compare Output mode, the out-
put is set on compare match and cleared at BOTTOM. Due to the single-slope operation, the
operating frequency of the fast PWM mode can be twice as high as the phase correct PWM
mode that uses dual-slope operation. This high frequency makes the fast PWM mode well suited
144 ATmega164P/324P/644P
8011D–AVR–02/07
for power regulation, rectification, and DAC applications. High frequency allows physically small
sized external components (coils, capacitors), and therefore reduces total system cost.
In fast PWM mode, the counter is incremented until the counter value matches the TOP value.
The counter is then cleared at the following timer clock cycle. The timing diagram for the fast
PWM mode is shown in Figure 15-6 on page 145. The TCNT2 value is in the timing diagram
shown as a histogram for illustrating the single-slope operation. The diagram includes non-
inverted and inverted PWM outputs. The small horizontal line marks on the TCNT2 slopes repre-
sent compare matches between OCR2x and TCNT2.
OCRnx Interrupt Flag Set
OCRnx Update and

TCNTn
OCnx (COMnx1:0 = 2)
OCnx (COMnx1:0 = 3)
Period 1 2 3 4 5 6 7
The Timer/Counter Overflow Flag (TOV2) is set each time the counter reaches TOP. If the inter-
rupt is enabled, the interrupt handler routine can be used for updating the compare value.
In fast PWM mode, the compare unit allows generation of PWM waveforms on the OC2x pin.
Setting the COM2x1:0 bits to two will produce a non-inverted PWM and an inverted PWM output
can be generated by setting the COM2x1:0 to three. TOP is defined as 0xFF when WGM2:0 = 3,
and OCR2A when WGM2:0 = 7 (See Table 15-3 on page 152). The actual OC2x value will only
be visible on the port pin if the data direction for the port pin is set as output. The PWM wave-
form is generated by setting (or clearing) the OC2x Register at the compare match between
OCR2x and TCNT2, and clearing (or setting) the OC2x Register at the timer clock cycle the
counter is cleared (changes from TOP to BOTTOM).
f clk_I/O
f OCnxPWM = -----------------
-
N ⋅ 256
waveform output in the fast PWM mode. If the OCR2A is set equal to BOTTOM, the output will
be a narrow spike for each MAX+1 timer clock cycle. Setting the OCR2A equal to MAX will result
in a constantly high or low output (depending on the polarity of the output set by the COM2A1:0
bits.)
145
8011D–AVR–02/07
ting OC2x to toggle its logical level on each compare match (COM2x1:0 = 1). The waveform
generated will have a maximum frequency of foc2 = fclk_I/O/2 when OCR2A is set to zero. This fea-
ture is similar to the OC2A toggle in CTC mode, except the double buffer feature of the Output
Compare unit is enabled in the fast PWM mode.
The phase correct PWM mode (WGM22:0 = 1 or 5) provides a high resolution phase correct
PWM waveform generation option. The phase correct PWM mode is based on a dual-slope
operation. The counter counts repeatedly from BOTTOM to TOP and then from TOP to BOT-
TOM. TOP is defined as 0xFF when WGM22:0 = 1, and OCR2A when MGM22:0 = 5. In non-
inverting Compare Output mode, the Output Compare (OC2x) is cleared on the compare match
between TCNT2 and OCR2x while upcounting, and set on the compare match while downcount-
ing. In inverting Output Compare mode, the operation is inverted. The dual-slope operation has
lower maximum operation frequency than single slope operation. However, due to the symmet-
ric feature of the dual-slope PWM modes, these modes are preferred for motor control
applications.
In phase correct PWM mode the counter is incremented until the counter value matches TOP.
When the counter reaches TOP, it changes the count direction. The TCNT2 value will be equal
to TOP for one timer clock cycle. The timing diagram for the phase correct PWM mode is shown
on Figure 15-7. The TCNT2 value is in the timing diagram shown as a histogram for illustrating
the dual-slope operation. The diagram includes non-inverted and inverted PWM outputs. The
small horizontal line marks on the TCNT2 slopes represent compare matches between OCR2x
and TCNT2.
OCRnx Update
TCNTn
OCnx (COMnx1:0 = 2)
OCnx (COMnx1:0 = 3)
Period 1 2 3
The Timer/Counter Overflow Flag (TOV2) is set each time the counter reaches BOTTOM. The
Interrupt Flag can be used to generate an interrupt each time the counter reaches the BOTTOM
value.
146 ATmega164P/324P/644P
8011D–AVR–02/07
In phase correct PWM mode, the compare unit allows generation of PWM waveforms on the
OC2x pin. Setting the COM2x1:0 bits to two will produce a non-inverted PWM. An inverted PWM
output can be generated by setting the COM2x1:0 to three. TOP is defined as 0xFF when
WGM2:0 = 3, and OCR2A when MGM2:0 = 7 (See Table 15-4 on page 152). The actual OC2x
value will only be visible on the port pin if the data direction for the port pin is set as output. The
PWM waveform is generated by clearing (or setting) the OC2x Register at the compare match
between OCR2x and TCNT2 when the counter increments, and setting (or clearing) the OC2x
Register at compare match between OCR2x and TCNT2 when the counter decrements. The
PWM frequency for the output when using phase correct PWM can be calculated by the follow-
ing equation:
f clk_I/O
f OCnxPCPWM = -----------------
-
N ⋅ 510
waveform output in the phase correct PWM mode. If the OCR2A is set equal to BOTTOM, the
output will be continuously low and if set equal to MAX the output will be continuously high for
non-inverted PWM mode. For inverted PWM the output will have the opposite logic values.
At the very start of period 2 in Figure 15-7 on page 146 OCnx has a transition from high to low
even though there is no Compare Match. The point of this transition is to guarantee symmetry
around BOTTOM. There are two cases that give a transition without Compare Match.
• OCR2A changes its value from MAX, like in Figure 15-7 on page 146. When the OCR2A value
is MAX the OCn pin value is the same as the result of a down-counting compare match. To
ensure symmetry around BOTTOM the OCn value at MAX must correspond to the result of an
up-counting Compare Match.
• The timer starts counting from a value higher than the one in OCR2A, and for that reason
misses the Compare Match and hence the OCn change that would have happened on the way
up.

The following figures show the Timer/Counter in synchronous mode, and the timer clock (clkT2)
is therefore shown as a clock enable signal. In asynchronous mode, clkI/O should be replaced by
the Timer/Counter Oscillator clock. The figures include information on when Interrupt Flags are
set. Figure 15-8 on page 148 contains timing data for basic Timer/Counter operation. The figure
shows the count sequence close to the MAX value in all modes other than phase correct PWM
mode.
147
8011D–AVR–02/07
clkI/O
clkTn
(clkI/O /1)
TOVn
Figure 15-9 on page 148 shows the same timing data, but with the prescaler enabled.
clkI/O
clkTn
(clkI/O /8)
TOVn
Figure 15-10 on page 148 shows the setting of OCF2A in all modes except CTC mode.
Figure 15-10. Timer/Counter Timing Diagram, Setting of OCF2A, with Prescaler (fclk_I/O/8)
clkI/O
clkTn
(clkI/O /8)
OCRnx OCRnx Value
OCFnx
148 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 15-11 on page 149 shows the setting of OCF2A and the clearing of TCNT2 in CTC mode.
Figure 15-11. Timer/Counter Timing Diagram, Clear Timer on Compare Match mode, with Pres-
caler (fclk_I/O/8)
clkI/O
clkTn
(clkI/O /8)
TCNTn
(CTC)
OCRnx TOP
OCFnx
15.9 Asynchronous Operation of Timer/Counter2

When Timer/Counter2 operates asynchronously, some considerations must be taken.
• Warning: When switching between asynchronous and synchronous clocking of
Timer/Counter2, the Timer Registers TCNT2, OCR2x, and TCCR2x might be corrupted. A safe
procedure for switching clock source is:
a. Disable the Timer/Counter2 interrupts by clearing OCIE2x and TOIE2.
b. Select clock source by setting AS2 as appropriate.
c. Write new values to TCNT2, OCR2x, and TCCR2x.
d. To switch to asynchronous operation: Wait for TCN2UB, OCR2xUB, and TCR2xUB.
e. Clear the Timer/Counter2 Interrupt Flags.
f. Enable interrupts, if needed.
• The CPU main clock frequency must be more than four times the Oscillator frequency.
• When writing to one of the registers TCNT2, OCR2x, or TCCR2x, the value is transferred to a
temporary register, and latched after two positive edges on TOSC1. The user should not write
a new value before the contents of the temporary register have been transferred to its
destination. Each of the five mentioned registers have their individual temporary register, which
means that e.g. writing to TCNT2 does not disturb an OCR2x write in progress. To detect that a
transfer to the destination register has taken place, the Asynchronous Status Register – ASSR
has been implemented.
• When entering Power-save or ADC Noise Reduction mode after having written to TCNT2,
OCR2x, or TCCR2x, the user must wait until the written register has been updated if
Timer/Counter2 is used to wake up the device. Otherwise, the MCU will enter sleep mode
before the changes are effective. This is particularly important if any of the Output Compare2
interrupt is used to wake up the device, since the Output Compare function is disabled during
writing to OCR2x or TCNT2. If the write cycle is not finished, and the MCU enters sleep mode
149
8011D–AVR–02/07
before the corresponding OCR2xUB bit returns to zero, the device will never receive a
compare match interrupt, and the MCU will not wake up.
• If Timer/Counter2 is used to wake the device up from Power-save or ADC Noise Reduction
mode, precautions must be taken if the user wants to re-enter one of these modes: The
interrupt logic needs one TOSC1 cycle to be reset. If the time between wake-up and re-
entering sleep mode is less than one TOSC1 cycle, the interrupt will not occur, and the device
will fail to wake up. If the user is in doubt whether the time before re-entering Power-save or
ADC Noise Reduction mode is sufficient, the following algorithm can be used to ensure that
one TOSC1 cycle has elapsed:
a. Write a value to TCCR2x, TCNT2, or OCR2x.
b. Wait until the corresponding Update Busy Flag in ASSR returns to zero.
c. Enter Power-save or ADC Noise Reduction mode.
• When the asynchronous operation is selected, the 32.768 kHz Oscillator for Timer/Counter2 is
always running, except in Power-down and Standby modes. After a Power-up Reset or wake-
up from Power-down or Standby mode, the user should be aware of the fact that this Oscillator
might take as long as one second to stabilize. The user is advised to wait for at least one
second before using Timer/Counter2 after power-up or wake-up from Power-down or Standby
mode. The contents of all Timer/Counter2 Registers must be considered lost after a wake-up
from Power-down or Standby mode due to unstable clock signal upon start-up, no matter
whether the Oscillator is in use or a clock signal is applied to the TOSC1 pin.
• Description of wake up from Power-save or ADC Noise Reduction mode when the timer is
clocked asynchronously: When the interrupt condition is met, the wake up process is started
on the following cycle of the timer clock, that is, the timer is always advanced by at least one
before the processor can read the counter value. After wake-up, the MCU is halted for four
cycles, it executes the interrupt routine, and resumes execution from the instruction following
SLEEP.
• Reading of the TCNT2 Register shortly after wake-up from Power-save may give an incorrect
result. Since TCNT2 is clocked on the asynchronous TOSC clock, reading TCNT2 must be
done through a register synchronized to the internal I/O clock domain. Synchronization takes
place for every rising TOSC1 edge. When waking up from Power-save mode, and the I/O clock
(clkI/O) again becomes active, TCNT2 will read as the previous value (before entering sleep)
until the next rising TOSC1 edge. The phase of the TOSC clock after waking up from Power-
save mode is essentially unpredictable, as it depends on the wake-up time. The recommended
procedure for reading TCNT2 is thus as follows:
a. Write any value to either of the registers OCR2x or TCCR2x.
b. Wait for the corresponding Update Busy Flag to be cleared.
c. Read TCNT2.
• During asynchronous operation, the synchronization of the Interrupt Flags for the
asynchronous timer takes 3 processor cycles plus one timer cycle. The timer is therefore
advanced by at least one before the processor can read the timer value causing the setting of
the Interrupt Flag. The Output Compare pin is changed on the timer clock and is not
synchronized to the processor clock.
150 ATmega164P/324P/644P
8011D–AVR–02/07
15.10 Timer/Counter Prescaler
Figure 15-12. Prescaler for Timer/Counter2
clkI/O clkT2S
10-BIT T/C PRESCALER
Clear
TOSC1
clkT2S/8
clkT2S/32
clkT2S/64
clkT2S/128
clkT2S/256
clkT2S/1024
AS2
PSRASY 0
CS20
CS21
CS22
TIMER/COUNTER2 CLOCK SOURCE

clkT2
The clock source for Timer/Counter2 is named clkT2S. clkT2S is by default connected to the main
system I/O clock clk IO. By setting the AS2 bit in ASSR, Timer/Counter2 is asynchronously
clocked from the TOSC1 pin. This enables use of Timer/Counter2 as a Real Time Counter
(RTC). When AS2 is set, pins TOSC1 and TOSC2 are disconnected from Port C. A crystal can
then be connected between the TOSC1 and TOSC2 pins to serve as an independent clock
source for Timer/Counter2. The Oscillator is optimized for use with a 32.768 kHz crystal. By set-
ting the EXCLK bit in the ASSR a 32 kHz external clock can be applied. See ”ASSR –
Asynchronous Status Register” on page 156 for details.
For Timer/Counter2, the possible prescaled selections are: clk T2S /8, clk T2S /32, clk T2S /64,
clkT2S/128, clkT2S/256, and clkT2S/1024. Additionally, clkT2S as well as 0 (stop) may be selected.
Setting the PSRASY bit in GTCCR resets the prescaler. This allows the user to operate with a
predictable prescaler.

15.11.1 TCCR2A – Timer/Counter Control Register A
Bit 7 6 5 4 3 2 1 0
(0xB0) COM2A1 COM2A0 COM2B1 COM2B0 – – WGM21 WGM20 TCCR2A
• Bits 7:6 – COM2A1:0: Compare Match Output A Mode

These bits control the Output Compare pin (OC2A) behavior. If one or both of the COM2A1:0
bits are set, the OC2A output overrides the normal port functionality of the I/O pin it is connected
to. However, note that the Data Direction Register (DDR) bit corresponding to the OC2A pin
151
8011D–AVR–02/07
When OC2A is connected to the pin, the function of the COM2A1:0 bits depends on the
WGM22:0 bit setting. Table 15-2 shows the COM2A1:0 bit functionality when the WGM22:0 bits

0 1 Toggle OC2A on Compare Match
1 0 Clear OC2A on Compare Match
1 1 Set OC2A on Compare Match
Table 15-3 shows the COM2A1:0 bit functionality when the WGM21:0 bits are set to fast PWM
mode.

0 1
Clear OC2A on Compare Match, set OC2A at BOTTOM,
1 0
Set OC2A on Compare Match, clear OC2A at BOTTOM,
1 1
(inverting mode).
Table 15-4 shows the COM2A1:0 bit functionality when the WGM22:0 bits are set to phase cor-
rect PWM mode.

0 1
Clear OC2A on Compare Match when up-counting. Set OC2A on
1 0
Set OC2A on Compare Match when up-counting. Clear OC2A on
1 1
• Bits 5:4 – COM2B1:0: Compare Match Output B Mode

These bits control the Output Compare pin (OC2B) behavior. If one or both of the COM2B1:0
bits are set, the OC2B output overrides the normal port functionality of the I/O pin it is connected
to. However, note that the Data Direction Register (DDR) bit corresponding to the OC2B pin
152 ATmega164P/324P/644P
8011D–AVR–02/07
When OC2B is connected to the pin, the function of the COM2B1:0 bits depends on the
WGM22:0 bit setting. Table 15-5 shows the COM2B1:0 bit functionality when the WGM22:0 bits

0 1 Toggle OC2B on Compare Match
1 0 Clear OC2B on Compare Match
1 1 Set OC2B on Compare Match
Table 15-6 shows the COM2B1:0 bit functionality when the WGM22:0 bits are set to fast PWM
mode.

0 1 Reserved
Clear OC2B on Compare Match, set OC2B at BOTTOM,
1 0
Set OC2B on Compare Match, clear OC2B at BOTTOM,
1 1
(inverting mode).
Table 15-7 shows the COM2B1:0 bit functionality when the WGM22:0 bits are set to phase cor-
rect PWM mode.

0 1 Reserved
Clear OC2B on Compare Match when up-counting. Set OC2B on
1 0
Set OC2B on Compare Match when up-counting. Clear OC2B on
1 1

153
8011D–AVR–02/07
• Bits 1:0 – WGM21:0: Waveform Generation Mode
Combined with the WGM22 bit found in the TCCR2B Register, these bits control the counting
form generation to be used, see Table 15-8. Modes of operation supported by the Timer/Counter
unit are: Normal mode (counter), Clear Timer on Compare Match (CTC) mode, and two types of
Pulse Width Modulation (PWM) modes (see ”Modes of Operation” on page 143).
Table 15-8. Waveform Generation Mode Bit Description

Timer/Counter
Mode of Update of TOV Flag
Mode WGM2 WGM1 WGM0 Operation TOP OCRx at Set on(1)(2)
0 0 0 0 Normal 0xFF Immediate MAX
PWM, Phase
1 0 0 1 0xFF TOP BOTTOM
Correct
2 0 1 0 CTC OCRA Immediate MAX
3 0 1 1 Fast PWM 0xFF BOTTOM MAX
4 1 0 0 Reserved – – –
PWM, Phase
5 1 0 1 OCRA TOP BOTTOM
Correct
6 1 1 0 Reserved – – –
7 1 1 1 Fast PWM OCRA BOTTOM TOP
Notes: 1. MAX= 0xFF

2. BOTTOM= 0x00
15.11.2 TCCR2B – Timer/Counter Control Register B
Bit 7 6 5 4 3 2 1 0
(0xB1) FOC2A FOC2B – – WGM22 CS22 CS21 CS20 TCCR2B
Read/Write W W R R R/W R/W R/W R/W
• Bit 7 – FOC2A: Force Output Compare A

The FOC2A bit is only active when the WGM bits specify a non-PWM mode.
TCCR2B is written when operating in PWM mode. When writing a logical one to the FOC2A bit,
an immediate Compare Match is forced on the Waveform Generation unit. The OC2A output is
changed according to its COM2A1:0 bits setting. Note that the FOC2A bit is implemented as a
strobe. Therefore it is the value present in the COM2A1:0 bits that determines the effect of the
forced compare.
A FOC2A strobe will not generate any interrupt, nor will it clear the timer in CTC mode using
OCR2A as TOP.
The FOC2A bit is always read as zero.
154 ATmega164P/324P/644P
8011D–AVR–02/07
• Bit 6 – FOC2B: Force Output Compare B
The FOC2B bit is only active when the WGM bits specify a non-PWM mode.
TCCR2B is written when operating in PWM mode. When writing a logical one to the FOC2B bit,
an immediate Compare Match is forced on the Waveform Generation unit. The OC2B output is
changed according to its COM2B1:0 bits setting. Note that the FOC2B bit is implemented as a
strobe. Therefore it is the value present in the COM2B1:0 bits that determines the effect of the
forced compare.
A FOC2B strobe will not generate any interrupt, nor will it clear the timer in CTC mode using
OCR2B as TOP.
The FOC2B bit is always read as zero.

• Bit 3 – WGM22: Waveform Generation Mode

See the description in the ”TCCR2A – Timer/Counter Control Register A” on page 151.
• Bit 2:0 – CS22:0: Clock Select

The three Clock Select bits select the clock source to be used by the Timer/Counter, see Table
15-9 on page 155.

CS22 CS21 CS20 Description
0 0 0 No clock source (Timer/Counter stopped).
0 0 1 clkT2S/(No prescaling)
0 1 0 clkT2S/8 (From prescaler)
If external pin modes are used for the Timer/Counter0, transitions on the T0 pin will clock the
counting.
15.11.3 TCNT2 – Timer/Counter Register
Bit 7 6 5 4 3 2 1 0
(0xB2) TCNT2[7:0] TCNT2
The Timer/Counter Register gives direct access, both for read and write operations, to the
Timer/Counter unit 8-bit counter. Writing to the TCNT2 Register blocks (removes) the Compare
155
8011D–AVR–02/07
Match on the following timer clock. Modifying the counter (TCNT2) while the counter is running,
introduces a risk of missing a Compare Match between TCNT2 and the OCR2x Registers.
15.11.4 OCR2A – Output Compare Register A
Bit 7 6 5 4 3 2 1 0
(0xB3) OCR2A[7:0] OCR2A
The Output Compare Register A contains an 8-bit value that is continuously compared with the
generate a waveform output on the OC2A pin.
15.11.5 OCR2B – Output Compare Register B
Bit 7 6 5 4 3 2 1 0
(0xB4) OCR2B[7:0] OCR2B
The Output Compare Register B contains an 8-bit value that is continuously compared with the
generate a waveform output on the OC2B pin.
15.11.6 ASSR – Asynchronous Status Register
Bit 7 6 5 4 3 2 1 0
(0xB6) – EXCLK AS2 TCN2UB OCR2AUB OCR2BUB TCR2AUB TCR2BUB ASSR
Read/Write R R/W R/W R R R R R
• Bit 6 – EXCLK: Enable External Clock Input

When EXCLK is written to one, and asynchronous clock is selected, the external clock input
buffer is enabled and an external clock can be input on Timer Oscillator 1 (TOSC1) pin instead
of a 32 kHz crystal. Writing to EXCLK should be done before asynchronous operation is
selected. Note that the crystal Oscillator will only run when this bit is zero.
• Bit 5 – AS2: Asynchronous Timer/Counter2

When AS2 is written to zero, Timer/Counter2 is clocked from the I/O clock, clkI/O. When AS2 is
written to one, Timer/Counter2 is clocked from a crystal Oscillator connected to the Timer Oscil-
lator 1 (TOSC1) pin. When the value of AS2 is changed, the contents of TCNT2, OCR2A,
OCR2B, TCCR2A and TCCR2B might be corrupted.
• Bit 4 – TCN2UB: Timer/Counter2 Update Busy

When Timer/Counter2 operates asynchronously and TCNT2 is written, this bit becomes set.
When TCNT2 has been updated from the temporary storage register, this bit is cleared by hard-
ware. A logical zero in this bit indicates that TCNT2 is ready to be updated with a new value.
• Bit 3 – OCR2AUB: Output Compare Register2 Update Busy

When Timer/Counter2 operates asynchronously and OCR2A is written, this bit becomes set.
When OCR2A has been updated from the temporary storage register, this bit is cleared by hard-
ware. A logical zero in this bit indicates that OCR2A is ready to be updated with a new value.
156 ATmega164P/324P/644P
8011D–AVR–02/07
• Bit 2 – OCR2BUB: Output Compare Register2 Update Busy
When Timer/Counter2 operates asynchronously and OCR2B is written, this bit becomes set.
When OCR2B has been updated from the temporary storage register, this bit is cleared by hard-
ware. A logical zero in this bit indicates that OCR2B is ready to be updated with a new value.
• Bit 1 – TCR2AUB: Timer/Counter Control Register2 Update Busy

When Timer/Counter2 operates asynchronously and TCCR2A is written, this bit becomes set.
When TCCR2A has been updated from the temporary storage register, this bit is cleared by
hardware. A logical zero in this bit indicates that TCCR2A is ready to be updated with a new
value.
• Bit 0 – TCR2BUB: Timer/Counter Control Register2 Update Busy

When Timer/Counter2 operates asynchronously and TCCR2B is written, this bit becomes set.
When TCCR2B has been updated from the temporary storage register, this bit is cleared by
hardware. A logical zero in this bit indicates that TCCR2B is ready to be updated with a new
value.
If a write is performed to any of the five Timer/Counter2 Registers while its update busy flag is
set, the updated value might get corrupted and cause an unintentional interrupt to occur.
The mechanisms for reading TCNT2, OCR2A, OCR2B, TCCR2A and TCCR2B are different.
When reading TCNT2, the actual timer value is read. When reading OCR2A, OCR2B, TCCR2A
and TCCR2B the value in the temporary storage register is read.
15.11.7 TIMSK2 – Timer/Counter2 Interrupt Mask Register
Bit 7 6 5 4 3 2 1 0
(0x70) – – – – – OCIE2B OCIE2A TOIE2 TIMSK2
• Bit 2 – OCIE2B: Timer/Counter2 Output Compare Match B Interrupt Enable

When the OCIE2B bit is written to one and the I-bit in the Status Register is set (one), the
Timer/Counter2 Compare Match B interrupt is enabled. The corresponding interrupt is executed
if a compare match in Timer/Counter2 occurs, i.e., when the OCF2B bit is set in the
• Bit 1 – OCIE2A: Timer/Counter2 Output Compare Match A Interrupt Enable

When the OCIE2A bit is written to one and the I-bit in the Status Register is set (one), the
Timer/Counter2 Compare Match A interrupt is enabled. The corresponding interrupt is executed
if a compare match in Timer/Counter2 occurs, i.e., when the OCF2A bit is set in the
• Bit 0 – TOIE2: Timer/Counter2 Overflow Interrupt Enable

When the TOIE2 bit is written to one and the I-bit in the Status Register is set (one), the
Timer/Counter2 Overflow interrupt is enabled. The corresponding interrupt is executed if an
overflow in Timer/Counter2 occurs, i.e., when the TOV2 bit is set in the Timer/Counter2 Interrupt
Flag Register – TIFR2.
157
8011D–AVR–02/07
15.11.8 TIFR2 – Timer/Counter2 Interrupt Flag Register
Bit 7 6 5 4 3 2 1 0
0x17 (0x37) – – – – – OCF2B OCF2A TOV2 TIFR2
• Bit 2 – OCF2B: Output Compare Flag 2 B

The OCF2B bit is set (one) when a compare match occurs between the Timer/Counter2 and the
data in OCR2B – Output Compare Register2. OCF2B is cleared by hardware when executing
the corresponding interrupt handling vector. Alternatively, OCF2B is cleared by writing a logic
one to the flag. When the I-bit in SREG, OCIE2B (Timer/Counter2 Compare match Interrupt
Enable), and OCF2B are set (one), the Timer/Counter2 Compare match Interrupt is executed.
• Bit 1 – OCF2A: Output Compare Flag 2 A

The OCF2A bit is set (one) when a compare match occurs between the Timer/Counter2 and the
data in OCR2A – Output Compare Register2. OCF2A is cleared by hardware when executing
the corresponding interrupt handling vector. Alternatively, OCF2A is cleared by writing a logic
one to the flag. When the I-bit in SREG, OCIE2A (Timer/Counter2 Compare match Interrupt
Enable), and OCF2A are set (one), the Timer/Counter2 Compare match Interrupt is executed.
• Bit 0 – TOV2: Timer/Counter2 Overflow Flag

The TOV2 bit is set (one) when an overflow occurs in Timer/Counter2. TOV2 is cleared by hard-
ware when executing the corresponding interrupt handling vector. Alternatively, TOV2 is cleared
by writing a logic one to the flag. When the SREG I-bit, TOIE2A (Timer/Counter2 Overflow Inter-
rupt Enable), and TOV2 are set (one), the Timer/Counter2 Overflow interrupt is executed. In
PWM mode, this bit is set when Timer/Counter2 changes counting direction at 0x00.
15.11.9 GTCCR – General Timer/Counter Control Register
Bit 7 6 5 4 3 2 1 0
0x23 (0x43) TSM – – – – – PSRASY PSRSYNC GTCCR
Read/Write R/W R R R R R R/W R/W
• Bit 1 – PSRASY: Prescaler Reset Timer/Counter2

When this bit is one, the Timer/Counter2 prescaler will be reset. This bit is normally cleared
immediately by hardware. If the bit is written when Timer/Counter2 is operating in asynchronous
mode, the bit will remain one until the prescaler has been reset. The bit will not be cleared by
hardware if the TSM bit is set. Refer to the description of the “Bit 7 – TSM: Timer/Counter Syn-
chronization Mode” on page 136 for a description of the Timer/Counter Synchronization mode.
158 ATmega164P/324P/644P
8011D–AVR–02/07
16. SPI – Serial Peripheral Interface

16.1 Features
• Full-duplex, Three-wire Synchronous Data Transfer
• Master or Slave Operation
• LSB First or MSB First Data Transfer
• Seven Programmable Bit Rates
• End of Transmission Interrupt Flag
• Write Collision Flag Protection
• Wake-up from Idle Mode
• Double Speed (CK/2) Master SPI Mode
16.2 Overview
The Serial Peripheral Interface (SPI) allows high-speed synchronous data transfer between the
ATmega164P/324P/644P and peripheral devices or between several AVR devices.
USART can also be used in Master SPI mode, see “USART in SPI Mode” on page 196.
The Power Reduction SPI bit, PRSPI, in ”PRR – Power Reduction Register” on page 48 on page
50 must be written to zero to enable SPI module.
Figure 16-1. SPI Block Diagram(1)
DIVIDER
/2/4/8/16/32/64/128
SPI2X
SPI2X
Note: 1. Refer to Figure 1-1 on page 2, and Table 12-6 on page 81 for SPI pin placement.
159
8011D–AVR–02/07
The interconnection between Master and Slave CPUs with SPI is shown in Figure 16-2. The sys-
tem consists of two shift Registers, and a Master clock generator. The SPI Master initiates the
communication cycle when pulling low the Slave Select SS pin of the desired Slave. Master and
Slave prepare the data to be sent in their respective shift Registers, and the Master generates
the required clock pulses on the SCK line to interchange data. Data is always shifted from Mas-
ter to Slave on the Master Out – Slave In, MOSI, line, and from Slave to Master on the Master In
– Slave Out, MISO, line. After each data packet, the Master will synchronize the Slave by pulling
high the Slave Select, SS, line.
When configured as a Master, the SPI interface has no automatic control of the SS line. This
must be handled by user software before communication can start. When this is done, writing a
byte to the SPI Data Register starts the SPI clock generator, and the hardware shifts the eight
bits into the Slave. After shifting one byte, the SPI clock generator stops, setting the end of
Transmission Flag (SPIF). If the SPI Interrupt Enable bit (SPIE) in the SPCR Register is set, an
interrupt is requested. The Master may continue to shift the next byte by writing it into SPDR, or
signal the end of packet by pulling high the Slave Select, SS line. The last incoming byte will be
kept in the Buffer Register for later use.
When configured as a Slave, the SPI interface will remain sleeping with MISO tri-stated as long
as the SS pin is driven high. In this state, software may update the contents of the SPI Data
Register, SPDR, but the data will not be shifted out by incoming clock pulses on the SCK pin
until the SS pin is driven low. As one byte has been completely shifted, the end of Transmission
Flag, SPIF is set. If the SPI Interrupt Enable bit, SPIE, in the SPCR Register is set, an interrupt
is requested. The Slave may continue to place new data to be sent into SPDR before reading
the incoming data. The last incoming byte will be kept in the Buffer Register for later use.
Figure 16-2. SPI Master-slave Interconnection
SHIFT
ENABLE
The system is single buffered in the transmit direction and double buffered in the receive direc-
tion. This means that bytes to be transmitted cannot be written to the SPI Data Register before
the entire shift cycle is completed. When receiving data, however, a received character must be
read from the SPI Data Register before the next character has been completely shifted in. Oth-
erwise, the first byte is lost.
In SPI Slave mode, the control logic will sample the incoming signal of the SCK pin. To ensure
correct sampling of the clock signal, the minimum low and high periods should be:
Low period: longer than 2 CPU clock cycles.
High period: longer than 2 CPU clock cycles.
160 ATmega164P/324P/644P
8011D–AVR–02/07
When the SPI is enabled, the data direction of the MOSI, MISO, SCK, and SS pins is overridden
according to Table 16-1. For more details on automatic port overrides, refer to ”Alternate Port
Functions” on page 77.
Table 16-1. SPI Pin Overrides(1)
Pin Direction, Master SPI Direction, Slave SPI
MOSI User Defined Input
MISO Input User Defined
SCK User Defined Input
SS User Defined Input
Note: 1. See ”Alternate Functions of Port B” on page 81 for a detailed description of how to define the
direction of the user defined SPI pins.
The following code examples show how to initialize the SPI as a Master and how to perform a
simple transmission. DDR_SPI in the examples must be replaced by the actual Data Direction
Register controlling the SPI pins. DD_MOSI, DD_MISO and DD_SCK must be replaced by the
actual data direction bits for these pins. E.g. if MOSI is placed on pin PB5, replace DD_MOSI
with DDB5 and DDR_SPI with DDRB.
161
8011D–AVR–02/07
SPI_MasterInit:
; Set MOSI and SCK output, all others input
ldi r17,(1<<DD_MOSI)|(1<<DD_SCK)
out DDR_SPI,r17
; Enable SPI, Master, set clock rate fck/16
ldi r17,(1<<SPE)|(1<<MSTR)|(1<<SPR0)
out SPCR,r17
ret
SPI_MasterTransmit:
; Start transmission of data (r16)
out SPDR,r16
Wait_Transmit:
; Wait for transmission complete
sbis SPSR,SPIF
rjmp Wait_Transmit
ret
C Code Example(1)
void SPI_MasterInit(void)
{
/* Set MOSI and SCK output, all others input */
DDR_SPI = (1<<DD_MOSI)|(1<<DD_SCK);
/* Enable SPI, Master, set clock rate fck/16 */
SPCR = (1<<SPE)|(1<<MSTR)|(1<<SPR0);
}
void SPI_MasterTransmit(char cData)

{
/* Start transmission */
SPDR = cData;
/* Wait for transmission complete */
while(!(SPSR & (1<<SPIF)))
;
}
162 ATmega164P/324P/644P
8011D–AVR–02/07
The following code examples show how to initialize the SPI as a Slave and how to perform a
simple reception.

SPI_SlaveInit:
; Set MISO output, all others input
ldi r17,(1<<DD_MISO)
out DDR_SPI,r17
; Enable SPI
ldi r17,(1<<SPE)
out SPCR,r17
ret
SPI_SlaveReceive:
; Wait for reception complete
sbis SPSR,SPIF
rjmp SPI_SlaveReceive
; Read received data and return
in r16,SPDR
ret
C Code Example(1)
void SPI_SlaveInit(void)
{
/* Set MISO output, all others input */
DDR_SPI = (1<<DD_MISO);
/* Enable SPI */
SPCR = (1<<SPE);
}
char SPI_SlaveReceive(void)
{
/* Wait for reception complete */
while(!(SPSR & (1<<SPIF)))
;
/* Return Data Register */
return SPDR;
}
163
8011D–AVR–02/07
16.3 SS Pin Functionality
16.3.1 Slave Mode
When the SPI is configured as a Slave, the Slave Select (SS) pin is always input. When SS is
held low, the SPI is activated, and MISO becomes an output if configured so by the user. All
other pins are inputs. When SS is driven high, all pins are inputs, and the SPI is passive, which
means that it will not receive incoming data. Note that the SPI logic will be reset once the SS pin
is driven high.
The SS pin is useful for packet/byte synchronization to keep the slave bit counter synchronous
with the master clock generator. When the SS pin is driven high, the SPI slave will immediately
reset the send and receive logic, and drop any partially received data in the Shift Register.
16.3.2 Master Mode
When the SPI is configured as a Master (MSTR in SPCR is set), the user can determine the
direction of the SS pin.
If SS is configured as an output, the pin is a general output pin which does not affect the SPI
system. Typically, the pin will be driving the SS pin of the SPI Slave.
If SS is configured as an input, it must be held high to ensure Master SPI operation. If the SS pin
is driven low by peripheral circuitry when the SPI is configured as a Master with the SS pin
defined as an input, the SPI system interprets this as another master selecting the SPI as a
slave and starting to send data to it. To avoid bus contention, the SPI system takes the following
actions:
1. The MSTR bit in SPCR is cleared and the SPI system becomes a Slave. As a result of
the SPI becoming a Slave, the MOSI and SCK pins become inputs.
2. The SPIF Flag in SPSR is set, and if the SPI interrupt is enabled, and the I-bit in SREG is
set, the interrupt routine will be executed.
Thus, when interrupt-driven SPI transmission is used in Master mode, and there exists a possi-
bility that SS is driven low, the interrupt should always check that the MSTR bit is still set. If the
MSTR bit has been cleared by a slave select, it must be set by the user to re-enable SPI Master
mode.
16.4 Data Modes

There are four combinations of SCK phase and polarity with respect to serial data, which are
determined by control bits CPHA and CPOL. The SPI data transfer formats are shown in Figure
16-3 on page 165 and Figure 16-4 on page 165. Data bits are shifted out and latched in on
opposite edges of the SCK signal, ensuring sufficient time for data signals to stabilize. This is
clearly seen by summarizing Table 16-3 on page 166 and Table 16-4 on page 166, as done in
Table 16-2 on page 165
164 ATmega164P/324P/644P
8011D–AVR–02/07
Table 16-2. SPI Modes

SPI Mode Conditions Leading Edge Trailing Edge
0 CPOL=0, CPHA=0 Sample (Rising) Setup (Falling)
1 CPOL=0, CPHA=1 Setup (Rising) Sample (Falling)
2 CPOL=1, CPHA=0 Sample (Falling) Setup (Rising)
3 CPOL=1, CPHA=1 Setup (Falling) Sample (Rising)
Figure 16-3. SPI Transfer Format with CPHA = 0
SCK (CPOL = 0)
mode 0
SCK (CPOL = 1)
mode 2
SAMPLE I
MOSI/MISO
CHANGE 0
MOSI PIN
CHANGE 0
MISO PIN
SS
MSB first (DORD = 0) MSB Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 LSB
LSB first (DORD = 1) LSB Bit 1 Bit 2 Bit 3 Bit 4 Bit 5 Bit 6 MSB
Figure 16-4. SPI Transfer Format with CPHA = 1
SCK (CPOL = 0)
mode 1
SCK (CPOL = 1)
mode 3
SAMPLE I
MOSI/MISO
CHANGE 0
MOSI PIN
CHANGE 0
MISO PIN
SS
MSB first (DORD = 0) MSB Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 LSB
LSB first (DORD = 1) LSB Bit 1 Bit 2 Bit 3 Bit 4 Bit 5 Bit 6 MSB
165
8011D–AVR–02/07
16.5.1 SPCR – SPI Control Register
Bit 7 6 5 4 3 2 1 0
0x2C (0x4C) SPIE SPE DORD MSTR CPOL CPHA SPR1 SPR0 SPCR
• Bit 7 – SPIE: SPI Interrupt Enable

This bit causes the SPI interrupt to be executed if SPIF bit in the SPSR Register is set and the if
the Global Interrupt Enable bit in SREG is set.
• Bit 6 – SPE: SPI Enable

When the SPE bit is written to one, the SPI is enabled. This bit must be set to enable any SPI
operations.
• Bit 5 – DORD: Data Order

When the DORD bit is written to one, the LSB of the data word is transmitted first.
When the DORD bit is written to zero, the MSB of the data word is transmitted first.
• Bit 4 – MSTR: Master/Slave Select

This bit selects Master SPI mode when written to one, and Slave SPI mode when written logic
zero. If SS is configured as an input and is driven low while MSTR is set, MSTR will be cleared,
and SPIF in SPSR will become set. The user will then have to set MSTR to re-enable SPI Mas-
ter mode.
• Bit 3 – CPOL: Clock Polarity

When this bit is written to one, SCK is high when idle. When CPOL is written to zero, SCK is low
when idle. Refer to Figure 16-3 and Figure 16-4 for an example. The CPOL functionality is sum-
marized below:
Table 16-3. CPOL Functionality

CPOL Leading Edge Trailing Edge
0 Rising Falling
1 Falling Rising
• Bit 2 – CPHA: Clock Phase

The settings of the Clock Phase bit (CPHA) determine if data is sampled on the leading (first) or
trailing (last) edge of SCK. Refer to Figure 16-3 and Figure 16-4 for an example. The CPOL
functionality is summarized below:
Table 16-4. CPHA Functionality

CPHA Leading Edge Trailing Edge
0 Sample Setup
1 Setup Sample
166 ATmega164P/324P/644P
8011D–AVR–02/07
• Bits 1:0 – SPR1, SPR0: SPI Clock Rate Select 1 and 0

These two bits control the SCK rate of the device configured as a Master. SPR1 and SPR0 have
no effect on the Slave. The relationship between SCK and the Oscillator Clock frequency fosc is
shown in the following table:
Table 16-5. Relationship Between SCK and the Oscillator Frequency

SPI2X SPR1 SPR0 SCK Frequency
0 0 0 fosc/4
0 0 1 fosc/16
0 1 0 fosc/64
0 1 1 fosc/128
1 0 0 fosc/2
1 0 1 fosc/8
1 1 0 fosc/32
1 1 1 fosc/64
16.5.2 SPSR – SPI Status Register
Bit 7 6 5 4 3 2 1 0
0x2D (0x4D) SPIF WCOL – – – – – SPI2X SPSR
Read/Write R R R R R R R R/W
• Bit 7 – SPIF: SPI Interrupt Flag

When a serial transfer is complete, the SPIF Flag is set. An interrupt is generated if SPIE in
SPCR is set and global interrupts are enabled. If SS is an input and is driven low when the SPI is
in Master mode, this will also set the SPIF Flag. SPIF is cleared by hardware when executing the
corresponding interrupt handling vector. Alternatively, the SPIF bit is cleared by first reading the
SPI Status Register with SPIF set, then accessing the SPI Data Register (SPDR).
• Bit 6 – WCOL: Write COLlision Flag

The WCOL bit is set if the SPI Data Register (SPDR) is written during a data transfer. The
WCOL bit (and the SPIF bit) are cleared by first reading the SPI Status Register with WCOL set,
and then accessing the SPI Data Register.

• Bit 0 – SPI2X: Double SPI Speed Bit

When this bit is written logic one the SPI speed (SCK Frequency) will be doubled when the SPI
is in Master mode (see Table 16-5). This means that the minimum SCK period will be two CPU
clock periods. When the SPI is configured as Slave, the SPI is only guaranteed to work at fosc/4
or lower.
The SPI interface on the ATmega164P/324P/644P is also used for program memory and
EEPROM downloading or uploading. See page 306 for serial programming and verification.
167
8011D–AVR–02/07
16.5.3 SPDR – SPI Data Register
Bit 7 6 5 4 3 2 1 0
0x2E (0x4E) MSB LSB SPDR
Initial Value X X X X X X X X Undefined
The SPI Data Register is a read/write register used for data transfer between the Register File
and the SPI Shift Register. Writing to the register initiates data transmission. Reading the regis-
ter causes the Shift Register Receive buffer to be read.
168 ATmega164P/324P/644P
8011D–AVR–02/07
17. USART
17.1 Features
• Full Duplex Operation (Independent Serial Receive and Transmit Registers)
• Asynchronous or Synchronous Operation
• Master or Slave Clocked Synchronous Operation
• High Resolution Baud Rate Generator
• Supports Serial Frames with 5, 6, 7, 8, or 9 Data Bits and 1 or 2 Stop Bits
• Odd or Even Parity Generation and Parity Check Supported by Hardware
• Data OverRun Detection
• Framing Error Detection
• Noise Filtering Includes False Start Bit Detection and Digital Low Pass Filter
• Three Separate Interrupts on TX Complete, TX Data Register Empty and RX Complete
• Multi-processor Communication Mode
• Double Speed Asynchronous Communication Mode
17.2 USART1 and USART0

The ATmega164P/324P/644P has two USART’s, USART0 and USART1.
The functionality for all USART’s is described below, most register and bit references in this sec-
tion are written in general form. A lower case “n” replaces the USART number.
USART0 and USART1 have different I/O registers as shown in ”Register Summary” on page
386.
17.3 Overview
The Universal Synchronous and Asynchronous serial Receiver and Transmitter (USART) is a
highly flexible serial communication device.
A simplified block diagram of the USART Transmitter is shown in Figure 17-1 on page 170. CPU
accessible I/O Registers and I/O pins are shown in bold.
The Power Reducion USART0 bit, PRUSART0, in ”PRR – Power Reduction Register” on page
48 must be disabled by writing a logical zero to it.
The Power Reducion USART1 bit, PRUSART1, in ”PRR – Power Reduction Register” on page
48 must be disabled by writing a logical zero to it.
169
8011D–AVR–02/07
Figure 17-1. USART Block Diagram(1)
Clock Generator
UBRR[H:L]
OSC
BAUD RATE GENERATOR
SYNC LOGIC PIN

XCK
CONTROL
Transmitter
TX
UDR (Transmit)
CONTROL
PARITY
GENERATOR
DATA BUS
PIN
TRANSMIT SHIFT REGISTER TxD
CONTROL
Receiver
CLOCK RX
RECOVERY CONTROL
DATA PIN
RECEIVE SHIFT REGISTER RxD
RECOVERY CONTROL
PARITY
UDR (Receive)
CHECKER
UCSRA UCSRB UCSRC
Note: 1. See Figure 1-1 on page 2 and ”Alternate Port Functions” on page 77 for USART pin
placement.
The dashed boxes in the block diagram separate the three main parts of the USART (listed from
the top): Clock Generator, Transmitter and Receiver. Control Registers are shared by all units.
The Clock Generation logic consists of synchronization logic for external clock input used by
synchronous slave operation, and the baud rate generator. The XCKn (Transfer Clock) pin is
only used by synchronous transfer mode. The Transmitter consists of a single write buffer, a
serial Shift Register, Parity Generator and Control logic for handling different serial frame for-
mats. The write buffer allows a continuous transfer of data without any delay between frames.
The Receiver is the most complex part of the USART module due to its clock and data recovery
units. The recovery units are used for asynchronous data reception. In addition to the recovery
units, the Receiver includes a Parity Checker, Control logic, a Shift Register and a two level
receive buffer (UDRn). The Receiver supports the same frame formats as the Transmitter, and
can detect Frame Error, Data OverRun and Parity Errors.
17.4 Clock Generation

The Clock Generation logic generates the base clock for the Transmitter and Receiver. The
USARTn supports four modes of clock operation: Normal asynchronous, Double Speed asyn-
chronous, Master synchronous and Slave synchronous mode. The UMSELn bit in USART
Control and Status Register C (UCSRnC) selects between asynchronous and synchronous
operation. Double Speed (asynchronous mode only) is controlled by the U2Xn found in the
170 ATmega164P/324P/644P
8011D–AVR–02/07
UCSRnA Register. When using synchronous mode (UMSELn = 1), the Data Direction Register
for the XCKn pin (DDR_XCKn) controls whether the clock source is internal (Master mode) or
external (Slave mode). The XCKn pin is only active when using synchronous mode.
Figure 17-2 shows a block diagram of the clock generation logic.
Figure 17-2. Clock Generation Logic, Block Diagram

UBRR
U2X
fosc
Prescaling UBRR+1
/2 /4 /2
Down-Counter 0
1
0
OSC txclk
1
DDR_XCK
Sync Edge
xcki Register Detector 0
XCK UMSEL
xcko 1
Pin
DDR_XCK UCPOL 1
rxclk
0
Signal description:
txclk Transmitter clock (Internal Signal).
rxclk Receiver base clock (Internal Signal).
xcki Input from XCK pin (internal Signal). Used for synchronous slave
operation.
xcko Clock output to XCK pin (Internal Signal). Used for synchronous master
operation.
fOSC XTAL pin frequency (System Clock).
17.4.1 Internal Clock Generation – The Baud Rate Generator
Internal clock generation is used for the asynchronous and the synchronous master modes of
operation. The description in this section refers to Figure 17-2 on page 171.
The USART Baud Rate Register (UBRRn) and the down-counter connected to it function as a
programmable prescaler or baud rate generator. The down-counter, running at system clock
(fosc), is loaded with the UBRRn value each time the counter has counted down to zero or when
the UBRRLn Register is written. A clock is generated each time the counter reaches zero. This
clock is the baud rate generator clock output (= fosc/(UBRRn+1)). The Transmitter divides the
baud rate generator clock output by 2, 8 or 16 depending on mode. The baud rate generator out-
put is used directly by the Receiver’s clock and data recovery units. However, the recovery units
use a state machine that uses 2, 8 or 16 states depending on mode set by the state of the
UMSELn, U2Xn and DDR_XCKn bits.
Table 17-1 on page 172 contains equations for calculating the baud rate (in bits per second) and
for calculating the UBRRn value for each mode of operation using an internally generated clock
source.
Note: 1. The baud rate is defined to be the transfer rate in bit per second (bps)
BAUD Baud rate (in bits per second, bps)
171
8011D–AVR–02/07
Table 17-1. Equations for Calculating Baud Rate Register Setting
Equation for Calculating Baud Equation for Calculating UBRR
Operating Mode Rate(1) Value
f OSC
-–1
UBRRn = -----------------------
16BAUD
Asynchronous Normal f OSC

mode (U2Xn = 0) BAUD = -----------------------------------------
-
16 ( UBRRn + 1 )
f OSC
-–1
UBRRn = -------------------
8BAUD
Asynchronous Double f OSC

Speed mode (U2Xn = 1) BAUD = --------------------------------------
-
8 ( UBRRn + 1 )
f OSC
-–1
UBRRn = -------------------
2BAUD
Synchronous Master f OSC

mode BAUD = --------------------------------------
-
2 ( UBRRn + 1 )
fOSC System Oscillator clock frequency

UBRRn Contents of the UBRRHn and UBRRLn Registers, (0-4095)
Some examples of UBRRn values for some system clock frequencies are found in Table 17-9 on
page 192.
17.4.2 Double Speed Operation (U2Xn)
The transfer rate can be doubled by setting the U2Xn bit in UCSRnA. Setting this bit only has
effect for the asynchronous operation. Set this bit to zero when using synchronous operation.
Setting this bit will reduce the divisor of the baud rate divider from 16 to 8, effectively doubling
the transfer rate for asynchronous communication. Note however that the Receiver will in this
case only use half the number of samples (reduced from 16 to 8) for data sampling and clock
recovery, and therefore a more accurate baud rate setting and system clock are required when
this mode is used. For the Transmitter, there are no downsides.
17.4.3 External Clock
External clocking is used by the synchronous slave modes of operation. The description in this
section refers to Figure 17-2 on page 171 for details.
External clock input from the XCKn pin is sampled by a synchronization register to minimize the
chance of meta-stability. The output from the synchronization register must then pass through
an edge detector before it can be used by the Transmitter and Receiver. This process intro-
172 ATmega164P/324P/644P
8011D–AVR–02/07
duces a two CPU clock period delay and therefore the maximum external XCKn clock frequency
is limited by the following equation:
f OSC
f XCK < -----------
4
Note that fosc depends on the stability of the system clock source. It is therefore recommended to
add some margin to avoid possible loss of data due to frequency variations.
17.4.4 Synchronous Clock Operation
When synchronous mode is used (UMSELn = 1), the XCKn pin will be used as either clock input
(Slave) or clock output (Master). The dependency between the clock edges and data sampling
or data change is the same. The basic principle is that data input (on RxDn) is sampled at the
opposite XCKn clock edge of the edge the data output (TxDn) is changed.
Figure 17-3. Synchronous Mode XCKn Timing.
UCPOL = 1 XCK
RxD / TxD
Sample
UCPOL = 0 XCK
RxD / TxD
Sample
The UCPOLn bit UCRSC selects which XCKn clock edge is used for data sampling and which is
used for data change. As Figure 17-3 on page 173 shows, when UCPOLn is zero the data will
be changed at rising XCKn edge and sampled at falling XCKn edge. If UCPOLn is set, the data
will be changed at falling XCKn edge and sampled at rising XCKn edge.
17.5 Frame Formats

A serial frame is defined to be one character of data bits with synchronization bits (start and stop
bits), and optionally a parity bit for error checking. The USART accepts all 30 combinations of
the following as valid frame formats:
• 1 start bit
• 5, 6, 7, 8, or 9 data bits
• no, even or odd parity bit
• 1 or 2 stop bits
A frame starts with the start bit followed by the least significant data bit. Then the next data bits,
up to a total of nine, are succeeding, ending with the most significant bit. If enabled, the parity bit
is inserted after the data bits, before the stop bits. When a complete frame is transmitted, it can
be directly followed by a new frame, or the communication line can be set to an idle (high) state.
Figure 17-4 on page 174 illustrates the possible combinations of the frame formats. Bits inside
brackets are optional.
173
8011D–AVR–02/07
Figure 17-4. Frame Formats
FRAME
(IDLE) St 0 1 2 3 4 [5] [6] [7] [8] [P] Sp1 [Sp2] (St / IDLE)
St Start bit, always low.

(n) Data bits (0 to 8).
P Parity bit. Can be odd or even.
Sp Stop bit, always high.
IDLE No transfers on the communication line (RxDn or TxDn). An IDLE line
must be high.
The frame format used by the USART is set by the UCSZn2:0, UPMn1:0 and USBSn bits in
UCSRnB and UCSRnC. The Receiver and Transmitter use the same setting. Note that changing
the setting of any of these bits will corrupt all ongoing communication for both the Receiver and
Transmitter.
The USART Character SiZe (UCSZn2:0) bits select the number of data bits in the frame. The
USART Parity mode (UPMn1:0) bits enable and set the type of parity bit. The selection between
one or two stop bits is done by the USART Stop Bit Select (USBSn) bit. The Receiver ignores
the second stop bit. An FE (Frame Error) will therefore only be detected in the cases where the
first stop bit is zero.
17.5.1 Parity Bit Calculation
The parity bit is calculated by doing an exclusive-or of all the data bits. If odd parity is used, the
result of the exclusive or is inverted. The relation between the parity bit and data bits is as
follows::
P even = d n – 1 ⊕ … ⊕ d 3 ⊕ d 2 ⊕ d 1 ⊕ d 0 ⊕ 0
P odd = d n – 1 ⊕ … ⊕ d 3 ⊕ d 2 ⊕ d 1 ⊕ d 0 ⊕ 1
Peven Parity bit using even parity

Podd Parity bit using odd parity
dn Data bit n of the character
If used, the parity bit is located between the last data bit and first stop bit of a serial frame.
17.6 USART Initialization

The USART has to be initialized before any communication can take place. The initialization pro-
cess normally consists of setting the baud rate, setting frame format and enabling the
Transmitter or the Receiver depending on the usage. For interrupt driven USART operation, the
Global Interrupt Flag should be cleared (and interrupts globally disabled) when doing the
initialization.
Before doing a re-initialization with changed baud rate or frame format, be sure that there are no
ongoing transmissions during the period the registers are changed. The TXCn Flag can be used
to check that the Transmitter has completed all transfers, and the RXC Flag can be used to
174 ATmega164P/324P/644P
8011D–AVR–02/07
check that there are no unread data in the receive buffer. Note that the TXCn Flag must be
cleared before each transmission (before UDRn is written) if it is used for this purpose.
The following simple USART initialization code examples show one assembly and one C func-
tion that are equal in functionality. The examples assume asynchronous operation using polling
(no interrupts enabled) and a fixed frame format. The baud rate is given as a function parameter.
For the assembly code, the baud rate parameter is assumed to be stored in the r17:r16
Registers.

USART_Init:
; Set baud rate
out UBRRHn, r17
out UBRRLn, r16
; Enable receiver and transmitter
ldi r16, (1<<RXENn)|(1<<TXENn)
out UCSRnB,r16
; Set frame format: 8data, 2stop bit
ldi r16, (1<<USBSn)|(3<<UCSZn0)
out UCSRnC,r16
ret
C Code Example(1)
void USART_Init( unsigned int baud )
{
/* Set baud rate */
UBRRHn = (unsigned char)(baud>>8);
UBRRLn = (unsigned char)baud;
/* Enable receiver and transmitter */
UCSRnB = (1<<RXENn)|(1<<TXENn);
/* Set frame format: 8data, 2stop bit */
UCSRnC = (1<<USBSn)|(3<<UCSZn0);
}

More advanced initialization routines can be made that include frame format as parameters, dis-
able interrupts and so on. However, many applications use a fixed setting of the baud and
control registers, and for these types of applications the initialization code can be placed directly
in the main routine, or be combined with initialization code for other I/O modules.
17.7 Data Transmission – The USART Transmitter

The USART Transmitter is enabled by setting the Transmit Enable (TXEN) bit in the UCSRnB
Register. When the Transmitter is enabled, the normal port operation of the TxDn pin is overrid-
den by the USART and given the function as the Transmitter’s serial output. The baud rate,
mode of operation and frame format must be set up once before doing any transmissions. If syn-
chronous operation is used, the clock on the XCKn pin will be overridden and used as
transmission clock.
175
8011D–AVR–02/07
17.7.1 Sending Frames with 5 to 8 Data Bit
A data transmission is initiated by loading the transmit buffer with the data to be transmitted. The
CPU can load the transmit buffer by writing to the UDRn I/O location. The buffered data in the
transmit buffer will be moved to the Shift Register when the Shift Register is ready to send a new
frame. The Shift Register is loaded with new data if it is in idle state (no ongoing transmission) or
immediately after the last stop bit of the previous frame is transmitted. When the Shift Register is
loaded with new data, it will transfer one complete frame at the rate given by the Baud Register,
U2Xn bit or by XCKn depending on mode of operation.
The following code examples show a simple USART transmit function based on polling of the
Data Register Empty (UDREn) Flag. When using frames with less than eight bits, the most sig-
nificant bits written to the UDRn are ignored. The USART has to be initialized before the function
can be used. For the assembly code, the data to be sent is assumed to be stored in Register
R16

USART_Transmit:
; Wait for empty transmit buffer
sbis UCSRnA,UDREn
rjmp USART_Transmit
; Put data (r16) into buffer, sends the data
out UDRn,r16
ret
C Code Example(1)
void USART_Transmit( unsigned char data )
{
/* Wait for empty transmit buffer */
while ( !( UCSRnA & (1<<UDREn)) )
;
/* Put data into buffer, sends the data */
UDRn = data;
}

The function simply waits for the transmit buffer to be empty by checking the UDREn Flag,
before loading it with new data to be transmitted. If the Data Register Empty interrupt is utilized,
the interrupt routine writes the data into the buffer.
17.7.2 Sending Frames with 9 Data Bit
If 9-bit characters are used (UCSZn = 7), the ninth bit must be written to the TXB8 bit in UCS-
RnB before the low byte of the character is written to UDRn. The following code examples show
a transmit function that handles 9-bit characters. For the assembly code, the data to be sent is
assumed to be stored in registers R17:R16.
176 ATmega164P/324P/644P
8011D–AVR–02/07
Assembly Code Example(1)(2)

USART_Transmit:
sbis UCSRnA,UDREn
rjmp USART_Transmit
; Copy 9th bit from r17 to TXB8
cbi UCSRnB,TXB8
sbrc r17,0
sbi UCSRnB,TXB8
; Put LSB data (r16) into buffer, sends the data
out UDRn,r16
ret
C Code Example(1)(2)
void USART_Transmit( unsigned int data )
{
while ( !( UCSRnA & (1<<UDREn))) )
;
/* Copy 9th bit to TXB8 */
UCSRnB &= ~(1<<TXB8);
if ( data & 0x0100 )
UCSRnB |= (1<<TXB8);
UDRn = data;
}
Notes: 1. These transmit functions are written to be general functions. They can be optimized if the con-
tents of the UCSRnB is static. For example, only the TXB8 bit of the UCSRnB Register is used
after initialization.
2. See “About Code Examples” on page 8.
The ninth bit can be used for indicating an address frame when using multi processor communi-
cation mode or for other protocol handling as for example synchronization.
17.7.3 Transmitter Flags and Interrupts
The USART Transmitter has two flags that indicate its state: USART Data Register Empty
(UDREn) and Transmit Complete (TXCn). Both flags can be used for generating interrupts.
The Data Register Empty (UDREn) Flag indicates whether the transmit buffer is ready to receive
new data. This bit is set when the transmit buffer is empty, and cleared when the transmit buffer
contains data to be transmitted that has not yet been moved into the Shift Register. For compat-
ibility with future devices, always write this bit to zero when writing the UCSRnA Register.
When the Data Register Empty Interrupt Enable (UDRIEn) bit in UCSRnB is written to one, the
USART Data Register Empty Interrupt will be executed as long as UDREn is set (provided that
global interrupts are enabled). UDREn is cleared by writing UDRn. When interrupt-driven data
177
8011D–AVR–02/07
transmission is used, the Data Register Empty interrupt routine must either write new data to
UDRn in order to clear UDREn or disable the Data Register Empty interrupt, otherwise a new
interrupt will occur once the interrupt routine terminates.
The Transmit Complete (TXCn) Flag bit is set one when the entire frame in the Transmit Shift
Register has been shifted out and there are no new data currently present in the transmit buffer.
The TXCn Flag bit is automatically cleared when a transmit complete interrupt is executed, or it
can be cleared by writing a one to its bit location. The TXCn Flag is useful in half-duplex commu-
nication interfaces (like the RS-485 standard), where a transmitting application must enter
receive mode and free the communication bus immediately after completing the transmission.
When the Transmit Compete Interrupt Enable (TXCIEn) bit in UCSRnB is set, the USART
Transmit Complete Interrupt will be executed when the TXCn Flag becomes set (provided that
global interrupts are enabled). When the transmit complete interrupt is used, the interrupt han-
dling routine does not have to clear the TXCn Flag, this is done automatically when the interrupt
is executed.
17.7.4 Parity Generator
The Parity Generator calculates the parity bit for the serial frame data. When parity bit is enabled
(UPMn1 = 1), the transmitter control logic inserts the parity bit between the last data bit and the
first stop bit of the frame that is sent.
17.7.5 Disabling the Transmitter
The disabling of the Transmitter (setting the TXEN to zero) will not become effective until ongo-
ing and pending transmissions are completed, i.e., when the Transmit Shift Register and
Transmit Buffer Register do not contain data to be transmitted. When disabled, the Transmitter
will no longer override the TxDn pin.
17.8 Data Reception – The USART Receiver

The USART Receiver is enabled by writing the Receive Enable (RXENn) bit in the
UCSRnB Register to one. When the Receiver is enabled, the normal pin operation of the RxDn
pin is overridden by the USART and given the function as the Receiver’s serial input. The baud
rate, mode of operation and frame format must be set up once before any serial reception can
be done. If synchronous operation is used, the clock on the XCKn pin will be used as transfer
clock.
17.8.1 Receiving Frames with 5 to 8 Data Bits
The Receiver starts data reception when it detects a valid start bit. Each bit that follows the start
bit will be sampled at the baud rate or XCKn clock, and shifted into the Receive Shift Register
until the first stop bit of a frame is received. A second stop bit will be ignored by the Receiver.
When the first stop bit is received, i.e., a complete serial frame is present in the Receive Shift
Register, the contents of the Shift Register will be moved into the receive buffer. The receive
buffer can then be read by reading the UDRn I/O location.
The following code example shows a simple USART receive function based on polling of the
Receive Complete (RXCn) Flag. When using frames with less than eight bits the most significant
bits of the data read from the UDRn will be masked to zero. The USART has to be initialized
before the function can be used.
178 ATmega164P/324P/644P
8011D–AVR–02/07

USART_Receive:
; Wait for data to be received
sbis UCSRnA, RXCn
rjmp USART_Receive
; Get and return received data from buffer
in r16, UDRn
ret
C Code Example(1)
unsigned char USART_Receive( void )
{
/* Wait for data to be received */
while ( !(UCSRnA & (1<<RXCn)) )
;
/* Get and return received data from buffer */
return UDRn;
}

The function simply waits for data to be present in the receive buffer by checking the RXCn Flag,
before reading the buffer and returning the value.
17.8.2 Receiving Frames with 9 Data Bits
If 9-bit characters are used (UCSZn=7) the ninth bit must be read from the RXB8n bit in UCS-
RnB before reading the low bits from the UDRn. This rule applies to the FEn, DORn and UPEn
Status Flags as well. Read status from UCSRnA, then data from UDRn. Reading the UDRn I/O
location will change the state of the receive buffer FIFO and consequently the TXB8n, FEn,
DORn and UPEn bits, which all are stored in the FIFO, will change.
The following code example shows a simple USART receive function that handles both nine bit
characters and the status bits.
179
8011D–AVR–02/07
USART_Receive:
sbis UCSRnA, RXCn
rjmp USART_Receive
; Get status and 9th bit, then data from buffer
in r18, UCSRnA
in r17, UCSRnB
in r16, UDRn
; If error, return -1
andi r18,(1<<FEn)|(1<<DORn)|(1<<UPEn)
breq USART_ReceiveNoError
ldi r17, HIGH(-1)
ldi r16, LOW(-1)
USART_ReceiveNoError:
; Filter the 9th bit, then return
lsr r17
andi r17, 0x01
ret
C Code Example(1)
unsigned int USART_Receive( void )
{
unsigned char status, resh, resl;
while ( !(UCSRnA & (1<<RXCn)) )
;
/* Get status and 9th bit, then data */
/* from buffer */
status = UCSRnA;
resh = UCSRnB;
resl = UDRn;
/* If error, return -1 */
if ( status & (1<<FEn)|(1<<DORn)|(1<<UPEn) )
return -1;
/* Filter the 9th bit, then return */
resh = (resh >> 1) & 0x01;
return ((resh << 8) | resl);
}
The receive function example reads all the I/O Registers into the Register File before any com-
putation is done. This gives an optimal receive buffer utilization since the buffer location read will
be free to accept new data as early as possible.
180 ATmega164P/324P/644P
8011D–AVR–02/07
17.8.3 Receive Compete Flag and Interrupt
The USART Receiver has one flag that indicates the Receiver state.
The Receive Complete (RXCn) Flag indicates if there are unread data present in the receive
buffer. This flag is one when unread data exist in the receive buffer, and zero when the receive
buffer is empty (i.e., does not contain any unread data). If the Receiver is disabled (RXENn = 0),
the receive buffer will be flushed and consequently the RXCn bit will become zero.
When the Receive Complete Interrupt Enable (RXCIEn) in UCSRnB is set, the USART Receive
Complete interrupt will be executed as long as the RXCn Flag is set (provided that global inter-
rupts are enabled). When interrupt-driven data reception is used, the receive complete routine
must read the received data from UDRn in order to clear the RXCn Flag, otherwise a new inter-
rupt will occur once the interrupt routine terminates.
17.8.4 Receiver Error Flags
The USART Receiver has three Error Flags: Frame Error (FEn), Data OverRun (DORn) and
Parity Error (UPEn). All can be accessed by reading UCSRnA. Common for the Error Flags is
that they are located in the receive buffer together with the frame for which they indicate the
error status. Due to the buffering of the Error Flags, the UCSRnA must be read before the
receive buffer (UDRn), since reading the UDRn I/O location changes the buffer read location.
Another equality for the Error Flags is that they can not be altered by software doing a write to
the flag location. However, all flags must be set to zero when the UCSRnA is written for upward
compatibility of future USART implementations. None of the Error Flags can generate interrupts.
The Frame Error (FEn) Flag indicates the state of the first stop bit of the next readable frame
stored in the receive buffer. The FEn Flag is zero when the stop bit was correctly read (as one),
and the FEn Flag will be one when the stop bit was incorrect (zero). This flag can be used for
detecting out-of-sync conditions, detecting break conditions and protocol handling. The FEn
Flag is not affected by the setting of the USBSn bit in UCSRnC since the Receiver ignores all,
except for the first, stop bits. For compatibility with future devices, always set this bit to zero
when writing to UCSRnA.
The Data OverRun (DORn) Flag indicates data loss due to a receiver buffer full condition. A
Data OverRun occurs when the receive buffer is full (two characters), it is a new character wait-
ing in the Receive Shift Register, and a new start bit is detected. If the DORn Flag is set there
was one or more serial frame lost between the frame last read from UDRn, and the next frame
read from UDRn. For compatibility with future devices, always write this bit to zero when writing
to UCSRnA. The DORn Flag is cleared when the frame received was successfully moved from
the Shift Register to the receive buffer.
The Parity Error (UPEn) Flag indicates that the next frame in the receive buffer had a Parity
Error when received. If Parity Check is not enabled the UPEn bit will always be read zero. For
compatibility with future devices, always set this bit to zero when writing to UCSRnA. For more
details see ”Parity Bit Calculation” on page 174 and ”Parity Checker” on page 181.
17.8.5 Parity Checker
The Parity Checker is active when the high USART Parity mode (UPMn1) bit is set. Type of Par-
ity Check to be performed (odd or even) is selected by the UPMn0 bit. When enabled, the Parity
Checker calculates the parity of the data bits in incoming frames and compares the result with
the parity bit from the serial frame. The result of the check is stored in the receive buffer together
181
8011D–AVR–02/07
with the received data and stop bits. The Parity Error (UPEn) Flag can then be read by software
to check if the frame had a Parity Error.
The UPEn bit is set if the next character that can be read from the receive buffer had a Parity
Error when received and the Parity Checking was enabled at that point (UPMn1 = 1). This bit is
valid until the receive buffer (UDRn) is read.
17.8.6 Disabling the Receiver
In contrast to the Transmitter, disabling of the Receiver will be immediate. Data from ongoing
receptions will therefore be lost. When disabled (i.e., the RXENn is set to zero) the Receiver will
no longer override the normal function of the RxDn port pin. The Receiver buffer FIFO will be
flushed when the Receiver is disabled. Remaining data in the buffer will be lost
17.8.7 Flushing the Receive Buffer
The receiver buffer FIFO will be flushed when the Receiver is disabled, i.e., the buffer will be
emptied of its contents. Unread data will be lost. If the buffer has to be flushed during normal
operation, due to for instance an error condition, read the UDRn I/O location until the RXCn Flag
is cleared. The following code example shows how to flush the receive buffer.

USART_Flush:
sbis UCSRnA, RXCn
ret
in r16, UDRn
rjmp USART_Flush
C Code Example(1)
void USART_Flush( void )
{
unsigned char dummy;
while ( UCSRnA & (1<<RXCn) ) dummy = UDRn;
}

17.9 Asynchronous Data Reception
The USART includes a clock recovery and a data recovery unit for handling asynchronous data
reception. The clock recovery logic is used for synchronizing the internally generated baud rate
clock to the incoming asynchronous serial frames at the RxDn pin. The data recovery logic sam-
ples and low pass filters each incoming bit, thereby improving the noise immunity of the
Receiver. The asynchronous reception operational range depends on the accuracy of the inter-
nal baud rate clock, the rate of the incoming frames, and the frame size in number of bits.
17.9.1 Asynchronous Clock Recovery
The clock recovery logic synchronizes internal clock to the incoming serial frames. Figure 17-5
illustrates the sampling process of the start bit of an incoming frame. The sample rate is 16 times
the baud rate for Normal mode, and eight times the baud rate for Double Speed mode. The hor-
izontal arrows illustrate the synchronization variation due to the sampling process. Note the
182 ATmega164P/324P/644P
8011D–AVR–02/07
larger time variation when using the Double Speed mode (U2Xn = 1) of operation. Samples
denoted zero are samples done when the RxDn line is idle (i.e., no communication activity).
Figure 17-5. Start Bit Sampling
RxD IDLE START BIT 0
Sample
(U2X = 0) 0 0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1 2 3
Sample
(U2X = 1) 0 1 2 3 4 5 6 7 8 1 2
When the clock recovery logic detects a high (idle) to low (start) transition on the RxDn line, the
start bit detection sequence is initiated. Let sample 1 denote the first zero-sample as shown in
the figure. The clock recovery logic then uses samples 8, 9, and 10 for Normal mode, and sam-
ples 4, 5, and 6 for Double Speed mode (indicated with sample numbers inside boxes on the
figure), to decide if a valid start bit is received. If two or more of these three samples have logical
high levels (the majority wins), the start bit is rejected as a noise spike and the Receiver starts
looking for the next high to low-transition. If however, a valid start bit is detected, the clock recov-
ery logic is synchronized and the data recovery can begin. The synchronization process is
repeated for each start bit.
17.9.2 Asynchronous Data Recovery
When the receiver clock is synchronized to the start bit, the data recovery can begin. The data
recovery unit uses a state machine that has 16 states for each bit in Normal mode and eight
states for each bit in Double Speed mode. Figure 17-6 shows the sampling of the data bits and
the parity bit. Each of the samples is given a number that is equal to the state of the recovery
unit.
Figure 17-6. Sampling of Data and Parity Bit
RxD BIT n
Sample
(U2X = 0) 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 1
Sample
(U2X = 1) 1 2 3 4 5 6 7 8 1
The decision of the logic level of the received bit is taken by doing a majority voting of the logic
value to the three samples in the center of the received bit. The center samples are emphasized
on the figure by having the sample number inside boxes. The majority voting process is done as
follows: If two or all three samples have high levels, the received bit is registered to be a logic 1.
If two or all three samples have low levels, the received bit is registered to be a logic 0. This
majority voting process acts as a low pass filter for the incoming signal on the RxDn pin. The
recovery process is then repeated until a complete frame is received. Including the first stop bit.
Note that the Receiver only uses the first stop bit of a frame.
Figure 17-7 on page 184 shows the sampling of the stop bit and the earliest possible beginning
of the start bit of the next frame.
183
8011D–AVR–02/07
Figure 17-7. Stop Bit Sampling and Next Start Bit Sampling
RxD STOP 1 (A) (B) (C)
Sample
(U2X = 0) 1 2 3 4 5 6 7 8 9 10 0/1 0/1 0/1
Sample
(U2X = 1) 1 2 3 4 5 6 0/1
The same majority voting is done to the stop bit as done for the other bits in the frame. If the stop
bit is registered to have a logic 0 value, the Frame Error (FEn) Flag will be set.
A new high to low transition indicating the start bit of a new frame can come right after the last of
the bits used for majority voting. For Normal Speed mode, the first low level sample can be at
point marked (A) in Figure 17-7 on page 184. For Double Speed mode the first low level must be
delayed to (B). (C) marks a stop bit of full length. The early start bit detection influences the
operational range of the Receiver.
17.9.3 Asynchronous Operational Range
The operational range of the Receiver is dependent on the mismatch between the received bit
rate and the internally generated baud rate. If the Transmitter is sending frames at too fast or too
slow bit rates, or the internally generated baud rate of the Receiver does not have a similar (see
Table 17-2 on page 185) base frequency, the Receiver will not be able to synchronize the
frames to the start bit.
The following equations can be used to calculate the ratio of the incoming data rate and internal
receiver baud rate.
( D + 1 )S ( D + 2 )S
R slow = ------------------------------------------
- R fast = -----------------------------------
S – 1 + D ⋅ S + SF ( D + 1 )S + S M
D Sum of character size and parity size (D = 5 to 10 bit)

S Samples per bit. S = 16 for Normal Speed mode and S = 8 for Double Speed
mode.
SF First sample number used for majority voting. SF = 8 for normal speed and SF = 4
for Double Speed mode.
SM Middle sample number used for majority voting. SM = 9 for normal speed and
SM = 5 for Double Speed mode.
Rslow is the ratio of the slowest incoming data rate that can be accepted in relation to the
receiver baud rate. Rfast is the ratio of the fastest incoming data rate that can be
accepted in relation to the receiver baud rate.
Table 17-2 on page 185 and Table 17-3 on page 185 list the maximum receiver baud rate error
that can be tolerated. Note that Normal Speed mode has higher toleration of baud rate
variations.
184 ATmega164P/324P/644P
8011D–AVR–02/07
Table 17-2. Recommended Maximum Receiver Baud Rate Error for Normal Speed Mode
(U2Xn = 0)
D Recommended Max
# (Data+Parity Bit) Rslow (%) Rfast (%) Max Total Error (%) Receiver Error (%)
5 93.20 106.67 +6.67/-6.8 ± 3.0
6 94.12 105.79 +5.79/-5.88 ± 2.5
7 94.81 105.11 +5.11/-5.19 ± 2.0
8 95.36 104.58 +4.58/-4.54 ± 2.0
9 95.81 104.14 +4.14/-4.19 ± 1.5
10 96.17 103.78 +3.78/-3.83 ± 1.5
Table 17-3. Recommended Maximum Receiver Baud Rate Error for Double Speed Mode
(U2Xn = 1)
D Recommended Max
# (Data+Parity Bit) Rslow (%) Rfast (%) Max Total Error (%) Receiver Error (%)
5 94.12 105.66 +5.66/-5.88 ± 2.5
6 94.92 104.92 +4.92/-5.08 ± 2.0
7 95.52 104,35 +4.35/-4.48 ± 1.5
8 96.00 103.90 +3.90/-4.00 ± 1.5
9 96.39 103.53 +3.53/-3.61 ± 1.5
10 96.70 103.23 +3.23/-3.30 ± 1.0
The recommendations of the maximum receiver baud rate error was made under the assump-
tion that the Receiver and Transmitter equally divides the maximum total error.
There are two possible sources for the receivers baud rate error. The Receiver’s system clock
(XTAL) will always have some minor instability over the supply voltage range and the tempera-
ture range. When using a crystal to generate the system clock, this is rarely a problem, but for a
resonator the system clock may differ more than 2% depending of the resonators tolerance. The
second source for the error is more controllable. The baud rate generator can not always do an
exact division of the system frequency to get the baud rate wanted. In this case an UBRR value
that gives an acceptable low error can be used if possible.
17.10 Multi-processor Communication Mode

Setting the Multi-processor Communication mode (MPCMn) bit in UCSRnA enables a filtering
function of incoming frames received by the USART Receiver. Frames that do not contain
address information will be ignored and not put into the receive buffer. This effectively reduces
the number of incoming frames that has to be handled by the CPU, in a system with multiple
MCUs that communicate via the same serial bus. The Transmitter is unaffected by the MPCMn
setting, but has to be used differently when it is a part of a system utilizing the Multi-processor
Communication mode.
If the Receiver is set up to receive frames that contain 5 to 8 data bits, then the first stop bit indi-
cates if the frame contains data or address information. If the Receiver is set up for frames with
nine data bits, then the ninth bit (RXB8n) is used for identifying address and data frames. When
185
8011D–AVR–02/07
the frame type bit (the first stop or the ninth bit) is one, the frame contains an address. When the
frame type bit is zero the frame is a data frame.
The Multi-processor Communication mode enables several slave MCUs to receive data from a
master MCU. This is done by first decoding an address frame to find out which MCU has been
addressed. If a particular slave MCU has been addressed, it will receive the following data
frames as normal, while the other slave MCUs will ignore the received frames until another
address frame is received.
17.10.1 Using MPCMn
For an MCU to act as a master MCU, it can use a 9-bit character frame format (UCSZn = 7). The
ninth bit (TXB8n) must be set when an address frame (TXB8n = 1) or cleared when a data frame
(TXB = 0) is being transmitted. The slave MCUs must in this case be set to use a 9-bit character
frame format.
The following procedure should be used to exchange data in Multi-processor Communication
mode:
1. All Slave MCUs are in Multi-processor Communication mode (MPCMn in UCSRnA is
set).
2. The Master MCU sends an address frame, and all slaves receive and read this frame. In
the Slave MCUs, the RXCn Flag in UCSRnA will be set as normal.
3. Each Slave MCU reads the UDRn Register and determines if it has been selected. If so,
it clears the MPCMn bit in UCSRnA, otherwise it waits for the next address byte and
keeps the MPCMn setting.
4. The addressed MCU will receive all data frames until a new address frame is received.
The other Slave MCUs, which still have the MPCMn bit set, will ignore the data frames.
5. When the last data frame is received by the addressed MCU, the addressed MCU sets
the MPCMn bit and waits for a new address frame from master. The process then
repeats from 2.
Using any of the 5- to 8-bit character frame formats is possible, but impractical since the
Receiver must change between using n and n+1 character frame formats. This makes full-
duplex operation difficult since the Transmitter and Receiver uses the same character size set-
ting. If 5- to 8-bit character frames are used, the Transmitter must be set to use two stop bit
(USBSn = 1) since the first stop bit is used for indicating the frame type.
Do not use Read-Modify-Write instructions (SBI and CBI) to set or clear the MPCMn bit. The
MPCMn bit shares the same I/O location as the TXCn Flag and this might accidentally be
cleared when using SBI or CBI instructions.
186 ATmega164P/324P/644P
8011D–AVR–02/07
17.11.1 UDRn – USART I/O Data Register n
Bit 7 6 5 4 3 2 1 0
RXB[7:0] UDRn (Read)
TXB[7:0] UDRn (Write)
The USART Transmit Data Buffer Register and USART Receive Data Buffer Registers share the
same I/O address referred to as USART Data Register or UDRn. The Transmit Data Buffer Reg-
ister (TXB) will be the destination for data written to the UDRn Register location. Reading the
UDRn Register location will return the contents of the Receive Data Buffer Register (RXB).
For 5-, 6-, or 7-bit characters the upper unused bits will be ignored by the Transmitter and set to
zero by the Receiver.
The transmit buffer can only be written when the UDREn Flag in the UCSRnA Register is set.
Data written to UDRn when the UDREn Flag is not set, will be ignored by the USART Transmit-
ter. When data is written to the transmit buffer, and the Transmitter is enabled, the Transmitter
will load the data into the Transmit Shift Register when the Shift Register is empty. Then the
data will be serially transmitted on the TxDn pin.
The receive buffer consists of a two level FIFO. The FIFO will change its state whenever the
receive buffer is accessed. Due to this behavior of the receive buffer, do not use Read-Modify-
Write instructions (SBI and CBI) on this location. Be careful when using bit test instructions
(SBIC and SBIS), since these also will change the state of the FIFO.
17.11.2 UCSRnA – USART Control and Status Register A
Bit 7 6 5 4 3 2 1 0
RXCn TXCn UDREn FEn DORn UPEn U2Xn MPCMn UCSRnA
Read/Write R R/W R R R R R/W R/W
• Bit 7 – RXCn: USART Receive Complete

This flag bit is set when there are unread data in the receive buffer and cleared when the receive
buffer is empty (i.e., does not contain any unread data). If the Receiver is disabled, the receive
buffer will be flushed and consequently the RXCn bit will become zero. The RXCn Flag can be
used to generate a Receive Complete interrupt (see description of the RXCIEn bit).
• Bit 6 – TXCn: USART Transmit Complete

This flag bit is set when the entire frame in the Transmit Shift Register has been shifted out and
there are no new data currently present in the transmit buffer (UDRn). The TXCn Flag bit is auto-
matically cleared when a transmit complete interrupt is executed, or it can be cleared by writing
a one to its bit location. The TXCn Flag can generate a Transmit Complete interrupt (see
description of the TXCIEn bit).
• Bit 5 – UDREn: USART Data Register Empty

The UDREn Flag indicates if the transmit buffer (UDRn) is ready to receive new data. If UDREn
is one, the buffer is empty, and therefore ready to be written. The UDREn Flag can generate a
187
8011D–AVR–02/07
Data Register Empty interrupt (see description of the UDRIEn bit).UDREn is set after a reset to
indicate that the Transmitter is ready.
• Bit 4 – FEn: Frame Error

This bit is set if the next character in the receive buffer had a Frame Error when received. I.e.,
when the first stop bit of the next character in the receive buffer is zero. This bit is valid until the
receive buffer (UDRn) is read. The FEn bit is zero when the stop bit of received data is one.
Always set this bit to zero when writing to UCSRnA.
• Bit 3 – DORn: Data OverRun

This bit is set if a Data OverRun condition is detected. A Data OverRun occurs when the receive
buffer is full (two characters), it is a new character waiting in the Receive Shift Register, and a
new start bit is detected. This bit is valid until the receive buffer (UDRn) is read. Always set this
bit to zero when writing to UCSRnA.
• Bit 2 – UPEn: USART Parity Error

This bit is set if the next character in the receive buffer had a Parity Error when received and the
Parity Checking was enabled at that point (UPMn1 = 1). This bit is valid until the receive buffer
(UDRn) is read. Always set this bit to zero when writing to UCSRnA.
• Bit 1 – U2Xn: Double the USART Transmission Speed

This bit only has effect for the asynchronous operation. Write this bit to zero when using syn-
chronous operation.
Writing this bit to one will reduce the divisor of the baud rate divider from 16 to 8 effectively dou-
bling the transfer rate for asynchronous communication.
• Bit 0 – MPCMn: Multi-processor Communication Mode

This bit enables the Multi-processor Communication mode. When the MPCMn bit is written to
one, all the incoming frames received by the USART Receiver that do not contain address infor-
mation will be ignored. The Transmitter is unaffected by the MPCMn setting. For more detailed
information see ”Multi-processor Communication Mode” on page 185.
17.11.3 UCSRnB – USART Control and Status Register n B
Bit 7 6 5 4 3 2 1 0
RXCIEn TXCIEn UDRIEn RXENn TXENn UCSZn2 RXB8n TXB8n UCSRnB
Read/Write R/W R/W R/W R/W R/W R/W R R/W
• Bit 7 – RXCIEn: RX Complete Interrupt Enable n

Writing this bit to one enables interrupt on the RXCn Flag. A USART Receive Complete interrupt
will be generated only if the RXCIEn bit is written to one, the Global Interrupt Flag in SREG is
written to one and the RXCn bit in UCSRnA is set.
• Bit 6 – TXCIEn: TX Complete Interrupt Enable n

Writing this bit to one enables interrupt on the TXCn Flag. A USART Transmit Complete interrupt
will be generated only if the TXCIEn bit is written to one, the Global Interrupt Flag in SREG is
written to one and the TXCn bit in UCSRnA is set.
188 ATmega164P/324P/644P
8011D–AVR–02/07
• Bit 5 – UDRIEn: USART Data Register Empty Interrupt Enable n
Writing this bit to one enables interrupt on the UDREn Flag. A Data Register Empty interrupt will
be generated only if the UDRIEn bit is written to one, the Global Interrupt Flag in SREG is written
to one and the UDREn bit in UCSRnA is set.
• Bit 4 – RXENn: Receiver Enable n

Writing this bit to one enables the USART Receiver. The Receiver will override normal port oper-
ation for the RxDn pin when enabled. Disabling the Receiver will flush the receive buffer
invalidating the FEn, DORn, and UPEn Flags.
• Bit 3 – TXENn: Transmitter Enable n

Writing this bit to one enables the USART Transmitter. The Transmitter will override normal port
operation for the TxDn pin when enabled. The disabling of the Transmitter (writing TXENn to
zero) will not become effective until ongoing and pending transmissions are completed, i.e.,
when the Transmit Shift Register and Transmit Buffer Register do not contain data to be trans-
mitted. When disabled, the Transmitter will no longer override the TxDn port.
• Bit 2 – UCSZn2: Character Size n

The UCSZn2 bits combined with the UCSZn1:0 bit in UCSRnC sets the number of data bits
(Character SiZe) in a frame the Receiver and Transmitter use.
• Bit 1 – RXB8n: Receive Data Bit 8 n

RXB8n is the ninth data bit of the received character when operating with serial frames with nine
data bits. Must be read before reading the low bits from UDRn.
• Bit 0 – TXB8n: Transmit Data Bit 8 n

TXB8n is the ninth data bit in the character to be transmitted when operating with serial frames
with nine data bits. Must be written before writing the low bits to UDRn.
17.11.4 UCSRnC – USART Control and Status Register n C
Bit 7 6 5 4 3 2 1 0
UMSELn1 UMSELn0 UPMn1 UPMn0 USBSn UCSZn1 UCSZn0 UCPOLn UCSRnC
• Bits 7:6 – UMSELn1:0 USART Mode Select

These bits select the mode of operation of the USARTn as shown in Table 17-4..
Table 17-4. UMSELn Bits Settings

UMSELn1 UMSELn0 Mode
0 0 Asynchronous USART
0 1 Synchronous USART
1 0 (Reserved)
1 1 Master SPI (MSPIM)(1)
Note: 1. See ”USART in SPI Mode” on page 196 for full description of the Master SPI Mode (MSPIM)
operation
189
8011D–AVR–02/07
• Bits 5:4 – UPMn1:0: Parity Mode
These bits enable and set type of parity generation and check. If enabled, the Transmitter will
automatically generate and send the parity of the transmitted data bits within each frame. The
Receiver will generate a parity value for the incoming data and compare it to the UPMn setting.
If a mismatch is detected, the UPEn Flag in UCSRnA will be set.
Table 17-5. UPMn Bits Settings

UPMn1 UPMn0 Parity Mode
0 0 Disabled
0 1 Reserved
1 0 Enabled, Even Parity
1 1 Enabled, Odd Parity
• Bit 3 – USBSn: Stop Bit Select

This bit selects the number of stop bits to be inserted by the Transmitter. The Receiver ignores
this setting.
Table 17-6. USBS Bit Settings

USBSn Stop Bit(s)
0 1-bit
1 2-bit
• Bit 2:1 – UCSZn1:0: Character Size

The UCSZn1:0 bits combined with the UCSZn2 bit in UCSRnB sets the number of data bits
(Character SiZe) in a frame the Receiver and Transmitter use.
Table 17-7. UCSZn Bits Settings

UCSZn2 UCSZn1 UCSZn0 Character Size
0 0 0 5-bit
0 0 1 6-bit
0 1 0 7-bit
0 1 1 8-bit
1 0 0 Reserved
1 0 1 Reserved
1 1 0 Reserved
1 1 1 9-bit
190 ATmega164P/324P/644P
8011D–AVR–02/07
• Bit 0 – UCPOLn: Clock Polarity

This bit is used for synchronous mode only. Write this bit to zero when asynchronous mode is
used. The UCPOLn bit sets the relationship between data output change and data input sample,
and the synchronous clock (XCKn).
Table 17-8. UCPOLn Bit Settings

Transmitted Data Changed (Output of Received Data Sampled (Input on RxDn
UCPOLn TxDn Pin) Pin)
0 Rising XCKn Edge Falling XCKn Edge
1 Falling XCKn Edge Rising XCKn Edge
17.11.5 UBRRnL and UBRRnH – USART Baud Rate Registers
Bit 15 14 13 12 11 10 9 8
– – – – UBRR[11:8] UBRRHn
UBRR[7:0] UBRRLn
7 6 5 4 3 2 1 0
0 0 0 0 0 0 0 0
• Bit 15:12 – Reserved Bits

These bits are reserved for future use. For compatibility with future devices, these bit must be
written to zero when UBRRH is written.
• Bit 11:0 – UBRR11:0: USART Baud Rate Register

This is a 12-bit register which contains the USART baud rate. The UBRRH contains the four
most significant bits, and the UBRRL contains the eight least significant bits of the USART baud
rate. Ongoing transmissions by the Transmitter and Receiver will be corrupted if the baud rate is
changed. Writing UBRRL will trigger an immediate update of the baud rate prescaler.
191
8011D–AVR–02/07
17.12 Examples of Baud Rate Setting
For standard crystal and resonator frequencies, the most commonly used baud rates for asyn-
chronous operation can be generated by using the UBRR settings in Table 17-9 to Table 17-12.
UBRR values which yield an actual baud rate differing less than 0.5% from the target baud rate,
are bold in the table. Higher error ratings are acceptable, but the Receiver will have less noise
resistance when the error ratings are high, especially for large serial frames (see ”Asynchronous
Operational Range” on page 184). The error values are calculated using the following equation:
BaudRate Closest Match

Error[%] = ⎛ -------------------------------------------------------
- – 1⎞ • 100%
⎝ BaudRate ⎠
Table 17-9. Examples of UBRRn Settings for Commonly Used Oscillator Frequencies
fosc = 1.0000 MHz fosc = 1.8432 MHz fosc = 2.0000 MHz
Baud
U2Xn = 0 U2Xn = 1 U2Xn = 0 U2Xn = 1 U2Xn = 0 U2Xn = 1
Rate
(bps) UBRR Error UBRR Error UBRR Error UBRR Error UBRR Error UBRR Error
2400 25 0.2% 51 0.2% 47 0.0% 95 0.0% 51 0.2% 103 0.2%
4800 12 0.2% 25 0.2% 23 0.0% 47 0.0% 25 0.2% 51 0.2%
9600 6 -7.0% 12 0.2% 11 0.0% 23 0.0% 12 0.2% 25 0.2%
14.4k 3 8.5% 8 -3.5% 7 0.0% 15 0.0% 8 -3.5% 16 2.1%
19.2k 2 8.5% 6 -7.0% 5 0.0% 11 0.0% 6 -7.0% 12 0.2%
28.8k 1 8.5% 3 8.5% 3 0.0% 7 0.0% 3 8.5% 8 -3.5%
38.4k 1 -18.6% 2 8.5% 2 0.0% 5 0.0% 2 8.5% 6 -7.0%
57.6k 0 8.5% 1 8.5% 1 0.0% 3 0.0% 1 8.5% 3 8.5%
76.8k – – 1 -18.6% 1 -25.0% 2 0.0% 1 -18.6% 2 8.5%
115.2k – – 0 8.5% 0 0.0% 1 0.0% 0 8.5% 1 8.5%
230.4k – – – – – – 0 0.0% – – – –
250k – – – – – – – – – – 0 0.0%
(1)
Max. 62.5 kbps 125 kbps 115.2 kbps 230.4 kbps 125 kbps 250 kbps
1. UBRR = 0, Error = 0.0%
192 ATmega164P/324P/644P
8011D–AVR–02/07
Table 17-10. Examples of UBRRn Settings for Commonly Used Oscillator Frequencies (Continued)
Baud
Rate
2400 95 0.0% 191 0.0% 103 0.2% 207 0.2% 191 0.0% 383 0.0%
4800 47 0.0% 95 0.0% 51 0.2% 103 0.2% 95 0.0% 191 0.0%
9600 23 0.0% 47 0.0% 25 0.2% 51 0.2% 47 0.0% 95 0.0%
14.4k 15 0.0% 31 0.0% 16 2.1% 34 -0.8% 31 0.0% 63 0.0%
19.2k 11 0.0% 23 0.0% 12 0.2% 25 0.2% 23 0.0% 47 0.0%
28.8k 7 0.0% 15 0.0% 8 -3.5% 16 2.1% 15 0.0% 31 0.0%
38.4k 5 0.0% 11 0.0% 6 -7.0% 12 0.2% 11 0.0% 23 0.0%
57.6k 3 0.0% 7 0.0% 3 8.5% 8 -3.5% 7 0.0% 15 0.0%
76.8k 2 0.0% 5 0.0% 2 8.5% 6 -7.0% 5 0.0% 11 0.0%
115.2k 1 0.0% 3 0.0% 1 8.5% 3 8.5% 3 0.0% 7 0.0%
230.4k 0 0.0% 1 0.0% 0 8.5% 1 8.5% 1 0.0% 3 0.0%
250k 0 -7.8% 1 -7.8% 0 0.0% 1 0.0% 1 -7.8% 3 -7.8%
0.5M – – 0 -7.8% – – 0 0.0% 0 -7.8% 1 -7.8%
1M – – – – – – – – – – 0 -7.8%
(1)
Max. 230.4 kbps 460.8 kbps 250 kbps 0.5 Mbps 460.8 kbps 921.6 kbps
1. UBRR = 0, Error = 0.0%
193
8011D–AVR–02/07
Baud
Rate
2400 207 0.2% 416 -0.1% 287 0.0% 575 0.0% 383 0.0% 767 0.0%
4800 103 0.2% 207 0.2% 143 0.0% 287 0.0% 191 0.0% 383 0.0%
9600 51 0.2% 103 0.2% 71 0.0% 143 0.0% 95 0.0% 191 0.0%
14.4k 34 -0.8% 68 0.6% 47 0.0% 95 0.0% 63 0.0% 127 0.0%
19.2k 25 0.2% 51 0.2% 35 0.0% 71 0.0% 47 0.0% 95 0.0%
28.8k 16 2.1% 34 -0.8% 23 0.0% 47 0.0% 31 0.0% 63 0.0%
38.4k 12 0.2% 25 0.2% 17 0.0% 35 0.0% 23 0.0% 47 0.0%
57.6k 8 -3.5% 16 2.1% 11 0.0% 23 0.0% 15 0.0% 31 0.0%
76.8k 6 -7.0% 12 0.2% 8 0.0% 17 0.0% 11 0.0% 23 0.0%
115.2k 3 8.5% 8 -3.5% 5 0.0% 11 0.0% 7 0.0% 15 0.0%
230.4k 1 8.5% 3 8.5% 2 0.0% 5 0.0% 3 0.0% 7 0.0%
250k 1 0.0% 3 0.0% 2 -7.8% 5 -7.8% 3 -7.8% 6 5.3%
0.5M 0 0.0% 1 0.0% – – 2 -7.8% 1 -7.8% 3 -7.8%
1M – – 0 0.0% – – – – 0 -7.8% 1 -7.8%
(1)
Max. 0.5 Mbps 1 Mbps 691.2 kbps 1.3824 Mbps 921.6 kbps 1.8432 Mbps
1. UBRR = 0, Error = 0.0%
194 ATmega164P/324P/644P
8011D–AVR–02/07
Baud
Rate
2400 416 -0.1% 832 0.0% 479 0.0% 959 0.0% 520 0.0% 1041 0.0%
4800 207 0.2% 416 -0.1% 239 0.0% 479 0.0% 259 0.2% 520 0.0%
9600 103 0.2% 207 0.2% 119 0.0% 239 0.0% 129 0.2% 259 0.2%
14.4k 68 0.6% 138 -0.1% 79 0.0% 159 0.0% 86 -0.2% 173 -0.2%
19.2k 51 0.2% 103 0.2% 59 0.0% 119 0.0% 64 0.2% 129 0.2%
28.8k 34 -0.8% 68 0.6% 39 0.0% 79 0.0% 42 0.9% 86 -0.2%
38.4k 25 0.2% 51 0.2% 29 0.0% 59 0.0% 32 -1.4% 64 0.2%
57.6k 16 2.1% 34 -0.8% 19 0.0% 39 0.0% 21 -1.4% 42 0.9%
76.8k 12 0.2% 25 0.2% 14 0.0% 29 0.0% 15 1.7% 32 -1.4%
115.2k 8 -3.5% 16 2.1% 9 0.0% 19 0.0% 10 -1.4% 21 -1.4%
230.4k 3 8.5% 8 -3.5% 4 0.0% 9 0.0% 4 8.5% 10 -1.4%
250k 3 0.0% 7 0.0% 4 -7.8% 8 2.4% 4 0.0% 9 0.0%
0.5M 1 0.0% 3 0.0% – – 4 -7.8% – – 4 0.0%
1M 0 0.0% 1 0.0% – – – – – – – –
(1)
Max. 1 Mbps 2 Mbps 1.152 Mbps 2.304 Mbps 1.25 Mbps 2.5 Mbps
1. UBRR = 0, Error = 0.0%
195
8011D–AVR–02/07
18. USART in SPI Mode
18.1 Features
• Full Duplex, Three-wire Synchronous Data Transfer
• Master Operation
• Supports all four SPI Modes of Operation (Mode 0, 1, 2, and 3)
• LSB First or MSB First Data Transfer (Configurable Data Order)
• Queued Operation (Double Buffered)
• High Resolution Baud Rate Generator
• High Speed Operation (fXCKmax = fCK/2)
• Flexible Interrupt Generation
18.2 Overview
The Universal Synchronous and Asynchronous serial Receiver and Transmitter (USART) can be
set to a master SPI compliant mode of operation.
Setting both UMSELn1:0 bits to one enables the USART in MSPIM logic. In this mode of opera-
tion the SPI master control logic takes direct control over the USART resources. These
resources include the transmitter and receiver shift register and buffers, and the baud rate gen-
erator. The parity generator and checker, the data and clock recovery logic, and the RX and TX
control logic is disabled. The USART RX and TX control logic is replaced by a common SPI
transfer control logic. However, the pin control logic and interrupt generation logic is identical in
both modes of operation.
The I/O register locations are the same in both modes. However, some of the functionality of the
control registers changes when using MSPIM.
18.3 Clock Generation

The Clock Generation logic generates the base clock for the Transmitter and Receiver. For
USART MSPIM mode of operation only internal clock generation (i.e. master operation) is sup-
ported. The Data Direction Register for the XCKn pin (DDR_XCKn) must therefore be set to one
(i.e. as output) for the USART in MSPIM to operate correctly. Preferably the DDR_XCKn should
be set up before the USART in MSPIM is enabled (i.e. TXENn and RXENn bit set to one).
The internal clock generation used in MSPIM mode is identical to the USART synchronous mas-
ter mode. The baud rate or UBRRn setting can therefore be calculated using the same
equations, see Table 18-1:
Table 18-1. Equations for Calculating Baud Rate Register Setting
Equation for Calculating Baud Equation for Calculating UBRRn
Operating Mode Rate(1) Value
Synchronous Master f OSC f OSC

mode BAUD = --------------------------------------
- -–1
UBRRn = -------------------
2 ( UBRRn + 1 ) 2BAUD
Note: 1. The baud rate is defined to be the transfer rate in bit per second (bps)
196 ATmega164P/324P/644P
8011D–AVR–02/07
BAUD Baud rate (in bits per second, bps)

fOSC System Oscillator clock frequency
UBRRn Contents of the UBRRnH and UBRRnL Registers, (0-4095)
18.4 SPI Data Modes and Timing

There are four combinations of XCKn (SCK) phase and polarity with respect to serial data, which
are determined by control bits UCPHAn and UCPOLn. The data transfer timing diagrams are
shown in Figure 18-1. Data bits are shifted out and latched in on opposite edges of the XCKn
signal, ensuring sufficient time for data signals to stabilize. The UCPOLn and UCPHAn function-
ality is summarized in Table 18-2. Note that changing the setting of any of these bits will corrupt
all ongoing communication for both the Receiver and Transmitter.
Table 18-2. UCPOLn and UCPHAn Functionality-

UCPOLn UCPHAn SPI Mode Leading Edge Trailing Edge
0 0 0 Sample (Rising) Setup (Falling)
0 1 1 Setup (Rising) Sample (Falling)
1 0 2 Sample (Falling) Setup (Rising)
1 1 3 Setup (Falling) Sample (Rising)
Figure 18-1. UCPHAn and UCPOLn data transfer timing diagrams.

UCPOL=0 UCPOL=1
UCPHA=1
XCK XCK
Data setup (TXD) Data setup (TXD)
Data sample (RXD) Data sample (RXD)

UCPHA=0
XCK XCK
Data setup (TXD) Data setup (TXD)
Data sample (RXD) Data sample (RXD)
18.5 Frame Formats

A serial frame for the MSPIM is defined to be one character of 8 data bits. The USART in MSPIM
mode has two valid frame formats:
• 8-bit data with MSB first
• 8-bit data with LSB first
A frame starts with the least or most significant data bit. Then the next data bits, up to a total of
eight, are succeeding, ending with the most or least significant bit accordingly. When a complete
frame is transmitted, a new frame can directly follow it, or the communication line can be set to
an idle (high) state.
197
8011D–AVR–02/07
The UDORDn bit in UCSRnC sets the frame format used by the USART in MSPIM mode. The
Receiver and Transmitter use the same setting. Note that changing the setting of any of these
bits will corrupt all ongoing communication for both the Receiver and Transmitter.
16-bit data transfer can be achieved by writing two data bytes to UDRn. A UART transmit com-
plete interrupt will then signal that the 16-bit value has been shifted out.
18.5.1 USART MSPIM Initialization
The USART in MSPIM mode has to be initialized before any communication can take place. The
initialization process normally consists of setting the baud rate, setting master mode of operation
(by setting DDR_XCKn to one), setting frame format and enabling the Transmitter and the
Receiver. Only the transmitter can operate independently. For interrupt driven USART opera-
tion, the Global Interrupt Flag should be cleared (and thus interrupts globally disabled) when
doing the initialization.
Note: To ensure immediate initialization of the XCKn output the baud-rate register (UBRRn) must be
zero at the time the transmitter is enabled. Contrary to the normal mode USART operation the
UBRRn must then be written to the desired value after the transmitter is enabled, but before the
first transmission is started. Setting UBRRn to zero before enabling the transmitter is not neces-
sary if the initialization is done immediately after a reset since UBRRn is reset to zero.
Before doing a re-initialization with changed baud rate, data mode, or frame format, be sure that
there is no ongoing transmissions during the period the registers are changed. The TXCn Flag
can be used to check that the Transmitter has completed all transfers, and the RXCn Flag can
be used to check that there are no unread data in the receive buffer. Note that the TXCn Flag
must be cleared before each transmission (before UDRn is written) if it is used for this purpose.
The following simple USART initialization code examples show one assembly and one C func-
tion that are equal in functionality. The examples assume polling (no interrupts enabled). The
baud rate is given as a function parameter. For the assembly code, the baud rate parameter is
assumed to be stored in the r17:r16 registers.
198 ATmega164P/324P/644P
8011D–AVR–02/07
USART_Init:
clr r18
out UBRRnH,r18
out UBRRnL,r18
; Setting the XCKn port pin as output, enables master mode.
sbi XCKn_DDR, XCKn
; Set MSPI mode of operation and SPI data mode 0.
ldi r18, (1<<UMSELn1)|(1<<UMSELn0)|(0<<UCPHAn)|(0<<UCPOLn)
out UCSRnC,r18
; Enable receiver and transmitter.
ldi r18, (1<<RXENn)|(1<<TXENn)
out UCSRnB,r18
; Set baud rate.
; IMPORTANT: The Baud Rate must be set after the transmitter is enabled!
out UBRRnH, r17
out UBRRnL, r18
ret
C Code Example(1)
void USART_Init( unsigned int baud )
{
UBRRn = 0;
/* Setting the XCKn port pin as output, enables master mode. */
XCKn_DDR |= (1<<XCKn);
/* Set MSPI mode of operation and SPI data mode 0. */
UCSRnC = (1<<UMSELn1)|(1<<UMSELn0)|(0<<UCPHAn)|(0<<UCPOLn);
/* Enable receiver and transmitter. */
UCSRnB = (1<<RXENn)|(1<<TXENn);
/* Set baud rate. */
/* IMPORTANT: The Baud Rate must be set after the transmitter is enabled
*/
UBRRn = baud;
}

18.6 Data Transfer
Using the USART in MSPI mode requires the Transmitter to be enabled, i.e. the TXENn bit in
the UCSRnB register is set to one. When the Transmitter is enabled, the normal port operation
of the TxDn pin is overridden and given the function as the Transmitter's serial output. Enabling
the receiver is optional and is done by setting the RXENn bit in the UCSRnB register to one.
When the receiver is enabled, the normal pin operation of the RxDn pin is overridden and given
the function as the Receiver's serial input. The XCKn will in both cases be used as the transfer
clock.
After initialization the USART is ready for doing data transfers. A data transfer is initiated by writ-
ing to the UDRn I/O location. This is the case for both sending and receiving data since the
199
8011D–AVR–02/07
transmitter controls the transfer clock. The data written to UDRn is moved from the transmit
buffer to the shift register when the shift register is ready to send a new frame.
Note: To keep the input buffer in sync with the number of data bytes transmitted, the UDRn register must
be read once for each byte transmitted. The input buffer operation is identical to normal USART
mode, i.e. if an overflow occurs the character last received will be lost, not the first data in the
buffer. This means that if four bytes are transferred, byte 1 first, then byte 2, 3, and 4, and the
UDRn is not read before all transfers are completed, then byte 3 to be received will be lost, and not
byte 1.
The following code examples show a simple USART in MSPIM mode transfer function based on
polling of the Data Register Empty (UDREn) Flag and the Receive Complete (RXCn) Flag. The
USART has to be initialized before the function can be used. For the assembly code, the data to
be sent is assumed to be stored in Register R16 and the data received will be available in the
same register (R16) after the function returns.
The function simply waits for the transmit buffer to be empty by checking the UDREn Flag,
before loading it with new data to be transmitted. The function then waits for data to be present
in the receive buffer by checking the RXCn Flag, before reading the buffer and returning the
value..

USART_MSPIM_Transfer:
sbis UCSRnA, UDREn
rjmp USART_MSPIM_Transfer
; Put data (r16) into buffer, sends the data
out UDRn,r16
USART_MSPIM_Wait_RXCn:
sbis UCSRnA, RXCn
rjmp USART_MSPIM_Wait_RXCn
; Get and return received data from buffer
in r16, UDRn
ret
C Code Example(1)
unsigned char USART_Receive( void )
{
while ( !( UCSRnA & (1<<UDREn)) );
UDRn = data;
while ( !(UCSRnA & (1<<RXCn)) );
/* Get and return received data from buffer */
return UDRn;
}
200 ATmega164P/324P/644P
8011D–AVR–02/07
18.6.1 Transmitter and Receiver Flags and Interrupts
The RXCn, TXCn, and UDREn flags and corresponding interrupts in USART in MSPIM mode
are identical in function to the normal USART operation. However, the receiver error status flags
(FE, DOR, and PE) are not in use and is always read as zero.
18.6.2 Disabling the Transmitter or Receiver
The disabling of the transmitter or receiver in USART in MSPIM mode is identical in function to
the normal USART operation.
18.7 AVR USART MSPIM vs. AVR SPI

The USART in MSPIM mode is fully compatible with the AVR SPI regarding:
• Master mode timing diagram.
• The UCPOLn bit functionality is identical to the SPI CPOL bit.
• The UCPHAn bit functionality is identical to the SPI CPHA bit.
• The UDORDn bit functionality is identical to the SPI DORD bit.
However, since the USART in MSPIM mode reuses the USART resources, the use of the
USART in MSPIM mode is somewhat different compared to the SPI. In addition to differences of
the control register bits, and that only master operation is supported by the USART in MSPIM
mode, the following features differ between the two modules:
• The USART in MSPIM mode includes (double) buffering of the transmitter. The SPI has no
buffer.
• The USART in MSPIM mode receiver includes an additional buffer level.
• The SPI WCOL (Write Collision) bit is not included in USART in MSPIM mode.
• The SPI double speed mode (SPI2X) bit is not included. However, the same effect is achieved
by setting UBRRn accordingly.
• Interrupt timing is not compatible.
• Pin control differs due to the master only operation of the USART in MSPIM mode.
A comparison of the USART in MSPIM mode and the SPI pins is shown in Table 18-3 on page
201.
Table 18-3. Comparison of USART in MSPIM mode and SPI pins.

USART_MSPIM SPI Comment
TxDn MOSI Master Out only
RxDn MISO Master In only
XCKn SCK (Functionally identical)
(N/A) SS Not supported by USART in MSPIM
201
8011D–AVR–02/07
The following section describes the registers used for SPI operation using the USART.
18.8.1 UDRn – USART MSPIM I/O Data Register
The function and bit description of the USART data register (UDRn) in MSPI mode is identical to
normal USART operation. See “UDRn – USART I/O Data Register n” on page 187.
18.8.2 UCSRnA – USART MSPIM Control and Status Register n A
Bit 7 6 5 4 3 2 1 0
RXCn TXCn UDREn - - - - - UCSRnA
Read/Write R/W R/W R/W R R R R R
• Bit 7 - RXCn: USART Receive Complete

This flag bit is set when there are unread data in the receive buffer and cleared when the receive
buffer is empty (i.e., does not contain any unread data). If the Receiver is disabled, the receive
buffer will be flushed and consequently the RXCn bit will become zero. The RXCn Flag can be
used to generate a Receive Complete interrupt (see description of the RXCIEn bit).
• Bit 6 - TXCn: USART Transmit Complete

This flag bit is set when the entire frame in the Transmit Shift Register has been shifted out and
there are no new data currently present in the transmit buffer (UDRn). The TXCn Flag bit is auto-
matically cleared when a transmit complete interrupt is executed, or it can be cleared by writing
a one to its bit location. The TXCn Flag can generate a Transmit Complete interrupt (see
description of the TXCIEn bit).
• Bit 5 - UDREn: USART Data Register Empty

The UDREn Flag indicates if the transmit buffer (UDRn) is ready to receive new data. If UDREn
is one, the buffer is empty, and therefore ready to be written. The UDREn Flag can generate a
Data Register Empty interrupt (see description of the UDRIE bit). UDREn is set after a reset to
indicate that the Transmitter is ready.
• Bit 4:0 - Reserved Bits in MSPI mode

When in MSPI mode, these bits are reserved for future use. For compatibility with future devices,
these bits must be written to zero when UCSRnA is written.
18.8.3 UCSRnB – USART MSPIM Control and Status Register n B
Bit 7 6 5 4 3 2 1 0
RXCIEn TXCIEn UDRIE RXENn TXENn - - - UCSRnB
Read/Write R/W R/W R/W R/W R/W R R R
• Bit 7 - RXCIEn: RX Complete Interrupt Enable

Writing this bit to one enables interrupt on the RXCn Flag. A USART Receive Complete interrupt
will be generated only if the RXCIEn bit is written to one, the Global Interrupt Flag in SREG is
written to one and the RXCn bit in UCSRnA is set.
202 ATmega164P/324P/644P
8011D–AVR–02/07
• Bit 6 - TXCIEn: TX Complete Interrupt Enable
Writing this bit to one enables interrupt on the TXCn Flag. A USART Transmit Complete interrupt
will be generated only if the TXCIEn bit is written to one, the Global Interrupt Flag in SREG is
written to one and the TXCn bit in UCSRnA is set.
• Bit 5 - UDRIE: USART Data Register Empty Interrupt Enable

Writing this bit to one enables interrupt on the UDREn Flag. A Data Register Empty interrupt will
be generated only if the UDRIE bit is written to one, the Global Interrupt Flag in SREG is written
to one and the UDREn bit in UCSRnA is set.
• Bit 4 - RXENn: Receiver Enable

Writing this bit to one enables the USART Receiver in MSPIM mode. The Receiver will override
normal port operation for the RxDn pin when enabled. Disabling the Receiver will flush the
receive buffer. Only enabling the receiver in MSPI mode (i.e. setting RXENn=1 and TXENn=0)
has no meaning since it is the transmitter that controls the transfer clock and since only master
mode is supported.
• Bit 3 - TXENn: Transmitter Enable

Writing this bit to one enables the USART Transmitter. The Transmitter will override normal port
operation for the TxDn pin when enabled. The disabling of the Transmitter (writing TXENn to
zero) will not become effective until ongoing and pending transmissions are completed, i.e.,
when the Transmit Shift Register and Transmit Buffer Register do not contain data to be trans-
mitted. When disabled, the Transmitter will no longer override the TxDn port.

these bits must be written to zero when UCSRnB is written.
18.8.4 UCSRnC – USART MSPIM Control and Status Register n C
Bit 7 6 5 4 3 2 1 0
UMSELn1 UMSELn0 - - - UDORDn UCPHAn UCPOLn UCSRnC
Read/Write R/W R/W R R R R/W R/W R/W
• Bit 7:6 - UMSELn1:0: USART Mode Select

These bits select the mode of operation of the USART as shown in Table 18-4. See ”UCSRnC –
USART Control and Status Register n C” on page 189 for full description of the normal USART
operation. The MSPIM is enabled when both UMSELn bits are set to one. The UDORDn,
UCPHAn, and UCPOLn can be set in the same write operation where the MSPIM is enabled.
Table 18-4. UMSELn Bits Settings

UMSELn1 UMSELn0 Mode
0 0 Asynchronous USART
0 1 Synchronous USART
1 0 (Reserved)
1 1 Master SPI (MSPIM)
203
8011D–AVR–02/07
these bits must be written to zero when UCSRnC is written.
• Bit 2 - UDORDn: Data Order

When set to one the LSB of the data word is transmitted first. When set to zero the MSB of the
data word is transmitted first. Refer to the Frame Formats section page 4 for details.
• Bit 1 - UCPHAn: Clock Phase

The UCPHAn bit setting determine if data is sampled on the leasing edge (first) or tailing (last)
edge of XCKn. Refer to the SPI Data Modes and Timing section page 4 for details.
• Bit 0 - UCPOLn: Clock Polarity

The UCPOLn bit sets the polarity of the XCKn clock. The combination of the UCPOLn and
UCPHAn bit settings determine the timing of the data transfer. Refer to the SPI Data Modes and
Timing section page 4 for details.
18.8.5 UBRRnL and UBRRnH –USART MSPIM Baud Rate Registers
The function and bit description of the baud rate registers in MSPI mode is identical to normal
USART operation. See ”UBRRnL and UBRRnH – USART Baud Rate Registers” on page 191.
204 ATmega164P/324P/644P
8011D–AVR–02/07
19. 2-wire Serial Interface

19.1 Features
• Simple Yet Powerful and Flexible Communication Interface, only two Bus Lines Needed
• Both Master and Slave Operation Supported
• Device can Operate as Transmitter or Receiver
• 7-bit Address Space Allows up to 128 Different Slave Addresses
• Multi-master Arbitration Support
• Up to 400 kHz Data Transfer Speed
• Slew-rate Limited Output Drivers
• Noise Suppression Circuitry Rejects Spikes on Bus Lines
• Fully Programmable Slave Address with General Call Support
• Address Recognition Causes Wake-up When AVR is in Sleep Mode
19.2 2-wire Serial Interface Bus Definition

The 2-wire Serial Interface (TWI) is ideally suited for typical microcontroller applications. The
TWI protocol allows the systems designer to interconnect up to 128 different devices using only
two bi-directional bus lines, one for clock (SCL) and one for data (SDA). The only external hard-
ware needed to implement the bus is a single pull-up resistor for each of the TWI bus lines. All
devices connected to the bus have individual addresses, and mechanisms for resolving bus
contention are inherent in the TWI protocol.
Figure 19-1. TWI Bus Interconnection

VCC
Device 1 Device 2 Device 3 ........ Device n R1 R2
SDA
SCL
19.2.1 TWI Terminology
The following definitions are frequently encountered in this section.
Table 19-1. TWI Terminology

Term Description
The device that initiates and terminates a transmission. The Master also generates the
Master
SCL clock.
Slave The device addressed by a Master.
Transmitter The device placing data on the bus.
Receiver The device reading data from the bus.
205
8011D–AVR–02/07
The Power Reduction TWI bit, PRTWI bit in ”PRR – Power Reduction Register” on page 48 must
be written to zero to enable the 2-wire Serial Interface.
19.2.2 Electrical Interconnection
As depicted in Figure 19-1, both bus lines are connected to the positive supply voltage through
pull-up resistors. The bus drivers of all TWI-compliant devices are open-drain or open-collector.
This implements a wired-AND function which is essential to the operation of the interface. A low
level on a TWI bus line is generated when one or more TWI devices output a zero. A high level
is output when all TWI devices trim-state their outputs, allowing the pull-up resistors to pull the
line high. Note that all AVR devices connected to the TWI bus must be powered in order to allow
any bus operation.
The number of devices that can be connected to the bus is only limited by the bus capacitance
limit of 400 pF and the 7-bit slave address space. A detailed specification of the electrical char-
acteristics of the TWI is given in ”SPI Timing Characteristics” on page 330. Two different sets of
specifications are presented there, one relevant for bus speeds below 100 kHz, and one valid for
bus speeds up to 400 kHz.
19.3 Data Transfer and Frame Format

19.3.1 Transferring Bits
Each data bit transferred on the TWI bus is accompanied by a pulse on the clock line. The level
of the data line must be stable when the clock line is high. The only exception to this rule is for
generating start and stop conditions.
Figure 19-2. Data Validity
SDA
SCL
Data Stable Data Stable
Data Change
19.3.2 START and STOP Conditions
The Master initiates and terminates a data transmission. The transmission is initiated when the
Master issues a START condition on the bus, and it is terminated when the Master issues a
STOP condition. Between a START and a STOP condition, the bus is considered busy, and no
other master should try to seize control of the bus. A special case occurs when a new START
condition is issued between a START and STOP condition. This is referred to as a REPEATED
START condition, and is used when the Master wishes to initiate a new transfer without relin-
quishing control of the bus. After a REPEATED START, the bus is considered busy until the next
STOP. This is identical to the START behavior, and therefore START is used to describe both
START and REPEATED START for the remainder of this datasheet, unless otherwise noted. As
206 ATmega164P/324P/644P
8011D–AVR–02/07
depicted below, START and STOP conditions are signalled by changing the level of the SDA
line when the SCL line is high.
Figure 19-3. START, REPEATED START and STOP conditions
SDA
SCL
START STOP START REPEATED START STOP

19.3.3 Address Packet Format
All address packets transmitted on the TWI bus are 9 bits long, consisting of 7 address bits, one
READ/WRITE control bit and an acknowledge bit. If the READ/WRITE bit is set, a read opera-
tion is to be performed, otherwise a write operation should be performed. When a Slave
recognizes that it is being addressed, it should acknowledge by pulling SDA low in the ninth SCL
(ACK) cycle. If the addressed Slave is busy, or for some other reason can not service the Mas-
ter’s request, the SDA line should be left high in the ACK clock cycle. The Master can then
transmit a STOP condition, or a REPEATED START condition to initiate a new transmission. An
address packet consisting of a slave address and a READ or a WRITE bit is called SLA+R or
SLA+W, respectively.
The MSB of the address byte is transmitted first. Slave addresses can freely be allocated by the
designer, but the address 0000 000 is reserved for a general call.
When a general call is issued, all slaves should respond by pulling the SDA line low in the ACK
cycle. A general call is used when a Master wishes to transmit the same message to several
slaves in the system. When the general call address followed by a Write bit is transmitted on the
bus, all slaves set up to acknowledge the general call will pull the SDA line low in the ack cycle.
The following data packets will then be received by all the slaves that acknowledged the general
call. Note that transmitting the general call address followed by a Read bit is meaningless, as
this would cause contention if several slaves started transmitting different data.
All addresses of the format 1111 xxx should be reserved for future purposes.
Figure 19-4. Address Packet Format
Addr MSB Addr LSB R/W ACK
SDA
SCL
1 2 7 8 9
START
207
8011D–AVR–02/07
19.3.4 Data Packet Format
All data packets transmitted on the TWI bus are nine bits long, consisting of one data byte and
an acknowledge bit. During a data transfer, the Master generates the clock and the START and
STOP conditions, while the Receiver is responsible for acknowledging the reception. An
Acknowledge (ACK) is signalled by the Receiver pulling the SDA line low during the ninth SCL
cycle. If the Receiver leaves the SDA line high, a NACK is signalled. When the Receiver has
received the last byte, or for some reason cannot receive any more bytes, it should inform the
Transmitter by sending a NACK after the final byte. The MSB of the data byte is transmitted first.
Figure 19-5. Data Packet Format

Data MSB Data LSB ACK
Aggregate
SDA
SDA from
Transmitter
SDA from
Receiver
SCL from
Master
1 2 7 8 9
STOP, REPEATED
SLA+R/W Data Byte START or Next
Data Byte
19.3.5 Combining Address and Data Packets into a Transmission
A transmission basically consists of a START condition, a SLA+R/W, one or more data packets
and a STOP condition. An empty message, consisting of a START followed by a STOP condi-
tion, is illegal. Note that the Wired-ANDing of the SCL line can be used to implement
handshaking between the Master and the Slave. The Slave can extend the SCL low period by
pulling the SCL line low. This is useful if the clock speed set up by the Master is too fast for the
Slave, or the Slave needs extra time for processing between the data transmissions. The Slave
extending the SCL low period will not affect the SCL high period, which is determined by the
Master. As a consequence, the Slave can reduce the TWI data transfer speed by prolonging the
SCL duty cycle.
Figure 19-6 on page 208 shows a typical data transmission. Note that several data bytes can be
transmitted between the SLA+R/W and the STOP condition, depending on the software protocol
implemented by the application software.
Figure 19-6. Typical Data Transmission
Addr MSB Addr LSB R/W ACK Data MSB Data LSB ACK
1 2 7 8 9 1 2 7 8 9
SLA+R/W Data Byte
208 ATmega164P/324P/644P
8011D–AVR–02/07
19.4 Multi-master Bus Systems, Arbitration and Synchronization
The TWI protocol allows bus systems with several masters. Special concerns have been taken
in order to ensure that transmissions will proceed as normal, even if two or more masters initiate
a transmission at the same time. Two problems arise in multi-master systems:
• An algorithm must be implemented allowing only one of the masters to complete the
transmission. All other masters should cease transmission when they discover that they have
lost the selection process. This selection process is called arbitration. When a contending
master discovers that it has lost the arbitration process, it should immediately switch to Slave
mode to check whether it is being addressed by the winning master. The fact that multiple
masters have started transmission at the same time should not be detectable to the slaves, i.e.
the data being transferred on the bus must not be corrupted.
• Different masters may use different SCL frequencies. A scheme must be devised to
synchronize the serial clocks from all masters, in order to let the transmission proceed in a
lockstep fashion. This will facilitate the arbitration process.
The wired-ANDing of the bus lines is used to solve both these problems. The serial clocks from
all masters will be wired-ANDed, yielding a combined clock with a high period equal to the one
from the Master with the shortest high period. The low period of the combined clock is equal to
the low period of the Master with the longest low period. Note that all masters listen to the SCL
line, effectively starting to count their SCL high and low time-out periods when the combined
SCL line goes high or low, respectively.
Figure 19-7. SCL Synchronization Between Multiple Masters

TA low TA high
SCL from
Master A
SCL from
Master B
SCL Bus
Line
TBlow TBhigh
Masters Start Masters Start

Counting Low Period Counting High Period
Arbitration is carried out by all masters continuously monitoring the SDA line after outputting
data. If the value read from the SDA line does not match the value the Master had output, it has
lost the arbitration. Note that a Master can only lose arbitration when it outputs a high SDA value
while another Master outputs a low value. The losing Master should immediately go to Slave
mode, checking if it is being addressed by the winning Master. The SDA line should be left high,
but losing masters are allowed to generate a clock signal until the end of the current data or
address packet. Arbitration will continue until only one Master remains, and this may take many
209
8011D–AVR–02/07
bits. If several masters are trying to address the same Slave, arbitration will continue into the
data packet.
Figure 19-8. Arbitration Between Two Masters

START Master A Loses
Arbitration, SDAA SDA
SDA from
Master A
SDA from
Master B
SDA Line
Synchronized
SCL Line
Note that arbitration is not allowed between:

• A REPEATED START condition and a data bit.
• A STOP condition and a data bit.
• A REPEATED START and a STOP condition.
It is the user software’s responsibility to ensure that these illegal arbitration conditions never
occur. This implies that in multi-master systems, all data transfers must use the same composi-
tion of SLA+R/W and data packets. In other words: All transmissions must contain the same
number of data packets, otherwise the result of the arbitration is undefined.
210 ATmega164P/324P/644P
8011D–AVR–02/07
19.5 Overview of the TWI Module
The TWI module is comprised of several submodules, as shown in Figure 19-9. All registers
drawn in a thick line are accessible through the AVR data bus.
Figure 19-9. Overview of the TWI Module
SCL SDA
Slew-rate Spike Slew-rate Spike
Control Filter Control Filter
Bus Interface Unit Bit Rate Generator

START / STOP
Spike Suppression Prescaler
Control
Address/Data Shift Bit Rate Register

Arbitration detection Ack
Register (TWDR) (TWBR)
Address Match Unit Control Unit
TWI Unit
Address Register Status Register Control Register
(TWAR) (TWSR) (TWCR)
State Machine and

Address Comparator
Status control
19.5.1 SCL and SDA Pins
These pins interface the AVR TWI with the rest of the MCU system. The output drivers contain a
slew-rate limiter in order to conform to the TWI specification. The input stages contain a spike
suppression unit removing spikes shorter than 50 ns. Note that the internal pull-ups in the AVR
pads can be enabled by setting the PORT bits corresponding to the SCL and SDA pins, as
explained in the I/O Port section. The internal pull-ups can in some systems eliminate the need
for external ones.
19.5.2 Bit Rate Generator Unit
This unit controls the period of SCL when operating in a Master mode. The SCL period is con-
trolled by settings in the TWI Bit Rate Register (TWBR) and the Prescaler bits in the TWI Status
Register (TWSR). Slave operation does not depend on Bit Rate or Prescaler settings, but the
CPU clock frequency in the Slave must be at least 16 times higher than the SCL frequency. Note
that slaves may prolong the SCL low period, thereby reducing the average TWI bus clock
period. The SCL frequency is generated according to the following equation:
211
8011D–AVR–02/07
CPU Clock frequency
SCL frequency = ----------------------------------------------------------
-
TWPS
16 + 2(TWBR) ⋅ 4
• TWBR = Value of the TWI Bit Rate Register.

• TWPS = Value of the prescaler bits in the TWI Status Register.
Note: Pull-up resistor values should be selected according to the SCL frequency and the capacitive bus
line load. See 2-wire Serial Bus Requirements in Table 26-9 on page 331 for value of pull-up
resistor.
19.5.3 Bus Interface Unit
This unit contains the Data and Address Shift Register (TWDR), a START/STOP Controller and
Arbitration detection hardware. The TWDR contains the address or data bytes to be transmitted,
or the address or data bytes received. In addition to the 8-bit TWDR, the Bus Interface Unit also
contains a register containing the (N)ACK bit to be transmitted or received. This (N)ACK Regis-
ter is not directly accessible by the application software. However, when receiving, it can be set
or cleared by manipulating the TWI Control Register (TWCR). When in Transmitter mode, the
value of the received (N)ACK bit can be determined by the value in the TWSR.
The START/STOP Controller is responsible for generation and detection of START, REPEATED
START, and STOP conditions. The START/STOP controller is able to detect START and STOP
conditions even when the AVR MCU is in one of the sleep modes, enabling the MCU to wake up
if addressed by a Master.
If the TWI has initiated a transmission as Master, the Arbitration Detection hardware continu-
ously monitors the transmission trying to determine if arbitration is in process. If the TWI has lost
an arbitration, the Control Unit is informed. Correct action can then be taken and appropriate
status codes generated.
19.5.4 Address Match Unit
The Address Match unit checks if received address bytes match the seven-bit address in the
TWI Address Register (TWAR). If the TWI General Call Recognition Enable (TWGCE) bit in the
TWAR is written to one, all incoming address bits will also be compared against the General Call
address. Upon an address match, the Control Unit is informed, allowing correct action to be
taken. The TWI may or may not acknowledge its address, depending on settings in the TWCR.
The Address Match unit is able to compare addresses even when the AVR MCU is in sleep
mode, enabling the MCU to wake up if addressed by a Master. If another interrupt (e.g., INT0)
occurs during TWI Power-down address match and wakes up the CPU, the TWI aborts opera-
tion and return to it’s idle state. If this cause any problems, ensure that TWI Address Match is the
only enabled interrupt when entering Power-down.
19.5.5 Control Unit
The Control unit monitors the TWI bus and generates responses corresponding to settings in the
TWI Control Register (TWCR). When an event requiring the attention of the application occurs
on the TWI bus, the TWI Interrupt Flag (TWINT) is asserted. In the next clock cycle, the TWI Sta-
tus Register (TWSR) is updated with a status code identifying the event. The TWSR only
contains relevant status information when the TWI Interrupt Flag is asserted. At all other times,
the TWSR contains a special status code indicating that no relevant status information is avail-
able. As long as the TWINT Flag is set, the SCL line is held low. This allows the application
software to complete its tasks before allowing the TWI transmission to continue.
212 ATmega164P/324P/644P
8011D–AVR–02/07
The TWINT Flag is set in the following situations:
• After the TWI has transmitted a START/REPEATED START condition.
• After the TWI has transmitted SLA+R/W.
• After the TWI has transmitted an address byte.
• After the TWI has lost arbitration.
• After the TWI has been addressed by own slave address or general call.
• After the TWI has received a data byte.
• After a STOP or REPEATED START has been received while still addressed as a Slave.
• When a bus error has occurred due to an illegal START or STOP condition.
19.6 Using the TWI
The AVR TWI is byte-oriented and interrupt based. Interrupts are issued after all bus events, like
reception of a byte or transmission of a START condition. Because the TWI is interrupt-based,
the application software is free to carry on other operations during a TWI byte transfer. Note that
the TWI Interrupt Enable (TWIE) bit in TWCR together with the Global Interrupt Enable bit in
SREG allow the application to decide whether or not assertion of the TWINT Flag should gener-
ate an interrupt request. If the TWIE bit is cleared, the application must poll the TWINT Flag in
order to detect actions on the TWI bus.
When the TWINT Flag is asserted, the TWI has finished an operation and awaits application
response. In this case, the TWI Status Register (TWSR) contains a value indicating the current
state of the TWI bus. The application software can then decide how the TWI should behave in
the next TWI bus cycle by manipulating the TWCR and TWDR Registers.
Figure 19-10 is a simple example of how the application can interface to the TWI hardware. In
this example, a Master wishes to transmit a single data byte to a Slave. This description is quite
abstract, a more detailed explanation follows later in this section. A simple code example imple-
menting the desired behavior is also presented.
Figure 19-10. Interfacing the Application to the TWI in a Typical Transmission
3. Check TWSR to see if START was 5. Check TWSR to see if SLA+W was
1. Application 7. Check TWSR to see if data was sent
sent. Application loads SLA+W into sent and ACK received.
Application
writes to TWCR to and ACK received.

TWDR, and loads appropriate control Application loads data into TWDR, and
Action
initiate Application loads appropriate control

signals into TWCR, makin sure that loads appropriate control signals into
transmission of signals to send STOP into TWCR,
TWINT is written to one, TWCR, making sure that TWINT is
START making sure that TWINT is written to one
and TWSTA is written to zero. written to one
TWI bus START SLA+W A Data A STOP
Indicates
4. TWINT set.
Hardware
2. TWINT set. 6. TWINT set. TWINT set

Action
Status code indicates

TWI
Status code indicates Status code indicates

SLA+W sent, ACK
START condition sent data sent, ACK received
received
1. The first step in a TWI transmission is to transmit a START condition. This is done by
writing a specific value into TWCR, instructing the TWI hardware to transmit a START
213
8011D–AVR–02/07
condition. Which value to write is described later on. However, it is important that the
TWINT bit is set in the value written. Writing a one to TWINT clears the flag. The TWI will
not start any operation as long as the TWINT bit in TWCR is set. Immediately after the
application has cleared TWINT, the TWI will initiate transmission of the START condition.
2. When the START condition has been transmitted, the TWINT Flag in TWCR is set, and
TWSR is updated with a status code indicating that the START condition has success-
fully been sent.
3. The application software should now examine the value of TWSR, to make sure that the
START condition was successfully transmitted. If TWSR indicates otherwise, the applica-
tion software might take some special action, like calling an error routine. Assuming that
the status code is as expected, the application must load SLA+W into TWDR. Remember
that TWDR is used both for address and data. After TWDR has been loaded with the
desired SLA+W, a specific value must be written to TWCR, instructing the TWI hardware
to transmit the SLA+W present in TWDR. Which value to write is described later on.
However, it is important that the TWINT bit is set in the value written. Writing a one to
TWINT clears the flag. The TWI will not start any operation as long as the TWINT bit in
TWCR is set. Immediately after the application has cleared TWINT, the TWI will initiate
transmission of the address packet.
4. When the address packet has been transmitted, the TWINT Flag in TWCR is set, and
TWSR is updated with a status code indicating that the address packet has successfully
been sent. The status code will also reflect whether a Slave acknowledged the packet or
not.
address packet was successfully transmitted, and that the value of the ACK bit was as
expected. If TWSR indicates otherwise, the application software might take some special
action, like calling an error routine. Assuming that the status code is as expected, the
application must load a data packet into TWDR. Subsequently, a specific value must be
written to TWCR, instructing the TWI hardware to transmit the data packet present in
TWDR. Which value to write is described later on. However, it is important that the
TWINT bit is set in the value written. Writing a one to TWINT clears the flag. The TWI will
not start any operation as long as the TWINT bit in TWCR is set. Immediately after the
application has cleared TWINT, the TWI will initiate transmission of the data packet.
6. When the data packet has been transmitted, the TWINT Flag in TWCR is set, and TWSR
is updated with a status code indicating that the data packet has successfully been sent.
The status code will also reflect whether a Slave acknowledged the packet or not.
data packet was successfully transmitted, and that the value of the ACK bit was as
expected. If TWSR indicates otherwise, the application software might take some special
action, like calling an error routine. Assuming that the status code is as expected, the
application must write a specific value to TWCR, instructing the TWI hardware to transmit
a STOP condition. Which value to write is described later on. However, it is important that
the TWINT bit is set in the value written. Writing a one to TWINT clears the flag. The TWI
will not start any operation as long as the TWINT bit in TWCR is set. Immediately after
the application has cleared TWINT, the TWI will initiate transmission of the STOP condi-
tion. Note that TWINT is NOT set after a STOP condition has been sent.
Even though this example is simple, it shows the principles involved in all TWI transmissions.
These can be summarized as follows:
• When the TWI has finished an operation and expects application response, the TWINT Flag is
set. The SCL line is pulled low until TWINT is cleared.
214 ATmega164P/324P/644P
8011D–AVR–02/07
• When the TWINT Flag is set, the user must update all TWI Registers with the value relevant for
the next TWI bus cycle. As an example, TWDR must be loaded with the value to be transmitted
in the next bus cycle.
• After all TWI Register updates and other pending application software tasks have been
completed, TWCR is written. When writing TWCR, the TWINT bit should be set. Writing a one
to TWINT clears the flag. The TWI will then commence executing whatever operation was
specified by the TWCR setting.
In the following an assembly and C implementation of the example is given. Note that the code
below assumes that several definitions have been made, for example by using include-files.
Assembly Code Example C Example Comments

ldi r16, (1<<TWINT)|(1<<TWSTA)| TWCR = (1<<TWINT)|(1<<TWSTA)|
1 (1<<TWEN) (1<<TWEN) Send START condition
out TWCR, r16
wait1: while (!(TWCR & (1<<TWINT)))
in r16,TWCR ; Wait for TWINT Flag set. This
2 indicates that the START condition
sbrs r16,TWINT has been transmitted
rjmp wait1
in r16,TWSR if ((TWSR & 0xF8) != START) Check value of TWI Status
andi r16, 0xF8 ERROR(); Register. Mask prescaler bits. If
cpi r16, START status different from START go to
ERROR
brne ERROR
3
ldi r16, SLA_W TWDR = SLA_W;
out TWDR, r16 TWCR = (1<<TWINT) | (1<<TWEN); Load SLA_W into TWDR Register.
Clear TWINT bit in TWCR to start
ldi r16, (1<<TWINT) | (1<<TWEN) transmission of address
out TWCR, r16
wait2: while (!(TWCR & (1<<TWINT))) Wait for TWINT Flag set. This
in r16,TWCR ; indicates that the SLA+W has been
4
sbrs r16,TWINT transmitted, and ACK/NACK has
been received.
rjmp wait2
in r16,TWSR if ((TWSR & 0xF8) != Check value of TWI Status
andi r16, 0xF8 MT_SLA_ACK) Register. Mask prescaler bits. If
cpi r16, MT_SLA_ACK ERROR(); status different from MT_SLA_ACK
go to ERROR
brne ERROR
5
ldi r16, DATA TWDR = DATA;
out TWDR, r16 TWCR = (1<<TWINT) | (1<<TWEN); Load DATA into TWDR Register.
Clear TWINT bit in TWCR to start
ldi r16, (1<<TWINT) | (1<<TWEN) transmission of data
out TWCR, r16
215
8011D–AVR–02/07
Assembly Code Example C Example Comments
wait3: while (!(TWCR & (1<<TWINT))) Wait for TWINT Flag set. This
in r16,TWCR ; indicates that the DATA has been
6
sbrs r16,TWINT transmitted, and ACK/NACK has
been received.
rjmp wait3
in r16,TWSR if ((TWSR & 0xF8) != Check value of TWI Status
andi r16, 0xF8 MT_DATA_ACK) Register. Mask prescaler bits. If
cpi r16, MT_DATA_ACK ERROR(); status different from
7 MT_DATA_ACK go to ERROR
brne ERROR
ldi r16, (1<<TWINT)|(1<<TWEN)| TWCR = (1<<TWINT)|(1<<TWEN)|
(1<<TWSTO) (1<<TWSTO); Transmit STOP condition
out TWCR, r16
19.7 Transmission Modes

The TWI can operate in one of four major modes. These are named Master Transmitter (MT),
Master Receiver (MR), Slave Transmitter (ST) and Slave Receiver (SR). Several of these
modes can be used in the same application. As an example, the TWI can use MT mode to write
data into a TWI EEPROM, MR mode to read the data back from the EEPROM. If other masters
are present in the system, some of these might transmit data to the TWI, and then SR mode
would be used. It is the application software that decides which modes are legal.
The following sections describe each of these modes. Possible status codes are described
along with figures detailing data transmission in each of the modes. These figures contain the
following abbreviations:
S: START condition
Rs: REPEATED START condition
R: Read bit (high level at SDA)
W: Write bit (low level at SDA)
A: Acknowledge bit (low level at SDA)
A: Not acknowledge bit (high level at SDA)
Data: 8-bit data byte
P: STOP condition
SLA: Slave Address
In Figure 19-12 on page 219 to Figure 19-18 on page 228, circles are used to indicate that the
TWINT Flag is set. The numbers in the circles show the status code held in TWSR, with the
prescaler bits masked to zero. At these points, actions must be taken by the application to con-
tinue or complete the TWI transfer. The TWI transfer is suspended until the TWINT Flag is
cleared by software.
When the TWINT Flag is set, the status code in TWSR is used to determine the appropriate soft-
ware action. For each status code, the required software action and details of the following serial
transfer are given in Table 19-2 on page 218 to Table 19-5 on page 227. Note that the prescaler
bits are masked to zero in these tables.
216 ATmega164P/324P/644P
8011D–AVR–02/07
19.7.1 Master Transmitter Mode
In the Master Transmitter mode, a number of data bytes are transmitted to a Slave Receiver
(see Figure 19-11 on page 217). In order to enter a Master mode, a START condition must be
transmitted. The format of the following address packet determines whether Master Transmitter
or Master Receiver mode is to be entered. If SLA+W is transmitted, MT mode is entered, if
SLA+R is transmitted, MR mode is entered. All the status codes mentioned in this section
assume that the prescaler bits are zero or are masked to zero.
Figure 19-11. Data Transfer in Master Transmitter Mode

VCC
Device 1 Device 2
MASTER SLAVE Device 3 ........ Device n R1 R2
TRANSMITTER RECEIVER
SDA
SCL
A START condition is sent by writing the following value to TWCR:

TWCR TWINT TWEA TWSTA TWSTO TWWC TWEN – TWIE
value 1 X 1 0 X 1 0 X
TWEN must be set to enable the 2-wire Serial Interface, TWSTA must be written to one to trans-
mit a START condition and TWINT must be written to one to clear the TWINT Flag. The TWI will
then test the 2-wire Serial Bus and generate a START condition as soon as the bus becomes
free. After a START condition has been transmitted, the TWINT Flag is set by hardware, and the
status code in TWSR will be 0x08 (see Table 19-2 on page 218). In order to enter MT mode,
SLA+W must be transmitted. This is done by writing SLA+W to TWDR. Thereafter the TWINT bit
should be cleared (by writing it to one) to continue the transfer. This is accomplished by writing
the following value to TWCR:
value 1 X 0 0 X 1 0 X
When SLA+W have been transmitted and an acknowledgement bit has been received, TWINT is
set again and a number of status codes in TWSR are possible. Possible status codes in Master
mode are 0x18, 0x20, or 0x38. The appropriate action to be taken for each of these status codes
is detailed in Table 19-2 on page 218.
When SLA+W has been successfully transmitted, a data packet should be transmitted. This is
done by writing the data byte to TWDR. TWDR must only be written when TWINT is high. If not,
the access will be discarded, and the Write Collision bit (TWWC) will be set in the TWCR Regis-
ter. After updating TWDR, the TWINT bit should be cleared (by writing it to one) to continue the
transfer. This is accomplished by writing the following value to TWCR:
value 1 X 0 0 X 1 0 X
217
8011D–AVR–02/07
This scheme is repeated until the last byte has been sent and the transfer is ended by generat-
ing a STOP condition or a repeated START condition. A STOP condition is generated by writing
the following value to TWCR:
value 1 X 0 1 X 1 0 X
A REPEATED START condition is generated by writing the following value to TWCR:

value 1 X 1 0 X 1 0 X
After a repeated START condition (state 0x10) the 2-wire Serial Interface can access the same
Slave again, or a new Slave without transmitting a STOP condition. Repeated START enables
the Master to switch between Slaves, Master Transmitter mode and Master Receiver mode with-
out losing control of the bus.
Table 19-2. Status codes for Master Transmitter Mode

Status Code Application Software Response
(TWSR) Status of the 2-wire Serial Bus To/from TWDR To TWCR
Prescaler Bits and 2-wire Serial Interface Hard-
are 0 ware STA STO TWINT TWEA Next Action Taken by TWI Hardware
0x08 A START condition has been Load SLA+W 0 0 1 X SLA+W will be transmitted;
transmitted ACK or NOT ACK will be received
0x10 A repeated START condition Load SLA+W or 0 0 1 X SLA+W will be transmitted;
has been transmitted ACK or NOT ACK will be received
Load SLA+R 0 0 1 X SLA+R will be transmitted;
Logic will switch to Master Receiver mode
0x18 SLA+W has been transmitted; Load data byte or 0 0 1 X Data byte will be transmitted and ACK or NOT ACK will
ACK has been received be received
No TWDR action or 1 0 1 X Repeated START will be transmitted
No TWDR action or 0 1 1 X STOP condition will be transmitted and
TWSTO Flag will be reset
No TWDR action 1 1 1 X STOP condition followed by a START condition will be
transmitted and TWSTO Flag will be reset
0x20 SLA+W has been transmitted; Load data byte or 0 0 1 X Data byte will be transmitted and ACK or NOT ACK will
NOT ACK has been received be received
0x28 Data byte has been transmitted; Load data byte or 0 0 1 X Data byte will be transmitted and ACK or NOT ACK will
ACK has been received be received
0x30 Data byte has been transmitted; Load data byte or 0 0 1 X Data byte will be transmitted and ACK or NOT ACK will
NOT ACK has been received be received
0x38 Arbitration lost in SLA+W or data No TWDR action or 0 0 1 X 2-wire Serial Bus will be released and not addressed
bytes Slave mode entered
No TWDR action 1 0 1 X A START condition will be transmitted when the bus be-
comes free
218 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 19-12. Formats and States in the Master Transmitter Mode
MT
Successfull
transmission S SLA W A DATA A P
to a slave
receiver
$08 $18 $28
Next transfer
started with a RS SLA W
repeated start
condition
$10
Not acknowledge R
received after the A P
slave address
$20
MR
Not acknowledge
received after a data A P
byte
$30
Arbitration lost in slave Other master Other master

address or data byte A or A continues A or A continues
$38 $38
Arbitration lost and Other master

addressed as slave A continues
To corresponding
$68 $78 $B0 states in slave mode
Any number of data bytes

From master to slave DATA A and their associated acknowledge bits
From slave to master This number (contained in TWSR) corresponds

n to a defined state of the Two-Wire Serial Bus. The
prescaler bits are zero or masked to zero
19.7.2 Master Receiver Mode
In the Master Receiver mode, a number of data bytes are received from a Slave Transmitter
(Slave see Figure 19-13 on page 220). In order to enter a Master mode, a START condition
must be transmitted. The format of the following address packet determines whether Master
Transmitter or Master Receiver mode is to be entered. If SLA+W is transmitted, MT mode is
entered, if SLA+R is transmitted, MR mode is entered. All the status codes mentioned in this
section assume that the prescaler bits are zero or are masked to zero.
219
8011D–AVR–02/07
Figure 19-13. Data Transfer in Master Receiver Mode
VCC
Device 1 Device 2
MASTER SLAVE Device 3 ........ Device n R1 R2
RECEIVER TRANSMITTER
SDA
SCL
A START condition is sent by writing the following value to TWCR:

value 1 X 1 0 X 1 0 X
TWEN must be written to one to enable the 2-wire Serial Interface, TWSTA must be written to
one to transmit a START condition and TWINT must be set to clear the TWINT Flag. The TWI
will then test the 2-wire Serial Bus and generate a START condition as soon as the bus
becomes free. After a START condition has been transmitted, the TWINT Flag is set by hard-
ware, and the status code in TWSR will be 0x08 (See Table 19-2 on page 218). In order to enter
MR mode, SLA+R must be transmitted. This is done by writing SLA+R to TWDR. Thereafter the
TWINT bit should be cleared (by writing it to one) to continue the transfer. This is accomplished
by writing the following value to TWCR:
value 1 X 0 0 X 1 0 X
When SLA+R have been transmitted and an acknowledgement bit has been received, TWINT is
set again and a number of status codes in TWSR are possible. Possible status codes in Master
mode are 0x38, 0x40, or 0x48. The appropriate action to be taken for each of these status codes
is detailed in Table 19-3 on page 221. Received data can be read from the TWDR Register
when the TWINT Flag is set high by hardware. This scheme is repeated until the last byte has
been received. After the last byte has been received, the MR should inform the ST by sending a
NACK after the last received data byte. The transfer is ended by generating a STOP condition or
a repeated START condition. A STOP condition is generated by writing the following value to
TWCR:
value 1 X 0 1 X 1 0 X
A REPEATED START condition is generated by writing the following value to TWCR:

value 1 X 1 0 X 1 0 X
After a repeated START condition (state 0x10) the 2-wire Serial Interface can access the same
Slave again, or a new Slave without transmitting a STOP condition. Repeated START enables
220 ATmega164P/324P/644P
8011D–AVR–02/07
the Master to switch between Slaves, Master Transmitter mode and Master Receiver mode with-
out losing control over the bus.
Table 19-3. Status codes for Master Receiver Mode

(TWSR) Status of the 2-wire Serial Bus To TWCR
Prescaler Bits and 2-wire Serial Interface Hard- To/from TWDR
0x08 A START condition has been Load SLA+R 0 0 1 X SLA+R will be transmitted
transmitted ACK or NOT ACK will be received
0x10 A repeated START condition Load SLA+R or 0 0 1 X SLA+R will be transmitted
has been transmitted ACK or NOT ACK will be received
Load SLA+W 0 0 1 X SLA+W will be transmitted
Logic will switch to Master Transmitter mode
0x38 Arbitration lost in SLA+R or NOT No TWDR action or 0 0 1 X 2-wire Serial Bus will be released and not addressed
ACK bit Slave mode will be entered
No TWDR action 1 0 1 X A START condition will be transmitted when the bus
becomes free
0x40 SLA+R has been transmitted; No TWDR action or 0 0 1 0 Data byte will be received and NOT ACK will be
ACK has been received returned
No TWDR action 0 0 1 1 Data byte will be received and ACK will be returned
0x48 SLA+R has been transmitted; No TWDR action or 1 0 1 X Repeated START will be transmitted
NOT ACK has been received No TWDR action or 0 1 1 X STOP condition will be transmitted and TWSTO Flag will
be reset
0x50 Data byte has been received; Read data byte or 0 0 1 0 Data byte will be received and NOT ACK will be
ACK has been returned returned
Read data byte 0 0 1 1 Data byte will be received and ACK will be returned
0x58 Data byte has been received; Read data byte or 1 0 1 X Repeated START will be transmitted
NOT ACK has been returned Read data byte or 0 1 1 X STOP condition will be transmitted and TWSTO Flag will
be reset
Read data byte 1 1 1 X STOP condition followed by a START condition will be
221
8011D–AVR–02/07
Figure 19-14. Formats and States in the Master Receiver Mode
MR
Successfull
reception S SLA R A DATA A DATA A P
from a slave
receiver
$08 $40 $50 $58
Next transfer
started with a RS SLA R
repeated start
condition
$10
Not acknowledge W
received after the A P
slave address
$48
MT
Arbitration lost in slave Other master Other master
address or data byte A or A continues A continues
$38 $38
Arbitration lost and Other master

addressed as slave A continues
To corresponding
$68 $78 $B0 states in slave mode


19.7.3 Slave Receiver Mode
In the Slave Receiver mode, a number of data bytes are received from a Master Transmitter
(see Figure 19-15). All the status codes mentioned in this section assume that the prescaler bits
are zero or are masked to zero.
Figure 19-15. Data transfer in Slave Receiver mode

VCC
Device 1 Device 2
SLAVE MASTER Device 3 ........ Device n R1 R2
RECEIVER TRANSMITTER
SDA
SCL
To initiate the Slave Receiver mode, TWAR and TWCR must be initialized as follows:
TWAR TWA6 TWA5 TWA4 TWA3 TWA2 TWA1 TWA0 TWGCE
value Device’s Own Slave Address
222 ATmega164P/324P/644P
8011D–AVR–02/07
The upper 7 bits are the address to which the 2-wire Serial Interface will respond when
addressed by a Master. If the LSB is set, the TWI will respond to the general call address (0x00),
otherwise it will ignore the general call address.
value 0 1 0 0 0 1 0 X
TWEN must be written to one to enable the TWI. The TWEA bit must be written to one to enable
the acknowledgement of the device’s own slave address or the general call address. TWSTA
and TWSTO must be written to zero.
When TWAR and TWCR have been initialized, the TWI waits until it is addressed by its own
slave address (or the general call address if enabled) followed by the data direction bit. If the
direction bit is “0” (write), the TWI will operate in SR mode, otherwise ST mode is entered. After
its own slave address and the write bit have been received, the TWINT Flag is set and a valid
status code can be read from TWSR. The status code is used to determine the appropriate soft-
ware action. The appropriate action to be taken for each status code is detailed in Table 19-4 on
page 224. The Slave Receiver mode may also be entered if arbitration is lost while the TWI is in
the Master mode (see states 0x68 and 0x78).
If the TWEA bit is reset during a transfer, the TWI will return a “Not Acknowledge” (“1”) to SDA
after the next received data byte. This can be used to indicate that the Slave is not able to
receive any more bytes. While TWEA is zero, the TWI does not acknowledge its own slave
address. However, the 2-wire Serial Bus is still monitored and address recognition may resume
at any time by setting TWEA. This implies that the TWEA bit may be used to temporarily isolate
the TWI from the 2-wire Serial Bus.
In all sleep modes other than Idle mode, the clock system to the TWI is turned off. If the TWEA
bit is set, the interface can still acknowledge its own slave address or the general call address by
using the 2-wire Serial Bus clock as a clock source. The part will then wake up from sleep and
the TWI will hold the SCL clock low during the wake up and until the TWINT Flag is cleared (by
writing it to one). Further data reception will be carried out as normal, with the AVR clocks run-
ning as normal. Observe that if the AVR is set up with a long start-up time, the SCL line may be
held low for a long time, blocking other data transmissions.
Note that the 2-wire Serial Interface Data Register – TWDR does not reflect the last byte present
on the bus when waking up from these Sleep modes.
223
8011D–AVR–02/07
Table 19-4. Status Codes for Slave Receiver Mode
(TWSR) Status of the 2-wire Serial Bus and To TWCR
Prescaler Bits 2-wire Serial Interface Hardware To/from TWDR
are 0 STA STO TWINT TWEA Next Action Taken by TWI Hardware
0x60 Own SLA+W has been received; No TWDR action or X 0 1 0 Data byte will be received and NOT ACK will be
ACK has been returned returned
No TWDR action X 0 1 1 Data byte will be received and ACK will be returned
0x68 Arbitration lost in SLA+R/W as No TWDR action or X 0 1 0 Data byte will be received and NOT ACK will be
Master; own SLA+W has been returned
received; ACK has been returned No TWDR action X 0 1 1 Data byte will be received and ACK will be returned
0x70 General call address has been No TWDR action or X 0 1 0 Data byte will be received and NOT ACK will be
received; ACK has been returned returned
No TWDR action X 0 1 1 Data byte will be received and ACK will be returned
0x78 Arbitration lost in SLA+R/W as No TWDR action or X 0 1 0 Data byte will be received and NOT ACK will be
Master; General call address has returned
been received; ACK has been No TWDR action X 0 1 1 Data byte will be received and ACK will be returned
returned
0x80 Previously addressed with own Read data byte or X 0 1 0 Data byte will be received and NOT ACK will be
SLA+W; data has been received; returned
ACK has been returned Read data byte X 0 1 1 Data byte will be received and ACK will be returned
0x88 Previously addressed with own Read data byte or 0 0 1 0 Switched to the not addressed Slave mode;
SLA+W; data has been received; no recognition of own SLA or GCA
NOT ACK has been returned Read data byte or 0 0 1 1 Switched to the not addressed Slave mode;
own SLA will be recognized;
GCA will be recognized if TWGCE = “1”
Read data byte or 1 0 1 0 Switched to the not addressed Slave mode;
no recognition of own SLA or GCA;
a START condition will be transmitted when the bus
becomes free
Read data byte 1 0 1 1 Switched to the not addressed Slave mode;
GCA will be recognized if TWGCE = “1”;
becomes free
0x90 Previously addressed with Read data byte or X 0 1 0 Data byte will be received and NOT ACK will be
general call; data has been re- returned
ceived; ACK has been returned Read data byte X 0 1 1 Data byte will be received and ACK will be returned
0x98 Previously addressed with Read data byte or 0 0 1 0 Switched to the not addressed Slave mode;
general call; data has been no recognition of own SLA or GCA
received; NOT ACK has been Read data byte or 0 0 1 1 Switched to the not addressed Slave mode;
returned own SLA will be recognized;
Read data byte or 1 0 1 0 Switched to the not addressed Slave mode;
becomes free
Read data byte 1 0 1 1 Switched to the not addressed Slave mode;
becomes free
0xA0 A STOP condition or repeated No action 0 0 1 0 Switched to the not addressed Slave mode;
START condition has been no recognition of own SLA or GCA
received while still addressed as 0 0 1 1 Switched to the not addressed Slave mode;
Slave own SLA will be recognized;
1 0 1 0 Switched to the not addressed Slave mode;
becomes free
1 0 1 1 Switched to the not addressed Slave mode;
becomes free
224 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 19-16. Formats and States in the Slave Receiver Mode
Reception of the own
slave address and one or S SLA W A DATA A DATA A P or S
more data bytes. All are
acknowledged
$60 $80 $80 $A0
Last data byte received

is not acknowledged A P or S
$88
Arbitration lost as master

and addressed as slave A
$68
Reception of the general call

address and one or more data General Call A DATA A DATA A P or S
bytes
$70 $90 $90 $A0
Last data byte received is

not acknowledged A P or S
$98
Arbitration lost as master and

addressed as slave by general call A
$78


19.7.4 Slave Transmitter Mode
In the Slave Transmitter mode, a number of data bytes are transmitted to a Master Receiver
(see Figure 19-17). All the status codes mentioned in this section assume that the prescaler bits
are zero or are masked to zero.
Figure 19-17. Data Transfer in Slave Transmitter Mode

VCC
Device 1 Device 2
SLAVE MASTER Device 3 ........ Device n R1 R2
TRANSMITTER RECEIVER
SDA
SCL
225
8011D–AVR–02/07
To initiate the Slave Transmitter mode, TWAR and TWCR must be initialized as follows:
TWAR TWA6 TWA5 TWA4 TWA3 TWA2 TWA1 TWA0 TWGCE
value Device’s Own Slave Address
The upper seven bits are the address to which the 2-wire Serial Interface will respond when
addressed by a Master. If the LSB is set, the TWI will respond to the general call address (0x00),
otherwise it will ignore the general call address.
value 0 1 0 0 0 1 0 X
TWEN must be written to one to enable the TWI. The TWEA bit must be written to one to enable
the acknowledgement of the device’s own slave address or the general call address. TWSTA
and TWSTO must be written to zero.
When TWAR and TWCR have been initialized, the TWI waits until it is addressed by its own
slave address (or the general call address if enabled) followed by the data direction bit. If the
direction bit is “1” (read), the TWI will operate in ST mode, otherwise SR mode is entered. After
its own slave address and the write bit have been received, the TWINT Flag is set and a valid
status code can be read from TWSR. The status code is used to determine the appropriate soft-
ware action. The appropriate action to be taken for each status code is detailed in Table 19-5 on
page 227. The Slave Transmitter mode may also be entered if arbitration is lost while the TWI is
in the Master mode (see state 0xB0).
If the TWEA bit is written to zero during a transfer, the TWI will transmit the last byte of the trans-
fer. State 0xC0 or state 0xC8 will be entered, depending on whether the Master Receiver
transmits a NACK or ACK after the final byte. The TWI is switched to the not addressed Slave
mode, and will ignore the Master if it continues the transfer. Thus the Master Receiver receives
all “1” as serial data. State 0xC8 is entered if the Master demands additional data bytes (by
transmitting ACK), even though the Slave has transmitted the last byte (TWEA zero and expect-
ing NACK from the Master).
While TWEA is zero, the TWI does not respond to its own slave address. However, the 2-wire
Serial Bus is still monitored and address recognition may resume at any time by setting TWEA.
This implies that the TWEA bit may be used to temporarily isolate the TWI from the 2-wire Serial
Bus.
In all sleep modes other than Idle mode, the clock system to the TWI is turned off. If the TWEA
bit is set, the interface can still acknowledge its own slave address or the general call address by
using the 2-wire Serial Bus clock as a clock source. The part will then wake up from sleep and
the TWI will hold the SCL clock will low during the wake up and until the TWINT Flag is cleared
(by writing it to one). Further data transmission will be carried out as normal, with the AVR clocks
running as normal. Observe that if the AVR is set up with a long start-up time, the SCL line may
be held low for a long time, blocking other data transmissions.
Note that the 2-wire Serial Interface Data Register – TWDR does not reflect the last byte present
on the bus when waking up from these sleep modes.
226 ATmega164P/324P/644P
8011D–AVR–02/07
Table 19-5. Status Codes for Slave Transmitter Mode

(TWSR) Status of the 2-wire Serial Bus and To TWCR
Prescaler Bits 2-wire Serial Interface Hardware To/from TWDR
are 0 STA STO TWINT TWEA Next Action Taken by TWI Hardware
0xA8 Own SLA+R has been received; Load data byte or X 0 1 0 Last data byte will be transmitted and NOT ACK should
ACK has been returned be received
Load data byte X 0 1 1 Data byte will be transmitted and ACK should be re-
ceived
0xB0 Arbitration lost in SLA+R/W as Load data byte or X 0 1 0 Last data byte will be transmitted and NOT ACK should
Master; own SLA+R has been be received
received; ACK has been returned Load data byte X 0 1 1 Data byte will be transmitted and ACK should be re-
ceived
0xB8 Data byte in TWDR has been Load data byte or X 0 1 0 Last data byte will be transmitted and NOT ACK should
transmitted; ACK has been be received
received Load data byte X 0 1 1 Data byte will be transmitted and ACK should be re-
ceived
0xC0 Data byte in TWDR has been No TWDR action or 0 0 1 0 Switched to the not addressed Slave mode;
transmitted; NOT ACK has been no recognition of own SLA or GCA
received No TWDR action or 0 0 1 1 Switched to the not addressed Slave mode;
No TWDR action or 1 0 1 0 Switched to the not addressed Slave mode;
becomes free
No TWDR action 1 0 1 1 Switched to the not addressed Slave mode;
becomes free
0xC8 Last data byte in TWDR has been No TWDR action or 0 0 1 0 Switched to the not addressed Slave mode;
transmitted (TWEA = “0”); ACK no recognition of own SLA or GCA
has been received No TWDR action or 0 0 1 1 Switched to the not addressed Slave mode;
No TWDR action or 1 0 1 0 Switched to the not addressed Slave mode;
becomes free
No TWDR action 1 0 1 1 Switched to the not addressed Slave mode;
becomes free
227
8011D–AVR–02/07
Figure 19-18. Formats and States in the Slave Transmitter Mode
Reception of the own
slave address and one or S SLA R A DATA A DATA A P or S
more data bytes
$A8 $B8 $C0
Arbitration lost as master

and addressed as slave A
$B0
Last data byte transmitted.

Switched to not addressed A All 1's P or S
slave (TWEA = '0')
$C8


19.7.5 Miscellaneous States
There are two status codes that do not correspond to a defined TWI state, see Table 19-6.
Status 0xF8 indicates that no relevant information is available because the TWINT Flag is not
set. This occurs between other states, and when the TWI is not involved in a serial transfer.
Status 0x00 indicates that a bus error has occurred during a 2-wire Serial Bus transfer. A bus
error occurs when a START or STOP condition occurs at an illegal position in the format frame.
Examples of such illegal positions are during the serial transfer of an address byte, a data byte,
or an acknowledge bit. When a bus error occurs, TWINT is set. To recover from a bus error, the
TWSTO Flag must set and TWINT must be cleared by writing a logic one to it. This causes the
TWI to enter the not addressed Slave mode and to clear the TWSTO Flag (no other bits in
TWCR are affected). The SDA and SCL lines are released, and no STOP condition is
transmitted.
Table 19-6. Miscellaneous States

(TWSR) Status of the 2-wire Serial Bus To TWCR
Prescaler Bits and 2-wire Serial Interface Hard- To/from TWDR
0xF8 No relevant state information No TWDR action No TWCR action Wait or proceed current transfer
available; TWINT = “0”
0x00 Bus error due to an illegal No TWDR action 0 1 1 X Only the internal hardware is affected, no STOP condi-
START or STOP condition tion is sent on the bus. In all cases, the bus is released
and TWSTO is cleared.
19.7.6 Combining Several TWI Modes
In some cases, several TWI modes must be combined in order to complete the desired action.
Consider for example reading data from a serial EEPROM. Typically, such a transfer involves
the following steps:
1. The transfer must be initiated.
2. The EEPROM must be instructed what location should be read.
3. The reading must be performed.
4. The transfer must be finished.
228 ATmega164P/324P/644P
8011D–AVR–02/07
Note that data is transmitted both from Master to Slave and vice versa. The Master must instruct
the Slave what location it wants to read, requiring the use of the MT mode. Subsequently, data
must be read from the Slave, implying the use of the MR mode. Thus, the transfer direction must
be changed. The Master must keep control of the bus during all these steps, and the steps
should be carried out as an atomical operation. If this principle is violated in a multimaster sys-
tem, another Master can alter the data pointer in the EEPROM between steps 2 and 3, and the
Master will read the wrong data location. Such a change in transfer direction is accomplished by
transmitting a REPEATED START between the transmission of the address byte and reception
of the data. After a REPEATED START, the Master keeps ownership of the bus. The following
figure shows the flow in this transfer.
Figure 19-19. Combining Several TWI Modes to Access a Serial EEPROM

Master Transmitter Master Receiver
S SLA+W A ADDRESS A Rs SLA+R A DATA A P
S = START Rs = REPEATED START P = STOP
Transmitted from master to slave Transmitted from slave to master
19.8 Multi-master Systems and Arbitration

If multiple masters are connected to the same bus, transmissions may be initiated simulta-
neously by one or more of them. The TWI standard ensures that such situations are handled in
such a way that one of the masters will be allowed to proceed with the transfer, and that no data
will be lost in the process. An example of an arbitration situation is depicted below, where two
masters are trying to transmit data to a Slave Receiver.
Figure 19-20. An Arbitration Example

VCC
Device 1 Device 2 Device 3

MASTER MASTER SLAVE ........ Device n R1 R2
TRANSMITTER TRANSMITTER RECEIVER
SDA
SCL
Several different scenarios may arise during arbitration, as described below:

• Two or more masters are performing identical communication with the same Slave. In this
case, neither the Slave nor any of the masters will know about the bus contention.
• Two or more masters are accessing the same Slave with different data or direction bit. In this
case, arbitration will occur, either in the READ/WRITE bit or in the data bits. The masters trying
to output a one on SDA while another Master outputs a zero will lose the arbitration. Losing
masters will switch to not addressed Slave mode or wait until the bus is free and transmit a new
START condition, depending on application software action.
229
8011D–AVR–02/07
• Two or more masters are accessing different slaves. In this case, arbitration will occur in the
SLA bits. Masters trying to output a one on SDA while another Master outputs a zero will lose
the arbitration. Masters losing arbitration in SLA will switch to Slave mode to check if they are
being addressed by the winning Master. If addressed, they will switch to SR or ST mode,
depending on the value of the READ/WRITE bit. If they are not being addressed, they will
switch to not addressed Slave mode or wait until the bus is free and transmit a new START
condition, depending on application software action.
This is summarized in Figure 19-21. Possible status values are given in circles.
Figure 19-21. Possible Status Codes Caused by Arbitration

START SLA Data STOP
Arbitration lost in SLA Arbitration lost in Data
Own No 38 TWI bus will be released and not addressed slave mode will be entered
Address / General Call
A START condition will be transmitted when the bus becomes free
received
Yes
Direction
Write 68/78 Data byte will be received and NOT ACK will be returned
Data byte will be received and ACK will be returned
Read Last data byte will be transmitted and NOT ACK should be received
B0 Data byte will be transmitted and ACK should be received

19.9.1 TWBR – TWI Bit Rate Register
Bit 7 6 5 4 3 2 1 0
(0xB8) TWBR7 TWBR6 TWBR5 TWBR4 TWBR3 TWBR2 TWBR1 TWBR0 TWBR
• Bits 7:0 – TWI Bit Rate Register

TWBR selects the division factor for the bit rate generator. The bit rate generator is a frequency
divider which generates the SCL clock frequency in the Master modes. See ”Bit Rate Generator
Unit” on page 211 for calculating bit rates.
19.9.2 TWCR – TWI Control Register
Bit 7 6 5 4 3 2 1 0
(0xBC) TWINT TWEA TWSTA TWSTO TWWC TWEN – TWIE TWCR
Read/Write R/W R/W R/W R/W R R/W R R/W
The TWCR is used to control the operation of the TWI. It is used to enable the TWI, to initiate a
Master access by applying a START condition to the bus, to generate a Receiver acknowledge,
to generate a stop condition, and to control halting of the bus while the data to be written to the
230 ATmega164P/324P/644P
8011D–AVR–02/07
bus are written to the TWDR. It also indicates a write collision if data is attempted written to
TWDR while the register is inaccessible.
• Bit 7 – TWINT: TWI Interrupt Flag

This bit is set by hardware when the TWI has finished its current job and expects application
software response. If the I-bit in SREG and TWIE in TWCR are set, the MCU will jump to the
TWI Interrupt Vector. While the TWINT Flag is set, the SCL low period is stretched. The TWINT
Flag must be cleared by software by writing a logic one to it. Note that this flag is not automati-
cally cleared by hardware when executing the interrupt routine. Also note that clearing this flag
starts the operation of the TWI, so all accesses to the TWI Address Register (TWAR), TWI Sta-
tus Register (TWSR), and TWI Data Register (TWDR) must be complete before clearing this
flag.
• Bit 6 – TWEA: TWI Enable Acknowledge Bit

The TWEA bit controls the generation of the acknowledge pulse. If the TWEA bit is written to
one, the ACK pulse is generated on the TWI bus if the following conditions are met:
1. The device’s own slave address has been received.
2. A general call has been received, while the TWGCE bit in the TWAR is set.
3. A data byte has been received in Master Receiver or Slave Receiver mode.
By writing the TWEA bit to zero, the device can be virtually disconnected from the 2-wire Serial
Bus temporarily. Address recognition can then be resumed by writing the TWEA bit to one
again.
• Bit 5 – TWSTA: TWI START Condition Bit

The application writes the TWSTA bit to one when it desires to become a Master on the 2-wire
Serial Bus. The TWI hardware checks if the bus is available, and generates a START condition
on the bus if it is free. However, if the bus is not free, the TWI waits until a STOP condition is
detected, and then generates a new START condition to claim the bus Master status. TWSTA
must be cleared by software when the START condition has been transmitted.
• Bit 4 – TWSTO: TWI STOP Condition Bit

Writing the TWSTO bit to one in Master mode will generate a STOP condition on the 2-wire
Serial Bus. When the STOP condition is executed on the bus, the TWSTO bit is cleared auto-
matically. In Slave mode, setting the TWSTO bit can be used to recover from an error condition.
This will not generate a STOP condition, but the TWI returns to a well-defined unaddressed
Slave mode and releases the SCL and SDA lines to a high impedance state.
• Bit 3 – TWWC: TWI Write Collision Flag

The TWWC bit is set when attempting to write to the TWI Data Register – TWDR when TWINT is
low. This flag is cleared by writing the TWDR Register when TWINT is high.
• Bit 2 – TWEN: TWI Enable Bit

The TWEN bit enables TWI operation and activates the TWI interface. When TWEN is written to
one, the TWI takes control over the I/O pins connected to the SCL and SDA pins, enabling the
slew-rate limiters and spike filters. If this bit is written to zero, the TWI is switched off and all TWI
transmissions are terminated, regardless of any ongoing operation.
• Bit 1 – Res: Reserved Bit

This bit is a reserved bit and will always read as zero.
231
8011D–AVR–02/07
• Bit 0 – TWIE: TWI Interrupt Enable
When this bit is written to one, and the I-bit in SREG is set, the TWI interrupt request will be acti-
vated for as long as the TWINT Flag is high.
19.9.3 TWSR – TWI Status Register
Bit 7 6 5 4 3 2 1 0
(0xB9) TWS7 TWS6 TWS5 TWS4 TWS3 – TWPS1 TWPS0 TWSR
Read/Write R R R R R R R/W R/W
• Bits 7:3 – TWS: TWI Status

These 5 bits reflect the status of the TWI logic and the 2-wire Serial Bus. The different status
codes are described later in this section. Note that the value read from TWSR contains both the
5-bit status value and the 2-bit prescaler value. The application designer should mask the pres-
caler bits to zero when checking the Status bits. This makes status checking independent of
prescaler setting. This approach is used in this datasheet, unless otherwise noted.

This bit is reserved and will always read as zero.
• Bits 1:0 – TWPS: TWI Prescaler Bits

These bits can be read and written, and control the bit rate prescaler.
Table 19-7. TWI Bit Rate Prescaler

TWPS1 TWPS0 Prescaler Value
0 0 1
0 1 4
1 0 16
1 1 64
To calculate bit rates, see ”Bit Rate Generator Unit” on page 211. The value of TWPS1..0 is
used in the equation.
19.9.4 TWDR – TWI Data Register
Bit 7 6 5 4 3 2 1 0
(0xBB) TWD7 TWD6 TWD5 TWD4 TWD3 TWD2 TWD1 TWD0 TWDR
In Transmit mode, TWDR contains the next byte to be transmitted. In Receive mode, the TWDR
contains the last byte received. It is writable while the TWI is not in the process of shifting a byte.
This occurs when the TWI Interrupt Flag (TWINT) is set by hardware. Note that the Data Regis-
ter cannot be initialized by the user before the first interrupt occurs. The data in TWDR remains
stable as long as TWINT is set. While data is shifted out, data on the bus is simultaneously
shifted in. TWDR always contains the last byte present on the bus, except after a wake up from
a sleep mode by the TWI interrupt. In this case, the contents of TWDR is undefined. In the case
232 ATmega164P/324P/644P
8011D–AVR–02/07
of a lost bus arbitration, no data is lost in the transition from Master to Slave. Handling of the
ACK bit is controlled automatically by the TWI logic, the CPU cannot access the ACK bit directly.
• Bits 7:0 – TWD: TWI Data Register

These eight bits constitute the next data byte to be transmitted, or the latest data byte received
on the 2-wire Serial Bus.
19.9.5 TWAR – TWI (Slave) Address Register
Bit 7 6 5 4 3 2 1 0
(0xBA) TWA6 TWA5 TWA4 TWA3 TWA2 TWA1 TWA0 TWGCE TWAR
The TWAR should be loaded with the 7-bit Slave address (in the seven most significant bits of
TWAR) to which the TWI will respond when programmed as a Slave Transmitter or Receiver,
and not needed in the Master modes. In multimaster systems, TWAR must be set in masters
which can be addressed as Slaves by other Masters.
The LSB of TWAR is used to enable recognition of the general call address (0x00). There is an
associated address comparator that looks for the slave address (or general call address if
enabled) in the received serial address. If a match is found, an interrupt request is generated.
• Bits 7:1 – TWA: TWI (Slave) Address Register

These seven bits constitute the slave address of the TWI unit.
• Bit 0 – TWGCE: TWI General Call Recognition Enable Bit

If set, this bit enables the recognition of a General Call given over the 2-wire Serial Bus.
19.9.6 TWAMR – TWI (Slave) Address Mask Register
Bit 7 6 5 4 3 2 1 0
(0xBD) TWAM[6:0] – TWAMR
Read/Write R/W R/W R/W R/W R/W R/W R/W R
• Bits 7:1 – TWAM: TWI Address Mask

The TWAMR can be loaded with a 7-bit Slave Address mask. Each of the bits in TWAMR can
mask (disable) the corresponding address bit in the TWI Address Register (TWAR). If the mask
bit is set to one then the address match logic ignores the compare between the incoming
address bit and the corresponding bit in TWAR. Figure 19-22 shows the address match logic in
detail.
233
8011D–AVR–02/07
Figure 19-22. TWI Address Match Logic, Block Diagram
TWAR0
Address
Address Match
Bit 0
TWAMR0
Address Bit Comparator 0
Address Bit Comparator 6..1

This bit is reserved and will always read as zero.
234 ATmega164P/324P/644P
8011D–AVR–02/07
20. AC - Analog Comparator

20.1 Overview
The Analog Comparator compares the input values on the positive pin AIN0 and negative pin
AIN1. When the voltage on the positive pin AIN0 is higher than the voltage on the negative pin
AIN1, the Analog Comparator output, ACO, is set. The comparator’s output can be set to trigger
the Timer/Counter1 Input Capture function. In addition, the comparator can trigger a separate
interrupt, exclusive to the Analog Comparator. The user can select Interrupt triggering on com-
parator output rise, fall or toggle. A block diagram of the comparator and its surrounding logic is
shown in Figure 20-1.
The Power Reduction ADC bit, PRADC, in ”PRR – Power Reduction Register” on page 48 must
be disabled by writing a logical zero to be able to use the ADC input MUX.
Figure 20-1. Analog Comparator Block Diagram(2)

BANDGAP
REFERENCE
ACBG
ACME
ADEN
ADC MULTIPLEXER
OUTPUT (1)
Notes: 1. See Table 20-1 on page 235.

2. Refer to Figure 1-1 on page 2 and Table 12-5 on page 80 for Analog Comparator pin
placement.
20.2 Analog Comparator Multiplexed Input

It is possible to select any of the ADC7..0 pins to replace the negative input to the Analog Com-
parator. The ADC multiplexer is used to select this input, and consequently, the ADC must be
switched off to utilize this feature. If the Analog Comparator Multiplexer Enable bit (ACME in
ADCSRB) is set and the ADC is switched off (ADEN in ADCSRA is zero), MUX2..0 in ADMUX
select the input pin to replace the negative input to the Analog Comparator, as shown in Table
20-1 on page 235. If ACME is cleared or ADEN is set, AIN1 is applied to the negative input to the
Analog Comparator.
Table 20-1. Analog Comparator Mulitiplexed Input

ACME ADEN MUX2..0 Analog Comparator Negative Input
0 x xxx AIN1
1 1 xxx AIN1
1 0 000 ADC0
235
8011D–AVR–02/07
Table 20-1. Analog Comparator Mulitiplexed Input
ACME ADEN MUX2..0 Analog Comparator Negative Input
1 0 001 ADC1
1 0 010 ADC2
1 0 011 ADC3
1 0 100 ADC4
1 0 101 ADC5
1 0 110 ADC6
1 0 111 ADC7

20.3.1 ADCSRB – ADC Control and Status Register B
Bit 7 6 5 4 3 2 1 0
(0x7B) – ACME – – MUX5 ADTS2 ADTS1 ADTS0 ADCSRB
Read/Write R R/W R R R/W R/W R/W R/W
• Bit 6 – ACME: Analog Comparator Multiplexer Enable

When this bit is written logic one and the ADC is switched off (ADEN in ADCSRA is zero), the
ADC multiplexer selects the negative input to the Analog Comparator. When this bit is written
logic zero, AIN1 is applied to the negative input of the Analog Comparator. For a detailed
description of this bit, see ”Analog Comparator Multiplexed Input” on page 235.
20.3.2 ACSR – Analog Comparator Control and Status Register
Bit 7 6 5 4 3 2 1 0
0x30 (0x50) ACD ACBG ACO ACI ACIE ACIC ACIS1 ACIS0 ACSR
Read/Write R/W R/W R R/W R/W R/W R/W R/W
Initial Value 0 0 N/A 0 0 0 0 0
• Bit 7 – ACD: Analog Comparator Disable

When this bit is written logic one, the power to the Analog Comparator is switched off. This bit
can be set at any time to turn off the Analog Comparator. This will reduce power consumption in
Active and Idle mode. When changing the ACD bit, the Analog Comparator Interrupt must be
disabled by clearing the ACIE bit in ACSR. Otherwise an interrupt can occur when the bit is
changed.
• Bit 6 – ACBG: Analog Comparator Bandgap Select

When this bit is set, a fixed bandgap reference voltage replaces the positive input to the Analog
Comparator. When this bit is cleared, AIN0 is applied to the positive input of the Analog Compar-
ator. When bandgap reference is used as input to the Analog Comparator, it will take a certain
time for the voltage to stabilize. If not stabilized, the first conversion may give wrong value. See
”Internal Voltage Reference” on page 54.
• Bit 5 – ACO: Analog Comparator Output

The output of the Analog Comparator is synchronized and then directly connected to ACO. The
synchronization introduces a delay of 1 - 2 clock cycles.
236 ATmega164P/324P/644P
8011D–AVR–02/07
• Bit 4 – ACI: Analog Comparator Interrupt Flag
This bit is set by hardware when a comparator output event triggers the interrupt mode defined
by ACIS1 and ACIS0. The Analog Comparator interrupt routine is executed if the ACIE bit is set
and the I-bit in SREG is set. ACI is cleared by hardware when executing the corresponding inter-
rupt handling vector. Alternatively, ACI is cleared by writing a logic one to the flag.
• Bit 3 – ACIE: Analog Comparator Interrupt Enable

When the ACIE bit is written logic one and the I-bit in the Status Register is set, the Analog Com-
parator interrupt is activated. When written logic zero, the interrupt is disabled.
• Bit 2 – ACIC: Analog Comparator Input Capture Enable

When written logic one, this bit enables the input capture function in Timer/Counter1 to be trig-
gered by the Analog Comparator. The comparator output is in this case directly connected to the
input capture front-end logic, making the comparator utilize the noise canceler and edge select
features of the Timer/Counter1 Input Capture interrupt. When written logic zero, no connection
between the Analog Comparator and the input capture function exists. To make the comparator
trigger the Timer/Counter1 Input Capture interrupt, the ICIE1 bit in the Timer Interrupt Mask
Register (TIMSK1) must be set.
• Bits 1:0 – ACIS1:ACIS0: Analog Comparator Interrupt Mode Select

These bits determine which comparator events that trigger the Analog Comparator interrupt. The
different settings are shown in Table 20-2.
Table 20-2. ACIS1/ACIS0 Settings

ACIS1 ACIS0 Interrupt Mode
0 0 Comparator Interrupt on Output Toggle.
0 1 Reserved
1 0 Comparator Interrupt on Falling Output Edge.
1 1 Comparator Interrupt on Rising Output Edge.
When changing the ACIS1/ACIS0 bits, the Analog Comparator Interrupt must be disabled by
clearing its Interrupt Enable bit in the ACSR Register. Otherwise an interrupt can occur when the
bits are changed.
20.3.3 DIDR1 – Digital Input Disable Register 1
Bit 7 6 5 4 3 2 1 0
(0x7F) – – – – – – AIN1D AIN0D DIDR1
Read/Write R R R R R R R/W R/W
• Bit 1:0 – AIN1D:AIN0D: AIN1:AIN0 Digital Input Disable

When this bit is written logic one, the digital input buffer on the AIN1/0 pin is disabled. The corre-
sponding PIN Register bit will always read as zero when this bit is set. When an analog signal is
applied to the AIN1/0 pin and the digital input from this pin is not needed, this bit should be writ-
ten logic one to reduce power consumption in the digital input buffer.
237
8011D–AVR–02/07
21. ADC - Analog-to-digital Converter
21.1 Features
• 10-bit Resolution
• 0.5 LSB Integral Non-linearity
• ±2 LSB Absolute Accuracy
• 13 - 260 µs Conversion Time
• Up to 15 kSPS at Maximum Resolution
• 8 Multiplexed Single Ended Input Channels
• Differential mode with selectable gain at 1x, 10x or 200x
• Optional Left adjustment for ADC Result Readout
• 0 - VCC ADC Input Voltage Range
• 2.7 - VCC Differential ADC Voltage Range
• Selectable 2.56V or 1.1V ADC Reference Voltage
• Free Running or Single Conversion Mode
• ADC Start Conversion by Auto Triggering on Interrupt Sources
• Interrupt on ADC Conversion Complete
• Sleep Mode Noise Canceler
Note: 1. The differential input channels are not tested for devices in PDIP Package. This feature is only
guaranteed to work for devices in TQFP and QFN/MLF Packages.
21.2 Overview
The ATmega164P/324P/644P features a 10-bit successive approximation ADC. The ADC is
connected to an 8-channel Analog Multiplexer which allows 8 single-ended voltage inputs con-
structed from the pins of Port A. The single-ended voltage inputs refer to 0V (GND).
The device also supports 16 differential voltage input combinations. Two of the differential inputs
(ADC1, ADC0 and ADC3, ADC2) are equipped with a programmable gain stage. This provides
amplification steps of 0 dB (1x), 20 dB (10x), or 46 dB (200x) on the differential input voltage
before the A/D conversion. Seven differential analog input channels share a common negative
terminal (ADC1), while any other ADC input can be selected as the positive input terminal. If 1x
or 10x gain is used, 8-bit resolution can be expected. If 200x gain is used, 6-bit resolution can be
expected. Note that internal references of 1.1V should not be used on 200x gain.
The ADC contains a Sample and Hold circuit which ensures that the input voltage to the ADC is
held at a constant level during conversion. A block diagram of the ADC is shown in Figure 21-1
on page 239.
The ADC has a separate analog supply voltage pin, AVCC. AVCC must not differ more than
±0.3 V from VCC. See the paragraph ”ADC Noise Canceler” on page 246 on how to connect this
pin.
Internal reference voltages of nominally 1.1V, 2.56V or AVCC are provided On-chip. The voltage
reference may be externally decoupled at the AREF pin by a capacitor for better noise perfor-
mance. If VCC is below 2.1V, internal voltage reference of 1.1V should not be used.
238 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 21-1. Analog-to-digital Converter Block Schematic
ADC CONVERSION
COMPLETE IRQ
INTERRUPT
FLAGS
ADTS[2:0]
8-BIT DATABUS
ADIE
ADIF
15 0
ADC MULTIPLEXER ADC CTRL & STATUS ADC CTRL & STATUS ADC DATA REGISTER
SELECT (ADMUX) REGISTER B (ADCSRB) REGISTER A (ADCSRA) (ADCH/ADCL)
ADLAR
ADPS[2:0]
ADIF
REFS[1:0]
MUX[4:0]
ADEN
ADSC
ADATE
ADC[9:0]
TRIGGER
SELECT
START
MUX DECODER PRESCALER
CHANNEL SELECTION
DIFF / GAIN SELECT
AVCC CONVERSION LOGIC
INTERNAL
REFERENCE
(1.1V/2.56V)
AREF 10-bit DAC -
+
SAMPLE & HOLD
COMPARATOR
NEG
ADC[2:0]
INPUT -
MUX
+
GAIN
AMPLIFIER
ADC[7:0]
POS ADC
BANDGAP (1.1V)
REFERENCE
INPUT MULTIPLEXER
MUX
OUTPUT
GND
21.3 Operation
The ADC converts an analog input voltage to a 10-bit digital value through successive approxi-
mation. The minimum value represents GND and the maximum value represents the voltage on
the AREF pin minus 1 LSB. Optionally, AVCC or an internal 2.56V reference voltage may be
connected to the AREF pin by writing to the REFSn bits in the ADMUX Register. The internal
voltage reference may thus be decoupled by an external capacitor at the AREF pin to improve
noise immunity.
The analog input channel and differential gain are selected by writing to the MUX bits in
ADMUX. Any of the ADC input pins, as well as GND and a fixed bandgap voltage reference, can
be selected as single ended inputs to the ADC. A selection of ADC input pins can be selected as
positive and negative inputs to the differential gain amplifier.
239
8011D–AVR–02/07
If differential channels are selected, the differential gain stage amplifies the voltage difference
between the selected input channel pair by the selected gain factor. This amplified value then
becomes the analog input to the ADC. If single ended channels are used, the gain amplifier is
bypassed altogether.
The ADC is enabled by setting the ADC Enable bit, ADEN in ADCSRA. Voltage reference and
input channel selections will not go into effect until ADEN is set. The ADC does not consume
power when ADEN is cleared, so it is recommended to switch off the ADC before entering power
saving sleep modes.
The ADC generates a 10-bit result which is presented in the ADC Data Registers, ADCH and
ADCL. By default, the result is presented right adjusted, but can optionally be presented left
adjusted by setting the ADLAR bit in ADMUX.
If the result is left adjusted and no more than 8-bit precision is required, it is sufficient to read
ADCH. Otherwise, ADCL must be read first, then ADCH, to ensure that the content of the Data
Registers belongs to the same conversion. Once ADCL is read, ADC access to Data Registers
is blocked. This means that if ADCL has been read, and a conversion completes before ADCH is
read, neither register is updated and the result from the conversion is lost. When ADCH is read,
ADC access to the ADCH and ADCL Registers is re-enabled.
The ADC has its own interrupt which can be triggered when a conversion completes. When ADC
access to the Data Registers is prohibited between reading of ADCH and ADCL, the interrupt
will trigger even if the result is lost.
21.4 Starting a Conversion

A single conversion is started by writing a logical one to the ADC Start Conversion bit, ADSC.
This bit stays high as long as the conversion is in progress and will be cleared by hardware
when the conversion is completed. If a different data channel is selected while a conversion is in
progress, the ADC will finish the current conversion before performing the channel change.
Alternatively, a conversion can be triggered automatically by various sources. Auto Triggering is
enabled by setting the ADC Auto Trigger Enable bit, ADATE in ADCSRA. The trigger source is
selected by setting the ADC Trigger Select bits, ADTS in ADCSRB (see description of the ADTS
bits for a list of the trigger sources). When a positive edge occurs on the selected trigger signal,
the ADC prescaler is reset and a conversion is started. This provides a method of starting con-
versions at fixed intervals. If the trigger signal still is set when the conversion completes, a new
conversion will not be started. If another positive edge occurs on the trigger signal during con-
version, the edge will be ignored. Note that an Interrupt Flag will be set even if the specific
interrupt is disabled or the global interrupt enable bit in SREG is cleared. A conversion can thus
be triggered without causing an interrupt. However, the Interrupt Flag must be cleared in order to
trigger a new conversion at the next interrupt event.
240 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 21-2. ADC Auto Trigger Logic
ADTS[2:0]
PRESCALER
START CLKADC
ADIF ADATE
SOURCE 1
. CONVERSION
. LOGIC
.
. EDGE
SOURCE n DETECTOR
ADSC
Using the ADC Interrupt Flag as a trigger source makes the ADC start a new conversion as soon
as the ongoing conversion has finished. The ADC then operates in Free Running mode, con-
stantly sampling and updating the ADC Data Register. The first conversion must be started by
writing a logical one to the ADSC bit in ADCSRA. In this mode the ADC will perform successive
conversions independently of whether the ADC Interrupt Flag, ADIF is cleared or not.
If Auto Triggering is enabled, single conversions can be started by writing ADSC in ADCSRA to
one. ADSC can also be used to determine if a conversion is in progress. The ADSC bit will be
read as one during a conversion, independently of how the conversion was started.
21.5 Prescaling and Conversion Timing
Figure 21-3. ADC Prescaler

ADEN
START Reset
7-BIT ADC PRESCALER
CK
CK/128
CK/64
CK/32
CK/16
CK/2
CK/4
CK/8
ADPS0
ADPS1
ADPS2
ADC CLOCK SOURCE
By default, the successive approximation circuitry requires an input clock frequency between 50
kHz and 200 kHz to get maximum resolution. If a lower resolution than 10 bits is needed, the
input clock frequency to the ADC can be higher than 200 kHz to get a higher sample rate.
The ADC module contains a prescaler, which generates an acceptable ADC clock frequency
from any CPU frequency above 100 kHz. The prescaling is set by the ADPS bits in ADCSRA.
The prescaler starts counting from the moment the ADC is switched on by setting the ADEN bit
241
8011D–AVR–02/07
in ADCSRA. The prescaler keeps running for as long as the ADEN bit is set, and is continuously
reset when ADEN is low.
When initiating a single ended conversion by setting the ADSC bit in ADCSRA, the conversion
starts at the following rising edge of the ADC clock cycle. See ”Differential Gain Channels” on
page 244 for details on differential conversion timing.
A normal conversion takes 13 ADC clock cycles. The first conversion after the ADC is switched
on (ADEN in ADCSRA is set) takes 25 ADC clock cycles in order to initialize the analog circuitry.
When the bandgap reference voltage is used as input to the ADC, it will take a certain time for
the voltage to stabilize. If not stabilized, the first value read after the first conversion may be
wrong.
The actual sample-and-hold takes place 1.5 ADC clock cycles after the start of a normal conver-
sion and 13.5 ADC clock cycles after the start of a first conversion. When a conversion is
complete, the result is written to the ADC Data Registers, and ADIF is set. In single conversion
mode, ADSC is cleared simultaneously. The software may then set ADSC again, and a new
conversion will be initiated on the first rising ADC clock edge.
When Auto Triggering is used, the prescaler is reset when the trigger event occurs. This assures
a fixed delay from the trigger event to the start of conversion. In this mode, the sample-and-hold
takes place 2 ADC clock cycles after the rising edge on the trigger source signal. Three addi-
tional CPU clock cycles are used for synchronization logic.
When using Differential mode, along with Auto Trigging from a source other than the ADC Con-
version Complete, each conversion will require 25 ADC clocks. This is because the ADC must
be disabled and re-enabled after every conversion.
In Free Running mode, a new conversion will be started immediately after the conversion com-
pletes, while ADSC remains high. For a summary of conversion times, see Table 21-1 on page
244.
Figure 21-4. ADC Timing Diagram, First Conversion (Single Conversion Mode)
Next
First Conversion Conversion
Cycle Number 1 2 12 13 14 15 16 17 18 19 20 21 22 23 24 25 1 2 3
ADC Clock
ADEN
ADSC
ADIF
ADCH MSB of Result
ADCL LSB of Result
MUX and REFS Conversion

Update Sample & Hold Complete MUX and REFS
Update
242 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 21-5. ADC Timing Diagram, Single Conversion
One Conversion Next Conversion
Cycle Number 1 2 3 4 5 6 7 8 9 10 11 12 13 1 2 3
ADC Clock
ADSC
ADIF
ADCH MSB of Result
ADCL LSB of Result
Sample & Hold Conversion

MUX and REFS Complete MUX and REFS
Update Update
Figure 21-6. ADC Timing Diagram, Auto Triggered Conversion

Cycle Number 1 2 3 4 5 6 7 8 9 10 11 12 13 1 2
ADC Clock
Trigger
Source
ADATE
ADIF
ADCH MSB of Result
ADCL LSB of Result
Sample & Hold Conversion Prescaler

Prescaler Reset
Complete
Reset
MUX and REFS
Update
Figure 21-7. ADC Timing Diagram, Free Running Conversion

11 12 13 1 2 3 4
Cycle Number
ADC Clock
ADSC
ADIF
ADCH MSB of Result
ADCL LSB of Result
Conversion Sample & Hold

Complete MUX and REFS
Update
243
8011D–AVR–02/07
Table 21-1. ADC Conversion Time
Sample & Hold (Cycles
Condition from Start of Conversion) Conversion Time (Cycles)
First conversion 14.5 25
Normal conversions, single ended 1.5 13
Auto Triggered conversions 2 13.5
Normal conversions, differential 1.5/2.5 13/14
21.5.1 Differential Gain Channels
When using differential gain channels, certain aspects of the conversion need to be taken into
consideration. Note that the differential channels should not be used with an AREF < 2V.
Differential conversions are synchronized to the internal clock CKADC2 equal to half the ADC
clock. This synchronization is done automatically by the ADC interface in such a way that the
sample-and-hold occurs at a specific phase of CKADC2. A conversion initiated by the user (i.e., all
single conversions, and the first free running conversion) when CKADC2 is low will take the same
amount of time as a single ended conversion (13 ADC clock cycles from the next prescaled
clock cycle). A conversion initiated by the user when CKADC2 is high will take 14 ADC clock
cycles due to the synchronization mechanism. In Free Running mode, a new conversion is initi-
ated immediately after the previous conversion completes, and since CKADC2 is high at this time,
all automatically started (i.e., all but the first) free running conversions will take 14 ADC clock
cycles.
The gain stage is optimized for a bandwidth of 4 kHz at all gain settings. Higher frequencies may
be subjected to non-linear amplification. An external low-pass filter should be used if the input
signal contains higher frequency components than the gain stage bandwidth. Note that the ADC
clock frequency is independent of the gain stage bandwidth limitation. For example, the ADC
clock period may be 6 µs, allowing a channel to be sampled at 12 kSPS, regardless of the band-
width of this channel.
If differential gain channels are used and conversions are started by Auto Triggering, the ADC
must be switched off between conversions. When Auto Triggering is used, the ADC prescaler is
reset before the conversion is started. Since the gain stage is dependent of a stable ADC clock
prior to the conversion, this conversion will not be valid. By disabling and then re-enabling the
ADC between each conversion (writing ADEN in ADCSRA to “0” then to “1”), only extended con-
versions are performed. The result from the extended conversions will be valid. See ”Prescaling
and Conversion Timing” on page 241 for timing details.
21.6 Changing Channel or Reference Selection

The MUXn and REFS1:0 bits in the ADMUX Register are single buffered through a temporary
register to which the CPU has random access. This ensures that the channels and reference
selection only takes place at a safe point during the conversion. The channel and reference
selection is continuously updated until a conversion is started. Once the conversion starts, the
channel and reference selection is locked to ensure a sufficient sampling time for the ADC. Con-
tinuous updating resumes in the last ADC clock cycle before the conversion completes (ADIF in
ADCSRA is set). Note that the conversion starts on the following rising ADC clock edge after
ADSC is written. The user is thus advised not to write new channel or reference selection values
to ADMUX until one ADC clock cycle after ADSC is written.
244 ATmega164P/324P/644P
8011D–AVR–02/07
If Auto Triggering is used, the exact time of the triggering event can be indeterministic. Special
care must be taken when updating the ADMUX Register, in order to control which conversion
will be affected by the new settings.
If both ADATE and ADEN is written to one, an interrupt event can occur at any time. If the
ADMUX Register is changed in this period, the user cannot tell if the next conversion is based
on the old or the new settings. ADMUX can be safely updated in the following ways:
1. When ADATE or ADEN is cleared.
2. During conversion, minimum one ADC clock cycle after the trigger event.
3. After a conversion, before the Interrupt Flag used as trigger source is cleared.
When updating ADMUX in one of these conditions, the new settings will affect the next ADC
conversion.
Special care should be taken when changing differential channels. Once a differential channel
has been selected, the gain stage may take as much as 125 µs to stabilize to the new value.
Thus conversions should not be started within the first 125 µs after selecting a new differential
channel. Alternatively, conversion results obtained within this period should be discarded.
The same settling time should be observed for the first differential conversion after changing
ADC reference (by changing the REFS1:0 bits in ADMUX).
21.6.1 ADC Input Channels
When changing channel selections, the user should observe the following guidelines to ensure
that the correct channel is selected:
In Single Conversion mode, always select the channel before starting the conversion. The chan-
nel selection may be changed one ADC clock cycle after writing one to ADSC. However, the
simplest method is to wait for the conversion to complete before changing the channel selection.
In Free Running mode, always select the channel before starting the first conversion. The chan-
nel selection may be changed one ADC clock cycle after writing one to ADSC. However, the
simplest method is to wait for the first conversion to complete, and then change the channel
selection. Since the next conversion has already started automatically, the next result will reflect
the previous channel selection. Subsequent conversions will reflect the new channel selection.
When switching to a differential gain channel, the first conversion result may have a poor accu-
racy due to the required settling time for the automatic offset cancellation circuitry. The user
should preferably disregard the first conversion result.
21.6.2 ADC Voltage Reference
The reference voltage for the ADC (VREF) indicates the conversion range for the ADC. Single
ended channels that exceed VREF will result in codes close to 0x3FF. VREF can be selected as
either AVCC, internal 2.56V reference, or external AREF pin.
AVCC is connected to the ADC through a passive switch. The internal 2.56V reference is gener-
ated from the internal bandgap reference (VBG) through an internal amplifier. In either case, the
external AREF pin is directly connected to the ADC, and the reference voltage can be made
more immune to noise by connecting a capacitor between the AREF pin and ground. VREF can
also be measured at the AREF pin with a high impedant voltmeter. Note that VREF is a high
impedant source, and only a capacitive load should be connected in a system.
245
8011D–AVR–02/07
If the user has a fixed voltage source connected to the AREF pin, the user may not use the other
reference voltage options in the application, as they will be shorted to the external voltage. If no
external voltage is applied to the AREF pin, the user may switch between AVCC and 2.56V as
reference selection. The first ADC conversion result after switching reference voltage source
may be inaccurate, and the user is advised to discard this result.
If differential channels are used, the selected reference should not be closer to AVCC than
indicated in Table 26-10 on page 333.
21.7 ADC Noise Canceler

The ADC features a noise canceler that enables conversion during sleep mode to reduce noise
induced from the CPU core and other I/O peripherals. The noise canceler can be used with ADC
Noise Reduction and Idle mode. To make use of this feature, the following procedure should be
used:
a. Make sure that the ADC is enabled and is not busy converting. Single Conversion
Mode must be selected and the ADC conversion complete interrupt must be enabled.
b. Enter ADC Noise Reduction mode (or Idle mode). The ADC will start a conversion
once the CPU has been halted.
c. If no other interrupts occur before the ADC conversion completes, the ADC interrupt
will wake up the CPU and execute the ADC Conversion Complete interrupt routine. If
another interrupt wakes up the CPU before the ADC conversion is complete, that
interrupt will be executed, and an ADC Conversion Complete interrupt request will be
generated when the ADC conversion completes. The CPU will remain in active mode
until a new sleep command is executed.
Note that the ADC will not be automatically turned off when entering other sleep modes than Idle
mode and ADC Noise Reduction mode. The user is advised to write zero to ADEN before enter-
ing such sleep modes to avoid excessive power consumption. If the ADC is enabled in such
sleep modes and the user wants to perform differential conversions, the user is advised to
switch the ADC off and on after waking up from sleep to prompt an extended conversion to get a
valid result.
21.7.1 Analog Input Circuitry
The Analog Input Circuitry for single ended channels is illustrated in Figure 21-8. An analog
source applied to ADCn is subjected to the pin capacitance and input leakage of that pin, regard-
less of whether that channel is selected as input for the ADC. When the channel is selected, the
source must drive the S/H capacitor through the series resistance (combined resistance in the
input path).
The ADC is optimized for analog signals with an output impedance of approximately 10 kΩ or
less. If such a source is used, the sampling time will be negligible. If a source with higher imped-
ance is used, the sampling time will depend on how long time the source needs to charge the
S/H capacitor, with can vary widely. The user is recommended to only use low impedant sources
with slowly varying signals, since this minimizes the required charge transfer to the S/H
capacitor.
If differential gain channels are used, the input circuitry looks somewhat different, although
source impedances of a few hundred kΩ or less is recommended.
246 ATmega164P/324P/644P
8011D–AVR–02/07
Signal components higher than the Nyquist frequency (fADC/2) should not be present for either
kind of channels, to avoid distortion from unpredictable signal convolution. The user is advised
to remove high frequency components with a low-pass filter before applying the signals as
inputs to the ADC.
Figure 21-8. Analog Input Circuitry
IIH
ADCn
1..100 kΩ
CS/H= 14 pF
IIL
VCC/2
247
8011D–AVR–02/07
21.7.2 Analog Noise Canceling Techniques
Digital circuitry inside and outside the device generates EMI which might affect the accuracy of
analog measurements. If conversion accuracy is critical, the noise level can be reduced by
applying the following techniques:
a. Keep analog signal paths as short as possible. Make sure analog tracks run over the
analog ground plane, and keep them well away from high-speed switching digital
tracks.
b. The AVCC pin on the device should be connected to the digital VCC supply voltage
via an LC network as shown in Figure 21-9.
c. Use the ADC noise canceler function to reduce induced noise from the CPU.
d. If any ADC port pins are used as digital outputs, it is essential that these do not
switch while a conversion is in progress.
Figure 21-9. ADC Power Connections
Analog Ground Plane

PA3 (ADC3)
PA0 (ADC0)
PA1 (ADC1)
PA2 (ADC2)
VCC
GND
PA4 (ADC4)
PA5 (ADC5)
PA6 (ADC6)
PA7 (ADC7)
10μH
AREF
GND
100nF
AVCC
PC7
21.7.3 Offset Compensation Schemes
The gain stage has a built-in offset cancellation circuitry that nulls the offset of differential mea-
surements as much as possible. The remaining offset in the analog path can be measured
directly by selecting the same channel for both differential inputs. This offset residue can be then
subtracted in software from the measurement results. Using this kind of software based offset
correction, offset on any channel can be reduced below one LSB.
248 ATmega164P/324P/644P
8011D–AVR–02/07
21.7.4 ADC Accuracy Definitions
An n-bit single-ended ADC converts a voltage linearly between GND and V REF in 2 n steps
(LSBs). The lowest code is read as 0, and the highest code is read as 2n-1.
Several parameters describe the deviation from the ideal behavior:
• Offset: The deviation of the first transition (0x000 to 0x001) compared to the ideal transition (at
0.5 LSB). Ideal value: 0 LSB.
Figure 21-10. Offset Error

Output Code
Ideal ADC
Actual ADC
Offset
Error
VREF Input Voltage
• Gain Error: After adjusting for offset, the Gain Error is found as the deviation of the last
transition (0x3FE to 0x3FF) compared to the ideal transition (at 1.5 LSB below maximum).
Ideal value: 0 LSB
Figure 21-11. Gain Error

Output Code Gain
Error
Ideal ADC
Actual ADC
VREF Input Voltage
• Integral Non-linearity (INL): After adjusting for offset and gain error, the INL is the maximum
deviation of an actual transition compared to an ideal transition for any code. Ideal value: 0
LSB.
249
8011D–AVR–02/07
Figure 21-12. Integral Non-linearity (INL)
Output Code
INL
Ideal ADC
Actual ADC
VREF Input Voltage
• Differential Non-linearity (DNL): The maximum deviation of the actual code width (the interval
between two adjacent transitions) from the ideal code width (1 LSB). Ideal value: 0 LSB.
Figure 21-13. Differential Non-linearity (DNL)

Output Code
0x3FF
1 LSB
DNL
0x000
0 VREF Input Voltage
• Quantization Error: Due to the quantization of the input voltage into a finite number of codes, a
range of input voltages (1 LSB wide) will code to the same value. Always ±0.5 LSB.
• Absolute Accuracy: The maximum deviation of an actual (unadjusted) transition compared to
an ideal transition for any code. This is the compound effect of Offset, Gain Error, Differential
Error, Non-linearity, and Quantization Error. Ideal value: ±0.5 LSB.
250 ATmega164P/324P/644P
8011D–AVR–02/07
21.8 ADC Conversion Result
After the conversion is complete (ADIF is high), the conversion result can be found in the ADC
Result Registers (ADCL, ADCH).
For single ended conversion, the result is
V IN ⋅ 1024
ADC = --------------------------
V REF
where VIN is the voltage on the selected input pin and VREF the selected voltage reference (see
Table 21-3 on page 253 and Table 21-4 on page 254). 0x000 represents analog ground, and
0x3FF represents the selected reference voltage minus one LSB.
If differential channels are used, the result is
( V POS – V NEG ) ⋅ GAIN ⋅ 512
ADC = -----------------------------------------------------------------------
-
V REF
where VPOS is the voltage on the positive input pin, VNEG the voltage on the negative input pin,
GAIN the selected gain factor, and VREF the selected voltage reference. The result is presented
in two’s complement form, from 0x200 (-512d) through 0x1FF (+511d). Note that if the user
wants to perform a quick polarity check of the results, it is sufficient to read the MSB of the result
(ADC9 in ADCH). If this bit is one, the result is negative, and if this bit is zero, the result is posi-
tive. Figure 21-14 on page 252 shows the decoding of the differential input range.
Table 21-2 on page 252 shows the resulting output codes if the differential input channel pair
(ADCn - ADCm) is selected with a gain of GAIN and a reference voltage of VREF.
251
8011D–AVR–02/07
Figure 21-14. Differential Measurement Range
Output Code
0x1FF
0x000
- V REF/GAIN 0 VREF/GAIN Differential Input

0x3FF
Voltage (Volts)
0x200
Table 21-2. Correlation between Input Voltage and Output Codes

VADCn Read code Corresponding Decimal Value
VADCm + VREF/GAIN 0x1FF 511
VADCm + 0.999 VREF/GAIN 0x1FF 511
VADCm + 0.998 VREF/GAIN 0x1FE 510
... ... ...
VADCm + 0.001 VREF/GAIN 0x001 1
VADCm 0x000 0
VADCm - 0.001 VREF/GAIN 0x3FF -1
... ... ...
VADCm - 0.999 VREF/GAIN 0x201 -511
VADCm - VREF/GAIN 0x200 -512
252 ATmega164P/324P/644P
8011D–AVR–02/07
Example:
ADMUX = 0xED (ADC3 - ADC2, 10x gain, 2.56V reference, left adjusted result)
Voltage on ADC3 is 300 mV, voltage on ADC2 is 500 mV.
ADCR = 512 * 10 * (300 - 500) / 2560 = -400 = 0x270
ADCL will thus read 0x00, and ADCH will read 0x9C. Writing zero to ADLAR right adjusts the
result: ADCL = 0x70, ADCH = 0x02.

21.9.1 ADMUX – ADC Multiplexer Selection Register
Bit 7 6 5 4 3 2 1 0
(0x7C) REFS1 REFS0 ADLAR MUX4 MUX3 MUX2 MUX1 MUX0 ADMUX
• Bit 7:6 – REFS1:0: Reference Selection Bits

These bits select the voltage reference for the ADC, as shown in Table 21-3. If these bits are
changed during a conversion, the change will not go in effect until this conversion is complete
(ADIF in ADCSRA is set). The internal voltage reference options may not be used if an external
reference voltage is being applied to the AREF pin.
Table 21-3. Voltage Reference Selections for ADC
REFS1 REFS0 Voltage Reference Selection
0 0 AREF, Internal Vref turned off
0 1 AVCC with external capacitor at AREF pin
1 0 Internal 1.1V Voltage Reference with external capacitor at AREF pin
1 1 Internal 2.56V Voltage Reference with external capacitor at AREF pin
Note: If 10x og 200x gain is selected, only 2.56V should be used as Internal Voltage Reference.
• Bit 5 – ADLAR: ADC Left Adjust Result

The ADLAR bit affects the presentation of the ADC conversion result in the ADC Data Register.
Write one to ADLAR to left adjust the result. Otherwise, the result is right adjusted. Changing the
ADLAR bit will affect the ADC Data Register immediately, regardless of any ongoing conver-
sions. For a complete description of this bit, see ”ADCL and ADCH – The ADC Data Register” on
page 256.
• Bits 4:0 – MUX4:0: Analog Channel and Gain Selection Bits

The value of these bits selects which combination of analog inputs are connected to the ADC.
These bits also select the gain for the differential channels. See Table 21-4 on page 254 for
details. If these bits are changed during a conversion, the change will not go in effect until this
conversion is complete (ADIF in ADCSRA is set).
253
8011D–AVR–02/07
Table 21-4. Input Channel and Gain Selections
Single Ended Positive Differential Negative Differential
MUX4..0 Input Input Input Gain
00000 ADC0
00001 ADC1
00010 ADC2
00011 ADC3 N/A
00100 ADC4
00101 ADC5
00110 ADC6
00111 ADC7
01000 ADC0 ADC0 10x
01001 ADC1 ADC0 10x
(1)
01010 ADC0 ADC0 200x
(1)
01011 ADC1 ADC0 200x
01100 ADC2 ADC2 10x
01101 ADC3 ADC2 10x
(1)
01110 ADC2 ADC2 200x
(1)
01111 ADC3 ADC2 200x
10000 ADC0 ADC1 1x
10001 ADC1 ADC1 1x
10010 N/A ADC2 ADC1 1x
10011 ADC3 ADC1 1x
10100 ADC4 ADC1 1x
10101 ADC5 ADC1 1x
10110 ADC6 ADC1 1x
10111 ADC7 ADC1 1x
11000 ADC0 ADC2 1x
11001 ADC1 ADC2 1x
11010 ADC2 ADC2 1x
11011 ADC3 ADC2 1x
11100 ADC4 ADC2 1x
11101 ADC5 ADC2 1x
11110 1.1V (VBG)
N/A
11111 0 V (GND)
Note: 1. The differential input channels are not tested for devices in PDIP Package. This feature is only
guaranteed to work for devices in TQFP and QFN/MLF Packages
254 ATmega164P/324P/644P
8011D–AVR–02/07
21.9.2 ADCSRA – ADC Control and Status Register A
Bit 7 6 5 4 3 2 1 0
(0x7A) ADEN ADSC ADATE ADIF ADIE ADPS2 ADPS1 ADPS0 ADCSRA
• Bit 7 – ADEN: ADC Enable

Writing this bit to one enables the ADC. By writing it to zero, the ADC is turned off. Turning the
ADC off while a conversion is in progress, will terminate this conversion.
• Bit 6 – ADSC: ADC Start Conversion

In Single Conversion mode, write this bit to one to start each conversion. In Free Running Mode,
write this bit to one to start the first conversion. The first conversion after ADSC has been written
after the ADC has been enabled, or if ADSC is written at the same time as the ADC is enabled,
will take 25 ADC clock cycles instead of the normal 13. This first conversion performs initializa-
tion of the ADC.
ADSC will read as one as long as a conversion is in progress. When the conversion is complete,
it returns to zero. Writing zero to this bit has no effect.
• Bit 5 – ADATE: ADC Auto Trigger Enable

When this bit is written to one, Auto Triggering of the ADC is enabled. The ADC will start a con-
version on a positive edge of the selected trigger signal. The trigger source is selected by setting
the ADC Trigger Select bits, ADTS in ADCSRB.
• Bit 4 – ADIF: ADC Interrupt Flag

This bit is set when an ADC conversion completes and the Data Registers are updated. The
ADC Conversion Complete Interrupt is executed if the ADIE bit and the I-bit in SREG are set.
ADIF is cleared by hardware when executing the corresponding interrupt handling vector. Alter-
natively, ADIF is cleared by writing a logical one to the flag. Beware that if doing a Read-Modify-
Write on ADCSRA, a pending interrupt can be disabled. This also applies if the SBI and CBI
instructions are used.
• Bit 3 – ADIE: ADC Interrupt Enable

When this bit is written to one and the I-bit in SREG is set, the ADC Conversion Complete Inter-
rupt is activated.
• Bits 2:0 – ADPS2:0: ADC Prescaler Select Bits

These bits determine the division factor between the XTAL frequency and the input clock to the
ADC.
Table 21-5. ADC Prescaler Selections

ADPS2 ADPS1 ADPS0 Division Factor
0 0 0 2
0 0 1 2
0 1 0 4
0 1 1 8
255
8011D–AVR–02/07
Table 21-5. ADC Prescaler Selections (Continued)
ADPS2 ADPS1 ADPS0 Division Factor
1 0 0 16
1 0 1 32
1 1 0 64
1 1 1 128
21.9.3 ADCL and ADCH – The ADC Data Register
ADLAR = 0
Bit 15 14 13 12 11 10 9 8
(0x79) – – – – – – ADC9 ADC8 ADCH
(0x78) ADC7 ADC6 ADC5 ADC4 ADC3 ADC2 ADC1 ADC0 ADCL
7 6 5 4 3 2 1 0
Read/Write R R R R R R R R
R R R R R R R R
0 0 0 0 0 0 0 0
ADLAR = 1
Bit 15 14 13 12 11 10 9 8
(0x79) ADC9 ADC8 ADC7 ADC6 ADC5 ADC4 ADC3 ADC2 ADCH
(0x78) ADC1 ADC0 – – – – – – ADCL
7 6 5 4 3 2 1 0
Read/Write R R R R R R R R
R R R R R R R R
0 0 0 0 0 0 0 0
When an ADC conversion is complete, the result is found in these two registers. If differential
channels are used, the result is presented in two’s complement form.
When ADCL is read, the ADC Data Register is not updated until ADCH is read. Consequently, if
the result is left adjusted and no more than 8-bit precision is required, it is sufficient to read
ADCH. Otherwise, ADCL must be read first, then ADCH.
The ADLAR bit in ADMUX, and the MUXn bits in ADMUX affect the way the result is read from
the registers. If ADLAR is set, the result is left adjusted. If ADLAR is cleared (default), the result
is right adjusted.
• ADC9:0: ADC Conversion Result

These bits represent the result from the conversion, as detailed in ”ADC Conversion Result” on
page 251.
21.9.4 ADCSRB – ADC Control and Status Register B
Bit 7 6 5 4 3 2 1 0
(0x7B) – ACME – – – ADTS2 ADTS1 ADTS0 ADCSRB
Read/Write R R/W R R R R/W R/W R/W
256 ATmega164P/324P/644P
8011D–AVR–02/07
• Bit 7, 5:3 – Res: Reserved Bits

These bits are reserved for future use in the ATmega164P/324P/644P. For ensuring compability
with future devices, these bits must be written zero when ADCSRB is written.
• Bit 2:0 – ADTS2:0: ADC Auto Trigger Source

If ADATE in ADCSRA is written to one, the value of these bits selects which source will trigger
an ADC conversion. If ADATE is cleared, the ADTS[2:0] settings will have no effect. A conver-
sion will be triggered by the rising edge of the selected Interrupt Flag. Note that switching from a
trigger source that is cleared to a trigger source that is set, will generate a positive edge on the
trigger signal. If ADEN in ADCSRA is set, this will start a conversion. Switching to Free Running
mode (ADTS[2:0]=0) will not cause a trigger event, even if the ADC Interrupt Flag is set.
Table 21-6. ADC Auto Trigger Source Selections
ADTS2 ADTS1 ADTS0 Trigger Source
0 0 0 Free Running mode
0 0 1 Analog Comparator
0 1 0 External Interrupt Request 0
0 1 1 Timer/Counter0 Compare Match
1 0 0 Timer/Counter0 Overflow
1 0 1 Timer/Counter1 Compare Match B
1 1 0 Timer/Counter1 Overflow
1 1 1 Timer/Counter1 Capture Event
21.9.5 DIDR0 – Digital Input Disable Register 0
Bit 7 6 5 4 3 2 1 0
(0x7E) ADC7D ADC6D ADC5D ADC4D ADC3D ADC2D ADC1D ADC0D DIDR0
• Bit 7..0 – ADC7D..ADC0D: ADC7..0 Digital Input Disable

When this bit is written logic one, the digital input buffer on the corresponding ADC pin is dis-
abled. The corresponding PIN Register bit will always read as zero when this bit is set. When an
analog signal is applied to the ADC7..0 pin and the digital input from this pin is not needed, this
bit should be written logic one to reduce power consumption in the digital input buffer.
257
8011D–AVR–02/07
22. JTAG Interface and On-chip Debug System
22.1 Features
• JTAG (IEEE std. 1149.1 Compliant) Interface
• Boundary-scan Capabilities According to the IEEE std. 1149.1 (JTAG) Standard
• Debugger Access to:
– All Internal Peripheral Units
– Internal and External RAM
– The Internal Register File
– Program Counter
– EEPROM and Flash Memories
• Extensive On-chip Debug Support for Break Conditions, Including
– AVR Break Instruction
– Break on Change of Program Memory Flow
– Single Step Break
– Program Memory Break Points on Single Address or Address Range
– Data Memory Break Points on Single Address or Address Range
• Programming of Flash, EEPROM, Fuses, and Lock Bits through the JTAG Interface
• On-chip Debugging Supported by AVR Studio®
22.2 Overview
The AVR IEEE std. 1149.1 compliant JTAG interface can be used for
• Testing PCBs by using the JTAG Boundary-scan capability
• Programming the non-volatile memories, Fuses and Lock bits
• On-chip debugging
A brief description is given in the following sections. Detailed descriptions for Programming via
the JTAG interface, and using the Boundary-scan Chain can be found in the sections ”Program-
ming via the JTAG Interface” on page 310 and ”IEEE 1149.1 (JTAG) Boundary-scan” on page
264, respectively. The On-chip Debug support is considered being private JTAG instructions,
and distributed within ATMEL and to selected third party vendors only.
Figure 22-1 shows a block diagram of the JTAG interface and the On-chip Debug system. The
TAP Controller is a state machine controlled by the TCK and TMS signals. The TAP Controller
selects either the JTAG Instruction Register or one of several Data Registers as the scan chain
(Shift Register) between the TDI – input and TDO – output. The Instruction Register holds JTAG
instructions controlling the behavior of a Data Register.
The ID-Register, Bypass Register, and the Boundary-scan Chain are the Data Registers used
for board-level testing. The JTAG Programming Interface (actually consisting of several physical
and virtual Data Registers) is used for serial programming via the JTAG interface. The Internal
Scan Chain and Break Point Scan Chain are used for On-chip debugging only.
22.3 TAP – Test Access Port

The JTAG interface is accessed through four of the AVR’s pins. In JTAG terminology, these pins
constitute the Test Access Port – TAP. These pins are:
• TMS: Test mode select. This pin is used for navigating through the TAP-controller state
machine.
258 ATmega164P/324P/644P
8011D–AVR–02/07
• TCK: Test Clock. JTAG operation is synchronous to TCK.
• TDI: Test Data In. Serial input data to be shifted in to the Instruction Register or Data Register
(Scan Chains).
• TDO: Test Data Out. Serial output data from Instruction Register or Data Register.
The IEEE std. 1149.1 also specifies an optional TAP signal; TRST – Test ReSeT – which is not
provided.
When the JTAGEN Fuse is unprogrammed, these four TAP pins are normal port pins, and the
TAP controller is in reset. When programmed, the input TAP signals are internally pulled high
and the JTAG is enabled for Boundary-scan and programming. The device is shipped with this
fuse programmed.
For the On-chip Debug system, in addition to the JTAG interface pins, the RESET pin is moni-
tored by the debugger to be able to detect external reset sources. The debugger can also pull
the RESET pin low to reset the whole system, assuming only open collectors on the reset line
are used in the application.
Figure 22-1. Block Diagram

I/O PORT 0
DEVICE BOUNDARY
BOUNDARY SCAN CHAIN
TDI
JTAG PROGRAMMING
TDO TAP INTERFACE
TCK CONTROLLER
TMS
AVR CPU
INTERNAL
FLASH Address SCAN PC
INSTRUCTION MEMORY Data CHAIN Instruction
REGISTER
ID
REGISTER BREAKPOINT
UNIT
M FLOW CONTROL
U BYPASS
UNIT
X REGISTER DIGITAL
ANALOG
PERIPHERAL Analog inputs
PERIPHERIAL
UNITS
UNITS
BREAKPOINT
SCAN CHAIN
JTAG / AVR CORE
COMMUNICATION
ADDRESS INTERFACE
DECODER OCD STATUS
AND CONTROL
Control & Clock lines
I/O PORT n
259
8011D–AVR–02/07
Figure 22-2. TAP Controller State Diagram
1 Test-Logic-Reset
0 1 1 1
Run-Test/Idle Select-DR Scan Select-IR Scan
0 0
1 1
Capture-DR Capture-IR
0 0
Shift-DR 0 Shift-IR 0
1 1
1 1
Exit1-DR Exit1-IR
0 0
Pause-DR 0 Pause-IR 0
1 1
0 0
Exit2-DR Exit2-IR
1 1
Update-DR Update-IR
1 0 1 0
22.4 TAP Controller

The TAP controller is a 16-state finite state machine that controls the operation of the Boundary-
scan circuitry, JTAG programming circuitry, or On-chip Debug system. The state transitions
depicted in Figure 22-2 depend on the signal present on TMS (shown adjacent to each state
transition) at the time of the rising edge at TCK. The initial state after a Power-on Reset is Test-
Logic-Reset.
As a definition in this document, the LSB is shifted in and out first for all Shift Registers.
Assuming Run-Test/Idle is the present state, a typical scenario for using the JTAG interface is:
• At the TMS input, apply the sequence 1, 1, 0, 0 at the rising edges of TCK to enter the Shift
Instruction Register – Shift-IR state. While in this state, shift the four bits of the JTAG
instructions into the JTAG Instruction Register from the TDI input at the rising edge of TCK.
The TMS input must be held low during input of the 3 LSBs in order to remain in the Shift-IR
state. The MSB of the instruction is shifted in when this state is left by setting TMS high. While
the instruction is shifted in from the TDI pin, the captured IR-state 0x01 is shifted out on the
TDO pin. The JTAG Instruction selects a particular Data Register as path between TDI and
TDO and controls the circuitry surrounding the selected Data Register.
260 ATmega164P/324P/644P
8011D–AVR–02/07
• Apply the TMS sequence 1, 1, 0 to re-enter the Run-Test/Idle state. The instruction is latched
onto the parallel output from the Shift Register path in the Update-IR state. The Exit-IR, Pause-
IR, and Exit2-IR states are only used for navigating the state machine.
• At the TMS input, apply the sequence 1, 0, 0 at the rising edges of TCK to enter the Shift Data
Register – Shift-DR state. While in this state, upload the selected Data Register (selected by
the present JTAG instruction in the JTAG Instruction Register) from the TDI input at the rising
edge of TCK. In order to remain in the Shift-DR state, the TMS input must be held low during
input of all bits except the MSB. The MSB of the data is shifted in when this state is left by
setting TMS high. While the Data Register is shifted in from the TDI pin, the parallel inputs to
the Data Register captured in the Capture-DR state is shifted out on the TDO pin.
• Apply the TMS sequence 1, 1, 0 to re-enter the Run-Test/Idle state. If the selected Data
Register has a latched parallel-output, the latching takes place in the Update-DR state. The
Exit-DR, Pause-DR, and Exit2-DR states are only used for navigating the state machine.
As shown in the state diagram, the Run-Test/Idle state need not be entered between selecting
JTAG instruction and using Data Registers, and some JTAG instructions may select certain
functions to be performed in the Run-Test/Idle, making it unsuitable as an Idle state.
Note: Independent of the initial state of the TAP Controller, the Test-Logic-Reset state can always be
entered by holding TMS high for five TCK clock periods.
For detailed information on the JTAG specification, refer to the literature listed in ”Bibliography”
on page 263.
22.5 Using the Boundary-scan Chain

A complete description of the Boundary-scan capabilities are given in the section ”IEEE 1149.1
(JTAG) Boundary-scan” on page 264.
22.6 Using the On-chip Debug System

As shown in Figure 22-1, the hardware support for On-chip Debugging consists mainly of
• A scan chain on the interface between the internal AVR CPU and the internal peripheral units.
• Break Point unit.
• Communication interface between the CPU and JTAG system.
All read or modify/write operations needed for implementing the Debugger are done by applying
AVR instructions via the internal AVR CPU Scan Chain. The CPU sends the result to an I/O
memory mapped location which is part of the communication interface between the CPU and the
JTAG system.
The Break Point Unit implements Break on Change of Program Flow, Single Step Break, two
Program Memory Break Points, and two combined Break Points. Together, the four Break
Points can be configured as either:
• 4 single Program Memory Break Points.
• 3 Single Program Memory Break Point + 1 single Data Memory Break Point.
• 2 single Program Memory Break Points + 2 single Data Memory Break Points.
• 2 single Program Memory Break Points + 1 Program Memory Break Point with mask (“range
Break Point”).
• 2 single Program Memory Break Points + 1 Data Memory Break Point with mask (“range Break
Point”).
261
8011D–AVR–02/07
A debugger, like the AVR Studio, may however use one or more of these resources for its inter-
nal purpose, leaving less flexibility to the end-user.
A list of the On-chip Debug specific JTAG instructions is given in ”On-chip Debug Specific JTAG
Instructions” on page 262.
The JTAGEN Fuse must be programmed to enable the JTAG Test Access Port. In addition, the
OCDEN Fuse must be programmed and no Lock bits must be set for the On-chip debug system
to work. As a security feature, the On-chip debug system is disabled when either of the LB1 or
LB2 Lock bits are set. Otherwise, the On-chip debug system would have provided a back-door
into a secured device.
The AVR Studio enables the user to fully control execution of programs on an AVR device with
On-chip Debug capability, AVR In-Circuit Emulator, or the built-in AVR Instruction Set Simulator.
AVR Studio® supports source level execution of Assembly programs assembled with Atmel Cor-
poration’s AVR Assembler and C programs compiled with third party vendors’ compilers.
AVR Studio runs under Microsoft® Windows® 95/98/2000 and Microsoft Windows NT®.
For a full description of the AVR Studio, please refer to the AVR Studio User Guide. Only high-
lights are presented in this document.
All necessary execution commands are available in AVR Studio, both on source level and on
disassembly level. The user can execute the program, single step through the code either by
tracing into or stepping over functions, step out of functions, place the cursor on a statement and
execute until the statement is reached, stop the execution, and reset the execution target. In
addition, the user can have an unlimited number of code Break Points (using the BREAK
instruction) and up to two data memory Break Points, alternatively combined as a mask (range)
Break Point.
22.7 On-chip Debug Specific JTAG Instructions

The On-chip debug support is considered being private JTAG instructions, and distributed within
ATMEL and to selected third party vendors only. Instruction opcodes are listed for reference.
22.7.1 PRIVATE0; 0x8
Private JTAG instruction for accessing On-chip debug system.

22.7.2 PRIVATE1; 0x9

22.7.3 PRIVATE2; 0xA

22.7.4 PRIVATE3; 0xB
22.8 Using the JTAG Programming Capabilities

Programming of AVR parts via JTAG is performed via the 4-pin JTAG port, TCK, TMS, TDI, and
TDO. These are the only pins that need to be controlled/observed to perform JTAG program-
ming (in addition to power pins). It is not required to apply 12V externally. The JTAGEN Fuse
must be programmed and the JTD bit in the MCUCR Register must be cleared to enable the
JTAG Test Access Port.
262 ATmega164P/324P/644P
8011D–AVR–02/07
The JTAG programming capability supports:
• Flash programming and verifying.
• EEPROM programming and verifying.
• Fuse programming and verifying.
• Lock bit programming and verifying.
The Lock bit security is exactly as in parallel programming mode. If the Lock bits LB1 or LB2 are
programmed, the OCDEN Fuse cannot be programmed unless first doing a chip erase. This is a
security feature that ensures no back-door exists for reading out the content of a secured
device.
The details on programming through the JTAG interface and programming specific JTAG
instructions are given in the section ”Programming via the JTAG Interface” on page 310.
22.9 Bibliography
For more information about general Boundary-scan, the following literature can be consulted:
• IEEE: IEEE Std. 1149.1-1990. IEEE Standard Test Access Port and Boundary-scan
Architecture, IEEE, 1993.
• Colin Maunder: The Board Designers Guide to Testable Logic Circuits, Addison-Wesley, 1992.
22.10.1 OCDR – On-chip Debug Register
Bit 7 6 5 4 3 2 1 0
0x31 (0x51) MSB/IDRD LSB OCDR
The OCDR Register provides a communication channel from the running program in the micro-
controller to the debugger. The CPU can transfer a byte to the debugger by writing to this
location. At the same time, an internal flag; I/O Debug Register Dirty – IDRD – is set to indicate
to the debugger that the register has been written. When the CPU reads the OCDR Register the
7 LSB will be from the OCDR Register, while the MSB is the IDRD bit. The debugger clears the
IDRD bit when it has read the information.
In some AVR devices, this register is shared with a standard I/O location. In this case, the OCDR
Register can only be accessed if the OCDEN Fuse is programmed, and the debugger enables
access to the OCDR Register. In all other cases, the standard I/O location is accessed.
Refer to the debugger documentation for further information on how to use this register.
263
8011D–AVR–02/07
23. IEEE 1149.1 (JTAG) Boundary-scan
23.1 Features
• JTAG (IEEE std. 1149.1 compliant) Interface
• Boundary-scan Capabilities According to the JTAG Standard
• Full Scan of all Port Functions as well as Analog Circuitry having Off-chip Connections
• Supports the Optional IDCODE Instruction
• Additional Public AVR_RESET Instruction to Reset the AVR
23.2 Overview
The Boundary-scan chain has the capability of driving and observing the logic levels on the digi-
tal I/O pins, as well as the boundary between digital and analog logic for analog circuitry having
off-chip connections. At system level, all ICs having JTAG capabilities are connected serially by
the TDI/TDO signals to form a long Shift Register. An external controller sets up the devices to
drive values at their output pins, and observe the input values received from other devices. The
controller compares the received data with the expected result. In this way, Boundary-scan pro-
vides a mechanism for testing interconnections and integrity of components on Printed Circuits
Boards by using the four TAP signals only.
The four IEEE 1149.1 defined mandatory JTAG instructions IDCODE, BYPASS, SAMPLE/PRE-
LOAD, and EXTEST, as well as the AVR specific public JTAG instruction AVR_RESET can be
used for testing the Printed Circuit Board. Initial scanning of the Data Register path will show the
ID-Code of the device, since IDCODE is the default JTAG instruction. It may be desirable to
have the AVR device in reset during test mode. If not reset, inputs to the device may be deter-
mined by the scan operations, and the internal software may be in an undetermined state when
exiting the test mode. Entering reset, the outputs of any port pin will instantly enter the high
impedance state, making the HIGHZ instruction redundant. If needed, the BYPASS instruction
can be issued to make the shortest possible scan chain through the device. The device can be
set in the reset state either by pulling the external RESET pin low, or issuing the AVR_RESET
instruction with appropriate setting of the Reset Data Register.
The EXTEST instruction is used for sampling external pins and loading output pins with data.
The data from the output latch will be driven out on the pins as soon as the EXTEST instruction
is loaded into the JTAG IR-Register. Therefore, the SAMPLE/PRELOAD should also be used for
setting initial values to the scan ring, to avoid damaging the board when issuing the EXTEST
instruction for the first time. SAMPLE/PRELOAD can also be used for taking a snapshot of the
external pins during normal operation of the part.
The JTAGEN Fuse must be programmed and the JTD bit in the I/O Register MCUCR must be
cleared to enable the JTAG Test Access Port.
When using the JTAG interface for Boundary-scan, using a JTAG TCK clock frequency higher
than the internal chip frequency is possible. The chip clock is not required to run.
264 ATmega164P/324P/644P
8011D–AVR–02/07
23.3 Data Registers
The Data Registers relevant for Boundary-scan operations are:
• Bypass Register
• Device Identification Register
• Reset Register
• Boundary-scan Chain
23.3.1 Bypass Register
The Bypass Register consists of a single Shift Register stage. When the Bypass Register is
selected as path between TDI and TDO, the register is reset to 0 when leaving the Capture-DR
controller state. The Bypass Register can be used to shorten the scan chain on a system when
the other devices are to be tested.
23.3.2 Device Identification Register
Figure 23-1 shows the structure of the Device Identification Register.
Figure 23-1. The Format of the Device Identification Register
MSB LSB
Bit 31 28 27 12 11 1 0
Device ID Version Part Number Manufacturer ID 1

4 bits 16 bits 11 bits 1-bit
Version Version is a 4-bit number identifying the revision of the component. The JTAG version number
follows the revision of the device. Revision A is 0x0, revision B is 0x1 and so on.
Part Number The part number is a 16-bit code identifying the component. The JTAG Part Number for
ATmega164P/324P/644P is listed in Table 25-6 on page 294.
Manufacturer ID The Manufacturer ID is a 11-bit code identifying the manufacturer. The JTAG manufacturer ID
for ATMEL is listed in Table 25-6 on page 294.
23.3.3 Reset Register
The Reset Register is a test Data Register used to reset the part. Since the AVR tri-states Port
Pins when reset, the Reset Register can also replace the function of the unimplemented optional
JTAG instruction HIGHZ.
A high value in the Reset Register corresponds to pulling the external Reset low. The part is
reset as long as there is a high value present in the Reset Register. Depending on the fuse set-
tings for the clock options, the part will remain reset for a reset time-out period (refer to ”Clock
Sources” on page 30) after releasing the Reset Register. The output from this Data Register is
not latched, so the reset will take place immediately, as shown in Figure 23-2.
265
8011D–AVR–02/07
Figure 23-2. Reset Register
To
TDO
From Other Internal and

External Reset Sources
From Internal reset

D Q
TDI
ClockDR · AVR_RESET
23.3.4 Boundary-scan Chain
The Boundary-scan Chain has the capability of driving and observing the logic levels on the dig-
ital I/O pins, as well as the boundary between digital and analog logic for analog circuitry having
off-chip connections.
See ”Boundary-scan Chain” on page 267 for a complete description.
23.4 Boundary-scan Specific JTAG Instructions

The Instruction Register is 4-bit wide, supporting up to 16 instructions. Listed below are the
JTAG instructions useful for Boundary-scan operation. Note that the optional HIGHZ instruction
is not implemented, but all outputs with tri-state capability can be set in high-impedant state by
using the AVR_RESET instruction, since the initial state for all port pins is tri-state.
As a definition in this datasheet, the LSB is shifted in and out first for all Shift Registers.
The OPCODE for each instruction is shown behind the instruction name in hex format. The text
describes which Data Register is selected as path between TDI and TDO for each instruction.
23.4.1 EXTEST; 0x0
Mandatory JTAG instruction for selecting the Boundary-scan Chain as Data Register for testing
circuitry external to the AVR package. For port-pins, Pull-up Disable, Output Control, Output
Data, and Input Data are all accessible in the scan chain. For Analog circuits having off-chip
connections, the interface between the analog and the digital logic is in the scan chain. The con-
tents of the latched outputs of the Boundary-scan chain is driven out as soon as the JTAG IR-
Register is loaded with the EXTEST instruction.
The active states are:
• Capture-DR: Data on the external pins are sampled into the Boundary-scan Chain.
• Shift-DR: The Internal Scan Chain is shifted by the TCK input.
• Update-DR: Data from the scan chain is applied to output pins.
266 ATmega164P/324P/644P
8011D–AVR–02/07
23.4.2 IDCODE; 0x1
Optional JTAG instruction selecting the 32 bit ID-Register as Data Register. The ID-Register
consists of a version number, a device number and the manufacturer code chosen by JEDEC.
This is the default instruction after power-up.
• Capture-DR: Data in the IDCODE Register is sampled into the Boundary-scan Chain.
• Shift-DR: The IDCODE scan chain is shifted by the TCK input.
23.4.3 SAMPLE_PRELOAD; 0x2
Mandatory JTAG instruction for pre-loading the output latches and taking a snap-shot of the
input/output pins without affecting the system operation. However, the output latches are not
connected to the pins. The Boundary-scan Chain is selected as Data Register.
• Capture-DR: Data on the external pins are sampled into the Boundary-scan Chain.
• Shift-DR: The Boundary-scan Chain is shifted by the TCK input.
• Update-DR: Data from the Boundary-scan chain is applied to the output latches. However, the
output latches are not connected to the pins.
23.4.4 AVR_RESET; 0xC
The AVR specific public JTAG instruction for forcing the AVR device into the Reset mode or
releasing the JTAG reset source. The TAP controller is not reset by this instruction. The one bit
Reset Register is selected as Data Register. Note that the reset will be active as long as there is
a logic “one” in the Reset Chain. The output from this chain is not latched.
• Shift-DR: The Reset Register is shifted by the TCK input.
23.4.5 BYPASS; 0xF
Mandatory JTAG instruction selecting the Bypass Register for Data Register.
• Capture-DR: Loads a logic “0” into the Bypass Register.
• Shift-DR: The Bypass Register cell between TDI and TDO is shifted.
23.5 Boundary-scan Chain
The Boundary-scan chain has the capability of driving and observing the logic levels on the digi-
tal I/O pins, as well as the boundary between digital and analog logic for analog circuitry having
off-chip connection.
23.5.1 Scanning the Digital Port Pins
Figure 23-3 shows the Boundary-scan Cell for a bi-directional port pin. The pull-up function is
disabled during Boundary-scan when the JTAG IC contains EXTEST or SAMPLE_PRELOAD.
The cell consists of a bi-directional pin cell that combines the three signals Output Control -
OCxn, Output Data - ODxn, and Input Data - IDxn, into only a two-stage Shift Register. The port
and pin indexes are not used in the following description
267
8011D–AVR–02/07
The Boundary-scan logic is not included in the figures in the datasheet. Figure 23-4 shows a
simple digital port pin as described in the section ”I/O-Ports” on page 71. The Boundary-scan
details from Figure 23-3 replaces the dashed box in Figure 23-4.
When no alternate port function is present, the Input Data - ID - corresponds to the PINxn Regis-
ter value (but ID has no synchronizer), Output Data corresponds to the PORT Register, Output
Control corresponds to the Data Direction - DD Register, and the Pull-up Enable - PUExn - cor-
responds to logic expression PUD · DDxn · PORTxn.
Digital alternate port functions are connected outside the dotted box in Figure 23-4 to make the
scan chain read the actual pin value. For analog function, there is a direct connection from the
external pin to the analog circuit. There is no scan chain on the interface between the digital and
the analog circuitry, but some digital control signal to analog circuitry are turned off to avoid driv-
ing contention on the pads.
When JTAG IR contains EXTEST or SAMPLE_PRELOAD the clock is not sent out on the port
pins even if the CKOUT fuse is programmed. Even though the clock is output when the JTAG IR
contains SAMPLE_PRELOAD, the clock is not sampled by the boundary scan.
Figure 23-3. Boundary-scan Cell for Bi-directional Port Pin with Pull-up Function.
ShiftDR To Next Cell EXTEST Vcc
Pull-up Enable (PUE)

0
Output Control (OC)
FF1 LD1 0
0
D Q D Q 1
1
G
Output Data (OD)
Port Pin (PXn)

0 FF0 LD0 0
0
1 D Q D Q 1
1
G
Input Data (ID)
From Last Cell ClockDR UpdateDR
268 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 23-4. General Port Pin Schematic Diagram
See Boundary-scan
Description for Details!
PUExn PUD
Q D
DDxn
Q CLR
WDx
RESET
OCxn
RDx
DATA BUS
Pxn Q D
ODxn PORTxn
Q CLR
IDxn WRx
RESET
SLEEP RRx
SYNCHRONIZER
RPx
D Q D Q
PINxn
L Q Q
CLK I/O
PUD: PULLUP DISABLE WDx: WRITE DDRx

PUExn: PULLUP ENABLE for pin Pxn RDx: READ DDRx
OCxn: OUTPUT CONTROL for pin Pxn WRx: WRITE PORTx
ODxn: OUTPUT DATA to pin Pxn RRx: READ PORTx REGISTER
IDxn: INPUT DATA from pin Pxn RPx: READ PORTx PIN
SLEEP: SLEEP CONTROL CLK I/O : I/O CLOCK
23.5.2 Scanning the RESET Pin
The RESET pin accepts 5V active low logic for standard reset operation, and 12V active high
logic for High Voltage Parallel programming. An observe-only cell as shown in Figure 23-5 is
inserted for the 5V reset signal.
Figure 23-5. Observe-only Cell

To
Next
ShiftDR Cell
From System Pin To System Logic
FF1
0
D Q
1
From ClockDR
Previous
Cell
269
8011D–AVR–02/07
23.6 ATmega164P/324P/644P Boundary-scan Order
Table 23-1 shows the Scan order between TDI and TDO when the Boundary-scan chain is
selected as data path. Bit 0 is the LSB; the first bit scanned in, and the first bit scanned out. The
scan order follows the pin-out order as far as possible. Therefore, the bits of Port A and Port K is
scanned in the opposite bit order of the other ports. Exceptions from the rules are the Scan
chains for the analog circuits, which constitute the most significant bits of the scan chain regard-
less of which physical pin they are connected to. In Figure 23-3, PXn. Data corresponds to FF0,
PXn. Control corresponds to FF1, PXn. Bit 4, 5, 6 and 7 of Port F is not in the scan chain, since
these pins constitute the TAP pins when the JTAG is enabled.
Table 23-1. ATmega164P/324P/644P Boundary-scan Order

Bit Number Signal Name Module
56 PB0.Data
55 PB0.Control
54 PB1.Data
53 PB1.Control
52 PB2.Data
51 PB2.Control
50 PB3.Data
49 PB3.Control
Port B
48 PB4.Data
47 PB4.Control
46 PB5.Data
45 PB5.Control
44 PB6.Data
43 PB6.Control
42 PB7.Data
41 PB7.Control
40 RSTT Reset Logic (Observe Only)
270 ATmega164P/324P/644P
8011D–AVR–02/07
Table 23-1. ATmega164P/324P/644P Boundary-scan Order (Continued)
39 PD0.Data
38 PD0.Control
37 PD1.Data
36 PD1.Control
35 PD2.Data
34 PD2.Control
Port D
33 PD3.Data
32 PD3.Control
31 PD4.Data
30 PD4.Control
29 PD5.Data
28 PD5.Control
27 PD6.Data
26 PD6.Control
25 PD7.Data
24 PD7.Control
23 PC0.Data
22 PC0.Control
21 PC1.Data
20 PC1.Control
19 PC6.Data Port C
18 PC6.Control
17 PC7.Data
16 PC7.Control
15 PA7.Data
271
8011D–AVR–02/07
Table 23-1. ATmega164P/324P/644P Boundary-scan Order (Continued)
14 PA7.Control
13 PA6.Data
12 PA6.Control
11 PA5.Data
10 PA5.Control
9 PA4.Data
8 PA4.Control
7 PA3.Data Port A
6 PA3.Control
5 PA2.Data
4 PA2.Control
3 PA1.Data
2 PA1.Control
1 PA0.Data
0 PA0.Control
23.7 Boundary-scan Description Language Files

Boundary-scan Description Language (BSDL) files describe Boundary-scan capable devices in
a standard format used by automated test-generation software. The order and function of bits in
the Boundary-scan Data Register are included in this description. BSDL files are available for
ATmega164P/324P/644P.
272 ATmega164P/324P/644P
8011D–AVR–02/07
The MCU Control Register contains control bits for general MCU functions.
Bit 7 6 5 4 3 2 1 0
• Bits 7 – JTD: JTAG Interface Disable

When this bit is zero, the JTAG interface is enabled if the JTAGEN Fuse is programmed. If this
bit is one, the JTAG interface is disabled. In order to avoid unintentional disabling or enabling of
the JTAG interface, a timed sequence must be followed when changing this bit: The application
software must write this bit to the desired value twice within four cycles to change its value. Note
that this bit must not be altered when using the On-chip Debug system.
23.8.2 MCUSR – MCU Status Register
The MCU Status Register provides information on which reset source caused an MCU reset.
Bit 7 6 5 4 3 2 1 0
0x34 (0x54) – – – JTRF WDRF BORF EXTRF PORF MCUSR
Initial Value 0 0 0 See Bit Description
• Bit 4 – JTRF: JTAG Reset Flag

This bit is set if a reset is being caused by a logic one in the JTAG Reset Register selected by
the JTAG instruction AVR_RESET. This bit is reset by a Power-on Reset, or by writing a logic
zero to the flag.
273
8011D–AVR–02/07
24. Boot Loader Support – Read-While-Write Self-Programming
24.1 Features
• Read-While-Write Self-Programming
• Flexible Boot Memory Size
• High Security (Separate Boot Lock Bits for a Flexible Protection)
• Separate Fuse to Select Reset Vector
• Optimized Page(1) Size
• Code Efficient Algorithm
• Efficient Read-Modify-Write Support
Note: 1. A page is a section in the Flash consisting of several bytes (see Table 25-7 on page 294) used
during programming. The page organization does not affect normal operation.
24.2 Overview
The Boot Loader Support provides a real Read-While-Write Self-Programming mechanism for
downloading and uploading program code by the MCU itself. This feature allows flexible applica-
tion software updates controlled by the MCU using a Flash-resident Boot Loader program. The
Boot Loader program can use any available data interface and associated protocol to read code
and write (program) that code into the Flash memory, or read the code from the program mem-
ory. The program code within the Boot Loader section has the capability to write into the entire
Flash, including the Boot Loader memory. The Boot Loader can thus even modify itself, and it
can also erase itself from the code if the feature is not needed anymore. The size of the Boot
Loader memory is configurable with fuses and the Boot Loader has two separate sets of Boot
Lock bits which can be set independently. This gives the user a unique flexibility to select differ-
ent levels of protection.
24.3 Application and Boot Loader Flash Sections

The Flash memory is organized in two main sections, the Application section and the Boot
Loader section (see Figure 24-2 on page 277). The size of the different sections is configured by
the BOOTSZ Fuses as shown in Table 24-7 on page 286 and Figure 24-2. These two sections
can have different level of protection since they have different sets of Lock bits.
24.3.1 Application Section
The Application section is the section of the Flash that is used for storing the application code.
The protection level for the Application section can be selected by the application Boot Lock bits
(Boot Lock bits 0), see Table 24-2 on page 278. The Application section can never store any
Boot Loader code since the SPM instruction is disabled when executed from the Application
section.
24.3.2 BLS – Boot Loader Section
While the Application section is used for storing the application code, the The Boot Loader soft-
ware must be located in the BLS since the SPM instruction can initiate a programming when
executing from the BLS only. The SPM instruction can access the entire Flash, including the
BLS itself. The protection level for the Boot Loader section can be selected by the Boot Loader
Lock bits (Boot Lock bits 1), see Table 24-3 on page 278.
274 ATmega164P/324P/644P
8011D–AVR–02/07
24.4 Read-While-Write and No Read-While-Write Flash Sections
Whether the CPU supports Read-While-Write or if the CPU is halted during a Boot Loader soft-
ware update is dependent on which address that is being programmed. In addition to the two
sections that are configurable by the BOOTSZ Fuses as described above, the Flash is also
divided into two fixed sections, the Read-While-Write (RWW) section and the No Read-While-
Write (NRWW) section. The limit between the RWW- and NRWW sections is given in Table 24-
1 and Figure 24-1 on page 276. The main difference between the two sections is:
• When erasing or writing a page located inside the RWW section, the NRWW section can be
read during the operation.
• When erasing or writing a page located inside the NRWW section, the CPU is halted during the
entire operation.
Note that the user software can never read any code that is located inside the RWW section dur-
ing a Boot Loader software operation. The syntax “Read-While-Write section” refers to which
section that is being programmed (erased or written), not which section that actually is being
read during a Boot Loader software update.
24.4.1 RWW – Read-While-Write Section
If a Boot Loader software update is programming a page inside the RWW section, it is possible
to read code from the Flash, but only code that is located in the NRWW section. During an on-
going programming, the software must ensure that the RWW section never is being read. If the
user software is trying to read code that is located inside the RWW section (i.e., by load program
memory, call, or jump instructions or an interrupt) during programming, the software might end
up in an unknown state. To avoid this, the interrupts should either be disabled or moved to the
Boot Loader section. The Boot Loader section is always located in the NRWW section. The
RWW Section Busy bit (RWWSB) in the Store Program Memory Control and Status Register
(SPMCSR) will be read as logical one as long as the RWW section is blocked for reading. After
a programming is completed, the RWWSB must be cleared by software before reading code
located in the RWW section. See Section “24.9.1” on page 289. for details on how to clear
RWWSB.
24.4.2 NRWW – No Read-While-Write Section
The code located in the NRWW section can be read when the Boot Loader software is updating
a page in the RWW section. When the Boot Loader code updates the NRWW section, the CPU
is halted during the entire Page Erase or Page Write operation.
Table 24-1. Read-While-Write Features

Which Section Can
Which Section does the Z-pointer be Read During Is the CPU Read-While-Write
Address During the Programming? Programming? Halted? Supported?
RWW Section NRWW Section No Yes
NRWW Section None Yes No
275
8011D–AVR–02/07
Figure 24-1. Read-While-Write vs. No Read-While-Write
Read-While-Write
(RWW) Section
Z-pointer
Addresses NRWW
Z-pointer Section
Addresses RWW No Read-While-Write
Section (NRWW) Section
CPU is Halted
During the Operation
Code Located in
NRWW Section
Can be Read During
the Operation
276 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 24-2. Memory Sections
Program Memory Program Memory
BOOTSZ = '11' BOOTSZ = '10'
0x0000 0x0000
Read-While-Write Section
Application Flash Section Application Flash Section
No Read-While-Write Section
End RWW End RWW
Start NRWW Start NRWW
End Application
End Application Start Boot Loader
Start Boot Loader Boot Loader Flash Section
Boot Loader Flash Section
Flashend Flashend
Program Memory Program Memory

BOOTSZ = '01' BOOTSZ = '00'
0x0000 0x0000
End RWW, End Application

End RWW
Start NRWW Start NRWW, Start Boot Loader
Application Flash Section
End Application
Start Boot Loader
Flashend Flashend
Note: 1. The parameters in the figure above are given in Table 24-7 on page 286.
24.5 Boot Loader Lock Bits
If no Boot Loader capability is needed, the entire Flash is available for application code. The
Boot Loader has two separate sets of Boot Lock bits which can be set independently. This gives
the user a unique flexibility to select different levels of protection.
The user can select:
• To protect the entire Flash from a software update by the MCU.
• To protect only the Boot Loader Flash section from a software update by the MCU.
• To protect only the Application Flash section from a software update by the MCU.
• Allow software update in the entire Flash.
See Table 24-2 on page 278 and Table 24-3 on page 278 for further details. The Boot Lock bits
can be set in software and in Serial or Parallel Programming mode, but they can be cleared by a
Chip Erase command only. The general Write Lock (Lock Bit mode 2) does not control the pro-
gramming of the Flash memory by SPM instruction. Similarly, the general Read/Write Lock
(Lock Bit mode 1) does not control reading nor writing by (E)LPM/SPM, if it is attempted.
277
8011D–AVR–02/07
Table 24-2. Boot Lock Bit0 Protection Modes (Application Section)(1)
BLB0 Mode BLB02 BLB01 Protection
No restrictions for SPM or (E)LPM accessing the Application
1 1 1
section.
2 1 0 SPM is not allowed to write to the Application section.
SPM is not allowed to write to the Application section, and
(E)LPM executing from the Boot Loader section is not allowed to
3 0 0 read from the Application section. If Interrupt Vectors are placed
in the Boot Loader section, interrupts are disabled while
executing from the Application section.
read from the Application section. If Interrupt Vectors are placed
4 0 1
Note: 1. “1” means unprogrammed, “0” means programmed
Table 24-3. Boot Lock Bit1 Protection Modes (Boot Loader Section)(1)
BLB1 Mode BLB12 BLB11 Protection
No restrictions for SPM or (E)LPM accessing the Boot Loader
1 1 1
section.
2 1 0 SPM is not allowed to write to the Boot Loader section.
SPM is not allowed to write to the Boot Loader section, and
(E)LPM executing from the Application section is not allowed to
3 0 0 read from the Boot Loader section. If Interrupt Vectors are
placed in the Application section, interrupts are disabled while
executing from the Boot Loader section.
read from the Boot Loader section. If Interrupt Vectors are
4 0 1

24.6 Entering the Boot Loader Program
Entering the Boot Loader takes place by a jump or call from the application program. This may
be initiated by a trigger such as a command received via USART, or SPI interface. Alternatively,
the Boot Reset Fuse can be programmed so that the Reset Vector is pointing to the Boot Flash
start address after a reset. In this case, the Boot Loader is started after a reset. After the applica-
tion code is loaded, the program can start executing the application code. Note that the fuses
cannot be changed by the MCU itself. This means that once the Boot Reset Fuse is pro-
grammed, the Reset Vector will always point to the Boot Loader Reset and the fuse can only be
changed through the serial or parallel programming interface.
Table 24-4. Boot Reset Fuse(1)
BOOTRST Reset Address
1 Reset Vector = Application Reset (address 0x0000)
0 Reset Vector = Boot Loader Reset (see Table 24-7 on page 286)
278 ATmega164P/324P/644P
8011D–AVR–02/07
24.7 Addressing the Flash During Self-Programming
The Z-pointer is used to address the SPM commands. The Z pointer consists of the Z-registers
ZL and ZH in the register file, and RAMPZ in the I/O space. The number of bits actually used is
implementation dependent. Note that the RAMPZ register is only implemented when the pro-
gram space is larger than 64K bytes.
Bit 23 22 21 20 19 18 17 16
15 14 13 12 11 10 9 8
RAMPZ RAMPZ7 RAMPZ6 RAMPZ5 RAMPZ4 RAMPZ3 RAMPZ2 RAMPZ1 RAMPZ0
ZH (R31) Z15 Z14 Z13 Z12 Z11 Z10 Z9 Z8
ZL (R30) Z7 Z6 Z5 Z4 Z3 Z2 Z1 Z0
7 6 5 4 3 2 1 0
Since the Flash is organized in pages (see Table 25-7 on page 294), the Program Counter can
be treated as having two different sections. One section, consisting of the least significant bits, is
addressing the words within a page, while the most significant bits are addressing the pages.
This is shown in Figure 24-3. Note that the Page Erase and Page Write operations are
addressed independently. Therefore it is of major importance that the Boot Loader software
addresses the same page in both the Page Erase and Page Write operation. Once a program-
ming operation is initiated, the address is latched and the Z-pointer can be used for other
operations.
The (E)LPM instruction use the Z-pointer to store the address. Since this instruction addresses
the Flash byte-by-byte, also bit Z0 of the Z-pointer is used.
Figure 24-3. Addressing the Flash During SPM(1)

BIT 15 ZPCMSB ZPAGEMSB 1 0
Z - REGISTER 0
PCMSB PAGEMSB
PROGRAM
PCPAGE PCWORD
COUNTER
PAGE ADDRESS WORD ADDRESS

WITHIN THE FLASH WITHIN A PAGE
PROGRAM MEMORY PAGE PCWORD[PAGEMSB:0]:

PAGE 00
INSTRUCTION WORD
01
02
PAGEEND
Note: 1. The different variables used in Figure 24-3 are listed in Table 24-9 on page 286.
279
8011D–AVR–02/07
24.8 Self-Programming the Flash
The program memory is updated in a page by page fashion. Before programming a page with
the data stored in the temporary page buffer, the page must be erased. The temporary page
buffer is filled one word at a time using SPM and the buffer can be filled either before the Page
Erase command or between a Page Erase and a Page Write operation:
Alternative 1, fill the buffer before a Page Erase
• Fill temporary page buffer
• Perform a Page Erase
• Perform a Page Write
Alternative 2, fill the buffer after Page Erase
• Perform a Page Erase
• Fill temporary page buffer
• Perform a Page Write
If only a part of the page needs to be changed, the rest of the page must be stored (for example
in the temporary page buffer) before the erase, and then be rewritten. When using alternative 1,
the Boot Loader provides an effective Read-Modify-Write feature which allows the user software
to first read the page, do the necessary changes, and then write back the modified data. If alter-
native 2 is used, it is not possible to read the old data while loading since the page is already
erased. The temporary page buffer can be accessed in a random sequence. It is essential that
the page address used in both the Page Erase and Page Write operation is addressing the same
page. See ”Simple Assembly Code Example for a Boot Loader” on page 284 for an assembly
code example.
24.8.1 Performing Page Erase by SPM
To execute Page Erase, set up the address in the Z-pointer, write “X0000011” to SPMCSR and
execute SPM within four clock cycles after writing SPMCSR. The data in R1 and R0 is ignored.
The page address must be written to PCPAGE in the Z-register. Other bits in the Z-pointer will
be ignored during this operation.
• Page Erase to the RWW section: The NRWW section can be read during the Page Erase.
• Page Erase to the NRWW section: The CPU is halted during the operation.
24.8.2 Filling the Temporary Buffer (Page Loading)
To write an instruction word, set up the address in the Z-pointer and data in R1:R0, write
“00000001” to SPMCSR and execute SPM within four clock cycles after writing SPMCSR. The
content of PCWORD in the Z-register is used to address the data in the temporary buffer. The
temporary buffer will auto-erase after a Page Write operation or by writing the RWWSRE bit in
SPMCSR. It is also erased after a system reset. Note that it is not possible to write more than
one time to each address without erasing the temporary buffer.
If the EEPROM is written in the middle of an SPM Page Load operation, all data loaded will be
lost.
24.8.3 Performing a Page Write
To execute Page Write, set up the address in the Z-pointer, write “X0000101” to SPMCSR and
execute SPM within four clock cycles after writing SPMCSR. The data in R1 and R0 is ignored.
280 ATmega164P/324P/644P
8011D–AVR–02/07
The page address must be written to PCPAGE. Other bits in the Z-pointer must be written to
zero during this operation.
• Page Write to the RWW section: The NRWW section can be read during the Page Write.
• Page Write to the NRWW section: The CPU is halted during the operation.
24.8.4 Using the SPM Interrupt
If the SPM interrupt is enabled, the SPM interrupt will generate a constant interrupt when the
SPMEN bit in SPMCSR is cleared. This means that the interrupt can be used instead of polling
the SPMCSR Register in software. When using the SPM interrupt, the Interrupt Vectors should
be moved to the BLS section to avoid that an interrupt is accessing the RWW section when it is
blocked for reading. How to move the interrupts is described in ”Interrupts” on page 61.
24.8.5 Consideration While Updating BLS
Special care must be taken if the user allows the Boot Loader section to be updated by leaving
Boot Lock bit11 unprogrammed. An accidental write to the Boot Loader itself can corrupt the
entire Boot Loader, and further software updates might be impossible. If it is not necessary to
change the Boot Loader software itself, it is recommended to program the Boot Lock bit11 to
protect the Boot Loader software from any internal software changes.
24.8.6 Prevent Reading the RWW Section During Self-Programming
During Self-Programming (either Page Erase or Page Write), the RWW section is always
blocked for reading. The user software itself must prevent that this section is addressed during
the self programming operation. The RWWSB in the SPMCSR will be set as long as the RWW
section is busy. During Self-Programming the Interrupt Vector table should be moved to the BLS
as described in ”Interrupts” on page 61, or the interrupts must be disabled. Before addressing
the RWW section after the programming is completed, the user software must clear the
RWWSB by writing the RWWSRE. See ”Simple Assembly Code Example for a Boot Loader” on
page 284 for an example.
24.8.7 Setting the Boot Loader Lock Bits by SPM
To set the Boot Loader Lock bits and general lock bits, write the desired data to R0, write
“X0001001” to SPMCSR and execute SPM within four clock cycles after writing SPMCSR.
Bit 7 6 5 4 3 2 1 0
R0 1 1 BLB12 BLB11 BLB02 BLB01 LB2 LB1
See Table 24-2 and Table 24-3 for how the different settings of the Boot Loader bits affect the
Flash access.
If bits 5..0 in R0 are cleared (zero), the corresponding Boot Lock bit will be programmed if an
SPM instruction is executed within four cycles after BLBSET and SPMEN are set in SPMCSR.
The Z-pointer is don’t care during this operation, but for future compatibility it is recommended to
load the Z-pointer with 0x0001 (same as used for reading the lOck bits). For future compatibility it
is also recommended to set bits 7 and 6 in R0 to “1” when writing the Lock bits. When program-
ming the Lock bits the entire Flash can be read during the operation.
24.8.8 EEPROM Write Prevents Writing to SPMCSR
Note that an EEPROM write operation will block all software programming to Flash. Reading the
Fuses and Lock bits from software will also be prevented during the EEPROM write operation. It
281
8011D–AVR–02/07
is recommended that the user checks the status bit (EEPE) in the EECR Register and verifies
that the bit is cleared before writing to the SPMCSR Register.
24.8.9 Reading the Fuse and Lock Bits from Software
It is possible to read both the Fuse and Lock bits from software. To read the Lock bits, load the
Z-pointer with 0x0001 and set the BLBSET and SPMEN bits in SPMCSR. When an (E)LPM
instruction is executed within three CPU cycles after the BLBSET and SPMEN bits are set in
SPMCSR, the value of the Lock bits will be loaded in the destination register. The BLBSET and
SPMEN bits will auto-clear upon completion of reading the Lock bits or if no (E)LPM instruction
is executed within three CPU cycles or no SPM instruction is executed within four CPU cycles.
When BLBSET and SPMEN are cleared, (E)LPM will work as described in the Instruction set
Manual.
Bit 7 6 5 4 3 2 1 0
Rd – – BLB12 BLB11 BLB02 BLB01 LB2 LB1
The algorithm for reading the Fuse Low byte is similar to the one described above for reading
the Lock bits. To read the Fuse Low byte, load the Z-pointer with 0x0000 and set the BLBSET
and SPMEN bits in SPMCSR. When an (E)LPM instruction is executed within three cycles after
the BLBSET and SPMEN bits are set in the SPMCSR, the value of the Fuse Low byte (FLB) will
be loaded in the destination register as shown below. Refer to Table 25-5 on page 293 for a
detailed description and mapping of the Fuse Low byte.
Bit 7 6 5 4 3 2 1 0
Rd FLB7 FLB6 FLB5 FLB4 FLB3 FLB2 FLB1 FLB0
Similarly, when reading the Fuse High byte, load 0x0003 in the Z-pointer. When an (E)LPM
instruction is executed within three cycles after the BLBSET and SPMEN bits are set in the
SPMCSR, the value of the Fuse High byte (FHB) will be loaded in the destination register as
shown below. Refer to Table 25-4 on page 293 for detailed description and mapping of the Fuse
High byte.
Bit 7 6 5 4 3 2 1 0
Rd FHB7 FHB6 FHB5 FHB4 FHB3 FHB2 FHB1 FHB0
When reading the Extended Fuse byte, load 0x0002 in the Z-pointer. When an (E)LPM instruc-
tion is executed within three cycles after the BLBSET and SPMEN bits are set in the SPMCSR,
the value of the Extended Fuse byte (EFB) will be loaded in the destination register as shown
below. Refer to Table 25-3 on page 292 for detailed description and mapping of the Extended
Fuse byte.
Bit 7 6 5 4 3 2 1 0
Rd – – – – – EFB2 EFB1 EFB0
Fuse and Lock bits that are programmed, will be read as zero. Fuse and Lock bits that are
unprogrammed, will be read as one.
24.8.10 Reading the Signature Row from Software
To read the Signature Row from software, load the Z-pointer with the signature byte address
given in Table 24-5 on page 283 and set the SIGRD and SPMEN bits in SPMCSR. When an
LPM instruction is executed within three CPU cycles after the SIGRD and SPMEN bits are set in
SPMCSR, the signature byte value will be loaded in the destination register. The SIGRD and
SPMEN bits will auto-clear upon completion of reading the Signature Row Lock bits or if no LPM
282 ATmega164P/324P/644P
8011D–AVR–02/07
instruction is executed within three CPU cycles. When SIGRD and SPMEN are cleared, LPM will
work as described in the Instruction set Manual.
Table 24-5. Signature Row Addressing

Signature Byte Z-Pointer Address
Device Signature Byte 1 0x0000
RC Oscillator Calibration Byte 0x0001
Note: All other addresses are reserved for future use.
24.8.11 Preventing Flash Corruption
During periods of low VCC, the Flash program can be corrupted because the supply voltage is
too low for the CPU and the Flash to operate properly. These issues are the same as for board
level systems using the Flash, and the same design solutions should be applied.
A Flash program corruption can be caused by two situations when the voltage is too low. First, a
regular write sequence to the Flash requires a minimum voltage to operate correctly. Secondly,
the CPU itself can execute instructions incorrectly, if the supply voltage for executing instructions
is too low.
Flash corruption can easily be avoided by following these design recommendations (one is
sufficient):
1. If there is no need for a Boot Loader update in the system, program the Boot Loader Lock
bits to prevent any Boot Loader software updates.
2. Keep the AVR RESET active (low) during periods of insufficient power supply voltage.
This can be done by enabling the internal Brown-out Detector (BOD) if the operating volt-
age matches the detection level. If not, an external low VCC reset protection circuit can be
used. If a reset occurs while a write operation is in progress, the write operation will be
completed provided that the power supply voltage is sufficient.
3. Keep the AVR core in Power-down sleep mode during periods of low VCC. This will pre-
vent the CPU from attempting to decode and execute instructions, effectively protecting
the SPMCSR Register and thus the Flash from unintentional writes.
24.8.12 Programming Time for Flash when Using SPM
The calibrated RC Oscillator is used to time Flash accesses. Table 24-6 on page 283 shows the
typical programming time for Flash accesses from the CPU.
Table 24-6. SPM Programming Time(1)

Symbol Min Programming Time Max Programming Time
Flash write (Page Erase, Page Write, and
3.7 ms 4.5 ms
write Lock bits by SPM)
Note: 1. Minimum and maximum programming times is per individual operation.
283
8011D–AVR–02/07
24.8.13 Simple Assembly Code Example for a Boot Loader
;-the routine writes one page of data from RAM to Flash

; the first data location in RAM is pointed to by the Y pointer
; the first data location in Flash is pointed to by the Z-pointer
;-error handling is not included
;-the routine must be placed inside the Boot space
; (at least the Do_spm sub routine). Only code inside NRWW section can
; be read during Self-Programming (Page Erase and Page Write).
;-registers used: r0, r1, temp1 (r16), temp2 (r17), looplo (r24),
; loophi (r25), spmcrval (r20)
; storing and restoring of registers is not included in the routine
; register usage can be optimized at the expense of code size
;-It is assumed that either the interrupt table is moved to the Boot
; loader section or that the interrupts are disabled.
.equ PAGESIZEB = PAGESIZE*2 ;PAGESIZEB is page size in BYTES, not words
.org SMALLBOOTSTART
Write_page:
; Page Erase
ldi spmcrval, (1<<PGERS) | (1<<SPMEN)
call Do_spm
; re-enable the RWW section

ldi spmcrval, (1<<RWWSRE) | (1<<SPMEN)
call Do_spm
; transfer data from RAM to Flash page buffer

ldi looplo, low(PAGESIZEB) ;init loop variable
ldi loophi, high(PAGESIZEB) ;not required for PAGESIZEB<=256
Wrloop:
ld r0, Y+
ld r1, Y+
ldi spmcrval, (1<<SPMEN)
call Do_spm
adiw ZH:ZL, 2
sbiw loophi:looplo, 2 ;use subi for PAGESIZEB<=256
brne Wrloop
; execute Page Write

subi ZL, low(PAGESIZEB) ;restore pointer
sbci ZH, high(PAGESIZEB) ;not required for PAGESIZEB<=256
ldi spmcrval, (1<<PGWRT) | (1<<SPMEN)
call Do_spm

call Do_spm
; read back and check, optional

ldi looplo, low(PAGESIZEB) ;init loop variable
ldi loophi, high(PAGESIZEB) ;not required for PAGESIZEB<=256
subi YL, low(PAGESIZEB) ;restore pointer
sbci YH, high(PAGESIZEB)
Rdloop:
elpm r0, Z+
ld r1, Y+
cpse r0, r1
jmp Error
284 ATmega164P/324P/644P
8011D–AVR–02/07
sbiw loophi:looplo, 1 ;use subi for PAGESIZEB<=256
brne Rdloop
; return to RWW section

; verify that RWW section is safe to read
Return:
in temp1, SPMCSR
sbrs temp1, RWWSB ; If RWWSB is set, the RWW section is not ready yet
ret
call Do_spm
rjmp Return
Do_spm:
; check for previous SPM complete
Wait_spm:
in temp1, SPMCSR
sbrc temp1, SPMEN
rjmp Wait_spm
; input: spmcrval determines SPM action
; disable interrupts if enabled, store status
in temp2, SREG
cli
; check that no EEPROM write access is present
Wait_ee:
sbic EECR, EEPE
rjmp Wait_ee
; SPM timed sequence
out SPMCSR, spmcrval
spm
; restore SREG (to enable interrupts if originally enabled)
out SREG, temp2
ret
285
8011D–AVR–02/07
24.8.14 ATmega164P Boot Loader Parameters
In Table 24-7 through Table 24-9, the parameters used in the description of the Self-Programming are given.
Table 24-7. Boot Size Configuration(1)

Boot Reset
End Address
Application Boot Loader Application (Start Boot
BOOTSZ1 BOOTSZ0 Boot Size Pages Flash Section Flash Section Section Loader Section)
1 1 128 words 2 0x0000 - 0x1F7F 0x1F80 - 0x1FFF 0x1F7F 0x1F80
1 0 256 words 4 0x0000 - 0x1EFF 0x1F00 - 0x1FFF 0x1EFF 0x1F00
0 1 512 words 8 0x0000 - 0x1DFF 0x1E00 - 0x1FFF 0x1DFF 0x1E00
0 0 1024 words 16 0x0000 - 0x1BFF 0x1C00 - 0x1FFF 0x1BFF 0x1C00
Note: 1. The different BOOTSZ Fuse configurations are shown in Figure 24-2 on page 277.
Table 24-8. Read-While-Write Limit(1)
Section Pages Address
Read-While-Write section (RWW) 112 0x0000 - 0x1BFF
No Read-While-Write section (NRWW) 16 0x1C00 - 0x7FFF
Note: 1. For details about these two section, see ”NRWW – No Read-While-Write Section” on page 275 and ”RWW – Read-While-
Write Section” on page 275.
Table 24-9. Explanation of different variables used in Figure 24-3 on page 279 and the mapping to the Z-pointer
Corresponding
Variable Z-value Description(1)
Most significant bit in the Program Counter. (The Program Counter is 13 bits
PCMSB 12
PC[12:0])
Most significant bit which is used to address the words within one page (128
PAGEMSB 5
words in a page requires seven bits PC [5:0]).
Bit in Z-pointer that is mapped to PCMSB. Because Z0 is not used, the
ZPCMSB Z13
ZPCMSB equals PCMSB + 1.
ZPAGEMSB Z6
ZPAGEMSB equals PAGEMSB + 1.
PCPAGE PC[12:6] Z14:Z7 Program Counter page address: Page select, for Page Erase and Page Write
Program Counter word address: Word select, for filling temporary buffer (must
PCWORD PC[5:0] Z6:Z1
be zero during Page Write operation)
Note: 1. Z0: should be zero for all SPM commands, byte select for the (E)LPM instruction.
See ”Addressing the Flash During Self-Programming” on page 279 for details about the use of Z-pointer during Self-
Programming.
286 ATmega164P/324P/644P
8011D–AVR–02/07

Boot Reset
End Address
Application Boot Loader Application (Start Boot
1 1 256 words 4 0x0000 - 0x3EFF 0x3F00 - 0x3FFF 0x3EFF 0x3F00
0 0 2048 words 32 0x0000 - 0x37FF 0x3800 - 0x3FFF 0x37FF 0x3800
Read-While-Write section (RWW) 224 0x0000 - 0x37FF
No Read-While-Write section (NRWW) 32 0x3800 - 0x3FFF
Table 24-12. Explanation of different variables used in Figure 24-3 on page 279 and the mapping to the Z-pointer
Corresponding
PCMSB 13
PC[13:0])
PAGEMSB 6
ZPCMSB Z14
ZPCMSB equals PCMSB + 1.
ZPAGEMSB Z7
Program Counter word address: Word select, for filling temporary buffer (must
be zero during Page Write operation)
Note: Z0: should be zero for all SPM commands, byte select for the (E)LPM instruction.
Programming.
287
8011D–AVR–02/07

Boot Reset
Address
Application Boot Loader End Application (Start Boot
0 1 2048 words 16 0x0000 - 0x77FF 0x7800 - 0x7FFF 0x77FF 0x7800
0 0 4096 words 32 0x0000 - 0x6FFF 0x7000 - 0x7FFF 0x6FFF 0x7000
Read-While-Write section (RWW) 224 0x0000 - 0xEFFF
No Read-While-Write section (NRWW) 32 0xF000 - 0xFFFF
Table 24-15.
Explanation of different variables used in Figure 24-3 on page 279 and the mapping to the Z-pointer
Correspondi
ng
PCMSB 14
PC[14:0])
PAGEMSB 7
Bit in Z-pointer that is mapped to PCMSB. Because Z0 is not used, the ZPCMSB
ZPCMSB Z15
equals PCMSB + 1.
ZPAGEMSB Z8
Program Counter word address: Word select, for filling temporary buffer (must be
zero during Page Write operation)
Note: 1. Z0: should be zero for all SPM commands, byte select for the (E)LPM instruction.
Programming.
288 ATmega164P/324P/644P
8011D–AVR–02/07
24.9.1 SPMCSR – Store Program Memory Control and Status Register
The Store Program Memory Control and Status Register contains the control bits needed to con-
trol the Boot Loader operations.
Bit 7 6 5 4 3 2 1 0
0x37 (0x57) SPMIE RWWSB SIGRD RWWSRE BLBSET PGWRT PGERS SPMEN SPMCSR
Read/Write R/W R R/W R/W R/W R/W R/W R/W
• Bit 7 – SPMIE: SPM Interrupt Enable

When the SPMIE bit is written to one, and the I-bit in the Status Register is set (one), the SPM
ready interrupt will be enabled. The SPM ready Interrupt will be executed as long as the SPMEN
bit in the SPMCSR Register is cleared.
• Bit 6 – RWWSB: Read-While-Write Section Busy

When a Self-Programming (Page Erase or Page Write) operation to the RWW section is initi-
ated, the RWWSB will be set (one) by hardware. When the RWWSB bit is set, the RWW section
cannot be accessed. The RWWSB bit will be cleared if the RWWSRE bit is written to one after a
Self-Programming operation is completed. Alternatively the RWWSB bit will automatically be
cleared if a page load operation is initiated.
• Bit 5 – SIGRD: Signature Row Read

If this bit is written to one at the same time as SPMEN, the next LPM instruction within three
clock cycles will read a byte from the signature row into the destination register. see ”Reading
the Signature Row from Software” on page 282 for details. An SPM instruction within four cycles
after SIGRD and SPMEN are set will have no effect. This operation is reserved for future use
and should not be used.
• Bit 4 – RWWSRE: Read-While-Write Section Read Enable

When programming (Page Erase or Page Write) to the RWW section, the RWW section is
blocked for reading (the RWWSB will be set by hardware). To re-enable the RWW section, the
user software must wait until the programming is completed (SPMEN will be cleared). Then, if
the RWWSRE bit is written to one at the same time as SPMEN, the next SPM instruction within
four clock cycles re-enables the RWW section. The RWW section cannot be re-enabled while
the Flash is busy with a Page Erase or a Page Write (SPMEN is set). If the RWWSRE bit is writ-
ten while the Flash is being loaded, the Flash load operation will abort and the data loaded will
be lost.
• Bit 3 – BLBSET: Boot Lock Bit Set

If this bit is written to one at the same time as SPMEN, the next SPM instruction within four clock
cycles sets Boot Lock bits, according to the data in R0. The data in R1 and the address in the Z-
pointer are ignored. The BLBSET bit will automatically be cleared upon completion of the Lock
bit set, or if no SPM instruction is executed within four clock cycles.
An (E)LPM instruction within three cycles after BLBSET and SPMEN are set in the SPMCSR
Register, will read either the Lock bits or the Fuse bits (depending on Z0 in the Z-pointer) into the
destination register. See ”Reading the Fuse and Lock Bits from Software” on page 282 for
details.
289
8011D–AVR–02/07
• Bit 2 – PGWRT: Page Write
cycles executes Page Write, with the data stored in the temporary buffer. The page address is
taken from the high part of the Z-pointer. The data in R1 and R0 are ignored. The PGWRT bit
will auto-clear upon completion of a Page Write, or if no SPM instruction is executed within four
clock cycles. The CPU is halted during the entire Page Write operation if the NRWW section is
addressed.
• Bit 1 – PGERS: Page Erase

cycles executes Page Erase. The page address is taken from the high part of the Z-pointer. The
data in R1 and R0 are ignored. The PGERS bit will auto-clear upon completion of a Page Erase,
or if no SPM instruction is executed within four clock cycles. The CPU is halted during the entire
Page Write operation if the NRWW section is addressed.
• Bit 0 – SPMEN: Store Program Memory Enable

This bit enables the SPM instruction for the next four clock cycles. If written to one together with
either RWWSRE, BLBSET, PGWRT’ or PGERS, the following SPM instruction will have a spe-
cial meaning, see description above. If only SPMEN is written, the following SPM instruction will
store the value in R1:R0 in the temporary page buffer addressed by the Z-pointer. The LSB of
the Z-pointer is ignored. The SPMEN bit will auto-clear upon completion of an SPM instruction,
or if no SPM instruction is executed within four clock cycles. During Page Erase and Page Write,
the SPMEN bit remains high until the operation is completed.
Writing any other combination than “10001”, “01001”, “00101”, “00011” or “00001” in the lower
five bits will have no effect.
Note: Only one SPM instruction should be active at any time.
290 ATmega164P/324P/644P
8011D–AVR–02/07
25. Memory Programming

25.1 Program And Data Memory Lock Bits
The ATmega164P/324P/644P provides six Lock bits which can be left unprogrammed (“1”) or
can be programmed (“0”) to obtain the additional features listed in Table 25-2. The Lock bits can
only be erased to “1” with the Chip Erase command.
Table 25-1. Lock Bit Byte(1)

Lock Bit Byte Bit No Description Default Value
7 – 1 (unprogrammed)
6 – 1 (unprogrammed)
BLB12 5 Boot Lock bit 1 (unprogrammed)
LB2 1 Lock bit 1 (unprogrammed)
LB1 0 Lock bit 1 (unprogrammed)
Table 25-2. Lock Bit Protection Modes(1)(2)

Memory Lock Bits Protection Type
LB Mode LB2 LB1
1 1 1 No memory lock features enabled.
Further programming of the Flash and EEPROM is disabled in
2 1 0 Parallel and Serial Programming mode. The Fuse bits are
locked in both Serial and Parallel Programming mode.(1)
Further programming and verification of the Flash and EEPROM
is disabled in Parallel and Serial Programming mode. The Boot
3 0 0
Lock bits and Fuse bits are locked in both Serial and Parallel
Programming mode.(1)
BLB0 Mode BLB02 BLB01
No restrictions for SPM or (E)LPM accessing the Application
1 1 1
section.
2 1 0 SPM is not allowed to write to the Application section.
SPM is not allowed to write to the Application section, and
3 0 0 read from the Application section. If Interrupt Vectors are placed
read from the Application section. If Interrupt Vectors are placed
4 0 1
291
8011D–AVR–02/07
Table 25-2. Lock Bit Protection Modes(1)(2) (Continued)
Memory Lock Bits Protection Type
BLB1 Mode BLB12 BLB11
No restrictions for SPM or (E)LPM accessing the Boot Loader
1 1 1
section.
2 1 0 SPM is not allowed to write to the Boot Loader section.
SPM is not allowed to write to the Boot Loader section, and
3 0 0 read from the Boot Loader section. If Interrupt Vectors are
read from the Boot Loader section. If Interrupt Vectors are
4 0 1
Notes: 1. Program the Fuse bits and Boot Lock bits before programming the LB1 and LB2.
2. “1” means unprogrammed, “0” means programmed
25.2 Fuse Bits
The ATmega164P/324P/644P has four Fuse bytes. Table 25-3 - Table 25-5 describe briefly the
functionality of all the fuses and how they are mapped into the Fuse bytes. Note that the fuses
are read as logical zero, “0”, if they are programmed.
Table 25-3. Extended Fuse Byte

Fuse Low Byte Bit No Description Default Value
– 7 – 1
– 6 – 1
– 5 – 1
– 4 – 1
– 3 – 1
BODLEVEL2(1) 2 Brown-out Detector trigger level 1 (unprogrammed)
BODLEVEL1(1) 1 Brown-out Detector trigger level 1 (unprogrammed)
(1)
BODLEVEL0 0 Brown-out Detector trigger level 1 (unprogrammed)
Note: 1. See ”System and Reset Characteristics” on page 329 for BODLEVEL Fuse decoding.
292 ATmega164P/324P/644P
8011D–AVR–02/07
Table 25-4. Fuse High Byte
Fuse High Byte Bit No Description Default Value
1 (unprogrammed, OCD
OCDEN(4) 7 Enable OCD
disabled)
JTAGEN 6 Enable JTAG 0 (programmed, JTAG enabled)
Enable Serial Program and Data 0 (programmed, SPI prog.
SPIEN(1) 5
Downloading enabled)
WDTON(3) 4 Watchdog Timer always on 1 (unprogrammed)
EEPROM memory is preserved 1 (unprogrammed, EEPROM
EESAVE 3
through the Chip Erase not preserved)
Select Boot Size (see Table 25-9 for
BOOTSZ1 2 0 (programmed)(2)
details)
Select Boot Size (see Table 25-9 for
BOOTSZ0 1 0 (programmed)(2)
details)
BOOTRST 0 Select Reset Vector 1 (unprogrammed)
Note: 1. The SPIEN Fuse is not accessible in serial programming mode.

2. The default value of BOOTSZ1..0 results in maximum Boot Size. See Table 24-7 on page 286
for details.
3. See ”WDTCSR – Watchdog Timer Control Register” on page 59 for details.
4. Never ship a product with the OCDEN Fuse programmed regardless of the setting of Lock bits
and JTAGEN Fuse. A programmed OCDEN Fuse enables some parts of the clock system to
be running in all sleep modes. This may increase the power consumption.
Table 25-5. Fuse Low Byte

Fuse Low Byte Bit No Description Default Value
CKDIV8(4) 7 Divide clock by 8 0 (programmed)
(3)
CKOUT 6 Clock output 1 (unprogrammed)
SUT1 5 Select start-up time 1 (unprogrammed)(1)
SUT0 4 Select start-up time 0 (programmed)(1)
CKSEL3 3 Select Clock source 0 (programmed)(2)
CKSEL1 1 Select Clock source 1 (unprogrammed)(2)
Note: 1. The default value of SUT1..0 results in maximum start-up time for the default clock source.
See ”System and Reset Characteristics” on page 329 for details.
2. The default setting of CKSEL3..0 results in internal RC Oscillator @ 8 MHz. See Table 7-1 on
3. The CKOUT Fuse allow the system clock to be output on PORTB1. See ”Clock Output Buffer”
on page 38 for details.
4. See ”System Clock Prescaler” on page 38 for details.
The status of the Fuse bits is not affected by Chip Erase. Note that the Fuse bits are locked if
Lock bit1 (LB1) is programmed. Program the Fuse bits before programming the Lock bits.
293
8011D–AVR–02/07
25.2.1 Latching of Fuses
The fuse values are latched when the device enters programming mode and changes of the
fuse values will have no effect until the part leaves Programming mode. This does not apply to
the EESAVE Fuse which will take effect once it is programmed. The fuses are also latched on
Power-up in Normal mode.
25.3 Signature Bytes

All Atmel microcontrollers have a three-byte signature code which identifies the device. This
code can be read in both serial and parallel mode, also when the device is locked. The three
bytes reside in a separate address space. For the ATmega164P/324P/644P the signature bytes
are given in Table 25-6.
Table 25-6. Device and JTAG ID

Signature Bytes Address JTAG
Part 0x000 0x001 0x002 Part Number Manufacture ID
ATmega164P 0x1E 0x94 0x0A 940A 0x1F
ATmega324P 0x1E 0x95 0x08 9508 0x1F
ATmega644P 0x1E 0x96 0x0A 960A 0x1F
25.4 Calibration Byte

The ATmega164P/324P/644P has a byte calibration value for the internal RC Oscillator. This
byte resides in the high byte of address 0x000 in the signature address space. During reset, this
byte is automatically written into the OSCCAL Register to ensure correct frequency of the cali-
brated RC Oscillator.
25.5 Page Size
Table 25-7. No. of Words in a Page and No. of Pages in the Flash
Device Flash Size Page Size PCWORD No. of Pages PCPAGE PCMSB
ATmega164P 8K words (16 Kbytes) 64 words PC[5:0] 128 PC[12:6] 12
Table 25-8. No. of Words in a Page and No. of Pages in the EEPROM
Device EEPROM Size Page Size PCWORD No. of Pages PCPAGE EEAMSB
ATmega164P 512 bytes 4 bytes EEA[1:0] 128 EEA[8:2] 8
ATmega324P 1 Kbytes 4 bytes EEA[1:0] 256 EEA[9:2] 9
ATmega644P 2 Kbytes 8 bytes EEA[2:0] 256 EEA[10:3] 10
294 ATmega164P/324P/644P
8011D–AVR–02/07
25.6 Parallel Programming Parameters, Pin Mapping, and Commands
This section describes how to parallel program and verify Flash Program memory, EEPROM
Data memory, Memory Lock bits, and Fuse bits in the ATmega164P/324P/644P. Pulses are
assumed to be at least 250 ns unless otherwise noted.
25.6.1 Signal Names
In this section, some pins of the ATmega164P/324P/644P are referenced by signal names
describing their functionality during parallel programming, see Figure 25-1 on page 295 and Fig-
ure 25-9 on page 295. Pins not described in the following table are referenced by pin names.
The XA1/XA0 pins determine the action executed when the XTAL1 pin is given a positive pulse.
The bit coding is shown in Table 25-12 on page 296.
When pulsing WR or OE, the command loaded determines the action executed. The different
commands are shown in Table 25-13 on page 296.
Figure 25-1. Parallel Programming(1)

+5V
RDY/BSY PD1
VCC
OE PD2 +5V
WR PD3
AVCC
BS1 PD4
XA0 PB7 - PB0 DATA

PD5
XA1 PD6
PAGEL PD7
+12 V RESET
BS2 PA0
XTAL1
GND
Note: 1. Unused Pins should be left floating.

Table 25-9. Pin Name Mapping
Signal Name in
Programming Mode Pin Name I/O Function
0: Device is busy programming, 1: Device is ready for
RDY/BSY PD1 O
new command.
OE PD2 I Output Enable (Active low).
WR PD3 I Write Pulse (Active low).
BS1 PD4 I Byte Select 1.
XA0 PD5 I XTAL Action Bit 0
XA1 PD6 I XTAL Action Bit 1
295
8011D–AVR–02/07
Table 25-9. Pin Name Mapping
Signal Name in
Programming Mode Pin Name I/O Function
PAGEL PD7 I Program Memory and EEPROM data Page Load.
BS2 PA0 I Byte Select 2.
DATA PB7-0 I/O Bi-directional Data bus (Output when OE is low).
Table 25-10. BS2 and BS1 Encoding

Flash Data
Flash / EEPROM Loading / Fuse Reading Fuse
BS2 BS1 Address Reading Programming and Lock Bits
0 0 Low Byte Low Byte Low Byte Fuse Low Byte
0 1 High Byte High Byte High Byte Lockbits
Extended High Extended Fuse
1 0 Reserved Extended Byte
Byte Byte
1 1 Reserved Reserved Reserved Fuse High Byte
,
Table 25-11. Pin Values Used to Enter Programming Mode
Pin Symbol Value
PAGEL Prog_enable[3] 0
XA1 Prog_enable[2] 0
XA0 Prog_enable[1] 0
BS1 Prog_enable[0] 0
Table 25-12. XA1 and XA0 Enoding

XA1 XA0 Action when XTAL1 is Pulsed
Load Flash or EEPROM Address (High or low address byte determined
0 0
by BS2 and BS1).
0 1 Load Data (High or Low data byte for Flash determined by BS1).
1 0 Load Command
1 1 No Action, Idle
Table 25-13. Command Byte Bit Encoding

Command Byte Command Executed
1000 0000 Chip Erase
0100 0000 Write Fuse bits
0010 0000 Write Lock bits
0001 0000 Write Flash
0001 0001 Write EEPROM
296 ATmega164P/324P/644P
8011D–AVR–02/07
Table 25-13. Command Byte Bit Encoding
Command Byte Command Executed
0000 1000 Read Signature Bytes and Calibration byte
0000 0100 Read Fuse and Lock bits
0000 0010 Read Flash
0000 0011 Read EEPROM
25.7 Parallel Programming

25.7.1 Enter Programming Mode
The following algorithm puts the device in parallel programming mode:

1. Apply 4.5 - 5.5V between VCC and GND.
2. Set RESET to “0” and toggle XTAL1 at least six times.
3. Set the Prog_enable pins listed in Table 25-11 on page 296 to “0000” and wait at least
100 ns.
4. Apply 11.5 - 12.5V to RESET. Any activity on Prog_enable pins within 100 ns after +12V
has been applied to RESET, will cause the device to fail entering programming mode.
5. Wait at least 50 µs before sending a new command.
25.7.2 Considerations for Efficient Programming
The loaded command and address are retained in the device during programming. For efficient
programming, the following should be considered.
• The command needs only be loaded once when writing or reading multiple memory locations.
• Skip writing the data value 0xFF, that is the contents of the entire EEPROM (unless the
EESAVE Fuse is programmed) and Flash after a Chip Erase.
• Address high byte needs only be loaded before programming or reading a new 256 word
window in Flash or 256 byte EEPROM. This consideration also applies to Signature bytes
reading.
25.7.3 Chip Erase
The Chip Erase will erase the Flash and EEPROM(1) memories plus Lock bits. The Lock bits are
not reset until the program memory has been completely erased. The Fuse bits are not
changed. A Chip Erase must be performed before the Flash and/or EEPROM are
reprogrammed.
Note: 1. The EEPRPOM memory is preserved during Chip Erase if the EESAVE Fuse is programmed.
Load Command “Chip Erase”
1. Set XA1, XA0 to “10”. This enables command loading.
2. Set BS1 to “0”.
3. Set DATA to “1000 0000”. This is the command for Chip Erase.
4. Give XTAL1 a positive pulse. This loads the command.
5. Give WR a negative pulse. This starts the Chip Erase. RDY/BSY goes low.
6. Wait until RDY/BSY goes high before loading a new command.
297
8011D–AVR–02/07
25.7.4 Programming the Flash
The Flash is organized in pages, see Table 25-7 on page 294. When programming the Flash,
the program data is latched into a page buffer. This allows one page of program data to be pro-
grammed simultaneously. The following procedure describes how to program the entire Flash
memory:
A. Load Command “Write Flash”
1. Set XA1, XA0 to “10”. This enables command loading.
2. Set BS1 to “0”.
3. Set DATA to “0001 0000”. This is the command for Write Flash.
4. Give XTAL1 a positive pulse. This loads the command.
B. Load Address Low byte (Address bits 7..0)
1. Set XA1, XA0 to “00”. This enables address loading.
2. Set BS2, BS1 to “00”. This selects the address low byte.
3. Set DATA = Address low byte (0x00 - 0xFF).
4. Give XTAL1 a positive pulse. This loads the address low byte.
C. Load Data Low Byte
1. Set XA1, XA0 to “01”. This enables data loading.
2. Set DATA = Data low byte (0x00 - 0xFF).
3. Give XTAL1 a positive pulse. This loads the data byte.
D. Load Data High Byte
1. Set BS1 to “1”. This selects high data byte.
2. Set XA1, XA0 to “01”. This enables data loading.
3. Set DATA = Data high byte (0x00 - 0xFF).
4. Give XTAL1 a positive pulse. This loads the data byte.
E. Latch Data
1. Set BS1 to “1”. This selects high data byte.
2. Give PAGEL a positive pulse. This latches the data bytes. (See Figure 25-3 on page 300
for signal waveforms)
F. Repeat B through E until the entire buffer is filled or until all data within the page is loaded.
While the lower bits in the address are mapped to words within the page, the higher bits address
the pages within the FLASH. This is illustrated in Figure 25-2 on page 299. Note that if less than
eight bits are required to address words in the page (pagesize < 256), the most significant bit(s)
in the address low byte are used to address the page when performing a Page Write.
G. Load Address High byte (Address bits15..8)
2. Set BS2, BS1 to “01”. This selects the address high byte.
3. Set DATA = Address high byte (0x00 - 0xFF).
4. Give XTAL1 a positive pulse. This loads the address high byte.
H. Load Address Extended High byte (Address bits 23..16)
2. Set BS2, BS1 to “10”. This selects the address extended high byte.
298 ATmega164P/324P/644P
8011D–AVR–02/07
3. Set DATA = Address extended high byte (0x00 - 0xFF).
4. Give XTAL1 a positive pulse. This loads the address high byte.
I. Program Page
1. Set BS2, BS1 to “00”
2. Give WR a negative pulse. This starts programming of the entire page of data. RDY/BSY
goes low.
3. Wait until RDY/BSY goes high (See Figure 25-3 on page 300 for signal waveforms).
J. Repeat B through I until the entire Flash is programmed or until all data has been
programmed.
K. End Page Programming
1. 1. Set XA1, XA0 to “10”. This enables command loading.
2. Set DATA to “0000 0000”. This is the command for No Operation.
3. Give XTAL1 a positive pulse. This loads the command, and the internal write signals are
reset.
Figure 25-2. Addressing the Flash Which is Organized in Pages(1)

PCMSB PAGEMSB
PROGRAM
PCPAGE PCWORD
COUNTER
PAGE ADDRESS WORD ADDRESS

WITHIN THE FLASH WITHIN A PAGE
PROGRAM MEMORY PAGE PCWORD[PAGEMSB:0]:

PAGE 00
INSTRUCTION WORD
01
02
PAGEEND
Note: 1. PCPAGE and PCWORD are listed in Table 25-7 on page 294.
299
8011D–AVR–02/07
Figure 25-3. Programming the Flash Waveforms(1)
F
A B C D E B C D E G H I
0x10 ADDR. LOW DATA LOW DATA HIGH XX ADDR. LOW DATA LOW DATA HIGH XX ADDR. HIGH ADDR. EXT.H XX
DATA
XA1
XA0
BS1
BS2
XTAL1
WR
RDY/BSY
RESET +12V
OE
PAGEL
Note: 1. “XX” is don’t care. The letters refer to the programming description above.
25.7.5 Programming the EEPROM
The EEPROM is organized in pages, see Table 25-8 on page 294. When programming the
EEPROM, the program data is latched into a page buffer. This allows one page of data to be
programmed simultaneously. The programming algorithm for the EEPROM data memory is as
follows (refer to ”Programming the Flash” on page 298 for details on Command, Address and
Data loading):
1. A: Load Command “0001 0001”.
2. G: Load Address High Byte (0x00 - 0xFF).
3. B: Load Address Low Byte (0x00 - 0xFF).
4. C: Load Data (0x00 - 0xFF).
5. E: Latch data (give PAGEL a positive pulse).
K: Repeat 3 through 5 until the entire buffer is filled.
L: Program EEPROM page
1. Set BS2, BS1 to “00”.
2. Give WR a negative pulse. This starts programming of the EEPROM page. RDY/BSY
goes low.
3. Wait until to RDY/BSY goes high before programming the next page (See Figure 25-4 for
signal waveforms).
300 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 25-4. Programming the EEPROM Waveforms
K
A G B C E B C E L
0x11 ADDR. HIGH ADDR. LOW DATA XX ADDR. LOW DATA XX
DATA
XA1
XA0
BS1
XTAL1
WR
RDY/BSY
RESET +12V
OE
PAGEL
BS2
25.7.6 Reading the Flash
The algorithm for reading the Flash memory is as follows (refer to ”Programming the Flash” on
page 298 for details on Command and Address loading):
1. A: Load Command “0000 0010”.
2. H: Load Address Extended Byte (0x00- 0xFF).
5. Set OE to “0”, and BS1 to “0”. The Flash word low byte can now be read at DATA.
6. Set BS to “1”. The Flash word high byte can now be read at DATA.
7. Set OE to “1”.
25.7.7 Reading the EEPROM
The algorithm for reading the EEPROM memory is as follows (refer to ”Programming the Flash”
on page 298 for details on Command and Address loading):
1. A: Load Command “0000 0011”.
4. Set OE to “0”, and BS1 to “0”. The EEPROM Data byte can now be read at DATA.
25.7.8 Programming the Fuse Low Bits
The algorithm for programming the Fuse Low bits is as follows (refer to ”Programming the Flash”
on page 298 for details on Command and Data loading):
1. A: Load Command “0100 0000”.
2. C: Load Data Low Byte. Bit n = “0” programs and bit n = “1” erases the Fuse bit.
3. Give WR a negative pulse and wait for RDY/BSY to go high.
301
8011D–AVR–02/07
25.7.9 Programming the Fuse High Bits
The algorithm for programming the Fuse High bits is as follows (refer to ”Programming the
Flash” on page 298 for details on Command and Data loading):
1. A: Load Command “0100 0000”.
2. C: Load Data Low Byte. Bit n = “0” programs and bit n = “1” erases the Fuse bit.
3. Set BS2, BS1 to “01”. This selects high data byte.
5. Set BS2, BS1 to “00”. This selects low data byte.
25.7.10 Programming the Extended Fuse Bits
The algorithm for programming the Extended Fuse bits is as follows (refer to ”Programming the
Flash” on page 298 for details on Command and Data loading):
1. 1. A: Load Command “0100 0000”.
2. 2. C: Load Data Low Byte. Bit n = “0” programs and bit n = “1” erases the Fuse bit.
3. 3. Set BS2, BS1 to “10”. This selects extended data byte.
4. 4. Give WR a negative pulse and wait for RDY/BSY to go high.
5. 5. Set BS2, BS1 to “00”. This selects low data byte.
Figure 25-5. Programming the FUSES Waveforms

Write Fuse Low byte Write Fuse high byte Write Extended Fuse byte
A C A C A C
0x40 DATA XX 0x40 DATA XX 0x40 DATA XX
DATA
XA1
XA0
BS1
BS2
XTAL1
WR
RDY/BSY
RESET +12V
OE
PAGEL
25.7.11 Programming the Lock Bits
The algorithm for programming the Lock bits is as follows (refer to ”Programming the Flash” on
page 298 for details on Command and Data loading):
1. A: Load Command “0010 0000”.
2. C: Load Data Low Byte. Bit n = “0” programs the Lock bit. If LB mode 3 is programmed
(LB1 and LB2 is programmed), it is not possible to program the Boot Lock bits by any
External Programming mode.
The Lock bits can only be cleared by executing Chip Erase.
302 ATmega164P/324P/644P
8011D–AVR–02/07
25.7.12 Reading the Fuse and Lock Bits
The algorithm for reading the Fuse and Lock bits is as follows (refer to ”Programming the Flash”
on page 298 for details on Command loading):
1. A: Load Command “0000 0100”.
2. Set OE to “0”, and BS2, BS1 to “00”. The status of the Fuse Low bits can now be read at
DATA (“0” means programmed).
3. Set OE to “0”, and BS2, BS1 to “11”. The status of the Fuse High bits can now be read at
DATA (“0” means programmed).
4. Set OE to “0”, and BS2, BS1 to “10”. The status of the Extended Fuse bits can now be
read at DATA (“0” means programmed).
5. Set OE to “0”, and BS2, BS1 to “01”. The status of the Lock bits can now be read at DATA
(“0” means programmed).
Figure 25-6. Mapping Between BS1, BS2 and the Fuse and Lock Bits During Read
Fuse Low Byte 0
0
Extended Fuse Byte 1
DATA
BS2
Lock Bits 0
1
BS1
Fuse High Byte 1
BS2
25.7.13 Reading the Signature Bytes
The algorithm for reading the Signature bytes is as follows (refer to ”Programming the Flash” on
1. A: Load Command “0000 1000”.
2. B: Load Address Low Byte (0x00 - 0x02).
3. Set OE to “0”, and BS to “0”. The selected Signature byte can now be read at DATA.
25.7.14 Reading the Calibration Byte
The algorithm for reading the Calibration byte is as follows (refer to ”Programming the Flash” on
1. A: Load Command “0000 1000”.
2. B: Load Address Low Byte, 0x00.
3. Set OE to “0”, and BS1 to “1”. The Calibration byte can now be read at DATA.
303
8011D–AVR–02/07
25.7.15 Parallel Programming Characteristics
Table 25-14. Parallel Programming Characteristics, VCC = 5V ± 10%

Symbol Parameter Min Typ Max Units
VPP Programming Enable Voltage 11.5 12.5 V
IPP Programming Enable Current 250 μA
tDVXH Data and Control Valid before XTAL1 High 67 ns
tXLXH XTAL1 Low to XTAL1 High 200 ns
tXHXL XTAL1 Pulse Width High 150 ns
tXLDX Data and Control Hold after XTAL1 Low 67 ns
tXLWL XTAL1 Low to WR Low 0 ns
tXLPH XTAL1 Low to PAGEL high 0 ns
tPLXH PAGEL low to XTAL1 high 150 ns
tBVPH BS1 Valid before PAGEL High 67 ns
tPHPL PAGEL Pulse Width High 150 ns
tPLBX BS1 Hold after PAGEL Low 67 ns
tWLBX BS2/1 Hold after WR Low 67 ns
tPLWL PAGEL Low to WR Low 67 ns
tBVWL BS2/1 Valid to WR Low 67 ns
tWLWH WR Pulse Width Low 150 ns
tWLRL WR Low to RDY/BSY Low 0 1 μs
tWLRH WR Low to RDY/BSY High(1) 3.7 4.5 ms
tWLRH_CE WR Low to RDY/BSY High for Chip Erase(2) 7.5 9 ms
tXLOL XTAL1 Low to OE Low 0 ns
tBVDV BS1 Valid to DATA valid 0 250 ns
tOLDV OE Low to DATA Valid 250 ns
tOHDZ OE High to DATA Tri-stated 250 ns
Notes: 1. tWLRH is valid for the Write Flash, Write EEPROM, Write Fuse bits and Write Lock bits
commands.
2. tWLRH_CE is valid for the Chip Erase command.
304 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 25-7. Parallel Programming Timing, Including some General Timing Requirements
tXLWL
tXHXL
XTAL1
tDVXH tXLDX
Data & Contol
(DATA, XA0/1, BS1, BS2)
tBVPH tPLBX t BVWL
tWLBX
PAGEL tPHPL
tWLWH
WR tPLWL
WLRL
RDY/BSY
tWLRH
Figure 25-8. Parallel Programming Timing, Loading Sequence with Timing Requirements(1)
LOAD ADDRESS LOAD DATA LOAD DATA LOAD DATA LOAD ADDRESS
(LOW BYTE) (LOW BYTE) (HIGH BYTE) (LOW BYTE)
t XLXH tXLPH
tPLXH
XTAL1
BS1
PAGEL
DATA ADDR0 (Low Byte) DATA (Low Byte) DATA (High Byte) ADDR1 (Low Byte)
XA0
XA1
Note: 1. The timing requirements shown in Figure 25-7 on page 305 (i.e., tDVXH, tXHXL, and tXLDX) also
apply to loading operation.
Figure 25-9. Parallel Programming Timing, Reading Sequence (within the Same Page) with
Timing Requirements(1)
LOAD ADDRESS READ DATA READ DATA LOAD ADDRESS

(LOW BYTE) (LOW BYTE) (HIGH BYTE) (LOW BYTE)
tXLOL
XTAL1
tBVDV
BS1
tOLDV
OE tOHDZ
DATA ADDR0 (Low Byte) DATA (Low Byte) DATA (High Byte) ADDR1 (Low Byte)
XA0
XA1
Note: 1. The timing requirements shown in Table 25-7 on page 305 (i.e., tDVXH, tXHXL, and tXLDX) also
apply to reading operation.
305
8011D–AVR–02/07
25.8 Serial Downloading
Both the Flash and EEPROM memory arrays can be programmed using a serial programming
bus while RESET is pulled to GND. The serial programming interface consists of pins SCK,
MOSI (input) and MISO (output). After RESET is set low, the Programming Enable instruction
needs to be executed first before program/erase operations can be executed. NOTE, in Table
25-15 on page 306, the pin mapping for serial programming is listed. Not all packages use the
SPI pins dedicated for the internal Serial Peripheral Interface - SPI.
25.8.1 Serial Programming Pin Mapping
Table 25-15. Pin Mapping Serial Programming

Pins Pins
Symbol (PDIP-40) (TQFP-44) I/O Description
MOSI PB5 PB5 I Serial Data in
MISO PB6 PB6 O Serial Data out
SCK PB7 PB7 I Serial Clock
Figure 25-10. Serial Programming and Verify(1)
+1.8 - 5.5V
VCC
+1.8 - 5.5V(2)
MOSI
AVCC
MISO
SCK
XTAL1
RESET
GND
Notes: 1. If the device is clocked by the internal Oscillator, it is no need to connect a clock source to the
XTAL1 pin.
2. VCC - 0.3V < AVCC < VCC + 0.3V, however, AVCC should always be within 1.8 - 5.5V
When programming the EEPROM, an auto-erase cycle is built into the self-timed programming
operation (in the Serial mode ONLY) and there is no need to first execute the Chip Erase
instruction. The Chip Erase operation turns the content of every memory location in both the
Program and EEPROM arrays into 0xFF.
Depending on CKSEL Fuses, a valid clock must be present. The minimum low and high periods
for the serial clock (SCK) input are defined as follows:
Low: > 2 CPU clock cycles for fck < 12 MHz, 3 CPU clock cycles for fck >= 12 MHz
High: > 2 CPU clock cycles for fck < 12 MHz, 3 CPU clock cycles for fck >= 12 MHz
306 ATmega164P/324P/644P
8011D–AVR–02/07
25.8.2 Serial Programming Algorithm
When writing serial data to the ATmega164P/324P/644P, data is clocked on the rising edge of
SCK.
When reading data from the ATmega164P/324P/644P, data is clocked on the falling edge of
SCK. See Figure 25-12 for timing details.
To program and verify the ATmega164P/324P/644P in the serial programming mode, the follow-
ing sequence is recommended (See four byte instruction formats in Table 25-17):
1. Power-up sequence:
Apply power between VCC and GND while RESET and SCK are set to “0”. In some sys-
tems, the programmer can not guarantee that SCK is held low during power-up. In this
case, RESET must be given a positive pulse of at least two CPU clock cycles duration
after SCK has been set to “0”.
2. Wait for at least 20 ms and enable serial programming by sending the Programming
Enable serial instruction to pin MOSI.
3. The serial programming instructions will not work if the communication is out of synchro-
nization. When in sync. the second byte (0x53), will echo back when issuing the third
byte of the Programming Enable instruction. Whether the echo is correct or not, all four
bytes of the instruction must be transmitted. If the 0x53 did not echo back, give RESET a
positive pulse and issue a new Programming Enable command.
4. The Flash is programmed one page at a time. The memory page is loaded one byte at a
time by supplying the 7 LSB of the address and data together with the Load Program
Memory Page instruction. To ensure correct loading of the page, the data low byte must
be loaded before data high byte is applied for a given address. The Program Memory
Page is stored by loading the Write Program Memory Page instruction with the address
lines 15..8. Before issuing this command, make sure the instruction Load Extended
Address Byte has been used to define the MSB of the address. The extended address
byte is stored until the command is re-issued, i.e., the command needs only be issued for
the first page, and when crossing the 64KWord boundary. If polling (RDY/BSY) is not
used, the user must wait at least tWD_FLASH before issuing the next page. (See Table 25-
16.) Accessing the serial programming interface before the Flash write operation com-
pletes can result in incorrect programming.
5. The EEPROM array is programmed one byte at a time by supplying the address and data
together with the appropriate Write instruction. An EEPROM memory location is first
automatically erased before new data is written. If polling is not used, the user must wait
at least tWD_EEPROM before issuing the next byte. (See Table 25-16.) In a chip erased
device, no 0xFFs in the data file(s) need to be programmed.
6. Any memory location can be verified by using the Read instruction which returns the con-
tent at the selected address at serial output MISO. When reading the Flash memory, use
the instruction Load Extended Address Byte to define the upper address byte, which is
not included in the Read Program Memory instruction. The extended address byte is
stored until the command is re-issued, i.e., the command needs only be issued for the
first page, and when crossing the 64KWord boundary.
7. At the end of the programming session, RESET can be set high to commence normal
operation.
8. Power-off sequence (if needed):
Set RESET to “1”.
Turn VCC power off.
307
8011D–AVR–02/07
Table 25-16. Minimum Wait Delay Before Writing the Next Flash or EEPROM Location
Symbol Minimum Wait Delay
tWD_FLASH 4.5 ms
tWD_EEPROM 9.0 ms
tWD_ERASE 9.0 ms
25.9 Serial Programming Instruction set

Table 25-17 on page 308 and Figure 25-11 on page 309 describes the Instruction set.
Table 25-17. Serial Programming Instruction Set (Hexadecimal values)

Instruction Format
Instruction/Operation Byte 1 Byte 2 Byte 3 Byte4
Programming Enable $AC $53 $00 $00
Chip Erase (Program Memory/EEPROM) $AC $80 $00 $00
Poll RDY/BSY $F0 $00 $00 data byte out
Load Instructions
Load Extended Address byte(1) $4D $00 Extended adr $00
Load Program Memory Page, High byte $48 $00 adr LSB high data byte in
Load Program Memory Page, Low byte $40 $00 adr LSB low data byte in
Load EEPROM Memory Page (page access) $C1 $00 0000 000aa data byte in
Read Instructions
Read Program Memory, High byte $28 adr MSB adr LSB high data byte out
Read Program Memory, Low byte $20 adr MSB adr LSB low data byte out
Read EEPROM Memory $A0 0000 00aa aaaa aaaa data byte out
Read Lock bits $58 $00 $00 data byte out
Read Signature Byte $30 $00 0000 000aa data byte out
Read Fuse bits $50 $00 $00 data byte out
Read Fuse High bits $58 $08 $00 data byte out
Read Extended Fuse Bits $50 $08 $00 data byte out
Read Calibration Byte $38 $00 $00 data byte out
Write Instructions(6)
Write Program Memory Page $4C adr MSB adr LSB $00
Write EEPROM Memory $C0 0000 00aa aaaa aaaa data byte in
Write EEPROM Memory Page (page access) $C2 0000 00aa aaaa aa00 $00
Write Lock bits $AC $E0 $00 data byte in
Write Fuse bits $AC $A0 $00 data byte in
Write Fuse High bits $AC $A8 $00 data byte in
Write Extended Fuse Bits $AC $A4 $00 data byte in
308 ATmega164P/324P/644P
8011D–AVR–02/07
Notes: 1. Not all instructions are applicable for all parts.
2. a = address.
3. Bits are programmed ‘0’, unprogrammed ‘1’.
4. To ensure future compatibility, unused Fuses and Lock bits should be unprogrammed (‘1’) .
5. Refer to the correspondig section for Fuse and Lock bits, Calibration and Signature bytes and
Page size.
6. Instructions accessing program memory use a word address. This address may be random
within the page range.
7. See htt://www.atmel.com/avr for Application Notes regarding programming and programmers.
If the LSB in RDY/BSY data byte out is ‘1’, a programming operation is still pending. Wait until
this bit returns ‘0’ before the next instruction is carried out.
Within the same page, the low data byte must be loaded prior to the high data byte.
After data is loaded to the page buffer, program the EEPROM page, see Figure 25-11 on page
309.
Figure 25-11. Serial Programming Instruction example

Serial Programming Instruction
Load Program Memory Page (High/Low Byte)/ Write Program Memory Page/
Load EEPROM Memory Page (page access) Write EEPROM Memory Page
Byte 1 Byte 2 Byte 3 Byte 4 Byte 1 Byte 2 Byte 3 Byte 4

Adr MSB
A Adr LSB Adr MSB Adrr LSB
B
Bit 15 B 0 Bit 15 B 0
Page Buffer
Page Offset
Page 0
Page 1
Page 2
Page Number
Page N-1
Program Memory/
EEPROM Memory
309
8011D–AVR–02/07
25.9.1 Serial Programming Characteristics
For characteristics of the Serial Programming module see “SPI Timing Characteristics” on page
330.
Figure 25-12. Serial Programming Waveforms

SERIAL DATA INPUT MSB LSB
(MOSI)
SERIAL DATA OUTPUT MSB LSB

(MISO)
SERIAL CLOCK INPUT

(SCK)
SAMPLE
25.10 Programming via the JTAG Interface

Programming through the JTAG interface requires control of the four JTAG specific pins: TCK,
TMS, TDI, and TDO. Control of the reset and clock pins is not required.
To be able to use the JTAG interface, the JTAGEN Fuse must be programmed. The device is
default shipped with the fuse programmed. In addition, the JTD bit in MCUCSR must be cleared.
Alternatively, if the JTD bit is set, the external reset can be forced low. Then, the JTD bit will be
cleared after two chip clocks, and the JTAG pins are available for programming. This provides a
means of using the JTAG pins as normal port pins in Running mode while still allowing In-Sys-
tem Programming via the JTAG interface. Note that this technique can not be used when using
the JTAG pins for Boundary-scan or On-chip Debug. In these cases the JTAG pins must be ded-
icated for this purpose.
During programming the clock frequency of the TCK Input must be less than the maximum fre-
quency of the chip. The System Clock Prescaler can not be used to divide the TCK Clock Input
into a sufficiently low frequency.
As a definition in this datasheet, the LSB is shifted in and out first of all Shift Registers.
25.10.1 Programming Specific JTAG Instructions
The Instruction Register is 4-bit wide, supporting up to 16 instructions. The JTAG instructions
useful for programming are listed below.
The OPCODE for each instruction is shown behind the instruction name in hex format. The text
describes which Data Register is selected as path between TDI and TDO for each instruction.
The Run-Test/Idle state of the TAP controller is used to generate internal clocks. It can also be
used as an idle state between JTAG sequences. The state machine sequence for changing the
instruction word is shown in Figure 25-13 on page 311.
310 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 25-13. State Machine Sequence for Changing the Instruction Word
1 Test-Logic-Reset
0 1 1 1
0 0
1 1
0 0
1 1
1 1
Exit1-DR Exit1-IR
0 0
1 1
0 0
Exit2-DR Exit2-IR
1 1
Update-DR Update-IR
1 0 1 0
25.10.2 AVR_RESET (0xC)
The AVR specific public JTAG instruction for setting the AVR device in the Reset mode or taking
the device out from the Reset mode. The TAP controller is not reset by this instruction. The one
bit Reset Register is selected as Data Register. Note that the reset will be active as long as there
is a logic “one” in the Reset Chain. The output from this chain is not latched.
• Shift-DR: The Reset Register is shifted by the TCK input.
25.10.3 PROG_ENABLE (0x4)
The AVR specific public JTAG instruction for enabling programming via the JTAG port. The 16-
bit Programming Enable Register is selected as Data Register. The active states are the
following:
• Shift-DR: The programming enable signature is shifted into the Data Register.
• Update-DR: The programming enable signature is compared to the correct value, and
Programming mode is entered if the signature is valid.
311
8011D–AVR–02/07
25.10.4 PROG_COMMANDS (0x5)
The AVR specific public JTAG instruction for entering programming commands via the JTAG
port. The 15-bit Programming Command Register is selected as Data Register. The active
states are the following:
• Capture-DR: The result of the previous command is loaded into the Data Register.
• Shift-DR: The Data Register is shifted by the TCK input, shifting out the result of the previous
command and shifting in the new command.
• Update-DR: The programming command is applied to the Flash inputs
• Run-Test/Idle: One clock cycle is generated, executing the applied command
25.10.5 PROG_PAGELOAD (0x6)
The AVR specific public JTAG instruction to directly load the Flash data page via the JTAG port.
An 8-bit Flash Data Byte Register is selected as the Data Register. This is physically the 8 LSBs
of the Programming Command Register. The active states are the following:
• Shift-DR: The Flash Data Byte Register is shifted by the TCK input.
• Update-DR: The content of the Flash Data Byte Register is copied into a temporary register. A
write sequence is initiated that within 11 TCK cycles loads the content of the temporary
register into the Flash page buffer. The AVR automatically alternates between writing the low
and the high byte for each new Update-DR state, starting with the low byte for the first Update-
DR encountered after entering the PROG_PAGELOAD command. The Program Counter is
pre-incremented before writing the low byte, except for the first written byte. This ensures that
the first data is written to the address set up by PROG_COMMANDS, and loading the last
location in the page buffer does not make the program counter increment into the next page.
25.10.6 PROG_PAGEREAD (0x7)
The AVR specific public JTAG instruction to directly capture the Flash content via the JTAG port.
An 8-bit Flash Data Byte Register is selected as the Data Register. This is physically the 8 LSBs
of the Programming Command Register. The active states are the following:
• Capture-DR: The content of the selected Flash byte is captured into the Flash Data Byte
Register. The AVR automatically alternates between reading the low and the high byte for each
new Capture-DR state, starting with the low byte for the first Capture-DR encountered after
entering the PROG_PAGEREAD command. The Program Counter is post-incremented after
reading each high byte, including the first read byte. This ensures that the first data is captured
from the first address set up by PROG_COMMANDS, and reading the last location in the page
makes the program counter increment into the next page.
• Shift-DR: The Flash Data Byte Register is shifted by the TCK input.
25.10.7 Data Registers
The Data Registers are selected by the JTAG instruction registers described in section ”Pro-
gramming Specific JTAG Instructions” on page 310. The Data Registers relevant for
programming operations are:
• Reset Register
• Programming Enable Register
• Programming Command Register
• Flash Data Byte Register
312 ATmega164P/324P/644P
8011D–AVR–02/07
25.10.8 Reset Register
The Reset Register is a Test Data Register used to reset the part during programming. It is
required to reset the part before entering Programming mode.
A high value in the Reset Register corresponds to pulling the external reset low. The part is reset
as long as there is a high value present in the Reset Register. Depending on the Fuse settings
for the clock options, the part will remain reset for a Reset Time-out period (refer to ”Clock
Sources” on page 30) after releasing the Reset Register. The output from this Data Register is
not latched, so the reset will take place immediately, as shown in Figure 23-2 on page 266.
25.10.9 Programming Enable Register
The Programming Enable Register is a 16-bit register. The contents of this register is compared
to the programming enable signature, binary code 0b1010_0011_0111_0000. When the con-
tents of the register is equal to the programming enable signature, programming via the JTAG
port is enabled. The register is reset to 0 on Power-on Reset, and should always be reset when
leaving Programming mode.
Figure 25-14. Programming Enable Register

TDI
0xA370
D
A
T
= D Q Programming Enable
ClockDR & PROG_ENABLE
TDO
25.10.10 Programming Command Register
The Programming Command Register is a 15-bit register. This register is used to serially shift in
programming commands, and to serially shift out the result of the previous command, if any. The
JTAG Programming Instruction Set is shown in Table 25-18 on page 315. The state sequence
when shifting in the programming commands is illustrated in Figure 25-16 on page 318.
313
8011D–AVR–02/07
Figure 25-15. Programming Command Register
TDI
S
T
R
O
B
E
S
Flash
EEPROM
A
Fuses
D
D
Lock Bits
R
E
S
S
/
D
A
T
A
TDO
314 ATmega164P/324P/644P
8011D–AVR–02/07
Table 25-18. JTAG Programming Instruction
Set a = address high bits, b = address low bits, c = address extended bits, H = 0 - Low byte, 1 - High Byte, o = data out,
i = data in, x = don’t care
Instruction TDI Sequence TDO Sequence Notes
0100011_10000000 xxxxxxx_xxxxxxxx
1a. Chip Erase
1b. Poll for Chip Erase Complete 0110011_10000000 xxxxxox_xxxxxxxx (2)
2a. Enter Flash Write 0100011_00010000 xxxxxxx_xxxxxxxx
2b. Load Address Extended High Byte 0001011_cccccccc xxxxxxx_xxxxxxxx (10)
2c. Load Address High Byte 0000111_aaaaaaaa xxxxxxx_xxxxxxxx
2d. Load Address Low Byte 0000011_bbbbbbbb xxxxxxx_xxxxxxxx
2e. Load Data Low Byte 0010011_iiiiiiii xxxxxxx_xxxxxxxx
2f. Load Data High Byte 0010111_iiiiiiii xxxxxxx_xxxxxxxx
2g. Latch Data 1110111_00000000 xxxxxxx_xxxxxxxx (1)
2h. Write Flash Page (1)
2i. Poll for Page Write Complete 0110111_00000000 xxxxxox_xxxxxxxx (2)
3a. Enter Flash Read 0100011_00000010 xxxxxxx_xxxxxxxx
3b. Load Address Extended High Byte 0001011_cccccccc xxxxxxx_xxxxxxxx (10)
3c. Load Address High Byte 0000111_aaaaaaaa xxxxxxx_xxxxxxxx
3d. Load Address Low Byte 0000011_bbbbbbbb xxxxxxx_xxxxxxxx
3e. Read Data Low and High Byte 0110110_00000000 xxxxxxx_oooooooo Low byte
0110111_00000000 xxxxxxx_oooooooo High byte
4a. Enter EEPROM Write 0100011_00010001 xxxxxxx_xxxxxxxx
4b. Load Address High Byte 0000111_aaaaaaaa xxxxxxx_xxxxxxxx (10)
4c. Load Address Low Byte 0000011_bbbbbbbb xxxxxxx_xxxxxxxx
4d. Load Data Byte 0010011_iiiiiiii xxxxxxx_xxxxxxxx
4e. Latch Data 1110111_00000000 xxxxxxx_xxxxxxxx (1)
4f. Write EEPROM Page (1)
4g. Poll for Page Write Complete 0110011_00000000 xxxxxox_xxxxxxxx (2)
315
8011D–AVR–02/07
Table 25-18. JTAG Programming Instruction (Continued)
Set (Continued) a = address high bits, b = address low bits, c = address extended bits, H = 0 - Low byte, 1 - High Byte,
o = data out, i = data in, x = don’t care
5a. Enter EEPROM Read 0100011_00000011 xxxxxxx_xxxxxxxx
5b. Load Address High Byte 0000111_aaaaaaaa xxxxxxx_xxxxxxxx (10)
5c. Load Address Low Byte 0000011_bbbbbbbb xxxxxxx_xxxxxxxx
0110011_bbbbbbbb xxxxxxx_xxxxxxxx
5d. Read Data Byte 0110010_00000000 xxxxxxx_xxxxxxxx
0110011_00000000 xxxxxxx_oooooooo
6a. Enter Fuse Write 0100011_01000000 xxxxxxx_xxxxxxxx
6b. Load Data Low Byte(6) 0010011_iiiiiiii xxxxxxx_xxxxxxxx (3)
6c. Write Fuse Extended Byte (1)
6d. Poll for Fuse Write Complete 0110111_00000000 xxxxxox_xxxxxxxx (2)
(7)
6e. Load Data Low Byte 0010011_iiiiiiii xxxxxxx_xxxxxxxx (3)
6f. Write Fuse High Byte (1)
6g. Poll for Fuse Write Complete 0110111_00000000 xxxxxox_xxxxxxxx (2)
6h. Load Data Low Byte(7) 0010011_iiiiiiii xxxxxxx_xxxxxxxx (3)
6i. Write Fuse Low Byte (1)
6j. Poll for Fuse Write Complete 0110011_00000000 xxxxxox_xxxxxxxx (2)
7a. Enter Lock Bit Write 0100011_00100000 xxxxxxx_xxxxxxxx
(9)
7b. Load Data Byte 0010011_11iiiiii xxxxxxx_xxxxxxxx (4)
7c. Write Lock Bits (1)
7d. Poll for Lock Bit Write complete 0110011_00000000 xxxxxox_xxxxxxxx (2)
8a. Enter Fuse/Lock Bit Read 0100011_00000100 xxxxxxx_xxxxxxxx
8b. Read Extended Fuse Byte(6)
8c. Read Fuse High Byte(7)
8d. Read Fuse Low Byte(8)
316 ATmega164P/324P/644P
8011D–AVR–02/07
Table 25-18. JTAG Programming Instruction (Continued)
Set (Continued) a = address high bits, b = address low bits, c = address extended bits, H = 0 - Low byte, 1 - High Byte,
o = data out, i = data in, x = don’t care
8e. Read Lock Bits(9) (5)
0110111_00000000 xxxxxxx_xxoooooo
0111010_00000000 xxxxxxx_xxxxxxxx (5)
0111110_00000000 xxxxxxx_oooooooo Fuse Ext. byte
8f. Read Fuses and Lock Bits 0110010_00000000 xxxxxxx_oooooooo Fuse High byte
0110110_00000000 xxxxxxx_oooooooo Fuse Low byte
0110111_00000000 xxxxxxx_oooooooo Lock bits
9a. Enter Signature Byte Read 0100011_00001000 xxxxxxx_xxxxxxxx
9b. Load Address Byte 0000011_bbbbbbbb xxxxxxx_xxxxxxxx
9c. Read Signature Byte
10a. Enter Calibration Byte Read 0100011_00001000 xxxxxxx_xxxxxxxx
10b. Load Address Byte 0000011_bbbbbbbb xxxxxxx_xxxxxxxx
10c. Read Calibration Byte
11a. Load No Operation Command
Notes: 1. This command sequence is not required if the seven MSB are correctly set by the previous command sequence (which is
normally the case).
2. Repeat until o = “1”.
3. Set bits to “0” to program the corresponding Fuse, “1” to unprogram the Fuse.
4. Set bits to “0” to program the corresponding Lock bit, “1” to leave the Lock bit unchanged.
5. “0” = programmed, “1” = unprogrammed.
6. The bit mapping for Fuses Extended byte is listed in Table 25-3 on page 292
7. The bit mapping for Fuses High byte is listed in Table 25-4 on page 293
8. The bit mapping for Fuses Low byte is listed in Table 25-5 on page 293
9. The bit mapping for Lock bits byte is listed in Table 25-1 on page 291
10. Address bits exceeding PCMSB and EEAMSB (Table 25-7 and Table 25-8) are don’t care
11. All TDI and TDO sequences are represented by binary digits (0b...).
317
8011D–AVR–02/07
Figure 25-16. State Machine Sequence for Changing/Reading the Data Word
1 Test-Logic-Reset
0 1 1 1
0 0
1 1
0 0
1 1
1 1
Exit1-DR Exit1-IR
0 0
1 1
0 0
Exit2-DR Exit2-IR
1 1
Update-DR Update-IR
1 0 1 0
25.10.11 Flash Data Byte Register
The Flash Data Byte Register provides an efficient way to load the entire Flash page buffer
before executing Page Write, or to read out/verify the content of the Flash. A state machine sets
up the control signals to the Flash and senses the strobe signals from the Flash, thus only the
data words need to be shifted in/out.
The Flash Data Byte Register actually consists of the 8-bit scan chain and a 8-bit temporary reg-
ister. During page load, the Update-DR state copies the content of the scan chain over to the
temporary register and initiates a write sequence that within 11 TCK cycles loads the content of
the temporary register into the Flash page buffer. The AVR automatically alternates between
writing the low and the high byte for each new Update-DR state, starting with the low byte for the
first Update-DR encountered after entering the PROG_PAGELOAD command. The Program
Counter is pre-incremented before writing the low byte, except for the first written byte. This
ensures that the first data is written to the address set up by PROG_COMMANDS, and loading
the last location in the page buffer does not make the Program Counter increment into the next
page.
During Page Read, the content of the selected Flash byte is captured into the Flash Data Byte
Register during the Capture-DR state. The AVR automatically alternates between reading the
low and the high byte for each new Capture-DR state, starting with the low byte for the first Cap-
318 ATmega164P/324P/644P
8011D–AVR–02/07
ture-DR encountered after entering the PROG_PAGEREAD command. The Program Counter is
post-incremented after reading each high byte, including the first read byte. This ensures that
the first data is captured from the first address set up by PROG_COMMANDS, and reading the
last location in the page makes the program counter increment into the next page.
Figure 25-17. Flash Data Byte Register
STROBES
State
Machine
TDI
ADDRESS
Flash
EEPROM
Fuses
Lock Bits
D
A
T
A
TDO
The state machine controlling the Flash Data Byte Register is clocked by TCK. During normal
operation in which eight bits are shifted for each Flash byte, the clock cycles needed to navigate
through the TAP controller automatically feeds the state machine for the Flash Data Byte Regis-
ter with sufficient number of clock pulses to complete its operation transparently for the user.
However, if too few bits are shifted between each Update-DR state during page load, the TAP
controller should stay in the Run-Test/Idle state for some TCK cycles to ensure that there are at
least 11 TCK cycles between each Update-DR state.
25.10.12 Programming Algorithm
All references below of type “1a”, “1b”, and so on, refer to Table 25-18 on page 315.
25.10.13 Entering Programming Mode
1. Enter JTAG instruction AVR_RESET and shift 1 in the Reset Register.

2. Enter instruction PROG_ENABLE and shift 0b1010_0011_0111_0000 in the Program-
ming Enable Register.
25.10.14 Leaving Programming Mode
1. Enter JTAG instruction PROG_COMMANDS.

2. Disable all programming instructions by using no operation instruction 11a.
3. Enter instruction PROG_ENABLE and shift 0b0000_0000_0000_0000 in the program-
ming Enable Register.
4. Enter JTAG instruction AVR_RESET and shift 0 in the Reset Register.
319
8011D–AVR–02/07
25.10.15 Performing Chip Erase

2. Start Chip Erase using programming instruction 1a.
3. Poll for Chip Erase complete using programming instruction 1b, or wait for tWLRH_CE (refer
to Table 25-14 on page 304).
25.10.16 Programming the Flash
Before programming the Flash a Chip Erase must be performed, see “Performing Chip Erase”
on page 320.
2. Enable Flash write using programming instruction 2a.
3. Load address Extended High byte using programming instruction 2b.
4. Load address High byte using programming instruction 2c.
5. Load address Low byte using programming instruction 2d.
6. Load data using programming instructions 2e, 2f and 2g.
7. Repeat steps 5 and 6 for all instruction words in the page.
8. Write the page using programming instruction 2h.
9. Poll for Flash write complete using programming instruction 2i, or wait for tWLRH (refer to
Table 25-14 on page 304).
10. Repeat steps 3 to 9 until all data have been programmed.
A more efficient data transfer can be achieved using the PROG_PAGELOAD instruction:
2. Enable Flash write using programming instruction 2a.
3. Load the page address using programming instructions 2b, 2c and 2d. PCWORD (refer
to Table 25-7 on page 294) is used to address within one page and must be written as 0.
4. Enter JTAG instruction PROG_PAGELOAD.
5. Load the entire page by shifting in all instruction words in the page byte-by-byte, starting
with the LSB of the first instruction in the page and ending with the MSB of the last
instruction in the page. Use Update-DR to copy the contents of the Flash Data Byte Reg-
ister into the Flash page location and to auto-increment the Program Counter before
each new word.
7. Write the page using programming instruction 2h.
8. Poll for Flash write complete using programming instruction 2i, or wait for tWLRH (refer to
25.10.17 Reading the Flash

2. Enable Flash read using programming instruction 3a.
3. Load address using programming instructions 3b, 3c and 3d.
4. Read data using programming instruction 3e.
5. Repeat steps 3 and 4 until all data have been read.
A more efficient data transfer can be achieved using the PROG_PAGEREAD instruction:
320 ATmega164P/324P/644P
8011D–AVR–02/07
2. Enable Flash read using programming instruction 3a.
3. Load the page address using programming instructions 3b, 3c and 3d. PCWORD (refer
to Table 25-7 on page 294) is used to address within one page and must be written as 0.
4. Enter JTAG instruction PROG_PAGEREAD.
5. Read the entire page (or Flash) by shifting out all instruction words in the page (or Flash),
starting with the LSB of the first instruction in the page (Flash) and ending with the MSB
of the last instruction in the page (Flash). The Capture-DR state both captures the data
from the Flash, and also auto-increments the program counter after each word is read.
Note that Capture-DR comes before the shift-DR state. Hence, the first byte which is
shifted out contains valid data.
7. Repeat steps 3 to 6 until all data have been read.
25.10.18 Programming the EEPROM
Before programming the EEPROM a Chip Erase must be performed, see “Performing Chip
Erase” on page 320.
2. Enable EEPROM write using programming instruction 4a.
3. Load address High byte using programming instruction 4b.
4. Load address Low byte using programming instruction 4c.
5. Load data using programming instructions 4d and 4e.
6. Repeat steps 4 and 5 for all data bytes in the page.
7. Write the data using programming instruction 4f.
8. Poll for EEPROM write complete using programming instruction 4g, or wait for tWLRH
(refer to Table 25-14 on page 304).
Note that the PROG_PAGELOAD instruction can not be used when programming the EEPROM.
25.10.19 Reading the EEPROM

2. Enable EEPROM read using programming instruction 5a.
3. Load address using programming instructions 5b and 5c.
4. Read data using programming instruction 5d.
5. Repeat steps 3 and 4 until all data have been read.
Note that the PROG_PAGEREAD instruction can not be used when reading the EEPROM.
25.10.20 Programming the Fuses

2. Enable Fuse write using programming instruction 6a.
3. Load data high byte using programming instructions 6b. A bit value of “0” will program the
corresponding fuse, a “1” will unprogram the fuse.
4. Write Fuse High byte using programming instruction 6c.
5. Poll for Fuse write complete using programming instruction 6d, or wait for tWLRH (refer to
321
8011D–AVR–02/07
6. Load data low byte using programming instructions 6e. A “0” will program the fuse, a “1”
will unprogram the fuse.
7. Write Fuse low byte using programming instruction 6f.
8. Poll for Fuse write complete using programming instruction 6g, or wait for tWLRH (refer to
25.10.21 Programming the Lock Bits

2. Enable Lock bit write using programming instruction 7a.
3. Load data using programming instructions 7b. A bit value of “0” will program the corre-
sponding lock bit, a “1” will leave the lock bit unchanged.
4. Write Lock bits using programming instruction 7c.
5. Poll for Lock bit write complete using programming instruction 7d, or wait for tWLRH (refer
to Table 25-14 on page 304).
25.10.22 Reading the Fuses and Lock Bits

2. Enable Fuse/Lock bit read using programming instruction 8a.
3. To read all Fuses and Lock bits, use programming instruction 8e.
To only read Fuse High byte, use programming instruction 8b.
To only read Fuse Low byte, use programming instruction 8c.
To only read Lock bits, use programming instruction 8d.
25.10.23 Reading the Signature Bytes

2. Enable Signature byte read using programming instruction 9a.
3. Load address 0x00 using programming instruction 9b.
4. Read first signature byte using programming instruction 9c.
5. Repeat steps 3 and 4 with address 0x01 and address 0x02 to read the second and third
signature bytes, respectively.
25.10.24 Reading the Calibration Byte

2. Enable Calibration byte read using programming instruction 10a.
3. Load address 0x00 using programming instruction 10b.
4. Read the calibration byte using programming instruction 10c.
322 ATmega164P/324P/644P
8011D–AVR–02/07
26. Electrical Characteristics
Absolute Maximum Ratings*

Operating Temperature.................................. -55°C to +125°C *NOTICE: Stresses beyond those listed under “Absolute
Maximum Ratings” may cause permanent dam-
Storage Temperature ..................................... -65°C to +150°C age to the device. This is a stress rating only and
functional operation of the device at these or
Voltage on any Pin except RESET other conditions beyond those indicated in the
with respect to Ground ................................-0.5V to VCC+0.5V operational sections of this specification is not
implied. Exposure to absolute maximum rating
Voltage on RESET with respect to Ground......-0.5V to +13.0V conditions for extended periods may affect
device reliability.
Maximum Operating Voltage ............................................ 6.0V
DC Current per I/O Pin ............................................... 40.0 mA
DC Current VCC and GND Pins................................ 200.0 mA
26.1 DC Characteristics
TA = -40°C to 85°C, VCC = 1.8V to 5.5V (unless otherwise noted)
Symbol Parameter Condition Min.(5) Typ. Max.(5) Units
(1)
Input Low Voltage,Except VCC = 1.8V - 2.4V -0.5 0.2VCC
VIL V
XTAL1 and Reset pin VCC = 2.4V - 5.5V -0.5 0.3VCC(1)
Input Low Voltage,
VIL1 VCC = 1.8V - 5.5V -0.5 0.1VCC(1) V
XTAL1 pin
Input Low Voltage,
VIL2 VCC = 1.8V - 5.5V -0.5 0.1VCC(1) V
RESET pin
Input High Voltage,
VCC = 1.8V - 2.4V 0.7VCC(2) VCC + 0.5
VIH Except XTAL1 and V
VCC = 2.4V - 5.5V 0.6VCC(2) VCC + 0.5
RESET pins
Input High Voltage, VCC = 1.8V - 2.4V 0.8VCC(2) VCC + 0.5
VIH1 V
XTAL1 pin VCC = 2.4V - 5.5V 0.7VCC(2) VCC + 0.5
Input High Voltage,
VIH2 VCC = 1.8V - 5.5V 0.9VCC(2) VCC + 0.5 V
RESET pin
IOL = 10 mA, VCC = 5V 0.9
VOL Output Low Voltage(3), V
IOL = 5 mA, VCC = 3V 0.6
IOH = -20 mA, VCC = 5V 4.2
VOH Output High Voltage(4), V
IOH = -10 mA, VCC = 3V 2.3
Input Leakage VCC = 5.5V, pin low
IIL 1 µA
Current I/O Pin (absolute value)
Input Leakage VCC = 5.5V, pin high
IIH 1 µA
Current I/O Pin (absolute value)
RRST Reset Pull-up Resistor 30 60 kΩ
RPU I/O Pin Pull-up Resistor 20 50 kΩ
323
8011D–AVR–02/07
TA = -40°C to 85°C, VCC = 1.8V to 5.5V (unless otherwise noted) (Continued)
Analog Comparator VCC = 5V

VACIO <10 40 mV
Input Offset Voltage Vin = VCC/2
Analog Comparator VCC = 5V
IACLK -50 50 nA
Input Leakage Current Vin = VCC/2
Analog Comparator VCC = 2.7V 750
tACID ns
Propagation Delay VCC = 4.0V 500
Note: 1. "Max" means the highest value where the pin is guaranteed to be read as low
2. "Min" means the lowest value where the pin is guaranteed to be read as high
3. Although each I/O port can sink more than the test conditions (20mA at VCC = 5V, 10mA at VCC = 3V) under steady state
conditions (non-transient), the following must be observed:
1.)The sum of all IOL, for ports PB0-PB7, XTAL2, PD0-PD7 should not exceed 100 mA.
2.)The sum of all IOL, for ports PA0-PA3, PC0-PC7 should not exceed 100 mA.
If IOL exceeds the test condition, VOL may exceed the related specification. Pins are not guaranteed to sink current greater
than the listed test condition.
4. Although each I/O port can source more than the test conditions (20mA at VCC = 5V, 10mA at VCC = 3V) under steady
state conditions (non-transient), the following must be observed:
1.)The sum of all IOH, for ports PB0-PB7, XTAL2, PD0-PD7 should not exceed 100 mA.
2.)The sum of all IOH, for ports PA0-PA3, PC0-PC7 should not exceed 100 mA.
If IOH exceeds the test condition, VOH may exceed the related specification. Pins are not guaranteed to source current
greater than the listed test condition.
5. All DC Characteristics contained in this datasheet are based on simulation and characterization of other AVR microcontrol-
lers manufactured in the same process technology. These values are preliminary values representing design targets, and
will be updated after characterization of actual silicon.
324 ATmega164P/324P/644P
8011D–AVR–02/07
26.1.1 ATmega164P DC Characteristics
Table 26-1. TA = -40°C to 85°C, VCC = 1.8V to 5.5V (unless otherwise noted)
Active 1 MHz, VCC = 2V 0.4 0.5 mA
Active 8 MHz, VCC = 5V 7.4 9 mA
Power Supply Current(2)
Idle 1 MHz, VCC = 2V 0.1 0.15 mA
ICC Idle 8 MHz, VCC = 5V) 1.9 4 mA

32 kHz TOSC enabled,
0.5 µA
VCC = 1.8V
Power-save mode(3)(4)
0.6 µA
VCC = 3V
WDT enabled, VCC = 3V 4.3 8 µA
Power-down mode(3)(4)
WDT disabled, VCC = 3V 0.2 2 µA
Notes: 1. All DC Characteristics contained in this datasheet are based on simulation and characterization of other AVR microcontrol-
will be updated after characterization of actual silicon
2. All bits set in the ”PRR – Power Reduction Register” on page 48.
3. Typical values at 25°C. Maximum values are characterized values and not test limits in production.
4. The current consumption values include input leakage current.
26.1.2 ATmega324P DC Characteristics
Table 26-2. TA = -40°C to 85°C, VCC = 1.8V to 5.5V (unless otherwise noted)
Active 8 MHz, VCC = 5V 8.7 9 mA
Power Supply Current(2)
ICC Idle 8 MHz, VCC = 5V) 2.5 4 mA

0.5 µA
VCC = 1.8V
(3)(4)
Power-save mode
0.6 µA
VCC = 3V
WDT enabled, VCC = 3V 4.3 8 µA
Power-down mode(3)(4)
WDT disabled, VCC = 3V 0.2 2 µA
Notes: 1. All DC Characteristics contained in this datasheet are based on simulation and characterization of other AVR microcontrol-
will be updated after characterization of actual silicon
2. All bits set in the ”PRR – Power Reduction Register” on page 48.
325
8011D–AVR–02/07
3. Typical values at 25°C. Maximum values are characterized values and not test limits in production.
4. The current consumption values include input leakage current
326 ATmega164P/324P/644P
8011D–AVR–02/07
26.1.3 ATmega644P DC Characteristics – TBD
26.2 Speed Grades

Maximum frequency is depending on VCC. As shown in Figure 26-1 and Figure 26-2, the Maxi-
mum Frequency vs. VCC curve is linear between 1.8V < VCC < 2.7V and between 2.7V < VCC <
4.5V.
Figure 26-1. Maximum Frequency vs. VCC, ATmega164PV/ATmega324PV/ATmega644PV
10 MHz
Safe Operating Area

4 MHz
1.8V 2.7V 5.5V
Figure 26-2. Maximum Frequency vs. VCC, ATmega164P/324P/644P
20 MHz
10 MHz
Safe Operating Area
2.7V 4.5V 5.5V
327
8011D–AVR–02/07
26.3 Clock Characteristics
Table 26-3. Calibration Accuracy of Internal RC Oscillator
Frequency VCC Temperature Calibration Accuracy
Factory
8.0 MHz 3V 25°C ±10%
Calibration
User 1.8V - 5.5V(1)
7.3 - 8.1 MHz -40°C - 85°C ±1%
Calibration 2.7V - 5.5V(2)
Notes: 1. Voltage range for ATmega164PV/324PV/644PV.

2. Voltage range for ATmega164P/324P/644P.
26.3.1 External Clock Drive Waveforms
Figure 26-3. External Clock Drive Waveforms
V IH1
V IL1
26.3.2 External Clock Drive
Table 26-4. External Clock Drive

VCC=1.8-5.5V VCC=2.7-5.5V VCC=4.5-5.5V
Symbol Parameter Min. Max. Min. Max. Min. Max. Units
Oscillator
1/tCLCL 0 2 0 8 0 16 MHz
Frequency
tCLCL Clock Period 500 125 62.5 ns
tCHCX High Time 200 50 25 ns
tCLCX Low Time 200 50 25 ns
tCLCH Rise Time 2.0 1.6 0.5 μs
tCHCL Fall Time 2.0 1.6 0.5 μs
Change in period
ΔtCLCL from one clock 2 2 2 %
cycle to the next
Note: All DC Characteristics contained in this datasheet are based on simulation and characterization of
other AVR microcontrollers manufactured in the same process technology. These values are pre-
liminary values representing design targets, and will be updated after characterization of actual
silicon.
328 ATmega164P/324P/644P
8011D–AVR–02/07
26.4 System and Reset Characteristics
Table 26-5. Reset, Brown-out and Internal Voltage Reference Characteristics(1)

Symbol Parameter Condition Min Typ Max Units
Power-on Reset Threshold Voltage (rising) 0.7 1.0 1.4 V
VPOT
Power-on Reset Threshold Voltage (falling)(2) 0.6 0.9 1.3 V
VRST RESET Pin Threshold Voltage 0.2VCC 0.9VCC V
tRST Minimum pulse width on RESET Pin 2.5 ns
VHYST Brown-out Detector Hysteresis 50 mV
tBOD Min Pulse Width on Brown-out Reset 2 ns
VBG Bandgap reference voltage VC C= 2.7V, TA = 25°C 1.0 1.1 1.2 V
tBG Bandgap reference start-up time VC C= 2.7V, TA = 25°C 40 70 µs
IBG Bandgap reference current consumption VC C= 2.7V, TA = 25°C 10 µA
Notes: 1. Values are guidelines only. Actual values are TBD.
2. The Power-on Reset will not work unless the supply voltage has been below VPOT (falling)
Table 26-6. BODLEVEL Fuse Coding(1)

BODLEVEL 2:0 Fuses Min VBOT Typ VBOT Max VBOT Units
111 BOD Disabled
110 1.7 1.8 2.0
101 2.5 2.7 2.9 V
100 4.1 4.3 4.5
011
010
Reserved
001
000
Note: 1. VBOT may be below nominal minimum operating voltage for some devices. For devices where this is the case, the device is
tested down to VCC = VBOT during the production test. This guarantees that a Brown-Out Reset will occur before VCC drops to
a voltage where correct operation of the microcontroller is no longer guaranteed. The test is performed using
BODLEVEL = 110 for ATmega164P/324P/644P and BODLEVEL = 101 for ATmega164P/324P/644PV.
26.5 External Interrupts Characteristics

Table 26-7. Asynchronous External Interrupt Characteristics
Symbol Parameter Condition Min Typ Max Units
tINT Minimum pulse width for asynchronous external interrupt 50 ns
329
8011D–AVR–02/07
26.6 SPI Timing Characteristics
See Figure 26-4 on page 330 and Figure 26-5 on page 331 for details.
Table 26-8. SPI Timing Parameters

Description Mode Min Typ Max
1 SCK period Master See Table 16-5
2 SCK high/low Master 50% duty cycle
3 Rise/Fall time Master TBD
4 Setup Master 10
5 Hold Master 10
6 Out to SCK Master 0.5 • tsck
7 SCK to out Master 10
8 SCK to out high Master 10
9 SS low to out Slave 15
ns
10 SCK period Slave 4 • tck
11 SCK high/low(1) Slave 2 • tck
12 Rise/Fall time Slave TBD
13 Setup Slave 10
14 Hold Slave tck
15 SCK to out Slave 15
16 SCK to SS high Slave 20
17 SS high to tri-state Slave 10
18 SS low to SCK Slave 20
Note: 1. In SPI Programming mode the minimum SCK high/low period is:
- 2 tCLCL for fCK < 12 MHz
- 3 tCLCL for fCK > 12 MHz
Figure 26-4. SPI Interface Timing Requirements (Master Mode)

SS
6 1
SCK
(CPOL = 0)
2 2
SCK
(CPOL = 1)
4 5 3
MISO
MSB ... LSB
(Data Input)
7 8
MOSI
MSB ... LSB
(Data Output)
330 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 26-5. SPI Interface Timing Requirements (Slave Mode)
SS
10 16
9
SCK
(CPOL = 0)
11 11
SCK
(CPOL = 1)
13 14 12
MOSI
MSB ... LSB
(Data Input)
15 17
MISO
MSB ... LSB X
(Data Output)
26.7 2-wire Serial Interface Characteristics

Table 26-9 describes the requirements for devices connected to the 2-wire Serial Bus. The ATmega164P/324P/644P 2-
wire Serial Interface meets or exceeds these requirements under the noted conditions.
Timing symbols refer to Figure 26-6.
Table 26-9. 2-wire Serial Bus Requirements
Symbol Parameter Condition Min Max Units
VIL Input Low-voltage -0.5 0.3 VCC V

VIH Input High-voltage 0.7 VCC VCC + 0.5 V
Vhys(1) Hysteresis of Schmitt Trigger Inputs 0.05 VCC(2) – V
VOL(1) Output Low-voltage 3 mA sink current 0 0.4 V
tr(1) Rise Time for both SDA and SCL 20 + 0.1Cb (2)(3)
300 ns
tof(1) Output Fall Time from VIHmin to VILmax 10 pF < Cb < 400 pF (3)
20 + 0.1Cb (2)(3)
250 ns
tSP(1) Spikes Suppressed by Input Filter 0 50(2) ns
Ii Input Current each I/O Pin 0.1VCC < Vi < 0.9VCC -10 10 µA
Ci(1) Capacitance for each I/O Pin – 10 pF
fSCL SCL Clock Frequency fCK(4) > max(16fSCL, 250kHz) (5)
0 400 kHz
fSCL ≤ 100 kHz V CC – 0,4V 1000ns
---------------------------- ------------------- Ω
3mA Cb
Rp Value of Pull-up resistor
fSCL > 100 kHz V CC – 0,4V 300ns
---------------------------- ---------------- Ω
3mA Cb
fSCL ≤ 100 kHz 4.0 – µs
tHD;STA Hold Time (repeated) START Condition
fSCL > 100 kHz 0.6 – µs
fSCL ≤ 100 kHz (6)
4.7 – µs
tLOW Low Period of the SCL Clock
(7)
fSCL > 100 kHz 1.3 – µs
331
8011D–AVR–02/07
Table 26-9. 2-wire Serial Bus Requirements (Continued)
Symbol Parameter Condition Min Max Units
fSCL ≤ 100 kHz 4.0 – µs

tHIGH High period of the SCL clock
fSCL > 100 kHz 0.6 – µs
fSCL ≤ 100 kHz 4.7 – µs
tSU;STA Set-up time for a repeated START condition
fSCL > 100 kHz 0.6 – µs
fSCL ≤ 100 kHz 0 3.45 µs
tHD;DAT Data hold time
fSCL > 100 kHz 0 0.9 µs
fSCL ≤ 100 kHz 250 – ns
tSU;DAT Data setup time
fSCL > 100 kHz 100 – ns
fSCL ≤ 100 kHz 4.0 – µs
tSU;STO Setup time for STOP condition
fSCL > 100 kHz 0.6 – µs
Bus free time between a STOP and START fSCL ≤ 100 kHz 4.7 – µs
tBUF
condition fSCL > 100 kHz 1.3 – µs
Notes: 1. In ATmega164P/324P/644P, this parameter is characterized and not 100% tested.
2. Required only for fSCL > 100 kHz.
3. Cb = capacitance of one bus line in pF.
4. fCK = CPU clock frequency
5. This requirement applies to all ATmega32 Two-wire Serial Interface operation. Other devices connected to the Two-wire
Serial Bus need only obey the general fSCL requirement.
6. The actual low period generated by the ATmega32 Two-wire Serial Interface is (1/fSCL - 2/fCK), thus fCK must be greater
than 6 MHz for the low time requirement to be strictly met at fSCL = 100 kHz.
7. The actual low period generated by the ATmega32 Two-wire Serial Interface is (1/fSCL - 2/fCK), thus the low time require-
ment will not be strictly met for fSCL > 308 kHz when fCK = 8 MHz. Still, ATmega32 devices connected to the bus may
communicate at full speed (400 kHz) with other ATmega32 devices, as well as any other device with a proper tLOW accep-
tance margin.
Figure 26-6. 2-wire Serial Bus Timing

tof tHIGH tr
tLOW tLOW
SCL
tSU;STA tHD;STA tHD;DAT tSU;DAT
tSU;STO
SDA
tBUF
332 ATmega164P/324P/644P
8011D–AVR–02/07
26.8 ADC Characteristics
Table 26-10. ADC Characteristics, Single Ended Channel
Symbol Parameter Condition Min(1) Typ(1) Max(1) Units
Resolution Single Ended Conversion 10 Bits
Single Ended Conversion
VREF = 4V, VCC = 4V, 2
ADC clock = 200 kHz
VREF = 4V, VCC = 4V, 2.75
ADC clock = 1 MHz
Absolute accuracy (Including
INL, DNL, quantization error, Single Ended Conversion
gain and offset error) VREF = 4V, VCC = 4V,
2
ADC clock = 200 kHz
Noise Reduction Mode
VREF = 4V, VCC = 4V,
2.75
ADC clock = 1 MHz
Noise Reduction Mode LSB

Integral Non-Linearity (INL) VREF = 4V, VCC = 4V, 1.2
ADC clock = 200 kHz
Differential Non-Linearity (DNL) VREF = 4V, VCC = 4V, 0.5
ADC clock = 200 kHz
Gain Error VREF = 4V, VCC = 4V, 2.5
ADC clock = 200 kHz
Offset Error VREF = 4V, VCC = 4V, 3
ADC clock = 200 kHz
Conversion Time Free Running Conversion 13 260 µs
Clock Frequency Single Ended Conversion 50 1000 kHz
AVCC Analog Supply Voltage VCC - 0.3 VCC + 0.3 V
VREF Reference Voltage 1.0 AVCC V
VIN Input Voltage GND VREF V
Input Bandwidth TBD kHz
VINT1 Internal Voltage Reference 1.1V 1.0 1.1 1.2 V
RREF Reference Input Resistance 32 kΩ
RAIN Analog Input Resistance 100 MΩ
Notes: 1. Values are guidelines only.
333
8011D–AVR–02/07
Table 26-11. ADC Characteristics, Differential Channels
Gain = 1x 10
Resolution Gain = 10x 10 Bits
Gain = 200x 7
Gain = 1x
VCC =5 V, VREF = 4V 18.5
ADC clock = 200 kHz
Absolute Accuracy (Including INL, Gain = 10x
DNL Quantization Error and Offset VCC =5 V, VREF = 4V 18.5
Error) ADC clock = 200 kHz
Gain = 200x
VCC =5 V, VREF = 4V 10
ADC clock = 200 kHz
Gain = 1x
ADC clock = 200 kHz
Gain = 10x
Integral Non-linearity (INL) VCC =5 V, VREF = 4V 4 LSB
ADC clock = 200 kHz
Gain = 200x
ADC clock = 200 kHz
Gain = 1x
VCC =5 V, VREF = 4V 0.75
ADC clock = 200 kHz
Gain = 10x
Differential Non-linearity (DNL) VCC =5 V, VREF = 4V 0.75
ADC clock = 200 kHz
Gain = 200x
ADC clock = 200 kHz
334 ATmega164P/324P/644P
8011D–AVR–02/07
Table 26-11. ADC Characteristics, Differential Channels (Continued)
Gain = 1x
ADC clock = 200 kHz
Gain = 10x
Gain Error VCC =5 V, VREF = 4V 18.5
ADC clock = 200 kHz
Gain = 200x
ADC clock = 200 kHz
LSB
Gain = 1x
ADC clock = 200 kHz
Gain = 10x
Offset Error VCC =5 V, VREF = 4V 0.75
ADC clock = 200 kHz
Gain = 200x
ADC clock = 200 kHz
Conversion Time 13 260 µs
Clock Frequency 50 1000 kHz
VCC - VCC +
AVCC Analog Supply Voltage V
0.3 0.3
AVCC -
VREF Reference Voltage 1.0 V
0.5
VIN Input Differential Voltage 0 AVCC V
ADC Conversion Output TBD TBD LSB
Input Bandwidth TBD kHz
RREF Reference Input Resistance 32 kΩ
Note: 1. Values are guidelines only.
335
8011D–AVR–02/07
27. Typical Characteristics
The following charts show typical behavior. These figures are not tested during manufacturing.
All current consumption measurements are performed with all I/O pins configured as inputs and
with internal pull-ups enabled. A sine wave generator with rail-to-rail output is used as clock
source.
All Active- and Idle current consumption measurements are done with all bits in the PRR regis-
ters set and thus, the corresponding I/O modules are turned off. Also the Analog Comparator is
disabled during these measurements. The power consumption in Power-down mode is indepen-
dent of clock selection.
The current consumption is a function of several factors such as: operating voltage, operating
frequency, loading of I/O pins, switching rate of I/O pins, code executed and ambient tempera-
ture. The dominating factors are operating voltage and frequency.
The current drawn from capacitive loaded pins may be estimated (for one pin) as CL*VCC*f where
CL = load capacitance, VCC = operating voltage and f = average switching frequency of I/O pin.
The parts are characterized at frequencies higher than test limits. Parts are not guaranteed to
function properly at frequencies higher than the ordering code indicates.
The difference between current consumption in Power-down mode with Watchdog Timer
enabled and Power-down mode with Watchdog Timer disabled represents the differential cur-
rent drawn by the Watchdog Timer.
27.1 ATmega164P Typical Characterization

27.1.1 Active Supply Current
Figure 27-1. Active Supply Current vs. Low Frequency (0.1 - 1.0 MHz).
ACTIVE SUPPLY CURRENT vs. LOW FREQUENCY

0.1 - 1.0 MHz
1,4
5.0 V
1,2
4.5 V
1
4.0 V
ICC (mA)
0,8
3.3 V
0,6
2.7 V
0,4
1.8 V
0,2
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Frequency (MHz)
336 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-2. Active Supply Current vs. Frequency (1 - 20 MHz).
ACTIVE SUPPLY CURRENT vs. FREQUENCY

1 - 20 MHz
20
5.5 V
18
16
5.0 V
14 4.5 V
12
ICC (mA)
10
4.0 V
8
6
3.3 V
4
2.7 V
2
1.8 V
0
0 2 4 6 8 10 12 14 16 18 20
Frequency (MHz)
Figure 27-3. Active Supply Current vs. VCC (Internal RC Oscillator, 8 MHz).
ACTIVE SUPPLY CURRENT vs. VCC
INTERNAL RC OSCILLATOR, 8 MHz
9
85 ˚C
8 25 ˚C
-40 ˚C
7
6
ICC (mA)
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
337
8011D–AVR–02/07

1,8
85˚C
1,6 25˚C
-40˚C
1,4
1,2
ICC (mA)
1
0,8
0,6
0,4
0,2
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-5. Active Supply Current vs. VCC (Internal RC Oscillator, 128 kHz).
INTERNAL RC OSCILLATOR, 128 kHz
0,3
-40 ˚C
0,25 85 ˚C
25 ˚C
0,2
I CC (mA)
0,15
0,1
0,05
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
338 ATmega164P/324P/644P
8011D–AVR–02/07
27.1.2 Idle Supply Current
Figure 27-6. Idle Supply Current vs.Low Frequency (0.1 - 1.0 MHz).
IDLE SUPPLY CURRENT vs. LOW FREQUENCY

0.1 - 1.0 MHz
0,4
0,35 5.5 V
0,3 5.0 V
4.5 V
0,25
4.0 V
ICC (mA)
0,2
3.3 V
0,15 2.7 V
0,1 1.8 V
0,05
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Frequency (MHz)
Figure 27-7. Idle Supply Current vs. Frequency (1 - 20 MHz).
IDLE SUPPLY CURRENT vs. FREQUENCY

1 - 20 MHz
6
5 5.5 V
5.0 V
4
4.5 V
ICC (mA)
4.0 V
2
3.3 V
1
2.7 V
1.8 V
0
0 2 4 6 8 10 12 14 16 18 20
Frequency (MHz)
339
8011D–AVR–02/07
Figure 27-8. Idle Supply Current vs. VCC (Internal RC Oscillator, 8 MHz).
IDLE SUPPLY CURRENT vs. VCC

2,5
85 ˚C
25 ˚C
2
-40 ˚C
1,5
ICC (mA)
0,5
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)

0,7
-40 ˚C
0,6
85 ˚C
25 ˚C
0,5
I CC (mA)
0,4
0,3
0,2
0,1
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
340 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-10. Idle Supply Current vs. VCC (Internal RC Oscillator, 128 kHz).

0,12
-40 ˚C
25 ˚C
0,1 85 ˚C
ICC (mA) 0,08
0,06
0,04
0,02
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
27.1.3 Supply Current of IO modules
The tables and formulas below can be used to calculate the additional current consumption for
the different I/O modules in Active and Idle mode. The enabling or disabling of the I/O modules
are controlled by the Power Reduction Register. See ”PRR – Power Reduction Register” on
Table 27-1. Additional Current Consumption for the different I/O modules (absolute values)
PRR bit Typical numbers
VCC = 2V, F = 1MHz VCC = 3V, F = 4MHz VCC = 5V, F = 8MHz
PRUSART1 6.0 uA 38.5 uA 150.0 uA
PRTWI 16.9 uA 116.2 uA 489.3 uA
PRTIM2 14.4 uA 95.8 uA 393.2 uA
PRTIM1 9.0 uA 57.3 uA 234.8 uA
PRTIM0 5.1uA 33.3 uA 132.5 uA
PRADC 18.1 uA 86.3 uA 335.3 uA
PRSPI 11.1 uA 70.5 uA 285.0 uA
341
8011D–AVR–02/07
Table 27-2. Additional Current Consumption (percentage) in Active and Idle mode
Additional Current consumption Additional Current consumption
compared to Active with external compared to Idle with external
clock (see Figure 27-1 on page clock (see Figure 27-6 on page
PRR bit 336 and Figure 27-2 on page 337) 339 and Figure 27-7 on page 339)
PRUSART1 1.8% 6.9%
PRUSART0 2.4% 9.1%
PRTWI 5.4% 21.2%
PRTIM2 4.6% 17.4%
PRTIM1 2.7% 10.5%
PRTIM0 1.6% 6.0%
PRADC 4.5% 16.8%
PRSPI 3.3% 12.9%
It is possible to calculate the typical current consumption based on the numbers from Table 27-1
on page 341 for other VCC and frequency settings than listed in Table 27-2 on page 342.
Example Calculate the expected current consumption in idle mode with TIMER1, ADC, and SPI enabled
at VCC = 2.0V and F = 1MHz. From Table 27-2 on page 342, third column, we see that we need
to add 10.5% for the TIMER1, 16.8 % for the ADC, and 12.9 % for the SPI module. Reading
from Figure 27-6 on page 339, we find that the idle current consumption is ~0,115 mA at VCC =
2.0V and F = 1MHz. The total current consumption in idle mode with TIMER1, ADC, and SPI
enabled, gives:
I CC total ≈ 0,115mA • ( 1 + 0,105 + 0,168 + 0,129 ) ≈ 0,161mA
27.1.4 Power-down Supply Current
Figure 27-11. Power-down Supply Current vs. VCC (Watchdog Timer Disabled).
POWER-DOWN SUPPLY CURRENT vs. VCC

WATCHDOG TIMER DISABLED
2
85 ˚C
1,8
1,6
1,4
1,2
ICC (uA)
1 -40 ˚C
0,8
25 ˚C
0,6
0,4
0,2
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
342 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-12. Power-down Supply Current vs. VCC (Watchdog Timer Enabled).

WATCHDOG TIMER ENABLED
10
-40 ˚C
85 ˚C
8 25 ˚C
6
ICC (uA)
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
27.1.5 Power-save Supply Current
Figure 27-13. Power-save Supply Current vs. VCC (Watchdog Timer Disabled and 32 kHz Crys-
tal Oscillator Running).
POWER-SAVE SUPPLY CURRENT vs. V CC
WATCHDOG TIMER DISABLED and 32 kHz CRYSTAL OSCILLATOR RUNNING
1,4
1,2
25 °C
1
I CC (uA)
0,8
0,6
0,4
0,2
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
343
8011D–AVR–02/07
27.1.6 Standby Supply Current
Figure 27-14. Standby Supply Current vs. VCC (Watchdog Timer Disabled).
STANDBY SUPPLY CURRENT vs. VCC
0,18
6MHz_xtal
0,16
6MHz_res
0,14
4MHz_xtal
0,12
4MHz_res
ICC (mA)
0,1
0,08 2MHz_res
2MHz_xtal
0,06 1MHz_res
450kHz_res
0,04
0,02
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
27.1.7 Pin Pull-up
Figure 27-15. I/O Pin Pull-up Resistor Current vs. Input Voltage (VCC = 1.8V).
I/O PIN PULL-UP RESISTOR CURRENT vs. INPUT VOLTAGE
VCC = 1.8V
50
45
40
35
30
IOP (uA)
25
20
15
10 25 ˚C
5 85 ˚C
0
-40 ˚C
0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2
VOP (V)
344 ATmega164P/324P/644P
8011D–AVR–02/07
VCC = 2.7V
80
70
60
IOP (uA) 50
40
30
20
25 ˚C
10 85 ˚C
-40 ˚C
0
0 0,5 1 1,5 2 2,5 3
VOP (V)
Figure 27-17. I/O Pin Pull-up Resistor Current vs. Input Voltage (VCC = 5V).

VCC = 5V
160
140
120
100
IOP (uA)
80
60
40
25 ˚C
20 85 ˚C
-40 ˚C
0
0 1 2 3 4 5 6
VOP (V)
345
8011D–AVR–02/07
Figure 27-18. Reset Pull-up Resistor Current vs. Reset Pin Voltage (VCC = 1.8V).
RESET PULL-UP RESISTOR CURRENT vs. RESET PIN VOLTAGE

VCC = 1.8V
40
35
30
IRESET(uA) 25
20
15
10
25 ˚C
5 -40 ˚C
85 ˚C
0
0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2
VRESET(V)
VCC = 2.7V
60
50
40
IRESET(uA)
30
20
10 25 ˚C
-40 ˚C
0
85 ˚C
0 0,5 1 1,5 2 2,5 3
VRESET(V)
346 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-20. Reset Pull-up Resistor Current vs. Reset Pin Voltage (VCC = 5V).
VCC = 5V
120
100
IRESET (uA) 80
60
40
20 25 ˚C
-40 ˚C
0 85 ˚C
0 1 2 3 4 5 6
VRESET(V)
27.1.8 Pin Driver Strength
Figure 27-21. I/O Pin Output Voltage vs. Sink Current (VCC = 3V).
I/O PIN OUTPUT VOLTAGE vs. SINK CURRENT
VCC = 3V
0,9
85 ˚C
0,8
0,7
25 ˚C
0,6
-40 ˚C
0,5
VOL (V)
0,4
0,3
0,2
0,1
0
0 5 10 15 20 25
IOL (mA)
347
8011D–AVR–02/07

VCC = 5V
0,6
85 ˚C
0,5
25 ˚C
0,4
VOL (V) -40 ˚C
0,3
0,2
0,1
0
0 5 10 15 20 25
IOL (mA)
I/O PIN OUTPUT VOLTAGE vs. SOURCE CURRENT
VCC = 3V
3,5
3
VOH (V)
2,5
-40 ˚C
25 ˚C
2
85 ˚C
1,5
0 5 10 15 20 25
IOH (mA)
348 ATmega164P/324P/644P
8011D–AVR–02/07
VCC = 5V
5,1
4,9
VOH (V) 4,8
4,7
4,6
-40 ˚C
4,5
25 ˚C
4,4
85 ˚C
4,3
0 5 10 15 20 25
IOH (mA)
27.1.9 Pin Threshold and Hysteresis
Figure 27-25. I/O Pin Input Threshold Voltage vs. VCC (VIH, I/O Pin Read as ‘1’).
I/O PIN INPUT THRESHOLD VOLTAGE vs. VCC
VIH, IO PIN READ AS '1'
3,5
85 ˚C
3
25 ˚C
2,5 -40 ˚C
Threshold (V)
1,5
0,5
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
349
8011D–AVR–02/07
Figure 27-26. I/O Pin Input Threshold Voltage vs. VCC (VIL, I/O Pin Read as ‘0’).
VIL, IO PIN READ AS '0'
2,5
85 ˚C
25 ˚C
2 -40 ˚C
Threshold (V) 1,5
0,5
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-27. I/O Pin Input Hysteresis vs. VCC.

I/O PIN INPUT HYSTERESIS vs. VCC
0,6
0,5
-40 ˚C
25 ˚C
Input Hysteresis (mV)
0,4 85 ˚C
0,3
0,2
0,1
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
350 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-28. Reset Input Threshold Voltage vs. VCC (VIH, I/O Pin Read as ‘1’).
RESET INPUT THRESHOLD VOLTAGE vs. VCC
2,5
-40 ˚C
25 ˚C
2
85 ˚C
Threshold (V)
1,5
0,5
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-29. Reset Input Threshold Voltage vs. VCC (VIL, I/O Pin Read as ‘0’).
2,5
85 ˚C
25 ˚C
2 -40 ˚C
Threshold (V)
1,5
0,5
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
351
8011D–AVR–02/07
Figure 27-30. Reset Input Hysteresis vs. VCC.
RESET PIN INPUT HYSTERESIS vs. VCC
0,7
0,6

0,5
0,4
0,3
0,2
-40 ˚C
0,1
25 ˚C
0 85 ˚C
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
27.1.10 BOD Threshold
Figure 27-31. BOD Threshold vs. Temperature (Bodlevel is 4.3V).

BOD THRESHOLDS vs. TEMPERATURE
BODLEVEL IS 4.3V
4,6
4,5
Threshold (V)
4,4
Ricing VCC
4,3
Falling VCC
4,2
4,1
4
-60 -40 -20 0 20 40 60 80 100
Temperature (C)
352 ATmega164P/324P/644P
8011D–AVR–02/07
BODLEVEL IS 2.7V
3
2,9
Ricing VCC
2,8
Threshold (V)
2,7
Falling VCC
2,6
2,5
2,4
-60 -40 -20 0 20 40 60 80 100
Temperature (C)

BODLEVEL IS 1.8V
2
1,95
1,9
Ricing VCC
Threshold (V)
1,85
1,8
Falling VCC
1,75
1,7
1,65
1,6
-60 -40 -20 0 20 40 60 80 100
Temperature (C)
353
8011D–AVR–02/07
27.1.11 Internal Oscillator Speed
Figure 27-34. Watchdog Oscillator Frequency vs. Temperature.

WATCHDOG OSCILLATOR FREQUENCY vs. TEMPERATURE
128
126
124
FRC (kHz)
122
1.8 V
120
2.7 V
3.3 V
118
4.0 V
5.5 V
116
-60 -40 -20 0 20 40 60 80 100
Temperature
Figure 27-35. Watchdog Oscillator Frequency vs. VCC.
WATCHDOG OSCILLATOR FREQUENCY vs. VCC
129
127
-40 ˚C
125
FRC (kHz)
123
25 ˚C
121
119
117 85 ˚C
115
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
354 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-36. Calibrated 8 MHz RC Oscillator Frequency vs. VCC.
CALIBRATED 8 MHz RC OSCILLATOR FREQUENCY vs. VCC
8,2
85 ˚C
8,1
8 25 ˚C
FRC (MHz)
7,9
7,8
-40 ˚C
7,7
7,6
7,5
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-37. Calibrated 8 MHz RC Oscillator Frequency vs. Temperature.
CALIBRATED 8 MHz RC OSCILLATOR FREQUENCY vs. TEMPERATURE
8,2
8,15
5.0 V
8,1
8,05
3.0 V
FRC (MHz)
7,95
7,9
7,85
7,8
7,75
7,7
-60 -40 -20 0 20 40 60 80 100
Temperature
355
8011D–AVR–02/07
Figure 27-38. Calibrated 8 MHz RC Oscillator Frequency vs. OSCCAL Value.
CALIBRATED 8 MHz RC OSCILLATOR FREQUENCY vs. OSCCAL VALUE
16
85 ˚C
14 25 ˚C
-40 ˚C
12
FRC (MHz) 10
0
0 16 32 48 64 80 96 112 128 144 160 176 192 208 224 240
OSCCAL (X1)
27.1.12 Current Consumption of Peripheral Units
Figure 27-39. ADC Current vs. VCC (AREF = AVCC).

ADC CURRENT vs. VCC
AREF = AVCC
400
-40 ˚C
350
85 ˚C
300
25 ˚C
250
ICC (uA)
200
150
100
50
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
356 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-40. Analog Comparator Current vs. VCC .
ANALOG COMPARATOR CURRENT vs. VCC
100
-40 ˚C
80
25 ˚C
85 ˚C
60
ICC (uA)
40
20
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-41. AREF External Reference Current vs. VCC .
AREF EXTERNAL REFERENCE CURRENT vs. VCC
200
-40 ˚C
180
25 ˚C
160 85 ˚C
140
120
ICC (uA)
100
80
60
40
20
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
357
8011D–AVR–02/07
Figure 27-42. Brownout Detector Current vs. VCC .
BROWNOUT DETECTOR CURRENT vs. VCC
30
25 85 ˚C
25 ˚C
-40 ˚C
20
ICC (uA)
15
10
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-43. Programming Current vs. VCC .

PROGRAMMING CURRENT vs. VCC
12
-40 ˚C
10
25 ˚C
8 85 ˚C
ICC (mA)
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
358 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-44. Watchdog Timer Current vs. VCC .
WATCHDOG TIMER CURRENT vs. VCC
10
8 -40 ˚C
7
25 ˚C
85 ˚C
ICC (uA) 6
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
27.1.13 Current Consumption in Reset and Reset Pulsewidth
Figure 27-45. Reset Supply Current vs. Low Frequency (0.1 - 1.0 MHz).
RESET SUPPLY CURRENT vs. LOW FREQUENCY
0.1 - 1.0 MHz
0,25
5.5 V
5.0 V
0,2
4.5 V
0,15 4.0 V
ICC (mA)
3.3 V
0,1
2.7 V
1.8 V
0,05
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Frequency (MHz)
359
8011D–AVR–02/07
Figure 27-46. Reset Supply Current vs. Frequency (1 - 20 MHz).
RESET SUPPLY CURRENT vs. FREQUENCY
1 - 20 MHz
4,5
5.5 V
4
3,5
5.0 V
3
4.5 V
ICC (mA)
2,5
4.0 V
2
1,5
1
3.3 V
2.7 V
0,5
1.8 V
0
0 2 4 6 8 10 12 14 16 18 20
Frequency (MHz)
Figure 27-47. Minimum Reset Pulsewidth vs. VCC.

MINIMUM RESET PULSE WIDTH vs. VCC
2500
2000
Pulsewidth (ns)
1500
1000
500 85 ˚C
25 ˚C
-40 ˚C
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
360 ATmega164P/324P/644P
8011D–AVR–02/07
27.2 ATmega324P Typical Characteristics
27.2.1 Active Supply Current
Figure 27-48. Active Supply Current vs. Low Frequency (0.1 - 1.0 MHz).
ACTIVE SUPPLY CURRENT vs. LOW FREQUENCY

0.1 - 1.0 MHz
1,8
1,6 5.5 V
1,4
5.0 V
1,2 4.5 V
ICC (mA)
1 4.0 V
0,8 3.3 V
0,6 2.7 V
0,4 1.8 V
0,2
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Frequency (MHz)
Figure 27-49. Active Supply Current vs. Frequency (1 - 20 MHz).
ACTIVE SUPPLY CURRENT vs. FREQUENCY

1 - 20 MHz
25
5.5 V
20
5.0 V
4.5 V
15
ICC (mA)
10
4.0 V
3.3 V
5
2.7 V
1.8 V
0
0 2 4 6 8 10 12 14 16 18 20
Frequency (MHz)
361
8011D–AVR–02/07

12
85 ˚C
10
25 ˚C
-40 ˚C
8
ICC (mA)
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
2
85 ˚C
25 ˚C
1,5
-40 ˚C
ICC (mA)
0,5
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
362 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-52. Active Supply Current vs. VCC (Internal RC Oscillator, 128 kHz).
0,3
-40 ˚C
0,25 25 ˚C
85 ˚C
ICC (mA) 0,2
0,15
0,1
0,05
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
27.2.2 Idle Supply Current
Figure 27-53. Idle Supply Current vs. VCC (0.1 - 1.0 MHz).
IDLE SUPPLY CURRENT vs. LOW FREQUENCY
0.1 - 1.0 MHz
0,45
5.5 V
0,4
5.0 V
0,35
4.5 V
0,3
4.0 V
ICC (mA)
0,25
3.3 V
0,2
2.7 V
0,15
1.8 V
0,1
0,05
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Frequency (MHz)
363
8011D–AVR–02/07
Figure 27-54. Idle Supply Current vs. VCC (1 - 20 MHz).
IDLE SUPPLY CURRENT vs. FREQUENCY
1 - 20 MHz
8
5.5 V
6 5.0 V
4.5 V
ICC (mA)
4
4.0 V
3.3 V
2
2.7 V
1.8 V
0
0 2 4 6 8 10 12 14 16 18 20
Frequency (MHz)
3,5
85 ˚C
3
25 ˚C
2,5 -40 ˚C
ICC (mA)
1,5
0,5
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
364 ATmega164P/324P/644P
8011D–AVR–02/07
0,8
0,7
-40 ˚C
0,6 85 ˚C
ICC (mA) 0,5
25 ˚C
0,4
0,3
0,2
0,1
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-57. Idle Supply Current vs. VCC (Internal RC Oscillator, 128 kHz).

0,14
-40 ˚C
0,12 25 ˚C
85 ˚C
0,1
ICC (mA)
0,08
0,06
0,04
0,02
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
365
8011D–AVR–02/07
27.2.3 Supply Current of I/O modules
The tables and formulas below can be used to calculate the additional current consumption for
the different I/O modules in Active and Idle mode. The enabling or disabling of the I/O modules
are controlled by the Power Reduction Register. See ”PRR – Power Reduction Register” on
Table 27-3. Additional Current Consumption for the different I/O modules (absolute values)
PRR bit Typical numbers
VCC = 2V, F = 1MHz VCC = 3V, F = 4MHz VCC = 5V, F = 8MHz
PRTWI 11.7 uA 73.3 uA 309.3 uA
PRTIM2 14.6 uA 91.3 uA 393.0 uA
PRTIM1 9.9 uA 64.0 uA 271.7 uA
PRTIM0 6.0 uA 39.2 uA 176.3 uA
PRADC 21.1 uA 114.5 uA 437.2 uA
PRSPI 10.7 uA 67.7 uA 288.2 uA
Table 27-4. Additional Current Consumption (percentage) in Active and Idle mode
Additional Current consumption Additional Current consumption
compared to Active with external compared to Idle with external
clock (see Figure 27-48 on page clock (see Figure 27-53 on page
361 and Figure 27-49 on page 363 and Figure 27-54 on page
PRR bit 361) 364)
PRUSART1 2.1 % 7.6%
PRUSART0 2.2% 8.0%
PRTWI 3.1% 11.2%
PRTIM2 3.8% 14.1%
PRTIM1 2.6% 9.7%
PRTIM0 1.6% 6.1%
PRADC 4.8% 17.7%
PRSPI 2.8% 10.4%
It is possible to calculate the typical current consumption based on the numbers from Table 27-3
on page 366 for other VCC and frequency settings than listed in Table 27-4 on page 366.
Example Calculate the expected current consumption in idle mode with TIMER1, ADC, and SPI enabled
at VCC = 2.0V and F = 1MHz. From Table 27-4 on page 366, third column, we see that we need
to add 9.7% for the TIMER1, 17.7 % for the ADC, and 10.4 % for the SPI module. Reading from
Figure 27-53 on page 363, we find that the idle current consumption is ~0,125 mA at VCC = 2.0V
and F = 1MHz. The total current consumption in idle mode with TIMER1, ADC, and SPI enabled,
gives:
I CC total ≈ 0,125mA • ( 1 + 0,097 + 0,177 + 0,104 ) ≈ 0,172mA
366 ATmega164P/324P/644P
8011D–AVR–02/07
27.2.4 Power-down Supply Current
Figure 27-58. Power-down Supply Current vs. VCC (Watchdog Timer Disabled).
POWER-DOWN SUPPLY CURRENT vs. V CC
2
85 °C
1,5
I CC (uA)
1 -40 °C
25 °C
0,5
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-59. Power-down Supply Current vs. VCC (Watchdog Timer Enabled).

WATCHDOG TIMER ENABLED
10
-40 ˚C
85 ˚C
8
25 ˚C
6
ICC (uA)
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
367
8011D–AVR–02/07
27.2.5 Power-save Supply Current
Figure 27-60. Power-save Supply Current vs. VCC (Watchdog Timer Disabled and 32 kHz Crys-
tal Oscillator Running).
POWER-SAVE SUPPLY CURRENT vs. V CC

WATCHDOG TIMER DISABLED and 32 kHz CRYSTAL OSCILLATOR RUNNING
1,4
1,2 25 °C
1
ICC (uA )
0,8
0,6
0,4
0,2
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
27.2.6 Standby Supply Current
Figure 27-61. Standby Supply Current vs. VCC (Watchdog Timer Disabled).
STANDBY SUPPLY CURRENT vs. VCC
0,18
0,16
6MHz_xtal
0,14 6MHz_res
0,12 4MHz_xtal
4MHz_res
ICC (mA)
0,1
0,08 2MHz_xtal
0,06
2MHz_res
450kHz_res
0,04
0,02
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
368 ATmega164P/324P/644P
8011D–AVR–02/07
27.2.7 Pin Pull-up

VCC = 1.8V
60
50
40
IOP (uA)
30
20
25 ˚C
10
85 ˚C
-40 ˚C
0
0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2
VOP (V)

VCC = 2.7V
80
70
60
50
IOP (uA)
40
30
20
25 ˚C
10 85 ˚C
-40 ˚C
0
0 0,5 1 1,5 2 2,5 3
VOP (V)
369
8011D–AVR–02/07
Figure 27-64. I/O Pin Pull-up Resistor Current vs. Input Voltage (VCC = 5V).
VCC = 5V
160
140
120
100
IOP (uA)
80
60
40 25 ˚C
20 85 ˚C
-40 ˚C
0
0 1 2 3 4 5 6
VOP (V)
VCC = 1.8V
40
35
30
25
IRESET(uA)
20
15
10
25 ˚C
5 -40 ˚C
85 ˚C
0
0 0,2 0,4 0,6 0,8 1 1,2 1,4 1,6 1,8 2
VRESET(V)
370 ATmega164P/324P/644P
8011D–AVR–02/07
VCC = 2.7V
60
50
40
IRESET(uA)
30
20
25 ˚C
10
-40 ˚C
85 ˚C
0
0 0,5 1 1,5 2 2,5 3
VRESET(V)
Figure 27-67. Reset Pull-up Resistor Current vs. Reset Pin Voltage (VCC = 5V).

VCC = 5V
120
100
80
IRESET(uA)
60
40
25 ˚C
20
-40 ˚C
85 ˚C
0
0 1 2 3 4 5 6
VRESET(V)
371
8011D–AVR–02/07
27.2.8 Pin Driver Strength
VCC = 3V
1
0,9
85 ˚C
0,8
0,7
25 ˚C
0,6
VOL (V)
-40 ˚C
0,5
0,4
0,3
0,2
0,1
0
0 5 10 15 20 25
IOL (mA)
VCC = 5V
0,6
85 ˚C
0,5
25 ˚C
0,4 -40 ˚C
VOL (V)
0,3
0,2
0,1
0
0 5 10 15 20 25
IOL (mA)
372 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-70. I/O Pin Output Voltage vs. Source Current (VCC = 3V).

VCC = 3V
3,5
2,5
-40 ˚C
2
25 ˚C
VOH (V)
85 ˚C
1,5
0,5
0
0 5 10 15 20 25
IOH (mA)
Figure 27-71. I/O Pin Output Voltage vs. Source Current (VCC = 5V).

VCC = 5V
5,1
4,9
4,8
VOH (V)
4,7
4,6
-40 ˚C
4,5
25 ˚C
4,4
85 ˚C
4,3
0 5 10 15 20 25
IOH (mA)
373
8011D–AVR–02/07
27.2.9 Pin Threshold and Hysteresis
Figure 27-72. I/O Pin Input Threshold vs. VCC (VIH , I/O Pin Read as ‘1’).

3,5
-40 ˚C
3 25 ˚C
2,5
85 ˚C
Threshold (V)
1,5
0,5
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-73. I/O Pin Input Threshold vs. VCC (VIL, I/O Pin Read as ‘0’).
2,5
85 ˚C
25 ˚C
2 -40 ˚C
Threshold (V)
1,5
0,5
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
374 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-74. I/O Pin Input Hysteresis vs. VCC
I/O PIN INPUT HYSTERESIS vs. VCC
0,6
-40 ˚C
0,5
25 ˚C

0,4
85 ˚C
0,3
0,2
0,1
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-75. Reset Pin Input Threshold vs. VCC (VIH , I/O Pin Read as ‘1’).
2,5
-40 ˚C
25 ˚C
2 85 ˚C
Threshold (V)
1,5
0,5
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
375
8011D–AVR–02/07
Figure 27-76. Reset Pin Input Threshold vs. VCC (VIL, I/O Pin Read as ‘0’).
2,5
85 ˚C
25 ˚C
2
-40 ˚C
Threshold (V) 1,5
0,5
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-77. Reset Pin Input Hysteresis vs. VCC

RESET PIN INPUT HYSTERESIS vs. VCC
0,7
0,6
0,5
0,4
0,3
0,2
-40 ˚C
0,1
25 ˚C
85 ˚C
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
376 ATmega164P/324P/644P
8011D–AVR–02/07
27.2.10 BOD Threshold
Figure 27-78. BOD Threshold vs. Temperature (VCC = 4.3V)

VCC = 4.3V
4,4
Ricing VCC
4,35
Threshold (V)
4,3
Falling VCC
4,25
4,2
-60 -40 -20 0 20 40 60 80 100
Temperature (C)

VCC = 2.7V
2,8
Ricing VCC
2,75
Threshold (V)
2,7
Falling VCC
2,65
2,6
-60 -40 -20 0 20 40 60 80 100
Temperature (C)
377
8011D–AVR–02/07
VCC = 1.8V
1,85
Ricing VCC
1,83
Threshold (V)
1,81
Falling VCC
1,79
1,77
1,75
-60 -40 -20 0 20 40 60 80 100
Temperature (C)
27.2.11 Internal Oscillator Speed
Figure 27-81. Watchdog Oscillator Frequency vs. Temperature
WATCHDOG OSCILLATOR FREQUENCY vs. TEMPERATURE
125
120
FRC (kHz)
2.1 V
2.7 V
3.3 V
115
4.0 V
5.5 V
110
-40 -20 0 20 40 60 80 100
Temperature
378 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-82. Watchdog Oscillator Frequency vs. VCC
WATCHDOG OSCILLATOR FREQUENCY vs. VCC
125
-40 ˚C
120
FRC (kHz)
25 ˚C
115
85 ˚C
110
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-83. Calibrated 8 MHz RC Oscillator vs. VCC

CALIBRATED 8 MHz RC OSCILLATOR FREQUENCY vs. VCC
8,5
8,3
85 ˚C
8,1
FRC (MHz)
25 ˚C
7,9
-40 ˚C
7,7
7,5
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
379
8011D–AVR–02/07
Figure 27-84. Calibrated 8 MHz RC Oscillator vs. Temperature
CALIBRATED 8 MHz RC OSCILLATOR FREQUENCY vs. TEMPERATURE
8,2
5.0 V
3.0 V
8,1
FRC (MHz)
7,9
7,8
-60 -40 -20 0 20 40 60 80 100
Temperature
Figure 27-85. Calibrated 8 MHz RC Oscillator vs. OSCCAL Value

CALIBRATED 8 MHz RC OSCILLATOR FREQUENCY vs. OSCCAL VALUE
16
85 ˚C
14 25 ˚C
12
-40 ˚C
10
FRC (MHz)
0
0 16 32 48 64 80 96 112 128 144 160 176 192 208 224 240 256
OSCCAL (X1)
380 ATmega164P/324P/644P
8011D–AVR–02/07
27.2.12 Current Consumption of Peripheral Units
Figure 27-86. ADC Current vs. VCC (AREF = AVCC)

ADC CURRENT vs. VCC
AREF = AV CC
350
85 ˚C
25 ˚C
300
-40 ˚C
250
ICC (uA)
200
150
100
50
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-87. Analog Comparator Current vs. VCC

ANALOG COMPARATOR CURRENT vs. VCC
100
90
-40 ˚C
80 25 ˚C
70 85 ˚C
60
ICC (uA)
50
40
30
20
10
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
381
8011D–AVR–02/07
Figure 27-88. AREF External Reference Current vs. VCC
AREF EXTERNAL REFERENCE CURRENT vs. VCC
200
85 ˚C
180
25 ˚C
160 -40 ˚C
140
120
ICC (uA)
100
80
60
40
20
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-89. Brownout Detector Current vs. VCC
BROWNOUT DETECTOR CURRENT vs. VCC
30
25 85 ˚C
25 ˚C
-40 ˚C
20
ICC (uA)
15
10
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
382 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-90. Programming Current vs. VCC
PROGRAMMING CURRENT vs. VCC
10
25 ˚C
9 85 ˚C
-40 ˚C
8
6
ICC (mA)
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
Figure 27-91. Watchdog Timer Current vs. VCC
WATCHDOG TIMER CURRENT vs. VCC
8 -40 ˚C
7 25 ˚C
85 ˚C
6
ICC (uA)
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
383
8011D–AVR–02/07
27.2.13 Current Consumption in Reset and Reset Pulsewidth
Figure 27-92. Reset Supply Current vs. Low Frequency

RESET SUPPLY CURRENT vs. LOW FREQUENCY
0.1 - 1.0 MHz
0,3
5.5 V
0,25
5.0 V
0,2
4.5 V
ICC (mA)
4.0 V
0,15
3.3 V
0,1
2.7 V
0,05
1.8 V
0
0 0,1 0,2 0,3 0,4 0,5 0,6 0,7 0,8 0,9 1
Frequency (MHz)
Figure 27-93. Reset Supply Current vs. Frequency
RESET SUPPLY CURRENT vs. FREQUENCY

1 - 20 MHz
6
5 5.5 V
5.0 V
4
4.5 V
ICC (mA)
4.0 V
2
3.3 V
1
2.7 V
0
1.8 V
0 2 4 6 8 10 12 14 16 18 20
Frequency (MHz)
384 ATmega164P/324P/644P
8011D–AVR–02/07
Figure 27-94. Reset Supply Current vs. Frequency
MINIMUM RESET PULSE WIDTH vs. VCC
2500
2000
Pulsewidth (ns)
1500
1000
500 85 ˚C
25 ˚C
-40 ˚C
0
1,5 2 2,5 3 3,5 4 4,5 5 5,5
VCC (V)
385
8011D–AVR–02/07
28. Register Summary
Address Name Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Bit 0 Page
(0xFF) Reserved - - - - - - -
(0xFE) Reserved - - - - - - - -
(0xFD) Reserved - - - - - - - -
(0xFC) Reserved - - - - - - - -
(0xFB) Reserved - - - - - - -
(0xFA) Reserved - - - - - - - -
(0xF9) Reserved - - - - - - -
(0xF8) Reserved - - - - - - - -
(0xF7) Reserved - - - - - - - -
(0xF6) Reserved - - - - - - - -
(0xF5) Reserved - - - - - - -
(0xF4) Reserved - - - - - - - -
(0xF3) Reserved - - - - - - - -
(0xF2) Reserved - - - - - - - -
(0xF1) Reserved - - - - - - -
(0xF0) Reserved - - - - - - - -
(0xEF) Reserved - - - - - - -
(0xEE) Reserved - - - - - - - -
(0xED) Reserved - - - - - - - -
(0xEC) Reserved - - - - - - - -
(0xEB) Reserved - - - - - - -
(0xEA) Reserved - - - - - - - -
(0xE9) Reserved - - - - - - - -
(0xE8) Reserved - - - - - - - -
(0xE7) Reserved - - - - - - -
(0xE6) Reserved - - - - - - - -
(0xE5) Reserved - - - - - - - -
(0xE4) Reserved - - - - - - - -
(0xE3) Reserved - - - - - - -
(0xE2) Reserved - - - - - - - -
(0xE1) Reserved - - - - - - -
(0xE0) Reserved - - - - - - -
(0xDF) Reserved - - - - - - - -
(0xDE) Reserved - - - - - - - -
(0xDD) Reserved - - - - - - - -
(0xDC) Reserved - - - - - - -
(0xDB) Reserved - - - - - - - -
(0xDA) Reserved - - - - - - - -
(0xD9) Reserved - - - - - - - -
(0xD8) Reserved - - - - - - - -
(0xD7) Reserved - - - - - - - -
(0xD6) Reserved - - - - - - - -
(0xD5) Reserved - - - - - - - -
(0xD4) Reserved - - - - - - - -
(0xD3) Reserved - - - - - - - -
(0xD2) Reserved - - - - - - - -
(0xD1) Reserved - - - - - - - -
(0xD0) Reserved - - - - - - - -
(0xCF) Reserved - - - - - - - -
(0xCE) UDR1 USART1 I/O Data Register 187
(0xCD) UBRR1H - - - - USART1 Baud Rate Register High Byte 191/204
(0xCC) UBRR1L USART1 Baud Rate Register Low Byte 191/204
(0xCB) Reserved - - - - - - - -
(0xCA) UCSR1C UMSEL11 UMSEL10 UPM11 UPM10 USBS1 UCSZ11 UCSZ10 UCPOL1 189/203
(0xC9) UCSR1B RXCIE1 TXCIE1 UDRIE1 RXEN1 TXEN1 UCSZ12 RXB81 TXB81 188/202
(0xC8) UCSR1A RXC1 TXC1 UDRE1 FE1 DOR1 UPE1 U2X1 MPCM1 187/202
(0xC7) Reserved - - - - - - - -
(0xC6) UDR0 USART0 I/O Data Register 187
(0xC5) UBRR0H - - - - USART0 Baud Rate Register High Byte 191/204
(0xC4) UBRR0L USART0 Baud Rate Register Low Byte 191/204
(0xC3) Reserved - - - - - - - -
(0xC2) UCSR0C UMSEL01 UMSEL00 UPM01 UPM00 USBS0 UCSZ01 UCSZ00 UCPOL0 189/203
(0xC1) UCSR0B RXCIE0 TXCIE0 UDRIE0 RXEN0 TXEN0 UCSZ02 RXB80 TXB80 188/202
386 ATmega164P/324P/644P
8011D–AVR–02/07
(0xC0) UCSR0A RXC0 TXC0 UDRE0 FE0 DOR0 UPE0 U2X0 MPCM0 187/202
(0xBF) Reserved - - - - - - - -
(0xBE) Reserved - - - - - - - -
(0xBD) TWAMR TWAM6 TWAM5 TWAM4 TWAM3 TWAM2 TWAM1 TWAM0 - 233
(0xBC) TWCR TWINT TWEA TWSTA TWSTO TWWC TWEN - TWIE 230
(0xBB) TWDR 2-wire Serial Interface Data Register 232
(0xBA) TWAR TWA6 TWA5 TWA4 TWA3 TWA2 TWA1 TWA0 TWGCE 233
(0xB9) TWSR TWS7 TWS6 TWS5 TWS4 TWS3 - TWPS1 TWPS0 232
(0xB8) TWBR 2-wire Serial Interface Bit Rate Register 230
(0xB7) Reserved - - - - - - - -
(0xB6) ASSR - EXCLK AS2 TCN2UB OCR2AUB OCR2BUB TCR2AUB TCR2BUB 156
(0xB5) Reserved - - - - - - - -
(0xB4) OCR2B Timer/Counter2 Output Compare Register B 156
(0xB3) OCR2A Timer/Counter2 Output Compare Register A 156
(0xB2) TCNT2 Timer/Counter2 (8 Bit) 155
(0xB1) TCCR2B FOC2A FOC2B - - WGM22 CS22 CS21 CS20 154
(0xB0) TCCR2A COM2A1 COM2A0 COM2B1 COM2B0 - - WGM21 WGM20 151
(0xAF) Reserved - - - - - - - -
(0xAE) Reserved - - - - - - - -
(0xAD) Reserved - - - - - - - -
(0xAC) Reserved - - - - - - - -
(0xAB) Reserved - - - - - - - -
(0xAA) Reserved - - - - - - - -
(0xA9) Reserved - - - - - - - -
(0xA8) Reserved - - - - - - - -
(0xA7) Reserved - - - - - - - -
(0xA6) Reserved - - - - - - - -
(0xA5) Reserved - - - - - - - -
(0xA4) Reserved - - - - - - - -
(0xA3) Reserved - - - - - - - -
(0xA2) Reserved - - - - - - - -
(0xA1) Reserved - - - - - - - -
(0xA0) Reserved - - - - - - - -
(0x9F) Reserved - - - - - - - -
(0x9E) Reserved - - - - - - - -
(0x9D) Reserved - - - - - - - -
(0x9C) Reserved - - - - - - - -
(0x9B) Reserved - - - - - - - -
(0x9A) Reserved - - - - - - - -
(0x99) Reserved - - - - - - - -
(0x98) Reserved - - - - - - - -
(0x97) Reserved - - - - - - - -
(0x96) Reserved - - - - - - - -
(0x95) Reserved - - - - - - - -
(0x94) Reserved - - - - - - - -
(0x93) Reserved - - - - - - - -
(0x92) Reserved - - - - - - - -
(0x91) Reserved - - - - - - - -
(0x90) Reserved - - - - - - - -
(0x8F) Reserved - - - - - - - -
(0x8E) Reserved - - - - - - - -
(0x8D) Reserved - - - - - - - -
(0x8C) Reserved - - - - - - - -
(0x8B) OCR1BH Timer/Counter1 - Output Compare Register B High Byte 135
(0x8A) OCR1BL Timer/Counter1 - Output Compare Register B Low Byte 135
(0x89) OCR1AH Timer/Counter1 - Output Compare Register A High Byte 135
(0x88) OCR1AL Timer/Counter1 - Output Compare Register A Low Byte 135
(0x87) ICR1H Timer/Counter1 - Input Capture Register High Byte 136
(0x86) ICR1L Timer/Counter1 - Input Capture Register Low Byte 136
(0x85) TCNT1H Timer/Counter1 - Counter Register High Byte 135
(0x84) TCNT1L Timer/Counter1 - Counter Register Low Byte 135
(0x83) Reserved - - - - - - - -
(0x82) TCCR1C FOC1A FOC1B - - - - - - 134
(0x81) TCCR1B ICNC1 ICES1 - WGM13 WGM12 CS12 CS11 CS10 133
(0x80) TCCR1A COM1A1 COM1A0 COM1B1 COM1B0 - - WGM11 WGM10 131
(0x7F) DIDR1 - - - - - - AIN1D AIN0D 237
387
8011D–AVR–02/07
(0x7E) DIDR0 ADC7D ADC6D ADC5D ADC4D ADC3D ADC2D ADC1D ADC0D 257
(0x7D) Reserved - - - - - - - -
(0x7C) ADMUX REFS1 REFS0 ADLAR MUX4 MUX3 MUX2 MUX1 MUX0 253
(0x7B) ADCSRB - ACME - - - ADTS2 ADTS1 ADTS0 236
(0x7A) ADCSRA ADEN ADSC ADATE ADIF ADIE ADPS2 ADPS1 ADPS0 255
(0x79) ADCH ADC Data Register High byte 256
(0x78) ADCL ADC Data Register Low byte 256
(0x77) Reserved - - - - - - - -
(0x76) Reserved - - - - - - - -
(0x75) Reserved - - - - - - - -
(0x74) Reserved - - - - - - - -
(0x73) PCMSK3 PCINT31 PCINT30 PCINT29 PCINT28 PCINT27 PCINT26 PCINT25 PCINT24 69
(0x72) Reserved - - - - - - - -
(0x71) Reserved - - - - - - - -
(0x70) TIMSK2 - - - - - OCIE2B OCIE2A TOIE2 157
(0x6F) TIMSK1 - - ICIE1 - - OCIE1B OCIE1A TOIE1 136
(0x6E) TIMSK0 - - - - - OCIE0B OCIE0A TOIE0 108
(0x6D) PCMSK2 PCINT23 PCINT22 PCINT21 PCINT20 PCINT19 PCINT18 PCINT17 PCINT16 69
(0x6C) PCMSK1 PCINT15 PCINT14 PCINT13 PCINT12 PCINT11 PCINT10 PCINT9 PCINT8 69
(0x6B) PCMSK0 PCINT7 PCINT6 PCINT5 PCINT4 PCINT3 PCINT2 PCINT1 PCINT0 70
(0x6A) Reserved - - - - - - - -
(0x69) EICRA - - ISC21 ISC20 ISC11 ISC10 ISC01 ISC00 66
(0x68) PCICR - - - - PCIE3 PCIE2 PCIE1 PCIE0 68
(0x67) Reserved - - - - - - - -
(0x66) OSCCAL Oscillator Calibration Register 40
(0x65) Reserved - - - - - - - -
(0x64) PRR PRTWI PRTIM2 PRTIM0 PRUSART1 PRTIM1 PRSPI PRUSART0 PRADC 48
(0x63) Reserved - - - - - - - -
(0x62) Reserved - - - - - - - -
(0x61) CLKPR CLKPCE - - - CLKPS3 CLKPS2 CLKPS1 CLKPS0 40
(0x60) WDTCSR WDIF WDIE WDP3 WDCE WDE WDP2 WDP1 WDP0 59
0x3F (0x5F) SREG I T H S V N Z C 10
0x3E (0x5E) SPH SP15 SP14 SP13 SP12 SP11 SP10 SP9 SP8 11
0x3D (0x5D) SPL SP7 SP6 SP5 SP4 SP3 SP2 SP1 SP0 11
0x3C (0x5C) Reserved - - - - - - - -
0x3B (0x5B) RAMPZ - - - - - - - RAMPZ0 14
0x3A (0x5A) Reserved - - - - - - - -
0x39 (0x59) Reserved - - - - - - - -
0x38 (0x58) Reserved - - - - - - - -
0x37 (0x57) SPMCSR SPMIE RWWSB SIGRD RWWSRE BLBSET PGWRT PGERS SPMEN 289
0x36 (0x56) Reserved - - - - - - - -
0x35 (0x55) MCUCR JTD BODS BODSE PUD - - IVSEL IVCE 90/273
0x34 (0x54) MCUSR - - - JTRF WDRF BORF EXTRF PORF 58/273
0x33 (0x53) SMCR - - - - SM2 SM1 SM0 SE 47
0x32 (0x52) Reserved - - - - - - - -
0x31 (0x51) OCDR On-Chip Debug Register 263
0x30 (0x50) ACSR ACD ACBG ACO ACI ACIE ACIC ACIS1 ACIS0 255
0x2F (0x4F) Reserved - - - - - - - -
0x2E (0x4E) SPDR SPI 0 Data Register 168
0x2D (0x4D) SPSR SPIF0 WCOL0 - - - - - SPI2X0 167
0x2C (0x4C) SPCR SPIE0 SPE0 DORD0 MSTR0 CPOL0 CPHA0 SPR01 SPR00 166
0x2B (0x4B) GPIOR2 General Purpose I/O Register 2 28
0x2A (0x4A) GPIOR1 General Purpose I/O Register 1 28
0x29 (0x49) Reserved - - - - - - - -
0x28 (0x48) OCR0B Timer/Counter0 Output Compare Register B 108
0x27 (0x47) OCR0A Timer/Counter0 Output Compare Register A 107
0x26 (0x46) TCNT0 Timer/Counter0 (8 Bit) 107
0x25 (0x45) TCCR0B FOC0A FOC0B - - WGM02 CS02 CS01 CS00 106
0x24 (0x44) TCCR0A COM0A1 COM0A0 COM0B1 COM0B0 - - WGM01 WGM00 108
0x23 (0x43) GTCCR TSM - - - - - PSR2 PSR54310 158
0x22 (0x42) EEARH - - - - EEPROM Address Register High Byte 23
0x21 (0x41) EEARL EEPROM Address Register Low Byte 23
0x20 (0x40) EEDR EEPROM Data Register 23
0x1F (0x3F) EECR - - EEPM1 EEPM0 EERIE EEMWE EEWE EERE 23
0x1E (0x3E) GPIOR0 General Purpose I/O Register 0 28
0x1D (0x3D) EIMSK - - - - - INT2 INT1 INT0 67
388 ATmega164P/324P/644P
8011D–AVR–02/07
0x1C (0x3C) EIFR - - - - - INTF2 INTF1 INTF0 67
0x1B (0x3B) PCIFR - - - - PCIF3 PCIF2 PCIF1 PCIF0 68
0x1A (0x3A) Reserved - - - - - - - -
0x19 (0x39) Reserved - - - - - - - -
0x18 (0x38) Reserved - - - - - - - -
0x17 (0x37) TIFR2 - - - - - OCF2b OCF2A TOV2 158
0x16 (0x36) TIFR1 - - ICF1 - - OCF1B OCF1A TOV1 137
0x15 (0x35) TIFR0 - - - - - OCF0B OCF0A TOV0 108
0x14 (0x34) Reserved - - - - - - - -
0x13 (0x33) Reserved - - - - - - - -
0x12 (0x32) Reserved - - - - - - - -
0x11 (0x31) Reserved - - - - - - - -
0x10 (0x30) Reserved - - - - - - - -
0x0F (0x2F) Reserved - - - - - - - -
0x0E (0x2E) Reserved - - - - - - - -
0x0D (0x2D) Reserved - - - - - - - -
0x0C (0x2C) Reserved - - - - - - - -
0x0B (0x2B) PORTD PORTD7 PORTD6 PORTD5 PORTD4 PORTD3 PORTD2 PORTD1 PORTD0 91
0x0A (0x2A) DDRD DDD7 DDD6 DDD5 DDD4 DDD3 DDD2 DDD1 DDD0 91
0x09 (0x29) PIND PIND7 PIND6 PIND5 PIND4 PIND3 PIND2 PIND1 PIND0 91
0x08 (0x28) PORTC PORTC7 PORTC6 PORTC5 PORTC4 PORTC3 PORTC2 PORTC1 PORTC0 91
0x07 (0x27) DDRC DDC7 DDC6 DDC5 DDC4 DDC3 DDC2 DDC1 DDC0 91
0x06 (0x26) PINC PINC7 PINC6 PINC5 PINC4 PINC3 PINC2 PINC1 PINC0 91
0x05 (0x25) PORTB PORTB7 PORTB6 PORTB5 PORTB4 PORTB3 PORTB2 PORTB1 PORTB0 90
0x04 (0x24) DDRB DDB7 DDB6 DDB5 DDB4 DDB3 DDB2 DDB1 DDB0 90
0x03 (0x23) PINB PINB7 PINB6 PINB5 PINB4 PINB3 PINB2 PINB1 PINB0 90
0x02 (0x22) PORTA PORTA7 PORTA6 PORTA5 PORTA4 PORTA3 PORTA2 PORTA1 PORTA0 90
0x01 (0x21) DDRA DDA7 DDA6 DDA5 DDA4 DDA3 DDA2 DDA1 DDA0 90
0x00 (0x20) PINA PINA7 PINA6 PINA5 PINA4 PINA3 PINA2 PINA1 PINA0 90
Notes: 1. For compatibility with future devices, reserved bits should be written to zero if accessed. Reserved I/O memory addresses
should never be written.
2. I/O registers within the address range $00 - $1F are directly bit-accessible using the SBI and CBI instructions. In these reg-
isters, the value of single bits can be checked by using the SBIS and SBIC instructions.
3. Some of the status flags are cleared by writing a logical one to them. Note that the CBI and SBI instructions will operate on
all bits in the I/O register, writing a one back into any flag read as set, thus clearing the flag. The CBI and SBI instructions
work with registers 0x00 to 0x1F only.
4. When using the I/O specific commands IN and OUT, the I/O addresses $00 - $3F must be used. When addressing I/O regis-
ters as data space using LD and ST instructions, $20 must be added to these addresses. The ATmega164P/324P/644P is a
complex microcontroller with more peripheral units than can be supported within the 64 location reserved in Opcode for the
IN and OUT instructions. For the Extended I/O space from $60 - $FF, only the ST/STS/STD and LD/LDS/LDD instructions
can be used.
389
8011D–AVR–02/07
29. Instruction Set Summary
Mnemonics Operands Description Operation Flags #Clocks
ARITHMETIC AND LOGIC INSTRUCTIONS
ADD Rd, Rr Add two Registers Rd ← Rd + Rr Z,C,N,V,H 1
ADC Rd, Rr Add with Carry two Registers Rd ← Rd + Rr + C Z,C,N,V,H 1
ADIW Rdl,K Add Immediate to Word Rdh:Rdl ← Rdh:Rdl + K Z,C,N,V,S 2
SUB Rd, Rr Subtract two Registers Rd ← Rd - Rr Z,C,N,V,H 1
SUBI Rd, K Subtract Constant from Register Rd ← Rd - K Z,C,N,V,H 1
SBC Rd, Rr Subtract with Carry two Registers Rd ← Rd - Rr - C Z,C,N,V,H 1
SBCI Rd, K Subtract with Carry Constant from Reg. Rd ← Rd - K - C Z,C,N,V,H 1
SBIW Rdl,K Subtract Immediate from Word Rdh:Rdl ← Rdh:Rdl - K Z,C,N,V,S 2
AND Rd, Rr Logical AND Registers Rd ← Rd • Rr Z,N,V 1
ANDI Rd, K Logical AND Register and Constant Rd ← Rd • K Z,N,V 1
OR Rd, Rr Logical OR Registers Rd ← Rd v Rr Z,N,V 1
ORI Rd, K Logical OR Register and Constant Rd ← Rd v K Z,N,V 1
EOR Rd, Rr Exclusive OR Registers Rd ← Rd ⊕ Rr Z,N,V 1
COM Rd One’s Complement Rd ← 0xFF − Rd Z,C,N,V 1
NEG Rd Two’s Complement Rd ← 0x00 − Rd Z,C,N,V,H 1
SBR Rd,K Set Bit(s) in Register Rd ← Rd v K Z,N,V 1
CBR Rd,K Clear Bit(s) in Register Rd ← Rd • (0xFF - K) Z,N,V 1
INC Rd Increment Rd ← Rd + 1 Z,N,V 1
DEC Rd Decrement Rd ← Rd − 1 Z,N,V 1
TST Rd Test for Zero or Minus Rd ← Rd • Rd Z,N,V 1
CLR Rd Clear Register Rd ← Rd ⊕ Rd Z,N,V 1
SER Rd Set Register Rd ← 0xFF None 1
MUL Rd, Rr Multiply Unsigned R1:R0 ← Rd x Rr Z,C 2
MULS Rd, Rr Multiply Signed R1:R0 ← Rd x Rr Z,C 2
MULSU Rd, Rr Multiply Signed with Unsigned R1:R0 ← Rd x Rr Z,C 2
FMUL Rd, Rr Fractional Multiply Unsigned R1:R0 ← (Rd x Rr) << 1 Z,C 2
FMULS Rd, Rr Fractional Multiply Signed R1:R0 ← (Rd x Rr) << 1 Z,C 2
FMULSU Rd, Rr Fractional Multiply Signed with Unsigned R1:R0 ← (Rd x Rr) << 1 Z,C 2
BRANCH INSTRUCTIONS
RJMP k Relative Jump PC ← PC + k + 1 None 2
IJMP Indirect Jump to (Z) PC ← Z None 2
JMP k Direct Jump PC ← k None 3
RCALL k Relative Subroutine Call PC ← PC + k + 1 None 4
ICALL Indirect Call to (Z) PC ← Z None 4
CALL k Direct Subroutine Call PC ← k None 5
RET Subroutine Return PC ← STACK None 5
RETI Interrupt Return PC ← STACK I 5
CPSE Rd,Rr Compare, Skip if Equal if (Rd = Rr) PC ← PC + 2 or 3 None 1/2/3
CP Rd,Rr Compare Rd − Rr Z, N,V,C,H 1
CPC Rd,Rr Compare with Carry Rd − Rr − C Z, N,V,C,H 1
CPI Rd,K Compare Register with Immediate Rd − K Z, N,V,C,H 1
SBRC Rr, b Skip if Bit in Register Cleared if (Rr(b)=0) PC ← PC + 2 or 3 None 1/2/3
SBRS Rr, b Skip if Bit in Register is Set if (Rr(b)=1) PC ← PC + 2 or 3 None 1/2/3
SBIC P, b Skip if Bit in I/O Register Cleared if (P(b)=0) PC ← PC + 2 or 3 None 1/2/3
SBIS P, b Skip if Bit in I/O Register is Set if (P(b)=1) PC ← PC + 2 or 3 None 1/2/3
BRBS s, k Branch if Status Flag Set if (SREG(s) = 1) then PC←PC+k + 1 None 1/2
BRBC s, k Branch if Status Flag Cleared if (SREG(s) = 0) then PC←PC+k + 1 None 1/2
BREQ k Branch if Equal if (Z = 1) then PC ← PC + k + 1 None 1/2
BRNE k Branch if Not Equal if (Z = 0) then PC ← PC + k + 1 None 1/2
BRCS k Branch if Carry Set if (C = 1) then PC ← PC + k + 1 None 1/2
BRCC k Branch if Carry Cleared if (C = 0) then PC ← PC + k + 1 None 1/2
BRSH k Branch if Same or Higher if (C = 0) then PC ← PC + k + 1 None 1/2
BRLO k Branch if Lower if (C = 1) then PC ← PC + k + 1 None 1/2
BRMI k Branch if Minus if (N = 1) then PC ← PC + k + 1 None 1/2
BRPL k Branch if Plus if (N = 0) then PC ← PC + k + 1 None 1/2
BRGE k Branch if Greater or Equal, Signed if (N ⊕ V= 0) then PC ← PC + k + 1 None 1/2
BRLT k Branch if Less Than Zero, Signed if (N ⊕ V= 1) then PC ← PC + k + 1 None 1/2
BRHS k Branch if Half Carry Flag Set if (H = 1) then PC ← PC + k + 1 None 1/2
BRHC k Branch if Half Carry Flag Cleared if (H = 0) then PC ← PC + k + 1 None 1/2
BRTS k Branch if T Flag Set if (T = 1) then PC ← PC + k + 1 None 1/2
BRTC k Branch if T Flag Cleared if (T = 0) then PC ← PC + k + 1 None 1/2
BRVS k Branch if Overflow Flag is Set if (V = 1) then PC ← PC + k + 1 None 1/2
390 ATmega164P/324P/644P
8011D–AVR–02/07
BRVC k Branch if Overflow Flag is Cleared if (V = 0) then PC ← PC + k + 1 None 1/2
BRIE k Branch if Interrupt Enabled if ( I = 1) then PC ← PC + k + 1 None 1/2
BRID k Branch if Interrupt Disabled if ( I = 0) then PC ← PC + k + 1 None 1/2
BIT AND BIT-TEST INSTRUCTIONS
SBI P,b Set Bit in I/O Register I/O(P,b) ← 1 None 2
CBI P,b Clear Bit in I/O Register I/O(P,b) ← 0 None 2
LSL Rd Logical Shift Left Rd(n+1) ← Rd(n), Rd(0) ← 0 Z,C,N,V 1
LSR Rd Logical Shift Right Rd(n) ← Rd(n+1), Rd(7) ← 0 Z,C,N,V 1
ROL Rd Rotate Left Through Carry Rd(0)←C,Rd(n+1)← Rd(n),C←Rd(7) Z,C,N,V 1
ROR Rd Rotate Right Through Carry Rd(7)←C,Rd(n)← Rd(n+1),C←Rd(0) Z,C,N,V 1
ASR Rd Arithmetic Shift Right Rd(n) ← Rd(n+1), n=0..6 Z,C,N,V 1
SWAP Rd Swap Nibbles Rd(3..0)←Rd(7..4),Rd(7..4)←Rd(3..0) None 1
BSET s Flag Set SREG(s) ← 1 SREG(s) 1
BCLR s Flag Clear SREG(s) ← 0 SREG(s) 1
BST Rr, b Bit Store from Register to T T ← Rr(b) T 1
BLD Rd, b Bit load from T to Register Rd(b) ← T None 1
SEC Set Carry C←1 C 1
CLC Clear Carry C←0 C 1
SEN Set Negative Flag N←1 N 1
CLN Clear Negative Flag N←0 N 1
SEZ Set Zero Flag Z←1 Z 1
CLZ Clear Zero Flag Z←0 Z 1
SEI Global Interrupt Enable I←1 I 1
CLI Global Interrupt Disable I←0 I 1
SES Set Signed Test Flag S←1 S 1
CLS Clear Signed Test Flag S←0 S 1
SEV Set Twos Complement Overflow. V←1 V 1
CLV Clear Twos Complement Overflow V←0 V 1
SET Set T in SREG T←1 T 1
CLT Clear T in SREG T←0 T 1
SEH Set Half Carry Flag in SREG H←1 H 1
CLH Clear Half Carry Flag in SREG H←0 H 1
DATA TRANSFER INSTRUCTIONS
MOV Rd, Rr Move Between Registers Rd ← Rr None 1
MOVW Rd, Rr Copy Register Word Rd+1:Rd ← Rr+1:Rr None 1
LDI Rd, K Load Immediate Rd ← K None 1
LD Rd, X Load Indirect Rd ← (X) None 2
LD Rd, X+ Load Indirect and Post-Inc. Rd ← (X), X ← X + 1 None 2
LD Rd, - X Load Indirect and Pre-Dec. X ← X - 1, Rd ← (X) None 2
LD Rd, Y Load Indirect Rd ← (Y) None 2
LD Rd, Y+ Load Indirect and Post-Inc. Rd ← (Y), Y ← Y + 1 None 2
LD Rd, - Y Load Indirect and Pre-Dec. Y ← Y - 1, Rd ← (Y) None 2
LDD Rd,Y+q Load Indirect with Displacement Rd ← (Y + q) None 2
LD Rd, Z Load Indirect Rd ← (Z) None 2
LD Rd, Z+ Load Indirect and Post-Inc. Rd ← (Z), Z ← Z+1 None 2
LD Rd, -Z Load Indirect and Pre-Dec. Z ← Z - 1, Rd ← (Z) None 2
LDD Rd, Z+q Load Indirect with Displacement Rd ← (Z + q) None 2
LDS Rd, k Load Direct from SRAM Rd ← (k) None 2
ST X, Rr Store Indirect (X) ← Rr None 2
ST X+, Rr Store Indirect and Post-Inc. (X) ← Rr, X ← X + 1 None 2
ST - X, Rr Store Indirect and Pre-Dec. X ← X - 1, (X) ← Rr None 2
ST Y, Rr Store Indirect (Y) ← Rr None 2
ST Y+, Rr Store Indirect and Post-Inc. (Y) ← Rr, Y ← Y + 1 None 2
ST - Y, Rr Store Indirect and Pre-Dec. Y ← Y - 1, (Y) ← Rr None 2
STD Y+q,Rr Store Indirect with Displacement (Y + q) ← Rr None 2
ST Z, Rr Store Indirect (Z) ← Rr None 2
ST Z+, Rr Store Indirect and Post-Inc. (Z) ← Rr, Z ← Z + 1 None 2
ST -Z, Rr Store Indirect and Pre-Dec. Z ← Z - 1, (Z) ← Rr None 2
STD Z+q,Rr Store Indirect with Displacement (Z + q) ← Rr None 2
STS k, Rr Store Direct to SRAM (k) ← Rr None 2
LPM Load Program Memory R0 ← (Z) None 3
LPM Rd, Z Load Program Memory Rd ← (Z) None 3
LPM Rd, Z+ Load Program Memory and Post-Inc Rd ← (Z), Z ← Z+1 None 3
ELPM Extended Load Program Memory R0 ← (RAMPZ:Z) None 3
ELPM Rd, Z Extended Load Program Memory Rd ← (Z) None 3
ELPM Rd, Z+ Extended Load Program Memory Rd ← (RAMPZ:Z), RAMPZ:Z ←RAMPZ:Z+1 None 3
391
8011D–AVR–02/07
SPM Store Program Memory (Z) ← R1:R0 None -
IN Rd, P In Port Rd ← P None 1
OUT P, Rr Out Port P ← Rr None 1
PUSH Rr Push Register on Stack STACK ← Rr None 2
POP Rd Pop Register from Stack Rd ← STACK None 2
MCU CONTROL INSTRUCTIONS
NOP No Operation None 1
SLEEP Sleep (see specific descr. for Sleep function) None 1
WDR Watchdog Reset (see specific descr. for WDR/timer) None 1
BREAK Break For On-chip Debug Only None N/A
392 ATmega164P/324P/644P
8011D–AVR–02/07
30. Ordering Information

30.1 ATmega164P
Speed (MHz)(3) Power Supply Ordering Code Package(1) Operational Range

(2)
ATmega164PV-10AU 44A
Industrial
10 1.8 - 5.5V ATmega164PV-10PU(2) 40P6
(-40oC to 85oC)
ATmega164PV-10MU(2) 44M1
ATmega164P-20AU(2) 44A
Industrial
20 2.7 - 5.5V ATmega164P-20PU(2) 40P6
(-40oC to 85oC)
ATmega164P-20MU(2) 44M1
Notes: 1. This device can also be supplied in wafer form. Please contact your local Atmel sales office for detailed ordering information
and minimum quantities.
2. Pb-free packaging, complies to the European Directive for Restriction of Hazardous Substances (RoHS directive). Also
Halide free and fully Green.
3. For Speed vs. VCC see ”Speed Grades” on page 327.
Package Type
44A 44-lead, Thin (1.0 mm) Plastic Gull Wing Quad Flat Package (TQFP)
40P6 40-pin, 0.600” Wide, Plastic Dual Inline Package (PDIP)
44M1 44-pad, 7 x 7 x 1.0 mm body, lead pitch 0.50 mm, Quad Flat No-Lead/Micro Lead Frame Package (QFN/MLF)
393
8011D–AVR–02/07
30.2 ATmega324P

(2)
Industrial
10 1.8 - 5.5V ATmega324PV-10PU(2) 40P6
(-40oC to 85oC)
Industrial
20 2.7 - 5.5V ATmega324P-20PU(2) 40P6
(-40oC to 85oC)
Package Type
394 ATmega164P/324P/644P
8011D–AVR–02/07
30.3 ATmega644P

(2)
Industrial
10 1.8 - 5.5V ATmega644PV-10PU(2) 40P6
(-40oC to 85oC)
Industrial
20 2.7 - 5.5V ATmega644P-20PU(2) 40P6
(-40oC to 85oC)
Package Type
395
8011D–AVR–02/07
31. Packaging Information
31.1 44A
PIN 1
B
PIN 1 IDENTIFIER
e E1 E
D1
D
C 0˚~7˚
A1 A2 A
L
COMMON DIMENSIONS
(Unit of Measure = mm)
SYMBOL MIN NOM MAX NOTE

A – – 1.20
A1 0.05 – 0.15
A2 0.95 1.00 1.05
D 11.75 12.00 12.25
D1 9.90 10.00 10.10 Note 2
E 11.75 12.00 12.25
Notes: 1. This package conforms to JEDEC reference MS-026, Variation ACB.
2. Dimensions D1 and E1 do not include mold protrusion. Allowable E1 9.90 10.00 10.10 Note 2
protrusion is 0.25 mm per side. Dimensions D1 and E1 are maximum B 0.30 – 0.45
plastic body size dimensions including mold mismatch.
C 0.09 – 0.20
3. Lead coplanarity is 0.10 mm maximum.
L 0.45 – 0.75
e 0.80 TYP
10/5/2001
TITLE DRAWING NO. REV.
2325 Orchard Parkway
44A, 44-lead, 10 x 10 mm Body Size, 1.0 mm Body Thickness,
R San Jose, CA 95131 44A B
0.8 mm Lead Pitch, Thin Profile Plastic Quad Flat Package (TQFP)
396 ATmega164P/324P/644P
8011D–AVR–02/07
31.2 40P6
D
PIN
1
E1
SEATING PLANE
A1
L
B
B1
e
COMMON DIMENSIONS
0º ~ 15º REF (Unit of Measure = mm)
C
SYMBOL MIN NOM MAX NOTE
eB A – – 4.826
A1 0.381 – –
D 52.070 – 52.578 Note 2
E 15.240 – 15.875
E1 13.462 – 13.970 Note 2
B 0.356 – 0.559
Notes: 1. This package conforms to JEDEC reference MS-011, Variation AC. B1 1.041 – 1.651
2. Dimensions D and E1 do not include mold Flash or Protrusion. L 3.048 – 3.556
Mold Flash or Protrusion shall not exceed 0.25 mm (0.010").
C 0.203 – 0.381
eB 15.494 – 17.526
e 2.540 TYP
09/28/01
40P6, 40-lead (0.600"/15.24 mm Wide) Plastic Dual 40P6 B
R San Jose, CA 95131 Inline Package (PDIP)
397
8011D–AVR–02/07
31.3 44M1
Marked Pin# 1 ID
SEATING PLANE
A1
TOP VIEW
A3
A
K
L
Pin #1 Corner SIDE VIEW
D2
1 Option A Pin #1
Triangle
2
3 COMMON DIMENSIONS
(Unit of Measure = mm)
E2 SYMBOL MIN NOM MAX NOTE

Option B
Pin #1 A 0.80 0.90 1.00
Chamfer
(C 0.30) A1 – 0.02 0.05
A3 0.25 REF
b 0.18 0.23 0.30
D 6.90 7.00 7.10
K Option C Pin #1
e Notch D2 5.00 5.20 5.40
b (0.20 R)
E 6.90 7.00 7.10
BOTTOM VIEW E2 5.00 5.20 5.40
e 0.50 BSC
L 0.59 0.64 0.69
Note: JEDEC Standard MO-220, Fig. 1 (SAW Singulation) VKKD-3.
K 0.20 0.26 0.41
5/27/06
44M1, 44-pad, 7 x 7 x 1.0 mm Body, Lead Pitch 0.50 mm, 44M1 G
R San Jose, CA 95131
5.20 mm Exposed Pad, Micro Lead Frame Package (MLF)
398 ATmega164P/324P/644P
8011D–AVR–02/07
32. Errata
32.1 ATmega164P Rev. A
No known Errata.

No known Errata.

No known Errata.
399
8011D–AVR–02/07
33. Datasheet Revision History
Please note that the referring page numbers in this section are referred to this document. The
referring revision in this section are referring to the document revision.
33.1 Rev. 8011D - 02/07
1. Updated ”Pinout ATmega164P/324P/644P” on page 2.

2. Updated ”Power-down Mode” on page 44.
3. Updated note in Table 11-1 on page 67.
4. Updated Table 23-1 on page 270.
5. Updated ”Boot Size Configuration(1)” on page 287.
6. Updated VOL limits in ”DC Characteristics” on page 323.
7. Updated note 3 and 4 in ”DC Characteristics” on page 323.
8. Added note to ”ATmega164P DC Characteristics” on page 325.
9. Added note to ”ATmega324P DC Characteristics” on page 325.
10. Updated Figure 27-13 on page 343 and Figure 27-60 on page 368.
33.2 Rev. 8011C - 10/06
1. Updated ”DC Characteristics” on page 323.
33.3 Rev. 8011B - 09/06
1. Updated ”DC Characteristics” on page 323.
33.4 Rev. 8011A - 08/06
1. Initial revision.
400 ATmega164P/324P/644P
8011D–AVR–02/07
Table of Contents
Features ..................................................................................................... 1
1 Pin Configurations ................................................................................... 2
2 Overview ................................................................................................... 3
2.1 Block Diagram ..........................................................................................................3
2.2 Comparison Between ATmega164P, ATmega324P and ATmega644P ..................4
2.3 Pin Descriptions .......................................................................................................4
3 Resources ................................................................................................. 7
4 About Code Examples ............................................................................. 8
5 AVR CPU Core .......................................................................................... 9

5.1 Overview ..................................................................................................................9
5.2 ALU – Arithmetic Logic Unit ...................................................................................10
5.3 Status Register .......................................................................................................10
5.4 General Purpose Register File ...............................................................................12
5.5 Stack Pointer ..........................................................................................................13
5.6 Instruction Execution Timing ..................................................................................14
5.7 Reset and Interrupt Handling .................................................................................15
6 AVR Memories ........................................................................................ 18

6.1 Overview ................................................................................................................18
6.2 In-System Reprogrammable Flash Program Memory ............................................18
6.3 SRAM Data Memory ..............................................................................................19
6.4 EEPROM Data Memory .........................................................................................21
6.5 I/O Memory .............................................................................................................22
6.6 Register Description ...............................................................................................23
7 System Clock and Clock Options ......................................................... 29

7.1 Clock Systems and their Distribution ......................................................................29
7.2 Clock Sources ........................................................................................................30
7.3 Low Power Crystal Oscillator ..................................................................................32
7.4 Full Swing Crystal Oscillator ...................................................................................33
7.5 Low Frequency Crystal Oscillator ...........................................................................34
7.6 Calibrated Internal RC Oscillator ............................................................................36
7.7 128 kHz Internal Oscillator .....................................................................................37
7.8 External Clock ........................................................................................................37
i
8011D–AVR–02/07
7.9 Timer/Counter Oscillator ........................................................................................38
7.10 Clock Output Buffer ..............................................................................................38
7.11 System Clock Prescaler .......................................................................................38
7.12 Register Description .............................................................................................40
8 Power Management and Sleep Modes ................................................. 42

8.1 Overview ................................................................................................................42
8.2 Sleep Modes ..........................................................................................................42
8.3 BOD Disable ...........................................................................................................43
8.4 Idle Mode ................................................................................................................43
8.5 ADC Noise Reduction Mode ..................................................................................43
8.6 Power-down Mode .................................................................................................44
8.7 Power-save Mode ..................................................................................................44
8.8 Standby Mode ........................................................................................................44
8.9 Extended Standby Mode ........................................................................................44
8.10 Power Reduction Register ....................................................................................45
8.11 Minimizing Power Consumption ...........................................................................45
9 System Control and Reset .................................................................... 50

9.1 Internal Voltage Reference .....................................................................................54
9.2 Watchdog Timer .....................................................................................................55
9.3 Register Description ...............................................................................................58
10 Interrupts ................................................................................................ 61
10.1 Overview ..............................................................................................................61
10.2 Interrupt Vectors in ATmega164P/324P/644P .....................................................61
11 External Interrupts ................................................................................. 66

11.1 Overview ..............................................................................................................66
12 I/O-Ports .................................................................................................. 71
12.1 Overview ..............................................................................................................71
12.2 Ports as General Digital I/O ..................................................................................72
12.3 Alternate Port Functions .......................................................................................77
13 8-bit Timer/Counter0 with PWM ............................................................ 92
ii ATmega164P/324P/644P
8011D–AVR–02/07
13.1 Features ...............................................................................................................92
13.2 Overview ..............................................................................................................92
13.3 Timer/Counter Clock Sources ..............................................................................93
13.4 Counter Unit .........................................................................................................93
13.5 Output Compare Unit ...........................................................................................94
13.6 Compare Match Output Unit .................................................................................96
13.7 Modes of Operation ..............................................................................................97
13.8 Timer/Counter Timing Diagrams ........................................................................101
13.9 Register Description ...........................................................................................103
14 16-bit Timer/Counter1 with PWM ........................................................ 110

14.1 Features .............................................................................................................110
14.2 Overview ............................................................................................................110
14.3 Accessing 16-bit Registers .................................................................................112
14.4 Timer/Counter Clock Sources ............................................................................115
14.5 Counter Unit .......................................................................................................116
14.6 Input Capture Unit ..............................................................................................117
14.7 Output Compare Units ........................................................................................119
14.8 Compare Match Output Unit ...............................................................................121
14.9 Modes of Operation ............................................................................................122
14.10 Timer/Counter Timing Diagrams ......................................................................129
14.11 Register Description .........................................................................................131
15 8-bit Timer/Counter2 with PWM and Asynchronous Operation ...... 138

15.1 Features .............................................................................................................138
15.2 Overview ............................................................................................................138
15.3 Timer/Counter Clock Sources ............................................................................139
15.4 Counter Unit .......................................................................................................140
15.5 Output Compare Unit .........................................................................................140
15.6 Compare Match Output Unit ...............................................................................142
15.7 Modes of Operation ............................................................................................143
15.8 Timer/Counter Timing Diagrams ........................................................................147
15.9 Asynchronous Operation of Timer/Counter2 ......................................................149
15.10 Timer/Counter Prescaler ..................................................................................151
16 SPI – Serial Peripheral Interface ......................................................... 159

16.1 Features .............................................................................................................159
iii
8011D–AVR–02/07
16.2 Overview ............................................................................................................159
16.3 SS Pin Functionality ...........................................................................................164
16.4 Data Modes ........................................................................................................164
17 USART ................................................................................................... 169

17.1 Features .............................................................................................................169
17.2 USART1 and USART0 .......................................................................................169
17.3 Overview ............................................................................................................169
17.4 Clock Generation ................................................................................................170
17.5 Frame Formats ...................................................................................................173
17.6 USART Initialization ...........................................................................................174
17.7 Data Transmission – The USART Transmitter ...................................................175
17.8 Data Reception – The USART Receiver ............................................................178
17.9 Asynchronous Data Reception ...........................................................................182
17.10 Multi-processor Communication Mode .............................................................185
17.12 Examples of Baud Rate Setting .......................................................................192
18 USART in SPI Mode ............................................................................. 196

18.1 Features .............................................................................................................196
18.2 Overview ............................................................................................................196
18.3 Clock Generation ................................................................................................196
18.4 SPI Data Modes and Timing ..............................................................................197
18.5 Frame Formats ...................................................................................................197
18.6 Data Transfer .....................................................................................................199
18.7 AVR USART MSPIM vs. AVR SPI .....................................................................201
19 2-wire Serial Interface .......................................................................... 205

19.1 Features .............................................................................................................205
19.2 2-wire Serial Interface Bus Definition .................................................................205
19.3 Data Transfer and Frame Format .......................................................................206
19.4 Multi-master Bus Systems, Arbitration and Synchronization ..............................209
19.5 Overview of the TWI Module ..............................................................................211
19.6 Using the TWI .....................................................................................................213
19.7 Transmission Modes ..........................................................................................216
19.8 Multi-master Systems and Arbitration .................................................................229
iv ATmega164P/324P/644P
8011D–AVR–02/07
20 AC - Analog Comparator ..................................................................... 235

20.1 Overview ............................................................................................................235
20.2 Analog Comparator Multiplexed Input ................................................................235
21 ADC - Analog-to-digital Converter ..................................................... 238

21.1 Features .............................................................................................................238
21.2 Overview ............................................................................................................238
21.3 Operation ............................................................................................................239
21.4 Starting a Conversion .........................................................................................240
21.5 Prescaling and Conversion Timing .....................................................................241
21.6 Changing Channel or Reference Selection ........................................................244
21.7 ADC Noise Canceler ..........................................................................................246
21.8 ADC Conversion Result .....................................................................................251
22 JTAG Interface and On-chip Debug System ..................................... 258

22.1 Features .............................................................................................................258
22.2 Overview ............................................................................................................258
22.3 TAP – Test Access Port .....................................................................................258
22.4 TAP Controller ....................................................................................................260
22.5 Using the Boundary-scan Chain .........................................................................261
22.6 Using the On-chip Debug System ......................................................................261
22.7 On-chip Debug Specific JTAG Instructions ........................................................262
22.8 Using the JTAG Programming Capabilities ........................................................262
22.9 Bibliography ........................................................................................................263
23 IEEE 1149.1 (JTAG) Boundary-scan ................................................... 264

23.1 Features .............................................................................................................264
23.2 Overview ............................................................................................................264
23.3 Data Registers ....................................................................................................265
23.4 Boundary-scan Specific JTAG Instructions ........................................................266
23.5 Boundary-scan Chain .........................................................................................267
23.6 ATmega164P/324P/644P Boundary-scan Order ...............................................270
23.7 Boundary-scan Description Language Files .......................................................272
v
8011D–AVR–02/07
24 Boot Loader Support – Read-While-Write Self-Programming ......... 274
24.1 Features .............................................................................................................274
24.2 Overview ............................................................................................................274
24.3 Application and Boot Loader Flash Sections ......................................................274
24.4 Read-While-Write and No Read-While-Write Flash Sections .............................275
24.5 Boot Loader Lock Bits ........................................................................................277
24.6 Entering the Boot Loader Program .....................................................................278
24.7 Addressing the Flash During Self-Programming ................................................279
24.8 Self-Programming the Flash ...............................................................................280
25 Memory Programming ......................................................................... 291

25.1 Program And Data Memory Lock Bits ................................................................291
25.2 Fuse Bits ............................................................................................................292
25.3 Signature Bytes ..................................................................................................294
25.4 Calibration Byte ..................................................................................................294
25.5 Page Size ...........................................................................................................294
25.6 Parallel Programming Parameters, Pin Mapping, and Commands ....................295
25.7 Parallel Programming .........................................................................................297
25.8 Serial Downloading ............................................................................................306
25.9 Serial Programming Instruction set ....................................................................308
25.10 Programming via the JTAG Interface ...............................................................310
26 Electrical Characteristics .................................................................... 323

26.1 DC Characteristics .............................................................................................323
26.2 Speed Grades ....................................................................................................327
26.3 Clock Characteristics ..........................................................................................328
26.4 System and Reset Characteristics .....................................................................329
26.5 External Interrupts Characteristics .....................................................................329
26.6 SPI Timing Characteristics .................................................................................330
26.7 2-wire Serial Interface Characteristics ................................................................331
26.8 ADC Characteristics ...........................................................................................333
27 Typical Characteristics ........................................................................ 336

27.1 ATmega164P Typical Characterization ..............................................................336
27.2 ATmega324P Typical Characteristics ................................................................361
28 Register Summary ............................................................................... 386
vi ATmega164P/324P/644P
8011D–AVR–02/07
29 Instruction Set Summary .................................................................... 390
30 Ordering Information ........................................................................... 393

30.1 ATmega164P .....................................................................................................393
30.2 ATmega324P .....................................................................................................394
30.3 ATmega644P .....................................................................................................395
31 Packaging Information ........................................................................ 396

31.1 44A .....................................................................................................................396
31.2 40P6 ...................................................................................................................397
31.3 44M1 ..................................................................................................................398
32 Errata ..................................................................................................... 399

32.1 ATmega164P Rev. A ..........................................................................................399
32.2 ATmega324P Rev. A ..........................................................................................399
32.3 ATmega644P Rev. A ..........................................................................................399
33 Datasheet Revision History ................................................................ 400

33.1 Rev. 8011D - 02/07 ............................................................................................400
33.2 Rev. 8011C - 10/06 ............................................................................................400
33.3 Rev. 8011B - 09/06 ............................................................................................400
33.4 Rev. 8011A - 08/06 ............................................................................................400
Table of Contents....................................................................................... i
vii
8011D–AVR–02/07
Atmel Corporation Atmel Operations
2325 Orchard Parkway Memory RF/Automotive
San Jose, CA 95131, USA 2325 Orchard Parkway Theresienstrasse 2
Tel: 1(408) 441-0311 San Jose, CA 95131, USA Postfach 3535
Fax: 1(408) 487-2600 Tel: 1(408) 441-0311 74025 Heilbronn, Germany
Fax: 1(408) 436-4314 Tel: (49) 71-31-67-0
Fax: (49) 71-31-67-2340
Regional Headquarters Microcontrollers
Europe 2325 Orchard Parkway 1150 East Cheyenne Mtn. Blvd.
Atmel Sarl San Jose, CA 95131, USA Colorado Springs, CO 80906, USA
Route des Arsenaux 41 Tel: 1(408) 441-0311 Tel: 1(719) 576-3300
Case Postale 80 Fax: 1(408) 436-4314 Fax: 1(719) 540-1759
CH-1705 Fribourg
Switzerland La Chantrerie Biometrics/Imaging/Hi-Rel MPU/
Tel: (41) 26-426-5555 BP 70602 High Speed Converters/RF Datacom
Fax: (41) 26-426-5500 44306 Nantes Cedex 3, France Avenue de Rochepleine
Tel: (33) 2-40-18-18-18 BP 123
Asia Fax: (33) 2-40-18-19-60 38521 Saint-Egreve Cedex, France
Room 1219 Tel: (33) 4-76-58-30-00
Chinachem Golden Plaza ASIC/ASSP/Smart Cards Fax: (33) 4-76-58-34-80
77 Mody Road Tsimshatsui Zone Industrielle
East Kowloon 13106 Rousset Cedex, France
Hong Kong Tel: (33) 4-42-53-60-00
Tel: (852) 2721-9778 Fax: (33) 4-42-53-60-01
Fax: (852) 2722-1369
1150 East Cheyenne Mtn. Blvd.
Japan Colorado Springs, CO 80906, USA
9F, Tonetsu Shinkawa Bldg. Tel: 1(719) 576-3300
1-24-8 Shinkawa Fax: 1(719) 540-1759
Chuo-ku, Tokyo 104-0033
Japan Scottish Enterprise Technology Park
Tel: (81) 3-3523-3551 Maxwell Building
Fax: (81) 3-3523-7581 East Kilbride G75 0QR, Scotland
Tel: (44) 1355-803-000
Fax: (44) 1355-242-743
Literature Requests
www.atmel.com/literature
Disclaimer: The information in this document is provided in connection with Atmel products. No license, express or implied, by estoppel or otherwise, to any
intellectual property right is granted by this document or in connection with the sale of Atmel products. EXCEPT AS SET FORTH IN ATMEL’S TERMS AND CONDI-
TIONS OF SALE LOCATED ON ATMEL’S WEB SITE, ATMEL ASSUMES NO LIABILITY WHATSOEVER AND DISCLAIMS ANY EXPRESS, IMPLIED OR STATUTORY
WARRANTY RELATING TO ITS PRODUCTS INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICULAR
PURPOSE, OR NON-INFRINGEMENT. IN NO EVENT SHALL ATMEL BE LIABLE FOR ANY DIRECT, INDIRECT, CONSEQUENTIAL, PUNITIVE, SPECIAL OR INCIDEN-
TAL DAMAGES (INCLUDING, WITHOUT LIMITATION, DAMAGES FOR LOSS OF PROFITS, BUSINESS INTERRUPTION, OR LOSS OF INFORMATION) ARISING OUT
OF THE USE OR INABILITY TO USE THIS DOCUMENT, EVEN IF ATMEL HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. Atmel makes no
representations or warranties with respect to the accuracy or completeness of the contents of this document and reserves the right to make changes to specifications
and product descriptions at any time without notice. Atmel does not make any commitment to update the information contained herein. Unless specifically provided
otherwise, Atmel products are not suitable for, and shall not be used in, automotive applications. Atmel’s products are not intended, authorized, or warranted for use
as components in applications intended to support or sustain life.
© 2007 Atmel Corporation. All rights reserved. ATMEL ®, logo and combinations thereof, Everywhere You Are ®, AVR ®, AVR Studio ®, and oth-
ers, are registered trademarks or trademarks of Atmel Corporation or its subsidiaries. Other terms and product names may be trademarks of oth-
ers.
8011D–AVR–02/07

Data Sheet

Uploaded by

Copyright:

Available Formats

Data Sheet

Uploaded by

Document Information

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Data Sheet

Uploaded by

Copyright:

Available Formats

Features

• High-performance, Low-power AVR® 8-bit Microcontroller

(PCINT8/XCK0/T0) PB0 PA0 (ADC0/PCINT0)

(PCINT13/MOSI) PB5 PA4 (ADC4/PCINT4)

2.1 Block Diagram

Figure 2-1. Block Diagram

Watchdog A/D Analog

JTAG/OCD 16bit T/C 1

PORT C (8) PORT D (8)

TOSC2/PC7 TOSC1/PC6 PC5..0 PD7..0

2.2 Comparison Between ATmega164P, ATmega324P and ATmega644P

Table 2-1. Differences between ATmega164P and ATmega644P

2.3 Pin Descriptions

Digital supply voltage.

Port A serves as analog inputs to the Analog-to-digital Converter.

Output from the inverting Oscillator amplifier.

This is the analog reference pin for the Analog-to-digital Converter.

5. AVR CPU Core

Figure 5-1. Block Diagram of the AVR Architecture

Data Bus 8-bit

Data I/O Module 2

5.2 ALU – Arithmetic Logic Unit

5.3 Status Register

The AVR Status Register – SREG – is defined as:

• Bit 7 – I: Global Interrupt Enable

• Bit 6 – T: Bit Copy Storage

• Bit 5 – H: Half Carry Flag

• Bit 4 – S: Sign Bit, S = N ⊕ V

• Bit 3 – V: Two’s Complement Overflow Flag

• Bit 2 – N: Negative Flag

• Bit 1 – Z: Zero Flag

• Bit 0 – C: Carry Flag

Figure 5-2. AVR CPU General Purpose Working Registers

Figure 5-3. The X-, Y-, and Z-registers

5.5 Stack Pointer

5.5.2 RAMPZ – Extended Z-pointer Register for ELPM/SPM

Figure 5-4. The Z-pointer used by ELPM and SPM

5.6 Instruction Execution Timing

Figure 5-6. Single Cycle ALU Operation

Register Operands Fetch

ALU Operation Execute

Result Write Back

5.7 Reset and Interrupt Handling

Assembly Code Example

Assembly Code Example

5.7.1 Interrupt Response Time

6.2 In-System Reprogrammable Flash Program Memory

Application Flash Section

Boot Flash Section

6.3 SRAM Data Memory

Figure 6-2. Data Memory Map

Figure 6-3. On-chip Data SRAM Access Cycles

Memory Access Instruction Next Instruction

• Bits 15:12 – Res: Reserved Bits

• Bits 11:0 – EEAR8:0: EEPROM Address

• Bits 7:0 – EEDR7:0: EEPROM Data

6.6.3 EECR – The EEPROM Control Register

• Bits 7:6 – Res: Reserved Bits

• Bits 5:4 – EEPM1 and EEPM0: EEPROM Programming Mode Bits

Table 6-1. EEPROM Mode Bits