90usb646 647 1286 1287

Features
• High Performance, Low Power AVR® 8-Bit Microcontroller

• Advanced RISC Architecture
– 135 Powerful Instructions – Most Single Clock Cycle Execution
– 32 x 8 General Purpose Working Registers
– Fully Static Operation
– Up to 16 MIPS Throughput at 16 MHz
– On-Chip 2-cycle Multiplier
• Non-volatile Program and Data Memories
– 64/128K Bytes of In-System Self-Programmable Flash 8-bit
• Endurance: 100,000 Write/Erase Cycles
– Optional Boot Code Section with Independent Lock Bits Microcontroller
• In-System Programming by On-chip Boot Program hardware activated after
reset
with
• True Read-While-Write Operation
– 2K/4K (64K/128K Flash version) Bytes EEPROM
64/128K Bytes
• Endurance: 100,000 Write/Erase Cycles of ISP Flash
– 4K/8K (64K/128K Flash version) Bytes Internal SRAM
– Up to 64K Bytes Optional External Memory Space and USB
– Programming Lock for Software Security
• JTAG (IEEE std. 1149.1 compliant) Interface Controller
– Boundary-scan Capabilities According to the JTAG Standard
– Extensive On-chip Debug Support
•
– Programming of Flash, EEPROM, Fuses, and Lock Bits through the JTAG Interface
USB 2.0 Full-speed/Low-speed Device and On-The-Go Module
AT90USB646
– Complies fully with: AT90USB647
– Universal Serial Bus Specification REV 2.0
– On-The-Go Supplement to the USB 2.0 Specification Rev 1.0 AT90USB1286
– Supports data transfer rates up to 12 Mbit/s and 1.5 Mbit/s
• USB Full-speed/Low Speed Device Module with Interrupt on Transfer Completion AT90USB1287
– Endpoint 0 for Control Transfers : up to 64-bytes
– 6 Programmable Endpoints with IN or Out Directions and with Bulk, Interrupt or
Isochronous Transfers
– Configurable Endpoints size up to 256 bytes in double bank mode
Preliminary
– Fully independant 832 bytes USB DPRAM for endpoint memory allocation
– Suspend/Resume Interrupts
– Power-on Reset and USB Bus Reset
– 48 MHz PLL for Full-speed Bus Operation
– USB Bus Disconnection on Microcontroller Request
• USB OTG:
– Supports Host Negotiation Protocol (HNP) and Session Request Protocol (SRP)
for OTG dual-role devices
– Provide Status and control signals for software implementation of HNP and SRP
– Provides programmable times required for HNP and SRP
• Peripheral Features
– Two 8-bit Timer/Counters with Separate Prescaler and Compare Mode
– Two16-bit Timer/Counter with Separate Prescaler, Compare- and Capture Mode
– Real Time Counter with Separate Oscillator
– Two 8-bit PWM Channels
7593D–AVR–07/06
– Six PWM Channels with Programmable Resolution from 2 to 16 Bits
– Output Compare Modulator
– 8-channels, 10-bit ADC
– Programmable Serial USART
– Master/Slave SPI Serial Interface
– Byte Oriented 2-wire Serial Interface
– Programmable Watchdog Timer with Separate On-chip Oscillator
– On-chip Analog Comparator
– Interrupt and Wake-up on Pin Change
• Special Microcontroller Features
– Power-on Reset and Programmable Brown-out Detection
– Internal Calibrated 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
– 48 Programmable I/O Lines
– 64-lead TQFP and 64-lead QFN
• Operating Voltages
– 2.7 - 5.5V
– 2.2 - 5.5V (Check availabilty)
• Operating temperature
– Industrial (-40°C to +85°C)
• Maximum Frequency
– 8 MHz at 2.7V - Industrial range
– 16 MHz at 4.5V - Industrial range
2 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
1. Pin Configurations
Figure 1-1. Pinout AT90USB64/128-TQFP
PF5 (ADC5/TMS)
PF6 (ADC6/TDO)
PF4 (ADC4/TCK)
PF7 (ADC7/TDI)
PF0 (ADC0)
PF1 (ADC1)
PF2 (ADC2)
PF3 (ADC3)
PA0 (AD0)
PA1 (AD1)
PA2 (AD2)
AVCC
AREF
GND
GND
VCC
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
(INT.6/AIN.0) PE6 1 48 PA3 (AD3)
(INT.7/AIN.1/UVcon) PE7 2 47 PA4 (AD4)
3 INDEX CORNER
UVcc 46 PA5 (AD5)
D- 4 45 PA6 (AD6)
D+ 5 44 PA7 (AD7)
UGnd 6 43 PE2 (ALE/HWB)
UCap 7 42 PC7 (A15/IC.3/CLK0)
VBus 8 AVR USB 41 PC6 (A14/OC.3A)
(IUID) PE3 9 TQFP64 40 PC5 (A13/OC.3B)
(SS/PCINT0) PB0 10 39 PC4 (A12/OC.3C)
(PCINT1/SCLK) PB1 11 38 PC3 (A11/T.3)
(PDI/PCINT2/MOSI) PB2 12 37 PC2 (A10)
(PDO/PCINT3/MISO) PB3 13 36 PC1 (A9)
(PCINT4/OC.2A) PB4 14 35 PC0 (A8)
(PCINT5/OC.1A) PB5 15 34 PE1 (RD)
(PCINT6/OC.1B) PB6 16 33 PE0 (WR)

17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
VCC
GND
XTAL2
XTAL1
(PCINT7/OC.0A/OC.1C) PB7
(INT4/TOSC1) PE4
(INT.5/TOSC2) PE5
RESET
(OC0B/SCL/INT0) PD0
(OC2B/SDA/INT1) PD1
(RXD1/INT2) PD2
(TXD1/INT3) PD3
(ICP1) PD4
(XCK1) PD5
(T1) PD6
(T0) PD7
3
7593D–AVR–07/06
Figure 1-2. Pinout AT90USB64/128-QFN
PF5 (ADC5/TMS)
PF6 (ADC6/TDO)
PF4 (ADC4/TCK)
PF7 (ADC7/TDI)
PF0 (ADC0)
PF1 (ADC1)
PF2 (ADC2)
PF3 (ADC3)
PA0 (AD0)
PA1 (AD1)
PA2 (AD2)
AVCC
AREF
GND
GND
VCC
64
63
62
61
60
59
58
57
56
55
54
53
52
51
50
49
(INT.6/AIN.0) PE6 1 48 PA3 (AD3)
(INT.7/AIN.1/UVcon) PE7 2 47 PA4 (AD4)
UVcc 3 46 PA5 (AD5)
INDEX CORNER
D- 4 45 PA6 (AD6)
D+ 5 44 PA7 (AD7)
UGnd 6 43 PE2 (ALE/HWB)
UCap 7 42 PC7 (A15/IC.3/CLK0)
VBus 8 AT90USB128 41 PC6 (A14/OC.3A)
(IUID) PE3 9 40 PC5 (A13/OC.3B)
(SS/PCINT0) PB0 10
(64-lead QFN top view) 39 PC4 (A12/OC.3C)
(PCINT1/SCLK) PB1 11 38 PC3 (A11/T.3)
(PDI/PCINT2/MOSI) PB2 12 37 PC2 (A10)
(PDO/PCINT3/MISO) PB3 13 36 PC1 (A9)
(PCINT4/OC.2A) PB4 14 35 PC0 (A8)
(PCINT5/OC.1A) PB5 15 34 PE1 (RD)
(PCINT6/OC.1B) PB6 16 33 PE0 (WR)
20
28
17
18
19
21
22
23
24
25
26
27
29
30
31
32
VCC
GND
XTAL2
XTAL1
(PCINT7/OC.0A/OC.1C) PB7
(INT4/TOSC1) PE4
(INT.5/TOSC2) PE5
RESET
(OC0B/SCL/INT0) PD0
(OC2B/SDA/INT1) PD1
(RXD1/INT2) PD2
(TXD1/INT3) PD3
(ICP1) PD4
(XCK1) PD5
(T1) PD6
(T0) PD7
Note: The large center pad underneath the MLF packages is made of metal and internally connected to
GND. It should be soldered or glued to the board to ensure good mechanical stability. If the center
pad is left unconnected, the package might loosen from the board.
1.1 Disclaimer
Typical values contained in this datasheet are based on simulations and characterization of
other AVR microcontrollers manufactured on the same process technology. Min and Max values
will be available after the device is characterized.
2. Overview
The AT90USB64/128 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
4 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
AT90USB64/128 achieves throughputs approaching 1 MIPS per MHz allowing the system
designer to optimize power consumption versus processing speed.
2.1 Block Diagram
Figure 2-1. Block Diagram
RESET
XTAL2
XTAL1
PF7 - PF0 PA7 - PA0 PC7 - PC0
VCC PORTF DRIVERS PORTA DRIVERS PORTC DRIVERS
GND
DATA REGISTER DATA DIR. DATA REGISTER DATA DIR. DATA REGISTER DATA DIR.
PORTF REG. PORTF PORTA REG. PORTA PORTC REG. PORTC
8-BIT DATA BUS
POR - BOD
RESET
AVCC INTERNAL
CALIB. OSC
OSCILLATOR
ADC
AGND
AREF OSCILLATOR
WATCHDOG
PROGRAM STACK TIMER
JTAG TAP
COUNTER POINTER
OSCILLATOR
PROGRAM SRAM MCU CONTROL TIMING AND

ON-CHIP DEBUG FLASH REGISTER CONTROL
BOUNDARY-
INSTRUCTION TIMER/
SCAN GENERAL
REGISTER COUNTERS
PURPOSE
REGISTERS
X
PROGRAMMING
INSTRUCTION Y INTERRUPT
LOGIC
DECODER Z UNIT
CONTROL
LINES ALU EEPROM
PLL
STATUS
REGISTER
TWO-WIRE SERIAL
USART0 SPI USB
INTERFACE
COMPARATOR
ANALOG
DATA REGISTER DATA DIR. DATA REGISTER DATA DIR. DATA REGISTER DATA DIR. DATA REG. DATA DIR.
PORTE REG. PORTE PORTB REG. PORTB PORTD REG. PORTD PORTG REG. PORTG
-
+
PORTE DRIVERS PORTB DRIVERS PORTD DRIVERS PORTG DRIVERS
PE7 - PE0 PB7 - PB0 PD7 - PD0 PG4 - PG0
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.
5
7593D–AVR–07/06
The AT90USB64/128 provides the following features: 64/128K bytes of In-System Programma-
ble Flash with Read-While-Write capabilities, 2K/4K bytes EEPROM, 4K/8K bytes SRAM, 48
general purpose I/O lines, 32 general purpose working registers, Real Time Counter (RTC), four
flexible Timer/Counters with compare modes and PWM, one USART, a byte oriented 2-wire
Serial Interface, a 8-channels, 10-bit ADC with optional differential input stage with programma-
ble 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 function-
ing. 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 asyn-
chronous 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 AT90USB64/128 is a powerful microcontroller that provides a highly flexible and cost
effective solution to many embedded control applications.
The AT90USB64/128 AVR is supported with a full suite of program and system development
tools including: C compilers, macro assemblers, program debugger/simulators, in-circuit emula-
tors, and evaluation kits.
2.2 Pin Descriptions
2.2.1 VCC
Digital supply voltage.
2.2.2 GND
Ground.
2.2.3 Port A (PA7..PA0)

Port A is 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 AT90USB64/128 as listed on
page 81.
6 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
2.2.4 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,
Port B has better driving capabilities than the other ports.
Port B also serves the functions of various special features of the AT90USB64/128 as listed on
page 82.
2.2.5 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 special features of the AT90USB64/128 as listed on page 85.
2.2.6 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 AT90USB64/128 as listed on
page 86.
2.2.7 Port E (PE7..PE0)

Port E is an 8-bit bi-directional I/O port with internal pull-up resistors (selected for each bit). The
Port E output buffers have symmetrical drive characteristics with both high sink and source
capability. As inputs, Port E pins that are externally pulled low will source current if the pull-up
resistors are activated. The Port E pins are tri-stated when a reset condition becomes active,
Port E also serves the functions of various special features of the AT90USB64/128 as listed on
page 89.
2.2.8 Port F (PF7..PF0)

Port F serves as analog inputs to the A/D Converter.
Port F also serves as an 8-bit bi-directional I/O port, if the A/D Converter is not used. Port pins
can provide internal pull-up resistors (selected for each bit). The Port F output buffers have sym-
metrical drive characteristics with both high sink and source capability. As inputs, Port F pins
that are externally pulled low will source current if the pull-up resistors are activated. The Port F
pins are tri-stated when a reset condition becomes active, even if the clock is not running. If the
JTAG interface is enabled, the pull-up resistors on pins PF7(TDI), PF5(TMS), and PF4(TCK) will
be activated even if a reset occurs.
Port F also serves the functions of the JTAG interface.
7
7593D–AVR–07/06
2.2.9 D-
USB Full speed / Low Speed Negative Data Upstream Port. Should be connected to the USB D-
connector pin with a serial 22 Ohms resistor.
2.2.10 D+
USB Full speed / Low Speed Positive Data Upstream Port. Should be connected to the USB D+
connector pin with a serial 22 Ohms resistor.
2.2.11 UGND
USB Pads Ground.
2.2.12 UVCC
USB Pads Internal Regulator Input supply voltage.
2.2.13 UCAP
USB Pads Internal Regulator Output supply voltage. Should be connected to an external capac-
itor (1µF).
2.2.14 VBUS
USB VBUS monitor and OTG negociations.
2.2.15 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 Table 8-1 on page
60. Shorter pulses are not guaranteed to generate a reset.
2.2.16 XTAL1
Input to the inverting Oscillator amplifier and input to the internal clock operating circuit.
2.2.17 XTAL2
Output from the inverting Oscillator amplifier.
2.2.18 AVCC
AVCC is the supply voltage pin for Port F and the A/D Converter. It should be externally con-
nected 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.
2.2.19 AREF
This is the analog reference pin for the A/D Converter.
3. 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.
These 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"
8 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
instructions must be replaced with instructions that allow access to extended I/O. Typically
"LDS" and "STS" combined with "SBRS", "SBRC", "SBR", and "CBR".
9
7593D–AVR–07/06
4. AVR CPU Core
4.1 Introduction
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.
4.2 Architectural Overview
Figure 4-1. 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.
10 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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-
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
AT90USB64/128 has Extended I/O space from 0x60 - 0x0FF in SRAM where only the
ST/STS/STD and LD/LDS/LDD instructions can be used.
4.3 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.
11
7593D–AVR–07/06
4.4 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
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.
The AVR Status Register – SREG – is defined as:
Bit 7 6 5 4 3 2 1 0
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.
12 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
• 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.
4.5 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 4-2 shows the structure of the 32 general purpose working registers in the CPU.
Figure 4-2. 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 4-2, 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.
4.5.1 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 4-3.
13
7593D–AVR–07/06
Figure 4-3. 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).
4.6 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.
Bit 15 14 13 12 11 10 9 8
SP15 SP14 SP13 SP12 SP11 SP10 SP9 SP8 SPH
SP7 SP6 SP5 SP4 SP3 SP2 SP1 SP0 SPL
7 6 5 4 3 2 1 0
R/W R/W R/W R/W R/W R/W R/W R/W
1 1 1 1 1 1 1 1
14 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
4.6.1 Extended Z-pointer Register for ELPM/SPM - RAMPZ
Bit 7 6 5 4 3 2 1 0
RAMPZ RAMPZ RAMPZ RAMPZ RAMPZ RAMPZ RAMPZ1 RAMPZ0 RAMPZ
7 6 5 4 3 2
For ELPM/SPM instructions, the Z-pointer is a concatenation of RAMPZ, ZH, and ZL, as shown
in Figure 4-4. Note that LPM is not affected by the RAMPZ setting.
Figure 4-4. 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.
4.7 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 4-5 shows the parallel instruction fetches and instruction executions enabled by the Har-
vard architecture and the fast-access Register File concept. This is the basic pipelining 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.
Figure 4-5. 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 4-6 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.
15
7593D–AVR–07/06
Figure 4-6. Single Cycle ALU Operation
T1 T2 T3 T4
clkCPU
Total Execution Time
Register Operands Fetch
ALU Operation Execute
Result Write Back
4.8 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 368 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 70. 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 70 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 368.
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
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.
16 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
17
7593D–AVR–07/06
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) */
4.8.1 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.
18 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
5. AVR AT90USB64/128 Memories
This section describes the different memories in the AT90USB64/128. The AVR architecture has
two main memory spaces, the Data Memory and the Program Memory space. In addition, the
AT90USB64/128 features an EEPROM Memory for data storage. All three memory spaces are
linear and regular.
Table 5-1. Memory Mapping.

Memory Mnemonic AT90USB64 AT90USB128
Size Flash size 64 K bytes 128K bytes
Start Address - 0x00000
Flash
0x0FFFF(1) 0x1FFFF(1)
End Address Flash end
0x7FFF(2) 0xFFFF(2)
Size - 32 bytes
32
Registers
End Address - 0x001F
Size - 64 bytes
I/O
Registers
End Address - 0x005F
Size - 160 bytes
Ext I/O
Registers
End Address - 0x00FF
Size ISRAM size 4 K bytes 8 K bytes
Internal
Start Address ISRAM start 0x0100
SRAM
End Address ISRAM end 0x10FF 0x20FF
Size XMem size 0-64 K bytes
External
Start Address XMem start 0x1100 0x2100
Memory
End Address XMem end 0xFFFF
Size E2 size 2 K bytes 4K bytes
EEPROM Start Address - 0x0000
End Address E2 end 0x07FF 0x0FFF
Notes: 1. Byte address.

2. Word (16-bit) address.
5.1 In-System Reprogrammable Flash Program Memory

The AT90USB64/128 contains 128K 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
64K 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 100,000 write/erase cycles. The
AT90USB64/128 Program Counter (PC) is 16 bits wide, thus addressing the 128K program
19
7593D–AVR–07/06
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 368. “Memory
Programming” on page 368 contains a detailed description on Flash data serial downloading
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 and ELPM - Extended Load Program Memory
instruction description).
Timing diagrams for instruction fetch and execution are presented in “Instruction Execution Tim-
ing” on page 15.
Figure 5-1. Program Memory Map

Program Memory
0x00000
Application Flash Section
Boot Flash Section

Flash End
5.2 SRAM Data Memory

Figure 5-2 shows how the AT90USB64/128 SRAM Memory is organized.
The AT90USB64/128 is a complex microcontroller with more peripheral units than can be sup-
ported within the 64 location reserved in the Opcode for the IN and OUT instructions. For the
Extended I/O space from $060 - $1FF in SRAM, only the ST/STS/STD and LD/LDS/LDD instruc-
tions can be used.
The first 4,608/8,704 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 416 locations of Extended I/O memory
and the next 8,192 locations address the internal data SRAM.
20 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
An optional external data SRAM can be used with the AT90USB64/128. This SRAM will occupy
an area in the remaining address locations in the 64K address space. This area starts at the
address following the internal SRAM. The Register file, I/O, Extended I/O and Internal SRAM
occupies the lowest 4,608/8,704 bytes, so when using 64KB (65,536 bytes) of External Memory,
60,478/56,832 Bytes of External Memory are available. See “External Memory Interface” on
page 30 for details on how to take advantage of the external memory map.
When the addresses accessing the SRAM memory space exceeds the internal data memory
locations, the external data SRAM is accessed using the same instructions as for the internal
data memory access. When the internal data memories are accessed, the read and write strobe
pins (PE0 and PE1) are inactive during the whole access cycle. External SRAM operation is
enabled by setting the SRE bit in the XMCRA Register.
Accessing external SRAM takes one additional clock cycle per byte compared to access of the
internal SRAM. This means that the commands LD, ST, LDS, STS, LDD, STD, PUSH, and POP
take one additional clock cycle. If the Stack is placed in external SRAM, interrupts, subroutine
calls and returns take three clock cycles extra because the three-byte program counter is
pushed and popped, and external memory access does not take advantage of the internal pipe-
line memory access. When external SRAM interface is used with wait-state, one-byte external
access takes two, three, or four additional clock cycles for one, two, and three wait-states
respectively. Interrupts, subroutine calls and returns will need five, seven, or nine clock cycles
more than specified in the instruction set manual for one, two, and three wait-states.
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.
The 32 general purpose working registers, 64 I/O registers, and the 8,192 bytes of internal data
SRAM in the AT90USB64/128 are all accessible through all these addressing modes. The Reg-
ister File is described in “General Purpose Register File” on page 13.
21
7593D–AVR–07/06
Figure 5-2. Data Memory Map
Data Memory
32 Registers $0000 - $001F
64 I/O Registers $0020 - $005F
160 E xt I/O Reg. $0060 - $00FF
ISRAM start
Internal S RAM
(8192 x 8)
ISRAM end
XMem start
External S RAM
(0 - 64K x 8)
$FFFF
5.2.1 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 5-3.
22 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 5-3. On-chip Data SRAM Access Cycles
T1 T2 T3
clkCPU
Address Compute Address Address valid
Data
Write
WR
Data
Read
RD
Memory Access Instruction Next Instruction
5.3 EEPROM Data Memory

The AT90USB64/128 contains 2K/4K bytes of data EEPROM memory. It is organized as a sep-
arate 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 382, page 387, and page 371 respectively.
5.3.1 EEPROM Read/Write Access

The EEPROM Access Registers are accessible in the I/O space.
The write access time for the EEPROM is given in Table 5-3. A self-timing function, however,
lets the user software detect when the next byte can be written. If the user code contains instruc-
tions 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 “Preventing EEPROM Corruption” on page 28. 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.
5.3.2 The EEPROM Address Register – EEARH and EEARL

Bit 15 14 13 12 11 10 9 8
– – – – EEAR11 EEAR10 EEAR9 EEAR8 EEARH
EEAR7 EEAR6 EEAR5 EEAR4 EEAR3 EEAR2 EEAR1 EEAR0 EEARL
23
7593D–AVR–07/06
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 AT90USB64/128 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.
5.3.3 The EEPROM Data Register – EEDR

Bit 7 6 5 4 3 2 1 0
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.
5.3.4 The EEPROM Control Register – EECR

Bit 7 6 5 4 3 2 1 0
– – 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 7..6 – Res: Reserved Bits

• 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 5-2. 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.
24 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 5-2. 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 368 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.
25
7593D–AVR–07/06
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 5-3 lists the typical pro-
gramming time for EEPROM access from the CPU.
Table 5-3. EEPROM Programming Time

Symbol Number of Calibrated RC Oscillator Cycles Typ Programming Time
EEPROM write
26,368 3.3 ms
(from CPU)
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.
26 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Assembly Code Example(1)

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.
27
7593D–AVR–07/06
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.

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;
}
5.3.5 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.
28 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
5.4 I/O Memory
The I/O space definition of the AT90USB64/128 is shown in “Register Summary” on page 414.
All AT90USB64/128 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 registers, 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 AT90USB64/128 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 -
0x1FF in SRAM, only the ST/STS/STD and LD/LDS/LDD instructions 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.
5.4.1 General Purpose I/O Registers

The AT90USB64/128 contains three General Purpose I/O Registers. 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.
5.4.2 General Purpose I/O Register 2 – GPIOR2

Bit 7 6 5 4 3 2 1 0
MSB LSB GPIOR2

Bit 7 6 5 4 3 2 1 0
MSB LSB GPIOR1

Bit 7 6 5 4 3 2 1 0
MSB LSB GPIOR0
29
7593D–AVR–07/06
5.5 External Memory Interface
With all the features the External Memory Interface provides, it is well suited to operate as an
interface to memory devices such as External SRAM and Flash, and peripherals such as LCD-
display, A/D, and D/A. The main features are:
• Four different wait-state settings (including no wait-state).
• Independent wait-state setting for different external Memory sectors (configurable sector size).
• The number of bits dedicated to address high byte is selectable.
• Bus keepers on data lines to minimize current consumption (optional).
5.5.1 Overview
When the eXternal MEMory (XMEM) is enabled, address space outside the internal SRAM
becomes available using the dedicated External Memory pins (see Figure 2-1 on page 5, Table
10-3 on page 81, and Table 10-9 on page 85). The memory configuration is shown in Figure 5-4.
Figure 5-4. External Memory with Sector Select

Memory Configuration A
0x0000
Internal memory
ISRAM end
XMem start
Lower sector
SRW01
SRW00
SRL[2..0]
External Memory Upper sector

(0-60K x 8)
SRW11
SRW10
0xFFFF
5.5.2 Using the External Memory Interface

The interface consists of:
• AD7:0: Multiplexed low-order address bus and data bus.
• A15:8: High-order address bus (configurable number of bits).
• ALE: Address latch enable.
• RD: Read strobe.
• WR: Write strobe.
The control bits for the External Memory Interface are located in two registers, the External
Memory Control Register A – XMCRA, and the External Memory Control Register B – XMCRB.
30 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
When the XMEM interface is enabled, the XMEM interface will override the setting in the data
direction registers that corresponds to the ports dedicated to the XMEM interface. For details
about the port override, see the alternate functions in section “I/O-Ports” on page 74. The XMEM
interface will auto-detect whether an access is internal or external. If the access is external, the
XMEM interface will output address, data, and the control signals on the ports according to Fig-
ure 5-6 (this figure shows the wave forms without wait-states). When ALE goes from high-to-low,
there is a valid address on AD7:0. ALE is low during a data transfer. When the XMEM interface
is enabled, also an internal access will cause activity on address, data and ALE ports, but the
RD and WR strobes will not toggle during internal access. When the External Memory Interface
is disabled, the normal pin and data direction settings are used. Note that when the XMEM inter-
face is disabled, the address space above the internal SRAM boundary is not mapped into the
internal SRAM. Figure 5-5 illustrates how to connect an external SRAM to the AVR using an
octal latch (typically “74 x 573” or equivalent) which is transparent when G is high.
5.5.3 Address Latch Requirements

Due to the high-speed operation of the XRAM interface, the address latch must be selected with
care for system frequencies above 8 MHz @ 4V and 4 MHz @ 2.7V. When operating at condi-
tions above these frequencies, the typical old style 74HC series latch becomes inadequate. The
External Memory Interface is designed in compliance to the 74AHC series latch. However, most
latches can be used as long they comply with the main timing parameters. The main parameters
for the address latch are:
• D to Q propagation delay (tPD).
• Data setup time before G low (tSU).
• Data (address) hold time after G low (TH).
The External Memory Interface is designed to guaranty minimum address hold time after G is
asserted low of th = 5 ns. Refer to tLAXX_LD/tLLAXX_ST in “External Data Memory Timing” Tables 30-
7 through Tables 30-13 on pages 408 - 411. The D-to-Q propagation delay (tPD) must be taken
into consideration when calculating the access time requirement of the external component. The
data setup time before G low (tSU) must not exceed address valid to ALE low (tAVLLC) minus PCB
wiring delay (dependent on the capacitive load).
Figure 5-5. External SRAM Connected to the AVR
D[7:0]
AD7:0 D Q A[7:0]
ALE G
SRAM
AVR
A15:8 A[15:8]
RD RD
WR WR
31
7593D–AVR–07/06
5.5.4 Pull-up and Bus-keeper
The pull-ups on the AD7:0 ports may be activated if the corresponding Port register is written to
one. To reduce power consumption in sleep mode, it is recommended to disable the pull-ups by
writing the Port register to zero before entering sleep.
The XMEM interface also provides a bus-keeper on the AD7:0 lines. The bus-keeper can be dis-
abled and enabled in software as described in “External Memory Control Register B – XMCRB”
on page 35. When enabled, the bus-keeper will keep the previous value on the AD7:0 bus while
these lines are tri-stated by the XMEM interface.
5.5.5 Timing
External Memory devices have different timing requirements. To meet these requirements, the
XMEM interface provides four different wait-states as shown in Table 5-5. It is important to con-
sider the timing specification of the External Memory device before selecting the wait-state. The
most important parameters are the access time for the external memory compared to the set-up
requirement. The access time for the External Memory is defined to be the time from receiving
the chip select/address until the data of this address actually is driven on the bus. The access
time cannot exceed the time from the ALE pulse must be asserted low until data is stable during
a read sequence (See tLLRL+ tRLRH - tDVRH in Tables 30-6 through Tables 30-13 on pages 408 -
411). The different wait-states are set up in software. As an additional feature, it is possible to
divide the external memory space in two sectors with individual wait-state settings. This makes it
possible to connect two different memory devices with different timing requirements to the same
XMEM interface. For XMEM interface timing details, please refer to Tables 30-6 through Tables
30-13 and Figure 30-7 to Figure 30-10 in the “External Data Memory Timing” on page 408.
Note that the XMEM interface is asynchronous and that the waveforms in the following figures
are related to the internal system clock. The skew between the internal and external clock
(XTAL1) is not guarantied (varies between devices temperature, and supply voltage). Conse-
quently, the XMEM interface is not suited for synchronous operation.
Figure 5-6. External Data Memory Cycles without Wait-state (SRWn1=0 and SRWn0=0)
T1 T2 T3 T4
System Clock (CLKCPU )
ALE
A15:8 Prev. addr. Address
Write
DA7:0 Prev. data Address XX Data
WR
DA7:0 (XMBK = 0) Prev. data Address Data

Read
RD
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).
32 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 5-7. External Data Memory Cycles with SRWn1 = 0 and SRWn0 = 1(1)
T1 T2 T3 T4 T5
ALE
Write
WR
Read
RD
SRW00 (lower sector).
The ALE pulse in period T5 is only present if the next instruction accesses the RAM (internal
or external).
T1 T2 T3 T4 T5
ALE
Write
WR
Read
RD
or external).
33
7593D–AVR–07/06
T1 T2 T3 T4 T5 T6
ALE
Write
WR
Read
RD
or external).
5.5.6 External Memory Control Register A – XMCRA

Bit 7 6 5 4 3 2 1 0
SRE SRL2 SRL1 SRL0 SRW11 SRW10 SRW01 SRW00 XMCRA
• Bit 7 – SRE: External SRAM/XMEM Enable

Writing SRE to one enables the External Memory Interface.The pin functions AD7:0, A15:8,
ALE, WR, and RD are activated as the alternate pin functions. The SRE bit overrides any pin
direction settings in the respective data direction registers. Writing SRE to zero, disables the
External Memory Interface and the normal pin and data direction settings are used.
• Bit 6..4 – SRL2:0: Wait-state Sector Limit

It is possible to configure different wait-states for different External Memory addresses. The
external memory address space can be divided in two sectors that have separate wait-state bits.
The SRL2, SRL1, and SRL0 bits select the split of the sectors, see Table 5-4 and Figure 5-4. By
default, the SRL2, SRL1, and SRL0 bits are set to zero and the entire external memory address
space is treated as one sector. When the entire SRAM address space is configured as one sec-
tor, the wait-states are configured by the SRW11 and SRW10 bits.
34 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 5-4. Sector limits with different settings of SRL2..0

SRL2 SRL1 SRL0 Sector Limits
Lower sector = N/A
0 0 x
Upper sector = 0x2100 - 0xFFFF
Lower sector = 0x2100 - 0x3FFF
0 1 0
0 1 1
1 0 0
1 0 1
Upper sector = 0xA000 - 0xFFFF
Lower sector = 0x2100 - 0xBFFF
1 1 0
Upper sector = 0xC000 - 0xFFFF
Lower sector = 0x2100 - 0xDFFF
1 1 1
Upper sector = 0xE000 - 0xFFFF
• Bit 3..2 – SRW11, SRW10: Wait-state Select Bits for Upper Sector
The SRW11 and SRW10 bits control the number of wait-states for the upper sector of the exter-
nal memory address space, see Table 5-5.
• Bit 1..0 – SRW01, SRW00: Wait-state Select Bits for Lower Sector
The SRW01 and SRW00 bits control the number of wait-states for the lower sector of the exter-
nal memory address space, see Table 5-5.
Table 5-5. Wait States(1)

SRWn1 SRWn0 Wait States
0 0 No wait-states
0 1 Wait one cycle during read/write strobe
1 0 Wait two cycles during read/write strobe
Wait two cycles during read/write and wait one cycle before driving out
1 1
new address
Note: 1. n = 0 or 1 (lower/upper sector).

For further details of the timing and wait-states of the External Memory Interface, see Figures
5-6 through Figures 5-9 for how the setting of the SRW bits affects the timing.
5.5.7 External Memory Control Register B – XMCRB

Bit 7 6 5 4 3 2 1 0
XMBK – – – – XMM2 XMM1 XMM0 XMCRB
Read/Write R/W R R R R R/W R/W R/W
• Bit 7– XMBK: External Memory Bus-keeper Enable

Writing XMBK to one enables the bus keeper on the AD7:0 lines. When the bus keeper is
enabled, AD7:0 will keep the last driven value on the lines even if the XMEM interface has tri-
stated the lines. Writing XMBK to zero disables the bus keeper. XMBK is not qualified with SRE,
35
7593D–AVR–07/06
so even if the XMEM interface is disabled, the bus keepers are still activated as long as XMBK is
one.
• Bit 6..3 – Res: Reserved Bits

These bits are reserved and will always read as zero. When writing to this address location,
write these bits to zero for compatibility with future devices.
• Bit 2..0 – XMM2, XMM1, XMM0: External Memory High Mask

When the External Memory is enabled, all Port C pins are default used for the high address byte.
If the full 60KB address space is not required to access the External Memory, some, or all, Port
C pins can be released for normal Port Pin function as described in Table 5-6. As described in
“Using all 64KB Locations of External Memory” on page 37, it is possible to use the XMMn bits to
access all 64KB locations of the External Memory.
Table 5-6. Port C Pins Released as Normal Port Pins when the External Memory is Enabled
XMM2 XMM1 XMM0 # Bits for External Memory Address Released Port Pins
0 0 0 8 (Full 56KB space) None
0 0 1 7 PC7
0 1 0 6 PC7 - PC6
0 1 1 5 PC7 - PC5
1 0 0 4 PC7 - PC4
1 0 1 3 PC7 - PC3
1 1 0 2 PC7 - PC2
1 1 1 No Address high bits Full Port C
5.5.8 Using all Locations of External Memory Smaller than 64 KB

Since the external memory is mapped after the internal memory as shown in Figure 5-4, the
external memory is not addressed when addressing the first 8,448/4,352 bytes (128/64Kbytes
version) of data space. It may appear that the first 8,448/4,352 bytes of the external memory are
inaccessible (external memory addresses 0x0000 to 0x10FF or 0x0000 to 0x20FF). However,
when connecting an external memory smaller than 64 KB, for example 32 KB, these locations
are easily accessed simply by addressing from address 0x8000 to 0xA1FF. Since the External
Memory Address bit A15 is not connected to the external memory, addresses 0x8000 to 0xA1FF
will appear as addresses 0x0000 to 0x21FF for the external memory. Addressing above address
0xA1FF is not recommended, since this will address an external memory location that is already
accessed by another (lower) address. To the Application software, the external 32 KB memory
will appear as one linear 32 KB address space from 0x2200 to 0xA1FF. This is illustrated in Fig-
ure 5-10.
36 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 5-10. Address Map with 32 KB External Memory
Memory Configuration A
AVR Memory Map External 32K S RAM
0x0000 0x0000
Internal Memory
0x20FF ISRAM end

0x2100 XMem start
0x7FFF External
0x7FFF
0x8000 Memory
ISRAM end + 0x8000

XMem start + 0x8000
(Unused)
0xFFFF
5.5.9 Using all 64KB Locations of External Memory

Since the External Memory is mapped after the Internal Memory as shown in Figure 5-4, only
56KB of External Memory is available by default (address space 0x0000 to 0x20FF is reserved
for internal memory). However, it is possible to take advantage of the entire External Memory by
masking the higher address bits to zero. This can be done by using the XMMn bits and control
by software the most significant bits of the address. By setting Port C to output 0x00, and releas-
ing the most significant bits for normal Port Pin operation, the Memory Interface will address
0x0000 - 0x2FFF. See the following code examples.
Care must be exercised using this option as most of the memory is masked away.
37
7593D–AVR–07/06
; OFFSET is defined to 0x4000 to ensure
; external memory access
; Configure Port C (address high byte) to
; output 0x00 when the pins are released
; for normal Port Pin operation
ldi r16, 0xFF
out DDRC, r16
ldi r16, 0x00
out PORTC, r16
; release PC7:6
ldi r16, (1<<XMM1)
sts XMCRB, r16
; write 0xAA to address 0x0001 of external
; memory
ldi r16, 0xaa
sts 0x0001+OFFSET, r16
; re-enable PC7:6 for external memory
ldi r16, (0<<XMM1)
sts XMCRB, r16
; store 0x55 to address (OFFSET + 1) of
; external memory
ldi r16, 0x55
sts 0x0001+OFFSET, r16
C Code Example(1)
#define OFFSET 0x4000
void XRAM_example(void)
{
unsigned char *p = (unsigned char *) (OFFSET + 1);
DDRC = 0xFF;
PORTC = 0x00;
XMCRB = (1<<XMM1);
*p = 0xaa;
XMCRB = 0x00;
*p = 0x55;
}
38 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
6. System Clock and Clock Options
6.1 Clock Systems and their Distribution

Figure 6-1 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 53. The clock systems are detailed below.
Figure 6-1. Clock Distribution

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

clkUSB (48MHz) Control Unit
USB PLL clkASY clkFLASH

X24
Reset Logic Watchdog Timer
clkPllin (2MHz)
Source clock Watchdog clock
System Clock
Prescaler
PLL Clock
Prescaler
Clock
Multiplexer
clkXTAL (2-16 MHz)
Timer/Counter Crystal Watchdog Calibrated RC

External Clock
Oscillator Oscillator Oscillator Oscillator
6.1.1 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.
6.1.2 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.
6.1.3 Flash Clock – clkFLASH

The Flash clock controls operation of the Flash interface. The Flash clock is usually active simul-
taneously with the CPU clock.
39
7593D–AVR–07/06
6.1.4 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.
6.1.5 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.
6.1.6 USB Clock – clkUSB

The USB is provided with a dedicated clock domain. This clock is generated with an on-chip PLL
running at 48MHz. The PLL always multiply its input frequency by 24. Thus the PLL clock regis-
ter should be programmed by software to generate a 2MHz clock on the PLL input.
6.2 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 6-1. Device Clocking Options Select(1)

Device Clocking Option CKSEL3..0
Low Power Crystal Oscillator 1111 - 1000
Reserved 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.
6.2.1 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.
6.2.2 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 58
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
selectable delays are shown in Table 6-2. The frequency of the Watchdog Oscillator is voltage
40 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
dependent as shown in “AT90USB64/128 Typical Characteristics – Preliminary Data” on page
429.
Table 6-2. 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.
6.3 Low Power Crystal Oscillator

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 6-2. Either a quartz crystal or a
ceramic resonator may be used.
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 “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 frequency of
the device, and if frequency stability at start-up is not important for the application.” on page 43.
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. Some initial guidelines for choosing capacitors for
use with crystals are given in Table 6-3. For ceramic resonators, the capacitor values given by
the manufacturer should be used.
41
7593D–AVR–07/06
Figure 6-2. Crystal Oscillator Connections
C2
XTAL2
C1
XTAL1
GND
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 6-3.
42 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 6-3. Low Power Crystal Oscillator Operating Modes(3)

Recommended Range for Capacitors
Frequency Range(1) (MHz) CKSEL3..1 C1 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
6-4.
Table 6-4. 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,
258 CK 14CK + 65 ms(1) 0 01
slowly rising power
Ceramic resonator,
1K CK 14CK(2) 0 10
BOD enabled
Ceramic resonator, fast
1K CK 14CK + 4.1 ms(2) 0 11
rising power
Ceramic resonator,
1K CK 14CK + 65 ms(2) 1 00
slowly 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,
16K CK 14CK + 65 ms 1 11
slowly 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.
43
7593D–AVR–07/06
Table 6-5. 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.
6.4 Low Frequency Crystal Oscillator

The device can utilize a 32.768 kHz watch crystal as clock source by a dedicated Low Fre-
quency Crystal Oscillator. The crystal should be connected as shown in Figure 6-2. When this
Oscillator is selected, start-up times are determined by the SUT Fuses and CKSEL0 as shown in
Table 6-6.
Table 6-6. Start-up Times for the Low Frequency Crystal Oscillator Clock Selection
Start-up Time from Additional Delay
Power-down and from Reset
Power Conditions Power-save (VCC = 5.0V) CKSEL0 SUT1..0
(1)
BOD enabled 1K CK 14CK 0 00
Fast rising power 1K CK 14CK + 4.1 ms(1) 0 01
(1)
Slowly rising power 1K CK 14CK + 65 ms 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.
6.5 Calibrated Internal RC Oscillator

The calibrated internal RC Oscillator by default provides a 8.0 MHz clock. The frequency is nom-
inal value at 3V and 25°C. The device is shipped with the CKDIV8 Fuse programmed. See
“System Clock Prescaler” on page 48 for more details. This clock may be selected as the system
clock by programming the CKSEL Fuses as shown in Table 6-7. If selected, it will operate with
no external components. During reset, hardware loads the calibration byte into the OSCCAL
Register and thereby automatically calibrates the RC Oscillator. At 3V and 25°C, this calibration
gives a frequency of 8 MHz ± 1%. The oscillator can be calibrated to any frequency in the range
7.3 - 8.1 MHz within ±1% accuracy, by changing the OSCCAL register. When this Oscillator is
used as the chip clock, the Watchdog Oscillator will still be used for the Watchdog Timer and for
44 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
the Reset Time-out. For more information on the pre-programmed calibration value, see the sec-
tion “Calibration Byte” on page 371
Table 6-7. 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.
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.
When this Oscillator is selected, start-up times are determined by the SUT Fuses as shown in
Table 6-5 on page 44.
45
7593D–AVR–07/06
Table 6-8. Start-up times for the internal calibrated RC Oscillator clock selection
Slowly rising power 6 CK 14CK + 65 ms(1) 10
Reserved 11
Note: 1. The device is shipped with this option selected.
6.5.1 Oscillator Calibration Register – OSCCAL

Bit 7 6 5 4 3 2 1 0
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. The factory-calibrated value is automat-
ically written to this register during chip reset, giving an oscillator frequency of 8.0 MHz at 25°C.
The application software can write this register to change the oscillator frequency. The oscillator
can be calibrated to any frequency in the range 7.3 - 8.1 MHz within ±1% accuracy. 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. Incrementing CAL6..0 by 1 will give a frequency increment of less than 2% in the fre-
quency range 7.3 - 8.1 MHz.
6.6 128 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 “11” as shown in Table 6-9.
Table 6-9. 128 kHz Internal Oscillator Operating Modes

Nominal Frequency CKSEL3..0
128 kHz 0011
Note: 1. The frequency is preliminary value. Actual value is TBD.
46 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
When this clock source is selected, start-up times are determined by the SUT Fuses as shown in
Table 6-10.
Table 6-10. 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
6.7 External Clock

The device can utilize a external clock source as shown in Figure 6-3. To run the device on an
external clock, the CKSEL Fuses must be programmed as shown in Table 6-1.
Figure 6-3. 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 6-11.
Table 6-11. Start-up Times for the External Clock Selection

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
48 for details.
47
7593D–AVR–07/06
6.8 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.
6.9 Timer/Counter Oscillator

The device can operate its Timer/Counter2 from an external 32.768 kHz watch crystal or a exter-
nal clock source. See Figure 6-2 on page 42 for crystal connection.
Applying an external clock source to TOSC1 requires EXCLK in the ASSR Register written to
logic one. See “Asynchronous operation of the Timer/Counter” on page 167 for further descrip-
tion on selecting external clock as input instead of a 32 kHz crystal.
6.10 System Clock Prescaler

The AVR USB has a system clock prescaler, and the system clock can be divided by setting the
“Clock Prescale Register – CLKPR” on page 48. This feature can be used to decrease the sys-
tem 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 6-12.
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
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.
6.10.1 Clock Prescale Register – CLKPR

Bit 7 6 5 4 3 2 1 0
CLK- – – – CLKPS CLKPS CLKPS CLKPS CLKPR
PCE 3 2 1 0
Read/Write R/W R R R R/W R/W R/W R/W
Initial Value 0 0 0 0 See Bit Description
48 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
• 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
Table 6-12.
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.
49
7593D–AVR–07/06
Table 6-12. 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
6.11 PLL
The PLL is used to generate internal high frequency (48 MHz) clock for USB interface, the PLL
input is generated from an external low-frequency (the crystal oscillator or external clock input
pin from XTAL1). The internal RC Oscillator can not be used for USB operations.
6.11.1 Internal PLL for USB interface

The internal PLL in AT90USB64/128 generates a clock frequency that is 24x multiplied from
nominally 2 MHz input. The source of the 2 MHz PLL input clock is the output of the internal PLL
clock prescaler that generates the 2 MHz (See Section 6.11.2 for PLL interface).
50 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 6-4. PLL Clocking System
PLLE PLOCK
Lock
Detector
clk
PLL clock 2MHz PLL
clkUSB (48MHz)
Prescaler 24x
XTAL1
OSCILLATORS System Clock
XTAL2
RC OSCILLATOR
8 MHz
Watchdog
OSCILLATOR
6.11.2 PLL Control and Status Register – PLLCSR

Bit 7 6 5 4 3 2 1 0
$29 ($29) PLLP2 PLLP1 PLLP0 PLLE PLOCK PLLCSR
Read/Write R R R R R R R/W R
Initial Value 0 0 0 0 0 0 0/1 0

These bits are reserved bits in the AT90USB64/128 and always read as zero.
• Bit 4..2 – PLLP2:0 PLL prescaler

These bits allow to configure the PLL input prescaler to generate the 2MHz input clock for the
PLL.
Table 6-13. PLL input prescaler configurations

Clock Division External XTAL required for USB
PLLP2 PLLP1 PLLP0 Factor operation (MHz)
0 0 0 Reserved -
0 0 1 Reserved -
0 1 0 Reserved -
0 1 1 4 8
1 0 0 Reserved -
1 0 1 Reserved -
1 1 0 8 16
1 1 1 Reserved -
• Bit 1 – PLLE: PLL Enable

When the PLLE is set, the PLL is started.
51
7593D–AVR–07/06
• Bit 0 – PLOCK: PLL Lock Detector
When the PLOCK bit is set, the PLL is locked to the reference clock, and it is safe to enable PCK
for Timer/Counter1. After the PLL is enabled, it takes about 100 ms for the PLL to lock.
To clear PLOCK, clear PLLE and PLLPx bits.
52 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
7. Power Management and Sleep Modes
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 consump-
tion to the application’s requirements.
To enter any of the five 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 (Idle, ADC Noise Reduction, Power-down, Power-save, or Standby) will be
activated by the SLEEP instruction. See Table 7-1 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.
Figure 6-1 on page 39 presents the different clock systems in the AT90USB64/128, and their
distribution. The figure is helpful in selecting an appropriate sleep mode.
7.0.1 Sleep Mode Control Register – SMCR

The Sleep Mode Control Register contains control bits for power management.
Bit 7 6 5 4 3 2 1 0
– – – – SM2 SM1 SM0 SE SMCR
• Bits 3, 2, 1 – SM2..0: Sleep Mode Select Bits 2, 1, and 0

These bits select between the six available sleep modes as shown in Table 7-1.
Table 7-1. 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 1 – 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.
53
7593D–AVR–07/06
7.1 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 USB, 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.
7.2 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 (including clkUSB).
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.
7.3 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 INT7:4, an external interrupt on INT3:0, a pin change interrupt or an asynchronous
USB interrupt sources (VBUSTI, WAKEUPI, IDTI and HWUPI), can wake up the MCU. This
sleep mode basically halts all generated clocks, allowing operation of asynchronous 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 96
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 40.
7.4 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:
54 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
7.5 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.
7.6 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.
Table 7-2. Active Clock Domains and Wake-up Sources in the Different Sleep Modes.
Active Clock Domains Oscillators Wake-up Sources
USB Asynchonous
USB Synchronous
EEPROM Ready
WDT Interrupt
TWI Address
Interrupts(4)
Pin Change
Main Clock
INT7:0 and
Timer Osc
Interrupts
Other I/O
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 X X
(2) (3) (2)
ADCNRM X X X X X X X X X X X X
(3)
Power-down X X X X
(2) (3)
Power-save X X X X X X X
Standby(1) X X(3) 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 Timer/Counter2 is running in asynchronous mode.
3. For INT7:4, only level interrupt.
4. Asynchronous USB interrupts are VBUSTI, WAKEUPI, IDTI, WAKEUPI and HWUPI.
55
7593D–AVR–07/06
7.7 Power Reduction Register
The Power Reduction Register, PRR, provides a method to stop the clock to individual peripher-
als 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. See “Supply Current of IO modules” on page 429 for examples. In all other
sleep modes, the clock is already stopped.
7.7.1 Power Reduction Register 0 - PRR0

Bit 7 6 5 4 3 2 1 0
PRTWI PRTIM2 PRTIM0 – PRTIM1 PRSPI - PRADC PRR0
Read/Write R/W R/W R/W R R/W R/W R R/W
• 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 - Res: Reserved bit

This bit is reserved and will always read as zero.

• 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 - Res: Reserved bit

These bits are reserved and will always read as zero.
• 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.
56 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
7.7.2 Power Reduction Register 1 - PRR1
Bit 7 6 5 4 3 2 1 0
PRUSB – – – PRTIM3 – – PRUSART1 PRR1
Read/Write R/W R R R R/W R R R/W
• Bit 7 - PRUSB: Power Reduction USB

Writing a logic one to this bit shuts down the USB by stopping the clock to the module. When
waking up the USB again, the USB should be re initialized to ensure proper operation.
• Bit 6..4 - Res: Reserved bits


• Bit 2..1 - Res: Reserved bits

• Bit 0 - 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 re initialized to ensure proper
operation.
7.8 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.
7.8.1 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 “Analog to Digital Converter - ADC” on page
316 for details on ADC operation.
7.8.2 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 “Analog Comparator” on page 313 for details on how to configure the Analog
Comparator.
57
7593D–AVR–07/06
7.8.3 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 61 for details
on how to configure the Brown-out Detector.
7.8.4 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 64 for details on the start-up time.
7.8.5 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 70 for details on how to configure the Watchdog Timer.
7.8.6 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 78 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 “Digital Input Disable Register 1 – DIDR1” on page 315 and “Digital Input Dis-
able Register 1 – DIDR1” on page 315 for details.
7.8.7 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.
58 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
8. System Control and Reset
8.0.1 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 8-1 shows the reset
logic. Table 8-1 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 40.
8.0.2 Reset Sources

The AT90USB64/128 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 341 for details.
59
7593D–AVR–07/06
Figure 8-1. 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]
Table 8-1. Reset Characteristics(1)

Symbol Parameter Condition Min Typ Max Units
Power-on Reset Threshold
TBD TBD TBD V
Voltage (rising)
VPOT
Power-on Reset Threshold
TBD TBD TBD V
Voltage (falling)(2)
VRST RESET Pin Threshold Voltage TBD TBD TBD V
Minimum pulse width on RESET
tRST TBD TBD TBD ns
Pin
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)
8.0.3 Power-on Reset

A Power-on Reset (POR) pulse is generated by an On-chip detection circuit. The detection level
is defined in Table 8-1. 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.
60 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 8-2. MCU Start-up, RESET Tied to VCC
VPOT
VCC
VRST
RESET
tTOUT
TIME-OUT
INTERNAL
RESET
Figure 8-3. MCU Start-up, RESET Extended Externally

VPOT
VCC
VRST
RESET
tTOUT
TIME-OUT
INTERNAL
RESET
8.0.4 External Reset

An External Reset is generated by a low level on the RESET pin. Reset pulses longer than the
minimum pulse width (see Table 8-1) 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 8-4. External Reset During Operation

CC
8.0.5 Brown-out Detection

AT90USB64/128 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
61
7593D–AVR–07/06
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 V BOT+ =
VBOT + VHYST/2 and VBOT- = VBOT - VHYST/2.
Table 8-2. BODLEVEL Fuse Coding(1)

BODLEVEL 2..0 Fuses Min VBOT Typ VBOT Max VBOT Units
111 BOD Disabled
110 2.0
101 2.2
100 2.4
011 2.6 V
010 3.4
001 3.5
000 4.3
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 guar-
antees 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 AT90USB64/128 and BODLEVEL = 101 for AT90USB64/128L.
Table 8-3. Brown-out Characteristics

Symbol Parameter Min Typ Max Units
VHYST Brown-out Detector Hysteresis 50 mV
tBOD Min Pulse Width on Brown-out Reset ns
When the BOD is enabled, and VCC decreases to a value below the trigger level (VBOT- in Figure
8-5), the Brown-out Reset is immediately activated. When VCC increases above the trigger level
(VBOT+ in Figure 8-5), 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 Table 8-1.
Figure 8-5. Brown-out Reset During Operation
VCC VBOT+
VBOT-
RESET
TIME-OUT tTOUT
INTERNAL
RESET
62 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
8.0.6 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 64 for details on operation of the Watchdog Timer.
Figure 8-6. Watchdog Reset During Operation

CC
CK
8.0.7 MCU Status Register – MCUSR

The MCU Status Register provides information on which reset source caused an MCU reset.
Bit 7 6 5 4 3 2 1 0
– – – JTRF WDRF BORF EXTRF PORF MCUSR
Read/Write R R R R/W R/W R/W R/W R/W
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.
63
7593D–AVR–07/06
8.1 Internal Voltage Reference
AT90USB64/128 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.
8.1.1 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 Table 8-4. 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.
Table 8-4. Internal Voltage Reference Characteristics(1)

VBG Bandgap reference voltage TBD TBD 1.1 TBD V
tBG Bandgap reference start-up time TBD 40 70 µs
Bandgap reference current
IBG TBD 10 TBD µA
consumption
Note: 1. Values are guidelines only. Actual values are TBD.
8.2 Watchdog Timer

AT90USB64/128 has an Enhanced Watchdog Timer (WDT). The main features are:
• Clocked from separate On-chip Oscillator
• 3 Operating modes
– Interrupt
– System Reset
– Interrupt and System Reset
64 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
• Selectable Time-out period from 16ms to 8s
• Possible Hardware fuse Watchdog always on (WDTON) for fail-safe mode
Figure 8-7. 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
The Watchdog Timer (WDT) is a timer counting 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.
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:
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.
65
7593D–AVR–07/06
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.
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.
66 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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: 1. The example code assumes that the part specific header file is included.
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.
8.2.1 Watchdog Timer Control Register - WDTCSR

Bit 7 6 5 4 3 2 1 0
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
67
7593D–AVR–07/06
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 8-5. 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 Interrupt, then go to
0 1 1
Reset Mode System 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.
• 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 8-6 on page 69.
68 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
.
Table 8-6. 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
69
7593D–AVR–07/06
9. Interrupts
This section describes the specifics of the interrupt handling as performed in AT90USB64/128.
For a general explanation of the AVR interrupt handling, refer to “Reset and Interrupt Handling”
on page 16.
9.1 Interrupt Vectors in AT90USB64/128
Table 9-1. 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
6 $000A INT4 External Interrupt Request 4
7 $000C INT5 External Interrupt Request 5
8 $000E INT6 External Interrupt Request 6
10 $0012 PCINT0 Pin Change Interrupt Request 0
11 $0014 USB General USB General Interrupt request
USB
12 $0016 USB ENdpoint/Pipe Interrupt request
Endpoint/Pipe
13 $0018 WDT Watchdog Time-out Interrupt
14 $001A TIMER2 COMPA Timer/Counter2 Compare Match A
15 $001C TIMER2 COMPB Timer/Counter2 Compare Match B
16 $001E TIMER2 OVF Timer/Counter2 Overflow
17 $0020 TIMER1 CAPT Timer/Counter1 Capture Event
18 $0022 TIMER1 COMPA Timer/Counter1 Compare Match A
19 $0024 TIMER1 COMPB Timer/Counter1 Compare Match B
20 $0026 TIMER1 COMPC Timer/Counter1 Compare Match C
21 $0028 TIMER1 OVF Timer/Counter1 Overflow
22 $002A TIMER0 COMPA Timer/Counter0 Compare Match A
23 $002C TIMER0 COMPB Timer/Counter0 Compare match B
24 $002E TIMER0 OVF Timer/Counter0 Overflow
25 $0030 SPI, STC SPI Serial Transfer Complete
26 $0032 USART1 RX USART1 Rx Complete
27 $0034 USART1 UDRE USART1 Data Register Empty
28 $0036 USART1TX USART1 Tx Complete
70 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 9-1. Reset and Interrupt Vectors (Continued)
Vector Program
No. Address(2) Source Interrupt Definition
29 $0038 ANALOG COMP Analog Comparator
30 $003A ADC ADC Conversion Complete
31 $003C EE READY EEPROM Ready
32 $003E TIMER3 CAPT Timer/Counter3 Capture Event
33 $0040 TIMER3 COMPA Timer/Counter3 Compare Match A
34 $0042 TIMER3 COMPB Timer/Counter3 Compare Match B
35 $0044 TIMER3 COMPC Timer/Counter3 Compare Match C
36 $0046 TIMER3 OVF Timer/Counter3 Overflow
37 $0048 TWI 2-wire Serial Interface
38 $004A SPM READY Store Program Memory Ready
Notes: 1. When the BOOTRST Fuse is programmed, the device will jump to the Boot Loader address at
reset, see “Memory Programming” on page 368.
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 9-2 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.
71
7593D–AVR–07/06
Table 9-2. 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 28-8 on page 366. For the BOOTRST Fuse “1”
means unprogrammed while “0” means programmed.
9.1.1 Moving Interrupts Between Application and Boot Space

The General Interrupt Control Register controls the placement of the Interrupt Vector table.
9.1.2 MCU Control Register – MCUCR

Bit 7 6 5 4 3 2 1 0
JTD – – PUD – – IVSEL IVCE MCUCR
Read/Write R/W R R R/W R R R/W R/W
• 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 368 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 pro-
grammed, 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 dis-
abled while executing from the Boot Loader section. Refer to the section “Memory Programming”
on page 368 for details on Boot Lock bits.
• Bit 0 – IVCE: Interrupt Vector Change Enable
72 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 Code Example below.

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);
}
73
7593D–AVR–07/06
10. I/O-Ports
10.1 Introduction
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 10-1. Refer to “Electrical Char-
acteristics” on page 400 for a complete list of parameters.
Figure 10-1. I/O Pin Equivalent Schematic
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 for I/O-Ports” on page 92.
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
75. 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 79. Refer to the individual module sections for a full description of the alter-
nate functions.
74 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
10.2 Ports as General Digital I/O

The ports are bi-directional I/O ports with optional internal pull-ups. Figure 10-2 shows a func-
tional description of one I/O-port pin, here generically called Pxn.
Figure 10-2. 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.
10.2.1 Configuring the Pin

Each port pin consists of three register bits: DDxn, PORTxn, and PINxn. As shown in “Register
Description for I/O-Ports” on page 92, 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.
75
7593D–AVR–07/06
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).
10.2.2 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.
10.2.3 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) occurs. Normally, the pull-up enabled state is fully acceptable, 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 10-1 summarizes the control signals for the pin value.
Table 10-1. Port Pin Configurations
PUD
DDxn PORTxn (in MCUCR) I/O Pull-up Comment
0 0 X Input No Tri-state (Hi-Z)
Pxn will source current if ext. pulled
0 1 0 Input Yes
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)
10.2.4 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 10-2, 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 10-3 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.
76 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 10-3. 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 10-4. 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 10-4. 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.
77
7593D–AVR–07/06
...
; 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.
10.2.5 Digital Input Enable and Sleep Modes

As shown in Figure 10-2, 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 79.
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.
10.2.6 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-
78 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
10.3 Alternate Port Functions

Most port pins have alternate functions in addition to being general digital I/Os. Figure 10-5
shows how the port pin control signals from the simplified Figure 10-2 can be overridden 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 10-5. 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.
79
7593D–AVR–07/06
Table 10-2 summarizes the function of the overriding signals. The pin and port indexes from Fig-
ure 10-5 are not shown in the succeeding tables. The overriding signals are generated internally
in the modules having the alternate function.
Table 10-2. 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
Pull-up Override
PUOE PUOV signal. If this signal is cleared, the pull-up is
Enable
enabled when {DDxn, PORTxn, PUD} = 0b010.
If PUOE is set, the pull-up is enabled/disabled when
Pull-up Override
PUOV PUOV is set/cleared, regardless of the setting of the
Value
DDxn, PORTxn, and PUD Register bits.
If this signal is set, the Output Driver Enable is controlled
Data Direction
DDOE by the DDOV signal. If this signal is cleared, the Output
Override Enable
driver is enabled by the DDxn Register bit.
If DDOE is set, the Output Driver is enabled/disabled
Data Direction
DDOV when DDOV is set/cleared, regardless of the setting of
Override Value
the DDxn Register bit.
If this signal is set and the Output Driver is enabled, the
Port Value port value is controlled by the PVOV signal. If PVOE is
PVOE
Override Enable cleared, and 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
PVOV
Override Value of the setting of the PORTxn Register bit.
Port Toggle
PTOE If PTOE is set, the PORTxn Register bit is inverted.
Override Enable
If this bit is set, the Digital Input Enable is controlled by
Digital Input
the DIEOV signal. If this signal is cleared, the Digital Input
DIEOE Enable Override
Enable is determined by MCU state (Normal mode, sleep
Enable
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
Value (Normal mode, sleep mode).
This is the Digital Input to alternate functions. In the
figure, the signal is connected to the output of the schmitt
DI Digital Input trigger but 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
Analog
AIO functions. The signal is connected directly to the pad, and
Input/Output
can be used bi-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.
80 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 75 for more details about this feature.
10.3.2 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 10-3. Port A Pins Alternate Functions

Port Pin Alternate Function
PA7 AD7 (External memory interface address and data bit 7)
Table 10-4 and Table 10-5 relates the alternate functions of Port A to the overriding signals
shown in Figure 10-5 on page 79.
Table 10-4. Overriding Signals for Alternate Functions in PA7..PA4

Signal
Name PA7/AD7 PA6/AD6 PA5/AD5 PA4/AD4
PUOE SRE SRE SRE SRE
(1)
~(WR | ADA ) • ~(WR | ADA) • ~(WR | ADA) • ~(WR | ADA) •
PUOV
PORTA7 • PUD PORTA6 • PUD PORTA5 • PUD PORTA4 • PUD
DDOE SRE SRE SRE SRE
DDOV WR | ADA WR | ADA WR | ADA WR | ADA
PVOE SRE SRE SRE SRE
A7 • ADA | D7 A6 • ADA | D6 A5 • ADA | D5 A4 • ADA | D4
PVOV
OUTPUT • WR OUTPUT • WR OUTPUT • WR OUTPUT • WR
DIEOE 0 0 0 0
DIEOV 0 0 0 0
DI D7 INPUT D6 INPUT D5 INPUT D4 INPUT
AIO – – – –
81
7593D–AVR–07/06
Note: 1. ADA is short for ADdress Active and represents the time when address is output. See “Exter-
nal Memory Interface” on page 30 for details.
Table 10-5. Overriding Signals for Alternate Functions in PA3..PA0

Signal
Name PA3/AD3 PA2/AD2 PA1/AD1 PA0/AD0
PUOE SRE SRE SRE SRE
~(WR | ADA) • ~(WR | ADA) • ~(WR | ADA) • ~(WR | ADA) •
PUOV
PORTA3 • PUD PORTA2 • PUD PORTA1 • PUD PORTA0 • PUD
DDOE SRE SRE SRE SRE
DDOV WR | ADA WR | ADA WR | ADA WR | ADA
PVOE SRE SRE SRE SRE
A3 • ADA | D3 A2• ADA | D2 A1 • ADA | D1 A0 • ADA | D0
PVOV
OUTPUT • WR OUTPUT • WR OUTPUT • WR OUTPUT • WR
DIEOE 0 0 0 0
DIEOV 0 0 0 0
DI D3 INPUT D2 INPUT D1 INPUT D0 INPUT
AIO – – – –
10.3.3 Alternate Functions of Port B

The Port B pins with alternate functions are shown in Table 10-6.
Table 10-6. Port B Pins Alternate Functions

Port Pin Alternate Functions
OC0A/OC1C/PCINT7 (Output Compare and PWM Output A for Timer/Counter0,
PB7
Output Compare and PWM Output C for Timer/Counter1 or Pin Change Interrupt 7)
OC1B/PCINT6 (Output Compare and PWM Output B for Timer/Counter1 or Pin
PB6
Change Interrupt 6)
OC1A/PCINT5 (Output Compare and PWM Output A for Timer/Counter1 or Pin
PB5
Change Interrupt 5)
OC2A/PCINT4 (Output Compare and PWM Output A for Timer/Counter2 or Pin
PB4
Change Interrupt 4)
PDO/MISO/PCINT3 (Programming Data Output or SPI Bus Master Input/Slave
PB3
Output or Pin Change Interrupt 3)
PDI/MOSI/PCINT2 (Programming Data Input orSPI Bus Master Output/Slave Input
PB2
or Pin Change Interrupt 2)
PB1 SCK/PCINT1 (SPI Bus Serial Clock or Pin Change Interrupt 1)
PB0 SS/PCINT0 (SPI Slave Select input or Pin Change Interrupt 0)
The alternate pin configuration is as follows:
• OC0A/OC1C/PCINT7, Bit 7
OC0A, Output Compare Match A output: The PB7 pin can serve as an external output for the
Timer/Counter0 Output Compare. The pin has to be configured as an output (DDB7 set “one”) to
serve this function. The OC0A pin is also the output pin for the PWM mode timer function.
82 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
OC1C, Output Compare Match C output: The PB7 pin can serve as an external output for the
Timer/Counter1 Output Compare C. The pin has to be configured as an output (DDB7 set (one))
to serve this function. The OC1C pin is also the output pin for the PWM mode timer function.
PCINT7, Pin Change Interrupt source 7: The PB7 pin can serve as an external interrupt source.
• OC1B/PCINT6, Bit 6
OC1B, Output Compare Match B output: The PB6 pin can serve as an external output for the
Timer/Counter1 Output Compare B. The pin has to be configured as an output (DDB6 set (one))
to serve this function. The OC1B pin is also the output pin for the PWM mode timer function.
• OC1A/PCINT5, Bit 5
OC1A, Output Compare Match A output: The PB5 pin can serve as an external output for the
Timer/Counter1 Output Compare A. The pin has to be configured as an output (DDB5 set (one))
to serve this function. The OC1A pin is also the output pin for the PWM mode timer function.
• OC2A/PCINT4, Bit 4
OC2A, Output Compare Match output: The PB4 pin can serve as an external output for the
Timer/Counter2 Output Compare. The pin has to be configured as an output (DDB4 set (one)) to
serve this function. The OC2A pin is also the output pin for the PWM mode timer function.
• PDO/MISO/PCINT3 – Port B, Bit 3

PDO, SPI Serial Programming Data Output. During Serial Program Downloading, this pin is
used as data output line for the AT90USB64/128.
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 DDB3. When the SPI is
enabled as a slave, the data direction of this pin is controlled by DDB3. When the pin is forced to
be an input, the pull-up can still be controlled by the PORTB3 bit.
• PDI/MOSI/PCINT2 – Port B, Bit 2

PDI, SPI Serial Programming Data Input. During Serial Program Downloading, this pin is used
as data input line for the AT90USB64/128.
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 DDB2. When the SPI is
enabled as a master, the data direction of this pin is controlled by DDB2. When the pin is forced
to be an input, the pull-up can still be controlled by the PORTB2 bit.
• SCK/PCINT1 – Port B, Bit 1

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 DDB1. When the SPI0 is
enabled as a master, the data direction of this pin is controlled by DDB1. When the pin is forced
to be an input, the pull-up can still be controlled by the PORTB1 bit.
83
7593D–AVR–07/06
• SS/PCINT0 – Port B, Bit 0

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 DDB0. 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 DDB0.
When the pin is forced to be an input, the pull-up can still be controlled by the PORTB0 bit.
Table 10-7 and Table 10-8 relate the alternate functions of Port B to the overriding signals
shown in Figure 10-5 on page 79. SPI MSTR INPUT and SPI SLAVE OUTPUT constitute the
MISO signal, while MOSI is divided into SPI MSTR OUTPUT and SPI SLAVE INPUT.
PCINT0, Pin Change Interrupt source 0: The PB7 pin can serve as an external interrupt source..
Table 10-7. Overriding Signals for Alternate Functions in PB7..PB4

Signal PB7/PCINT7/OC0A/ PB6/PCINT6/OC PB5/PCINT5/OC PB4/PCINT4/OC
Name OC1C 1B 1A 2A
PUOE 0 0 0 0
PUOV 0 0 0 0
DDOE 0 0 0 0
DDOV 0 0 0 0
PVOE OC0/OC1C ENABLE OC1B ENABLE OC1A ENABLE OC2A ENABLE
PVOV OC0/OC1C OC1B OC1A OC2A
DIEOE PCINT7 • PCIE0 PCINT6 • PCIE0 PCINT5 • PCIE0 PCINT4 • PCIE0
DIEOV 1 1 1 1
DI PCINT7 INPUT PCINT6 INPUT PCINT5 INPUT PCINT4 INPUT
AIO – – – –
Table 10-8. Overriding Signals for Alternate Functions in PB3..PB0

Signal PB3/PD0/PCINT3/ PB2/PDI/PCINT2/ PB1/PCINT1/ PB0/PCINT0/
Name MISO MOSI SCK SS
PUOE SPE • MSTR SPE • MSTR SPE • MSTR SPE • MSTR
PUOV PORTB3 • PUD PORTB2 • PUD PORTB1 • PUD PORTB0 • PUD
DDOE SPE • MSTR SPE • MSTR SPE • MSTR SPE • MSTR
DDOV 0 0 0 0
PVOE SPE • MSTR SPE • MSTR SPE • MSTR 0
PVOV SPI SLAVE OUTPUT SPI MSTR OUTPUT SCK OUTPUT 0
PCINT1 • PCINT0 •
DIEOE PCINT3 • PCIE0 PCINT2 • PCIE0
PCIE0 PCIE0
DIEOV 1 1 1 1
SPI MSTR INPUT SPI SLAVE INPUT SCK INPUT SPI SS
DI
PCINT3 INPUT PCINT2 INPUT PCINT1 INPUT PCINT0 INPUT
AIO – – – –
84 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
10.3.4 Alternate Functions of Port C
The Port C alternate function is as follows:
Table 10-9. Port C Pins Alternate Functions

A15/IC.3/CLKO(External Memory interface address bit 15 or
PC7
Input Capture Timer 3 or CLK0 (Divided System Clock)
A14/OC.3A(External Memory interface address bit 14 or
PC6
Output Compare and PWM output A for Timer/Counter3)
A13/OC.3B(External Memory interface address bit 13 or
PC5
Output Compare and PWM output B for Timer/Counter3)
A12/OC.3C(External Memory interface address bit 12 or
PC4
Output Compare and PWM output C for Timer/Counter3)
A11/T.3(External Memory interface address bit 11or
PC3
Timer/Counter3 Clok Input)
PC2 A10(External Memory interface address bit 10)
Table 10-10 and Table 10-11 relate the alternate functions of Port C to the overriding signals
Table 10-10. Overriding Signals for Alternate Functions in PC7..PC4

Signal PC7/A15/IC.3/CLK
Name O PC6/A14/OC.3A PC5/A13/OC.3B PC4/A12/OC.3C
SRE • SRE • SRE •
PUOE SRE • (XMM<1) (XMM<2)|OC3A (XMM<3)|OC3B (XMM<4)|OC3C
enable enable enable
PUOV 0 0 0 0
DDOE SRE • (XMM<1) SRE • (XMM<2) SRE • (XMM<3) SRE • (XMM<4)
DDOV 1 1 1 1
PVOE SRE • (XMM<1) SRE • (XMM<2) SRE • (XMM<3) SRE • (XMM<4)
if (SRE.XMM<2) if (SRE.XMM<2) if (SRE.XMM<2)
PVOV A15 then A14 then A13 then A12
else OC3A else OC3B else OC3C
DIEOE 0 0 0 0
DIEOV 0 0 0 0
DI ICP3 input – – –
AIO – – – –
85
7593D–AVR–07/06
Table 10-11. Overriding Signals for Alternate Functions in PC3..PC0
Signal
Name PC3/A11/T.3 PC2/A10 PC1/A9 PC0/A8
PUOE SRE • (XMM<5) SRE • (XMM<6) SRE • (XMM<7) SRE • (XMM<7)
PUOV 0 0 0 0
DDOE SRE • (XMM<5) SRE • (XMM<6) SRE • (XMM<7) SRE • (XMM<7)
DDOV 1 1 1 1
PVOE SRE • (XMM<5) SRE • (XMM<6) SRE • (XMM<7) SRE • (XMM<7)
PVOV A11 A10 A9 A8
DIEOE 0 0 0 0
DIEOV 0 0 0 0
DI T3 input – – –
AIO – – – –
10.3.5 Alternate Functions of Port D

The Port D pins with alternate functions are shown in Table 10-12.
Table 10-12. Port D Pins Alternate Functions

PD7 T0 (Timer/Counter0 Clock Input)
PD6 T1 (Timer/Counter1 Clock Input)
PD5 XCK1 (USART1 External Clock Input/Output)
PD4 ICP1 (Timer/Counter1 Input Capture Trigger)
PD3 INT3/TXD1 (External Interrupt3 Input or USART1 Transmit Pin)
PD2 INT2/RXD1 (External Interrupt2 Input or USART1 Receive Pin)
INT1/SDA/OC2B (External Interrupt1 Input or TWI Serial DAta or Output
PD1
Compare for Timer/Counter2)
INT0/SCL/OC0B (External Interrupt0 Input or TWI Serial CLock or Output
PD0
Compare for Timer/Counter0)
The alternate pin configuration is as follows:
• T0 – Port D, Bit 7
T0, Timer/Counter0 counter source.
• T1 – Port D, Bit 6
T1, Timer/Counter1 counter source.
• XCK1 – Port D, Bit 5

XCK1, USART1 External clock. The Data Direction Register (DDD5) controls whether the clock
is output (DDD5 set) or input (DDD5 cleared). The XCK1 pin is active only when the USART1
operates in Synchronous mode.
86 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
• ICP1 – Port D, Bit 4
ICP1 – Input Capture Pin 1: The PD4 pin can act as an input capture pin for Timer/Counter1.
• INT3/TXD1 – Port D, Bit 3

INT3, External Interrupt source 3: 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.
• INT2/RXD1 – Port D, Bit 2

INT2, External Interrupt source 2. The PD2 pin can serve as an External Interrupt source to the
MCU.
RXD1, 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.
• INT1/SDA/OC2B – Port D, Bit 1

INT1, External Interrupt source 1. The PD1 pin can serve as an external interrupt source to the
MCU.
SDA, 2-wire Serial Interface Data: When the TWEN bit in TWCR is set (one) to enable the 2-wire
Serial Interface, pin PD1 is disconnected from the port and becomes the Serial Data I/O pin for
the 2-wire Serial Interface. In this mode, there is a spike filter on the pin to suppress spikes
shorter than 50 ns on the input signal, and the pin is driven by an open drain driver with slew-
rate limitation.
• INT0/SCL/OC0B – Port D, Bit 0

INT0, External Interrupt source 0. The PD0 pin can serve as an external interrupt source to the
MCU.
SCL, 2-wire Serial Interface Clock: When the TWEN bit in TWCR is set (one) to enable the 2-
wire Serial Interface, pin PD0 is disconnected from the port and becomes the Serial Clock I/O
pin for the 2-wire Serial Interface. In this mode, there is a spike filter on the pin to suppress
spikes shorter than 50 ns on the input signal, and the pin is driven by an open drain driver with
slew-rate limitation.
Table 10-13 and Table 10-14 relates the alternate functions of Port D to the overriding signals
87
7593D–AVR–07/06
Table 10-13. Overriding Signals for Alternate Functions PD7..PD4
Signal Name PD7/T0 PD6/T1 PD5/XCK1 PD4/ICP1
PUOE 0 0 0 0
PUOV 0 0 0 0
DDOE 0 0 XCK1 OUTPUT ENABLE 0
DDOV 0 0 1 0
PVOE 0 0 XCK1 OUTPUT ENABLE 0
PVOV 0 0 XCK1 OUTPUT 0
DIEOE 0 0 0 0
DIEOV 0 0 0 0
DI T0 INPUT T1 INPUT XCK1 INPUT ICP1 INPUT
AIO – – – –
Table 10-14. Overriding Signals for Alternate Functions in PD3..PD0(1)

PD1/INT1/SDA/ PD0/INT0/SCL/
Signal Name PD3/INT3/TXD1 PD2/INT2/RXD1 OC2B OC0B
PUOE TXEN1 RXEN1 TWEN TWEN
PUOV 0 PORTD2 • PUD PORTD1 • PUD PORTD0 • PUD
DDOE TXEN1 RXEN1 TWEN TWEN
DDOV 1 0 SDA_OUT SCL_OUT
TWEN | OC2B TWEN | OC0B
PVOE TXEN1 0
ENABLE ENABLE
PVOV TXD1 0 OC2B OC0B
DIEOE INT3 ENABLE INT2 ENABLE INT1 ENABLE INT0 ENABLE
DIEOV 1 1 1 1
DI INT3 INPUT INT2 INPUT/RXD1 INT1 INPUT INT0 INPUT
AIO – – SDA INPUT SCL INPUT
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.
88 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
10.3.6 Alternate Functions of Port E
The Port E pins with alternate functions are shown in Table 10-15.
Table 10-15. Port E Pins Alternate Functions

INT7/AIN.1/UVCON (External Interrupt 7 Input, Analog Comparator Positive Input
PE7
or VBUS Control)
PE6 INT6/AIN.0 (External Interrupt 6 Input or Analog Comparator Positive Input)
PE5 INT5/TOSC2 (External Interrupt 5 Input or RTC Oscillator Timer/Counter2))
PE4 INT4/TOSC2 (External Interrupt4 Input or RTC Oscillator Timer/Counter2)
PE3 UID
PE2 ALE/HWB (Address latch to extenal memory or Hardware bootloader activation)
PE1 RD (Read strobe to external memory)
PE0 WR (Write strobe to external memory)
• INT7/AIN.1/UVCON – Port E, Bit 7

INT7, External Interrupt source 7: The PE7 pin can serve as an external interrupt source.
AIN1 – Analog Comparator Negative input. This pin is directly connected to the negative input of
the Analog Comparator.
UVCON - When using USB host mode, this pin allows to control an external VBUS generator
(active high).
• INT6/AIN.0 – Port E, Bit 6

INT6, External Interrupt source 6: The PE6 pin can serve as an external interrupt source.
AIN0 – Analog Comparator Negative input. This pin is directly connected to the negative input of
the Analog Comparator.
• INT5/TOSC2 – Port E, Bit 5

INT5, External Interrupt source 5: The PE5 pin can serve as an External Interrupt source.
TOSC2, Timer/Counter2 Oscillator pin1. When the AS2 bit in ASSR is set to enable asynchro-
nous clocking of Timer/Counter2, pin PE5 is disconnected from the port, and becomes the ouput
of the inverting Oscillator amplifier. In this mode, a crystal is connected to this pin, and the pin
can not be used as an I/O pin.
• INT4/TOSC1 – Port E, Bit 4

INT4, External Interrupt source 4: The PE4 pin can serve as an External Interrupt source.
TOSC1, Timer/Counter2 Oscillator pin2. When the AS2 bit in ASSR is set to enable asynchro-
nous clocking of Timer/Counter2, pin PE4 is disconnected from the port, and becomes the input
of the inverting Oscillator amplifier. In this mode, a crystal is connected to this pin, and the pin
can not be used as an I/O pin.
• UID – Port E, Bit 3

ID pin of the USB bus.
• ALE/HWB – Port E, Bit 2
89
7593D–AVR–07/06
ALE is the external data memory Address latch enable.
HWB allows to execute the bootloader section after reset when tied to ground during external
reset pulse. The HWB mode of this pin is active only when the HWBE fuse is enable.
• RD – Port E, Bit 1
RD is the external data memory read control enable.
• WR – Port E, Bit 0
WR is the external data memory write control enable.
Table 10-16. Overriding Signals for Alternate Functions PE7..PE4

Signal PE7/INT7/AIN.1/ PE5/INT5/ PE4/INT4/
Name UVCON PE6/INT6/AIN.0 TOSC1 TOSC2
PUOE 0 0 0 0
PUOV 0 0 0 0
DDOE UVCONE 0 0 0
DDOV UVCONE 0 0 0
PVOE UVCONE 0 0 0
PVOV UVCON 0 0 0
DIEOE INT7 ENABLE INT6 ENABLE INT5 ENABLE INT4 ENABLE
DIEOV 1 1 1 1
DI INT7 INPUT INT6 INPUT INT5 INPUT INT4 INPUT
AIO AIN1 INPUT AIN0 INPUT – –
Table 10-17. Overriding Signals for Alternate Functions in PE3..PE0

Signal
Name PE3/UID PE2/ALE/HWB PE1/RD PE0/WR
PUOE UIDE 0 SRE SRE
PUOV 1 0 0 0
DDOE UIDE SRE SRE SRE
DDOV 0 1 1 0
PVOE 0 SRE SRE SRE
PVOV 0 ALE RD WR
DIEOE UIDE 0 0 0
DIEOV 1 0 0 1
DI UID HWB – –
PE0 0 0 0 0
AIO – – – –
90 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
10.3.7 Alternate Functions of Port F
The Port F has an alternate function as analog input for the ADC as shown in Table 10-18. If
some Port F pins are configured as outputs, it is essential that these do not switch when a con-
version is in progress. This might corrupt the result of the conversion. If the JTAG interface is
enabled, the pull-up resistors on pins PF7(TDI), PF5(TMS), and PF4(TCK) will be activated even
if a Reset occurs.
Table 10-18. Port F Pins Alternate Functions

PF7 ADC7/TDI (ADC input channel 7 or JTAG Test Data Input)
PF6 ADC6/TDO (ADC input channel 6 or JTAG Test Data Output)
PF5 ADC5/TMS (ADC input channel 5 or JTAG Test Mode Select)
PF4 ADC4/TCK (ADC input channel 4 or JTAG Test ClocK)
PF3 ADC3 (ADC input channel 3)
• TDI, ADC7 – Port F, Bit 7

ADC7, Analog to Digital Converter, Channel 7.
TDI, JTAG Test Data In: Serial input data to be shifted in to the Instruction Register or Data Reg-
ister (scan chains). When the JTAG interface is enabled, this pin can not be used as an I/O pin.
• TDO, ADC6 – Port F, Bit 6

TDO, JTAG Test Data Out: Serial output data from Instruction Register or Data Register. When
the JTAG interface is enabled, this pin can not be used as an I/O pin.
The TDO pin is tri-stated unless TAP states that shift out data are entered.
• TMS, ADC5 – Port F, Bit 5

TMS, JTAG Test Mode Select: This pin is used for navigating through the TAP-controller state
machine. When the JTAG interface is enabled, this pin can not be used as an I/O pin.
• TCK, ADC4 – Port F, Bit 4

TCK, JTAG Test Clock: JTAG operation is synchronous to TCK. When the JTAG interface is
enabled, this pin can not be used as an I/O pin.
• ADC3 – ADC0 – Port F, Bit 3..0
91
7593D–AVR–07/06
Analog to Digital Converter, Channel 3..0.
Table 10-19. Overriding Signals for Alternate Functions in PF7..PF4

Signal
Name PF7/ADC7/TDI PF6/ADC6/TDO PF5/ADC5/TMS PF4/ADC4/TCK
PUOE JTAGEN JTAGEN JTAGEN JTAGEN
PUOV 1 0 1 1
DDOE JTAGEN JTAGEN JTAGEN JTAGEN
SHIFT_IR +
DDOV 0 0 0
SHIFT_DR
PVOE 0 JTAGEN 0 0
PVOV 0 TDO 0 0
DIEOE JTAGEN JTAGEN JTAGEN JTAGEN
DIEOV 0 0 0 0
DI – – – –
TMS/ADC5 TCK/ADC4
AIO TDI/ADC7 INPUT ADC6 INPUT
INPUT INPUT
Table 10-20. Overriding Signals for Alternate Functions in PF3..PF0

Signal Name PF3/ADC3 PF2/ADC2 PF1/ADC1 PF0/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
DIEOE 0 0 0 0
DIEOV 0 0 0 0
DI – – – –
AIO ADC3 INPUT ADC2 INPUT ADC1 INPUT ADC0 INPUT
10.4 Register Description for I/O-Ports
10.4.1 Port A Data Register – PORTA

Bit 7 6 5 4 3 2 1 0
PORTA PORTA PORTA PORTA PORTA PORTA PORTA PORTA PORTA
7 6 5 4 3 2 1 0
92 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
10.4.2 Port A Data Direction Register – DDRA
Bit 7 6 5 4 3 2 1 0
DDA7 DDA6 DDA5 DDA4 DDA3 DDA2 DDA1 DDA0 DDRA
10.4.3 Port A Input Pins Address – PINA

Bit 7 6 5 4 3 2 1 0
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
10.4.4 Port B Data Register – PORTB

Bit 7 6 5 4 3 2 1 0
PORTB PORTB PORTB PORTB PORTB PORTB PORTB PORTB PORTB
7 6 5 4 3 2 1 0
10.4.5 Port B Data Direction Register – DDRB

Bit 7 6 5 4 3 2 1 0
DDB7 DDB6 DDB5 DDB4 DDB3 DDB2 DDB1 DDB0 DDRB
10.4.6 Port B Input Pins Address – PINB

Bit 7 6 5 4 3 2 1 0
PINB7 PINB6 PINB5 PINB4 PINB3 PINB2 PINB1 PINB0 PINB
10.4.7 Port C Data Register – PORTC

Bit 7 6 5 4 3 2 1 0
PORTC PORTC PORTC PORTC PORTC PORTC PORTC PORTC PORTC
7 6 5 4 3 2 1 0
10.4.8 Port C Data Direction Register – DDRC

Bit 7 6 5 4 3 2 1 0
DDC7 DDC6 DDC5 DDC4 DDC3 DDC2 DDC1 DDC0 DDRC
10.4.9 Port C Input Pins Address – PINC

Bit 7 6 5 4 3 2 1 0
PINC7 PINC6 PINC5 PINC4 PINC3 PINC2 PINC1 PINC0 PINC
93
7593D–AVR–07/06
10.4.10 Port D Data Register – PORTD
Bit 7 6 5 4 3 2 1 0
PORTD PORTD PORTD PORTD PORTD PORTD PORTD PORTD PORTD
7 6 5 4 3 2 1 0
10.4.11 Port D Data Direction Register – DDRD

Bit 7 6 5 4 3 2 1 0
DDD7 DDD6 DDD5 DDD4 DDD3 DDD2 DDD1 DDD0 DDRD
10.4.12 Port D Input Pins Address – PIND

Bit 7 6 5 4 3 2 1 0
PIND7 PIND6 PIND5 PIND4 PIND3 PIND2 PIND1 PIND0 PIND
10.4.13 Port E Data Register – PORTE

Bit 7 6 5 4 3 2 1 0
PORTE PORTE PORTE PORTE PORTE PORTE PORTE PORTE PORTE
7 6 5 4 3 2 1 0
10.4.14 Port E Data Direction Register – DDRE

Bit 7 6 5 4 3 2 1 0
DDE7 DDE6 DDE5 DDE4 DDE3 DDE2 DDE1 DDE0 DDRE
10.4.15 Port E Input Pins Address – PINE

Bit 7 6 5 4 3 2 1 0
PINE7 PINE6 PINE5 PINE4 PINE3 PINE2 PINE1 PINE0 PINE
10.4.16 Port F Data Register – PORTF

Bit 7 6 5 4 3 2 1 0
PORTF PORTF PORTF PORTF PORTF PORTF PORTF PORTF PORTF
7 6 5 4 3 2 1 0
10.4.17 Port F Data Direction Register – DDRF

Bit 7 6 5 4 3 2 1 0
DDF7 DDF6 DDF5 DDF4 DDF3 DDF2 DDF1 DDF0 DDRF
94 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
10.4.18 Port F Input Pins Address – PINF
Bit 7 6 5 4 3 2 1 0
PINF7 PINF6 PINF5 PINF4 PINF3 PINF2 PINF1 PINF0 PINF
95
7593D–AVR–07/06
11. External Interrupts
The External Interrupts are triggered by the INT7:0 pin or any of the PCINT23..0 pins. Observe
that, if enabled, the interrupts will trigger even if the INT7:0 or PCINT23..0 pins are configured as
outputs. This feature provides a way of generating a software interrupt.
The Pin change interrupt PCI0 will trigger if any enabled PCINT7:0 pin toggles. PCMSK0 Regis-
ter control which pins contribute to the pin change interrupts. Pin change interrupts on PCINT7
..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 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 (INT3:0)
and EICRB (INT7:4). When the external interrupt is enabled and is configured as level triggered,
the interrupt will trigger as long as the pin is held low. Note that recognition of falling or rising
edge interrupts on INT7:4 requires the presence of an I/O clock, described in “System Clock and
Clock Options” on page 39. Low level interrupts and the edge interrupt on INT3:0 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 39.
11.0.1 External Interrupt Control Register A – EICRA

The External Interrupt Control Register A contains control bits for interrupt sense control.
Bit 7 6 5 4 3 2 1 0
ISC31 ISC30 ISC21 ISC20 ISC11 ISC10 ISC01 ISC00 EICRA
• Bits 7..0 – ISC31, ISC30 – ISC00, ISC00: External Interrupt 3 - 0 Sense Control Bits
The External Interrupts 3 - 0 are activated by the external pins INT3:0 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 INT3..INT0 are registered asynchro-
nously. Pulses on INT3:0 pins wider than the minimum pulse width given in Table 11-2 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 completion of the currently executing
instruction to generate an interrupt. If enabled, a level triggered interrupt will generate an inter-
rupt 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.
96 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 = 3, 2, 1or 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.
Table 11-2. Asynchronous External Interrupt Characteristics

Minimum pulse width for
tINT 50 ns
asynchronous external interrupt
11.0.2 External Interrupt Control Register B – EICRB

Bit 7 6 5 4 3 2 1 0
ISC71 ISC70 ISC61 ISC60 ISC51 ISC50 ISC41 ISC40 EICRB
• Bits 7..0 – ISC71, ISC70 - ISC41, ISC40: External Interrupt 7 - 4 Sense Control Bits
The External Interrupts 7 - 4 are activated by the external pins INT7:4 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 11-3. The value on the INT7:4 pins are sampled
before detecting edges. If edge or toggle interrupt is selected, pulses that last longer than one
clock period will generate an interrupt. Shorter pulses are not guaranteed to generate an inter-
rupt. Observe that CPU clock frequency can be lower than the XTAL frequency if the XTAL
divider is enabled. If low level interrupt is selected, the low level must be held until the comple-
tion 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.
Table 11-3. Interrupt Sense Control(1)

ISCn1 ISCn0 Description
0 0 The low level of INTn generates an interrupt request.
0 1 Any logical change on INTn generates an interrupt request
The falling edge between two samples of INTn generates an interrupt
1 0
request.
The rising edge between two samples of INTn generates an interrupt
1 1
request.
Note: 1. n = 7, 6, 5 or 4.
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.
11.0.3 External Interrupt Mask Register – EIMSK

Bit 7 6 5 4 3 2 1 0
97
7593D–AVR–07/06
INT7 INT6 INT5 INT4 INT3 INT2 INT1 IINT0 EIMSK
• Bits 7..0 – INT7 – INT0: External Interrupt Request 7 - 0 Enable

When an INT7 – INT0 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 Registers – EICRA and EICRB – defines whether the external inter-
rupt 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.
11.0.4 External Interrupt Flag Register – EIFR

Bit 7 6 5 4 3 2 1 0
INTF7 INTF6 INTF5 INTF4 INTF3 INTF2 INTF1 IINTF0 EIFR
• Bits 7..0 – INTF7 - INTF0: External Interrupt Flags 7 - 0

When an edge or logic change on the INT7:0 pin triggers an interrupt request, INTF7:0 becomes
set (one). If the I-bit in SREG and the corresponding interrupt enable bit, INT7:0 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 INT7:0 are configured as level interrupt. Note that when entering sleep
mode with the INT3:0 interrupts disabled, the input buffers on these pins will be disabled. This
may cause a logic change in internal signals which will set the INTF3:0 flags. See “Digital Input
Enable and Sleep Modes” on page 78 for more information.
11.0.5 Pin Change Interrupt Control Register - PCICR

Bit 7 6 5 4 3 2 1 0
– – – – – PCIE0 PCICR
Read/Write R R R R R R R R/W
• Bit 0 – PCIE0: Pin Change Interrupt Enable 0

When the PCIE0 bit is set (one) and the I-bit in the Status Register (SREG) is set (one), pin
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.
11.0.6 Pin Change Interrupt Flag Register – PCIFR

Bit 7 6 5 4 3 2 1 0
– – – – – PCIF0 PCIFR
• Bit 0 – PCIF0: Pin Change Interrupt Flag 0

When a logic change on any PCINT7..0 pin triggers an interrupt request, PCIF0 becomes set
(one). If the I-bit in SREG and the PCIE0 bit in EIMSK are set (one), the MCU will jump to the
98 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
11.0.7 Pin Change Mask Register 0 – PCMSK0

Bit 7 6 5 4 3 2 1 0
PCINT7 PCINT6 PCINT5 PCINT4 PCINT3 PCINT2 PCINT1 PCINT0 PCMSK0
• Bit 7..0 – PCINT7..0: Pin Change Enable Mask 7..0

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
corresponding I/O pin. If PCINT7..0 is cleared, pin change interrupt on the corresponding I/O pin
is disabled.
99
7593D–AVR–07/06
12. Timer/Counter0, Timer/Counter1, and Timer/Counter3 Prescalers
Timer/Counter0, 1, and 3 share the same prescaler module, but the Timer/Counters can have
different prescaler settings. The description below applies to all Timer/Counters. Tn is used as a
general name, n = 0, 1 or 3.
12.1 Internal Clock Source

The Timer/Counter can be clocked directly by the system clock (by setting the CSn2:0 = 1). This
provides the fastest operation, with a maximum Timer/Counter clock frequency equal to system
clock frequency (fCLK_I/O). Alternatively, one of four taps from the prescaler can be used as a
clock source. The prescaled clock has a frequency of either fCLK_I/O/8, fCLK_I/O/64, fCLK_I/O/256, or
fCLK_I/O/1024.
12.2 Prescaler Reset

The prescaler is free running, i.e., operates independently of the Clock Select logic of the
Timer/Counter, and it is shared by the Timer/Counter Tn. Since the prescaler is not affected by
the Timer/Counter’s clock select, the state of the prescaler will have implications for situations
where a prescaled clock is used. One example of prescaling artifacts occurs when the timer is
enabled and clocked by the prescaler (6 > CSn2:0 > 1). The number of system clock cycles from
when the timer is enabled to the first count occurs can be from 1 to N+1 system clock cycles,
where N equals the prescaler divisor (8, 64, 256, or 1024).
It is possible to use the prescaler reset for synchronizing the Timer/Counter to program execu-
tion. However, care must be taken if the other Timer/Counter that shares the same prescaler
also uses prescaling. A prescaler reset will affect the prescaler period for all Timer/Counters it is
connected to.
12.3 External Clock Source

An external clock source applied to the Tn pin can be used as Timer/Counter clock (clkTn). The
Tn pin is sampled once every system clock cycle by the pin synchronization logic. The synchro-
nized (sampled) signal is then passed through the edge detector. Figure 1 shows a functional
equivalent block diagram of the Tn synchronization and edge detector logic. The registers are
clocked at the positive edge of the internal system clock (clkI/O). The latch is transparent in the
high period of the internal system clock.
The edge detector generates one clkTn pulse for each positive (CSn2:0 = 7) or negative (CSn2:0
= 6) edge it detects.
Figure 1. Tn/T0 Pin Sampling
Tn D Q D Q D Q Tn_sync
(To Clock
Select Logic)
LE
clk I/O
Synchronization Edge Detector
The synchronization and edge detector logic introduces a delay of 2.5 to 3.5 system clock cycles
from an edge has been applied to the Tn pin to the counter is updated.
Enabling and disabling of the clock input must be done when Tn has been stable for at least one
system clock cycle, otherwise it is a risk that a false Timer/Counter clock pulse is generated.
100 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Each half period of the external clock applied must be longer than one system clock cycle to
ensure correct sampling. The external clock must be guaranteed to have less than half the sys-
tem clock frequency (fExtClk < fclk_I/O/2) given a 50/50% duty cycle. Since the edge detector uses
sampling, the maximum frequency of an external clock it can detect is half the sampling fre-
quency (Nyquist sampling theorem). However, due to variation of the system clock frequency
and duty cycle caused by Oscillator source (crystal, resonator, and capacitors) tolerances, it is
recommended that maximum frequency of an external clock source is less than fclk_I/O/2.5.
An external clock source can not be prescaled.
Figure 2. Prescaler for synchronous Timer/Counters

clk I/O
Clear
PSR10
Tn
Synchronization
Tn
Synchronization
CSn0 CSn0
CSn1 CSn1
CSn2 CSn2
TIMER/COUNTERn CLOCK SOURCE TIMER/COUNTERn CLOCK SOURCE

clkTn clkTn
12.4 General Timer/Counter Control Register – GTCCR

Bit 7 6 5 4 3 2 1 0
TSM – – – – – PSRA- PSRSY GTCCR
SY NC
Read/Write R/W R R R R R R/W R/W
• Bit 7 – TSM: Timer/Counter Synchronization Mode

Writing the TSM bit to one activates the Timer/Counter Synchronization mode. In this mode, the
value that is written to the PSRASY and PSRSYNC bits is kept, hence keeping the correspond-
ing prescaler reset signals asserted. This ensures that the corresponding Timer/Counters are
halted and can be configured to the same value without the risk of one of them advancing during
configuration. When the TSM bit is written to zero, the PSRASY and PSRSYNC bits are cleared
by hardware, and the Timer/Counters start counting simultaneously.
• Bit 0 – PSRSYNC: Prescaler Reset for Synchronous Timer/Counters
When this bit is one, Timer/Counter0 and Timer/Counter1 and Timer/Counter3 prescaler will be
Reset. This bit is normally cleared immediately by hardware, except if the TSM bit is set. Note
that Timer/Counter0, Timer/Counter1 and Timer/Counter3 share the same prescaler and a reset
of this prescaler will affect all timers.
101
7593D–AVR–07/06
13. 8-bit Timer/Counter0 with PWM
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. The main features are:
• 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)
13.1 Overview
A simplified block diagram of the 8-bit Timer/Counter is shown in Figure 13-1. For the actual
placement of I/O pins, refer to “Pinout AT90USB64/128-TQFP” on page 3. CPU accessible I/O
Registers, including I/O bits and I/O pins, are shown in bold. The device-specific I/O Register
and bit locations are listed in the “8-bit Timer/Counter Register Description” on page 112.
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
13.1.1 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
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).
102 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 “Output Compare Unit” on page 104. 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.
13.1.2 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 the table below are also used extensively throughout the document.
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 dependent on the mode of operation.
13.2 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/Counter0, Timer/Counter1, and Timer/Counter3 Prescalers” on page 100.
13.3 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):

count Increment or decrement TCNT0 by 1.
direction Select between increment and decrement.
clear Clear TCNT0 (set all bits to zero).
103
7593D–AVR–07/06
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 107.
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.
13.4 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 107).
Figure 13-3 shows a block diagram of the Output Compare unit.
104 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
13.4.1 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).
13.4.2 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.
13.4.3 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
105
7593D–AVR–07/06
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.
13.5 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 “8-bit Timer/Counter Register Description” on page 112.
106 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
13.5.1 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 13-1 on page 113. For fast PWM mode, refer to Table 13-2
on page 113, and for phase correct PWM refer to Table 13-3 on page 113.
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.
13.6 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 “Compare Match Output Unit” on page 106.).
For detailed timing information see “Timer/Counter Timing Diagrams” on page 111.
13.6.1 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.
13.6.2 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.
107
7593D–AVR–07/06
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.
13.6.3 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
108 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 13-2 on page 113). 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
109
7593D–AVR–07/06
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.
13.6.4 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
110 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 113). 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.
13.7 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.
111
7593D–AVR–07/06
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
13.8 8-bit Timer/Counter Register Description
13.8.1 Timer/Counter Control Register A – TCCR0A

Bit 7 6 5 4 3 2 1 0
COM0A COM0A COM0B COM0B – – WGM0 WGM0 TCCR0A
1 0 1 0 1 0
Read/Write R/W R/W R/W R/W R R R/W R/W
112 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
• Bits 7:6 – COM01A: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 13-1 shows the COM0A1:0 bit functionality when the WGM02:0 bits
are set to a normal or CTC mode (non-PWM).
Table 13-1. 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 13-2 shows the COM0A1:0 bit functionality when the WGM01:0 bits are set to fast PWM
mode.
Table 13-2. Compare Output Mode, Fast PWM Mode(1)

WGM02 = 0: Normal Port Operation, OC0A Disconnected.
0 1
WGM02 = 1: Toggle OC0A on Compare Match.
1 0 Clear OC0A on Compare Match, set OC0A at TOP
1 1 Set OC0A on Compare Match, clear OC0A at TOP
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 TOP. See “Fast PWM Mode” on page 108
for more details.
Table 13-3 shows the COM0A1:0 bit functionality when the WGM02:0 bits are set to phase cor-
rect PWM mode.
Table 13-3. 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
113
7593D–AVR–07/06
pare Match is ignored, but the set or clear is done at TOP. See “Phase Correct PWM Mode” on
page 110 for more details.
• 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

COM01 COM00 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 13-2 shows the COM0B1:0 bit functionality when the WGM02:0 bits are set to fast PWM
mode.

COM01 COM00 Description
0 1 Reserved
1 0 Clear OC0B on Compare Match, set OC0B at TOP
1 1 Set OC0B on Compare Match, clear OC0B at TOP
Note: 1. A special case occurs when OCR0B equals TOP and COM0B1 is set. In this case, the Com-
for more details.
Table 13-3 shows the COM0B1:0 bit functionality when the WGM02:0 bits are set to phase cor-
rect 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
114 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
• Bits 3, 2 – Res: Reserved Bits

• 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 13-7. 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 107).
Table 13-7. 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 TOP 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 TOP TOP
Notes: 1. MAX = 0xFF

2. BOTTOM = 0x00
13.8.2 Timer/Counter Control Register B – TCCR0B

Bit 7 6 5 4 3 2 1 0
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.
115
7593D–AVR–07/06
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.
• Bits 5:4 – Res: Reserved Bits

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

See the description in the “Timer/Counter Control Register A – TCCR0A” on page 112.
• Bits 2:0 – CS02:0: Clock Select

The three Clock Select bits select the clock source to be used by the Timer/Counter.
Table 13-8. 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.
13.8.3 Timer/Counter Register – TCNT0

Bit 7 6 5 4 3 2 1 0
TCNT0[7:0] TCNT0
116 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
13.8.4 Output Compare Register A – OCR0A

Bit 7 6 5 4 3 2 1 0
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.
13.8.5 Output Compare Register B – OCR0B

Bit 7 6 5 4 3 2 1 0
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.
13.8.6 Timer/Counter Interrupt Mask Register – TIMSK0

Bit 7 6 5 4 3 2 1 0
– – – – – OCIE0B OCIE0A TOIE0 TIMSK0
Read/Write R R R R R R/W R/W R/W
• Bits 7..3, 0 – Res: Reserved Bits

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.
117
7593D–AVR–07/06
13.8.7 Timer/Counter 0 Interrupt Flag Register – TIFR0
Bit 7 6 5 4 3 2 1 0
– – – – – OCF0B OCF0A TOV0 TIFR0
• Bits 7..3, 0 – Res: Reserved Bits

• 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 13-7, “Waveform
Generation Mode Bit Description” on page 115.
118 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
14. 16-bit Timer/Counter (Timer/Counter1 and Timer/Counter3)
The 16-bit Timer/Counter unit allows accurate program execution timing (event management),
wave generation, and signal timing measurement. The main features are:
• True 16-bit Design (i.e., Allows 16-bit PWM)
• Three 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
• Twenty independent interrupt sources (TOV1, OCF1A, OCF1B, OCF1C, ICF1, TOV3, OCF3A,
OCF3B, OCF3C, ICF3, TOV4, OCF4A, OCF4B, OCF4C, ICF4, TOV5, OCF5A, OCF5B, OCF5C and
ICF5)
14.1 Overview
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
placement of I/O pins, see “Pinout AT90USB64/128-TQFP” on page 3. CPU accessible I/O Reg-
isters, including I/O bits and I/O pins, are shown in bold. The device-specific I/O Register and bit
locations are listed in the “16-bit Timer/Counter (Timer/Counter1 and Timer/Counter3)” on page
119.
The Power Reduction Timer/Counter1 bit, PRTIM1, in “Power Reduction Register 0 - PRR0” on
page 56 must be written to zero to enable Timer/Counter1 module.
119
7593D–AVR–07/06
Figure 14-1. 16-bit Timer/Counter Block Diagram(1)
Count TOVn
Clear (Int.Req.)
Control Logic
Direction TCLK Clock Select
Edge
Tn
Detector
TOP BOTTOM
( From Prescaler )
Timer/Counter
TCNTn
= =0
OCFnA
(Int.Req.)
Waveform
= OCnA
Generation
OCRnA
Fixed OCFnB
TOP (Int.Req.)
Values
Waveform
= OCnB
Generation
DATABUS
OCRnB
OCFnC
(Int.Req.)
Waveform
= OCnC
Generation
OCRnC
( From Analog
Comparator Ouput )
ICFn (Int.Req.)
Edge Noise
ICRn
Detector Canceler
ICPn
TCCRnA TCCRnB TCCRnC
Note: 1. Refer to Figure 1-1 on page 3, Table 10-6 on page 82, and Table 10-9 on page 85 for
Timer/Counter1 and 3 and 3 pin placement and description.
14.1.1 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 121. The Timer/Counter Control Registers (TCCRnA/B/C) are 8-bit registers and have no
CPU access restrictions. Interrupt requests (shorten as Int.Req.) signals are all visible in the
Timer Interrupt Flag Register (TIFRn). All interrupts are individually masked with the Timer Inter-
rupt Mask Register (TIMSKn). TIFRn and TIMSKn are not shown in the figure since these
registers are shared by other timer units.
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).
120 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
See “Output Compare Units” on page 128.. 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
“Analog Comparator” on page 313.) The Input Capture unit includes a digital filtering unit (Noise
Canceler) 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.
14.1.2 Definitions
The following definitions are used extensively throughout the document:
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:
TOP
0x00FF, 0x01FF, or 0x03FF, or to the value stored in the OCRnA or ICRn
Register. The assignment is dependent of the mode of operation.
14.2 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
Temporary 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.
121
7593D–AVR–07/06
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;
...

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.
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.
122 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 */
/* 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.
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.
123
7593D–AVR–07/06
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 */
/* 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.
14.2.1 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/Counter0, Timer/Counter1, and Timer/Counter3 Prescalers” on page 100.
14.4 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.
124 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 131.
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.
125
7593D–AVR–07/06
14.5 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, for the Timer/Counter1 only, via the
Analog Comparator unit. The time-stamps can then be used to calculate frequency, duty-cycle,
and other features of the signal 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
Note: The Analog Comparator Output (ACO) can only trigger the Timer/Counter1 ICP – not
Timer/Counter3, 4 or 5.
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 (TICIEn =
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.
126 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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-
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 121.
14.5.1 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 1 on page 100). 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.
14.5.2 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.
14.5.3 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
127
7593D–AVR–07/06
Register has been read. After a change of the edge, the Input Capture Flag (ICFn) must be
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).
14.6 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 “Modes of Operation” on page 131.)
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
128 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 121.

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).
14.6.2 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.

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.
129
7593D–AVR–07/06
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
DATA BUS D Q
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 14-1, Table 14-2 and Table 14-3 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 “16-bit Timer/Counter (Timer/Counter1 and Timer/Counter3)” on page 119.
The COMnx1:0 bits have no effect on the Input Capture unit.

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 14-1 on page 142. For fast PWM mode refer to Table 14-2 on
page 142, and for phase correct and phase and frequency correct PWM refer to Table 14-3 on
page 143.
130 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 “Compare Match Output Unit” on page 129.)
For detailed timing information refer to “Timer/Counter Timing Diagrams” on page 138.
14.8.1 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.
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.
131
7593D–AVR–07/06
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 set on
the compare match between TCNTn and OCRnx, and cleared at TOP. In inverting Compare
Output mode output is cleared on compare match and set at TOP. Due to the single-slope oper-
ation, the operating frequency of the fast PWM mode can be twice as high as the phase correct
and phase and frequency correct PWM modes that use dual-slope operation. This high fre-
quency 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.
132 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 OCnA Interrupt
Flag Set or ICFn
Interrupt Flag Set
(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
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
133
7593D–AVR–07/06
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 142). 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
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
134 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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:
log ( TOP + 1 )
R PCPWM = -----------------------------------
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
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-
135
7593D–AVR–07/06
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 14-3 on page 143).
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 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.
14.8.5 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
136 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
137
7593D–AVR–07/06
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 14-3 on
page 143). 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
138 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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)
139
7593D–AVR–07/06
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)
14.10.1 Timer/Counter1 Control Register A – TCCR1A
Bit 7 6 5 4 3 2 1 0
COM1A COM1A COM1B COM1B COM1C COM1C WGM11 WGM1 TCCR1
1 0 1 0 1 0 0 A
14.10.2 Timer/Counter3 Control Register A – TCCR3A
Bit 7 6 5 4 3 2 1 0
COM3A COM3A COM3B COM3B COM3C COM3C WGM3 WGM3 TCCR3
1 0 1 0 1 0 1 0 A
• Bit 7:6 – COMnA1:0: Compare Output Mode for Channel A

• Bit 5:4 – COMnB1:0: Compare Output Mode for Channel B
• Bit 3:2 – COMnC1:0: Compare Output Mode for Channel C
The COMnA1:0, COMnB1:0, and COMnC1:0 control the output compare pins (OCnA, OCnB,
and OCnC respectively) 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 bits are written to one, the OCnB output overrides the normal port func-
tionality of the I/O pin it is connected to. If one or both of the COMnC1:0 bits are written to one,
the OCnC output overrides the normal port functionality of the I/O pin it is connected to. How-
ever, note that the Data Direction Register (DDR) bit corresponding to the OCnA, OCnB or
OCnC pin must be set in order to enable the output driver.
140 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
When the OCnA, OCnB or OCnC is connected to the pin, the function of the COMnx1:0 bits is
dependent of the WGMn3:0 bits setting. Table 14-1 shows the COMnx1:0 bit functionality when
the WGMn3:0 bits are set to a normal or a CTC mode (non-PWM).
141
7593D–AVR–07/06
.
Table 14-1. Compare Output Mode, non-PWM

COMnA1/COMnB1/ COMnA0/COMnB0/
COMnC1 COMnC0 Description
Normal port operation, OCnA/OCnB/OCnC
0 0
disconnected.
Toggle OCnA/OCnB/OCnC on compare
0 1
match.
Clear OCnA/OCnB/OCnC on compare
1 0
match (set output to low level).
Set OCnA/OCnB/OCnC on compare match
1 1
(set output to high level).
Table 14-2 shows the COMnx1:0 bit functionality when the WGMn3:0 bits are set to the fast
PWM mode.
Table 14-2. Compare Output Mode, Fast PWM

COMnA1/COMnB1/ COMnA0/COMnB0/
0 0
disconnected.
WGM13:0 = 14 or 15: Toggle OC1A on
Compare Match, OC1B and OC1C
0 1 disconnected (normal port operation). For all
other WGM1 settings, normal port operation,
OC1A/OC1B/OC1C disconnected.
1 0
match, set OCnA/OCnB/OCnC at TOP
Set OCnA/OCnB/OCnC on compare match,
1 1
clear OCnA/OCnB/OCnC at TOP
Note: A special case occurs when OCRnA/OCRnB/OCRnC equals TOP and

COMnA1/COMnB1/COMnC1 is set. In this case the compare match is ignored, but the set or clear
is done at TOP. See “Fast PWM Mode” on page 108. for more details.
Table 14-3 shows the COMnx1:0 bit functionality when the WGMn3:0 bits are set to the phase
correct and frequency correct PWM mode.
142 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 14-3. Compare Output Mode, Phase Correct and Phase and Frequency Correct PWM
COMnA1/COMnB/ COMnA0/COMnB0/
0 0
disconnected.
WGM13:0 = 8, 9 10 or 11: Toggle OC1A on
Compare Match, OC1B and OC1C
0 1 disconnected (normal port operation). For all
other WGM1 settings, normal port operation,
OC1A/OC1B/OC1C disconnected.
match when up-counting. Set
1 0
OCnA/OCnB/OCnC on compare match
when downcounting.
Set OCnA/OCnB/OCnC on compare match
when up-counting. Clear
1 1
OCnA/OCnB/OCnC on compare match
when downcounting.
Note: A special case occurs when OCRnA/OCRnB/OCRnC equals TOP and
COMnA1/COMnB1//COMnC1 is set. See “Phase Correct PWM Mode” on page 110. 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
unit are: Normal mode (counter), Clear Timer on Compare match (CTC) mode, and three types
of Pulse Width Modulation (PWM) modes. (See “Modes of Operation” on page 107.).
143
7593D–AVR–07/06
Table 14-4. 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 TOP 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 TOP TOP
15 1 1 1 1 Fast PWM OCRnA TOP 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.
14.10.3 Timer/Counter1 Control Register B – TCCR1B

Bit 7 6 5 4 3 2 1 0
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
14.10.4 Timer/Counter3 Control Register B – TCCR3B

Bit 7 6 5 4 3 2 1 0
ICNC3 ICES3 – WGM33 WGM32 CS32 CS31 CS30 TCCR3B
• 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
144 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
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
13-8 and Figure 13-9.
145
7593D–AVR–07/06
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.
14.10.5 Timer/Counter1 Control Register C – TCCR1C

Bit 7 6 5 4 3 2 1 0
FOC1A FOC1B FOC1C – – – – – TCCR1C
Read/Write W W W R R R R R
14.10.6 Timer/Counter3 Control Register C – TCCR3C

Bit 7 6 5 4 3 2 1 0
FOC3A FOC3B FOC3C – – – – – TCCR3C
Read/Write W W W R R R R R
• Bit 7 – FOCnA: Force Output Compare for Channel A

• Bit 6 – FOCnB: Force Output Compare for Channel B
• Bit 5 – FOCnC: Force Output Compare for Channel C
The FOCnA/FOCnB/FOCnC bits are only active when the WGMn3:0 bits specifies a non-PWM
mode. When writing a logical one to the FOCnA/FOCnB/FOCnC bit, an immediate compare
match is forced on the waveform generation unit. The OCnA/OCnB/OCnC output is changed
according to its COMnx1:0 bits setting. Note that the FOCnA/FOCnB/FOCnC bits are imple-
mented as strobes. Therefore it is the value present in the COMnx1:0 bits that determine the
effect of the forced compare.
A FOCnA/FOCnB/FOCnC 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/FOCnB bits are always read as zero.
• Bit 4:0 – Reserved Bits
These bits are reserved for future use. For ensuring compatibility with future devices, these bits
must be written to zero when TCCRnC is written.
14.10.7 Timer/Counter1 – TCNT1H and TCNT1L

Bit 7 6 5 4 3 2 1 0
146 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
TCNT1[15:8] TCNT1H
TCNT1[7:0] TCNT1L
14.10.8 Timer/Counter3 – TCNT3H and TCNT3L

Bit 7 6 5 4 3 2 1 0
TCNT3[15:8] TCNT3H
TCNT3[7:0] TCNT3L
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 “Accessing 16-bit
Registers” on page 121.
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.
14.10.9 Output Compare Register 1 A – OCR1AH and OCR1AL

Bit 7 6 5 4 3 2 1 0
OCR1A[15:8] OCR1AH
OCR1A[7:0] OCR1AL
14.10.10 Output Compare Register 1 B – OCR1BH and OCR1BL

Bit 7 6 5 4 3 2 1 0
OCR1B[15:8] OCR1BH
OCR1B[7:0] OCR1BL
14.10.11 Output Compare Register 1 C – OCR1CH and OCR1CL

Bit 7 6 5 4 3 2 1 0
OCR1C[15:8] OCR1CH
OCR1C[7:0] OCR1CL
14.10.12 Output Compare Register 3 A – OCR3AH and OCR3AL

Bit 7 6 5 4 3 2 1 0
OCR3A[15:8] OCR3AH
OCR3A[7:0] OCR3AL
147
7593D–AVR–07/06
14.10.13 Output Compare Register 3 B – OCR3BH and OCR3BL
Bit 7 6 5 4 3 2 1 0
OCR3B[15:8] OCR3BH
OCR3B[7:0] OCR3BL
14.10.14 Output Compare Register 3 C – OCR3CH and OCR3CL

Bit 7 6 5 4 3 2 1 0
OCR3C[15:8] OCR3CH
OCR3C[7:0] OCR3CL
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 “Accessing 16-bit Registers” on page 121.
14.10.15 Input Capture Register 1 – ICR1H and ICR1L

Bit 7 6 5 4 3 2 1 0
ICR1[15:8] ICR1H
ICR1[7:0] ICR1L
14.10.16 Input Capture Register 3 – ICR3H and ICR3L

Bit 7 6 5 4 3 2 1 0
ICR3[15:8] ICR3H
ICR3[7:0] ICR3L
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 “Accessing 16-bit Registers” on page 121.
14.10.17 Timer/Counter1 Interrupt Mask Register – TIMSK1

Bit 7 6 5 4 3 2 1 0
– – ICIE1 – OCIE1 OCIE1B OCIE1A TOIE1 TIMSK1
C
Read/Write R R R/W R R/W R/W R/W R/W
148 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Bit 7 6 5 4 3 2 1 0
– – ICIE3 – OCIE3 OCIE3B OCIE3A TOIE3 TIMSK3
C
• Bit 5 – ICIEn: Timer/Countern, 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/Countern Input Capture interrupt is enabled. The corresponding Interrupt
Vector (See “Interrupts” on page 70.) is executed when the ICFn Flag, located in TIFRn, is set.
• Bit 3 – OCIEnC: Timer/Countern, Output Compare C Match Interrupt Enable
enabled), the Timer/Countern Output Compare C Match interrupt is enabled. The corresponding
Interrupt Vector (See “Interrupts” on page 70.) is executed when the OCFnC Flag, located in
TIFRn, is set.
• Bit 2 – OCIEnB: Timer/Countern, Output Compare B Match Interrupt Enable

enabled), the Timer/Countern Output Compare B Match interrupt is enabled. The corresponding
Interrupt Vector (See “Interrupts” on page 70.) is executed when the OCFnB Flag, located in
TIFRn, is set.
• Bit 1 – OCIEnA: Timer/Countern, Output Compare A Match Interrupt Enable

enabled), the Timer/Countern Output Compare A Match interrupt is enabled. The corresponding
Interrupt Vector (See “Interrupts” on page 70.) is executed when the OCFnA Flag, located in
TIFRn, is set.
• Bit 0 – TOIEn: Timer/Countern, Overflow Interrupt Enable

enabled), the Timer/Countern Overflow interrupt is enabled. The corresponding Interrupt Vector
(See “Interrupts” on page 70.) is executed when the TOVn Flag, located in TIFRn, is set.
14.10.19 Timer/Counter1 Interrupt Flag Register – TIFR1

Bit 7 6 5 4 3 2 1 0
– – ICF1 – OCF1C OCF1B OCF1A TOV1 TIFR1

Bit 7 6 5 4 3 2 1 0
– – ICF3 – OCF3C OCF3B OCF3A TOV3 TIFR3
• Bit 5 – ICFn: Timer/Countern, Input Capture Flag
149
7593D–AVR–07/06
This flag is set when a capture event occurs on the ICPn pin. When the Input Capture Register
(ICRn) is set by the WGMn3:0 to be used as the TOP value, the ICFn Flag is set when the
counter reaches the TOP value.
ICFn is automatically cleared when the Input Capture Interrupt Vector is executed. Alternatively,
ICFn can be cleared by writing a logic one to its bit location.
• Bit 3– OCFnC: Timer/Countern, Output Compare C Match Flag

This flag is set in the timer clock cycle after the counter (TCNTn) value matches the Output
Compare Register C (OCRnC).
Note that a Forced Output Compare (FOCnC) strobe will not set the OCFnC Flag.
OCFnC is automatically cleared when the Output Compare Match C Interrupt Vector is exe-
cuted. Alternatively, OCFnC can be cleared by writing a logic one to its bit location.
• Bit 2 – OCFnB: Timer/Counter1, Output Compare B Match Flag

This flag is set in the timer clock cycle after the counter (TCNTn) value matches the Output
Compare Register B (OCRnB).
Note that a Forced Output Compare (FOCnB) strobe will not set the OCFnB Flag.
OCFnB is automatically cleared when the Output Compare Match B Interrupt Vector is exe-
cuted. Alternatively, OCFnB 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 (TCNTn value matches the Output Com-
pare Register A (OCRnA).
Note that a Forced Output Compare (FOCnA) strobe will not set the OCFnA Flag.
OCFnA is automatically cleared when the Output Compare Match A Interrupt Vector is exe-
cuted. Alternatively, OCFnA can be cleared by writing a logic one to its bit location.
• Bit 0 – TOVn: Timer/Countern, Overflow Flag

The setting of this flag is dependent of the WGMn3:0 bits setting. In Normal and CTC modes,
the TOVn Flag is set when the timer overflows. Refer to Table 14-4 on page 144 for the TOVn
Flag behavior when using another WGMn3:0 bit setting.
TOVn is automatically cleared when the Timer/Countern Overflow Interrupt Vector is executed.
Alternatively, TOVn can be cleared by writing a logic one to its bit location.
150 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
15. 8-bit Timer/Counter2 with PWM and Asynchronous Operation
Timer/Counter2 is a general purpose, single channel, 8-bit Timer/Counter module. The main
features are:
• 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
15.1 Overview
A simplified block diagram of the 8-bit Timer/Counter is shown in Figure 15-1.. For the actual
placement of I/O pins, see “Pin Configurations” on page 3. 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 “8-bit Timer/Counter Register Description” on page 162.
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
151
7593D–AVR–07/06
15.1.1 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 “Output Compare Unit” on page 153. 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.
15.1.2 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 the table below 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 dependent 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 “Asyn-
chronous Status Register – ASSR” on page 167. For details on clock sources and prescaler, see
“Timer/Counter Prescaler” on page 171.
15.3 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.
152 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 156.
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.
15.4 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
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 156).
Figure 14-10 on page 138 shows a block diagram of the Output Compare unit.
153
7593D–AVR–07/06
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).
15.4.2 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.

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.
154 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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-
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 “8-bit Timer/Counter Register Description” on page 162.
155
7593D–AVR–07/06
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 15-4 on page 163. For fast PWM mode, refer to Table 15-5 on
page 164, and for phase correct PWM refer to Table 15-6 on page 164.
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 “Compare Match Output Unit” on page 155.).
For detailed timing information refer to “Timer/Counter Timing Diagrams” on page 160.
15.6.1 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.
The timing diagram for the CTC mode is shown in Table 15-5. The counter value (TCNT2)
increases until a compare match occurs between TCNT2 and OCR2A, and then counter
(TCNT2) is cleared.
156 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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
for power regulation, rectification, and DAC applications. High frequency allows physically small
sized external components (coils, capacitors), and therefore reduces total system cost.
157
7593D–AVR–07/06
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. The TCNT2 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 TCNT2 slopes represent 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 15-2 on page 163). 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.)
ting OC2x to toggle its logical level on each compare match (COM2x1:0 = 1). The waveform
158 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
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
159
7593D–AVR–07/06
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 15-3 on page 163). 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 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.
• OCR2A changes its value from MAX, like in Figure 15-7. 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 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.
clkI/O
clkTn
(clkI/O /1)
TOVn
160 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
clkI/O
clkTn
(clkI/O /8)
TOVn
Figure 15-10 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
Figure 15-11 shows the setting of OCF2A and the clearing of TCNT2 in CTC mode.
161
7593D–AVR–07/06
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
15.8.1 Timer/Counter Control Register A – TCCR2A

Bit 7 6 5 4 3 2 1 0
COM2A COM2A COM2B COM2B – – WGM2 WGM2 TCCR2A
1 0 1 0 1 0
• 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
When OC2A is connected to the pin, the function of the COM2A1:0 bits depends on the

0 1 Toggle OC2A on Compare Match
1 0 Clear OC2A on Compare Match
1 1 Set OC2A on Compare Match
162 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 15-2 shows the COM2A1:0 bit functionality when the WGM21:0 bits are set to fast PWM
mode.

0 1
1 0 Clear OC2A on Compare Match, set OC2A at TOP
1 1 Set OC2A on Compare Match, clear OC2A at TOP
for more details.
Table 15-3 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
When OC2B is connected to the pin, the function of the COM2B1:0 bits depends on the
WGM22:0 bit setting. Table 15-4 shows the COM2B1:0 bit functionality when the WGM22:0 bits

COM2B1 COM2B0 Description
0 1 Toggle OC2B on Compare Match
1 0 Clear OC2B on Compare Match
1 1 Set OC2B on Compare Match
163
7593D–AVR–07/06
Table 15-5 shows the COM2B1:0 bit functionality when the WGM22:0 bits are set to fast PWM
mode.

0 1 Reserved
1 0 Clear OC2B on Compare Match, set OC2B at TOP
1 1 Set OC2B on Compare Match, clear OC2B at TOP
for more details.
Table 15-6 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
• Bits 3, 2 – Res: Reserved Bits

• Bits 1:0 – WGM21:0: Waveform Generation Mode

Combined with the WGM22 bit found in the TCCR2B Register, these bits control the counting
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 156).

Timer/Counter
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
164 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Timer/Counter
3 0 1 1 Fast PWM 0xFF TOP 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 TOP TOP
Notes: 1. MAX= 0xFF

2. BOTTOM= 0x00
15.8.2 Timer/Counter Control Register B – TCCR2B

Bit 7 6 5 4 3 2 1 0
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.
• 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.
• Bits 5:4 – Res: Reserved Bits

165
7593D–AVR–07/06
• Bit 3 – WGM22: Waveform Generation Mode
See the description in the “Timer/Counter Control Register A – TCCR2A” on page 162.
• 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
15-8.

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.
15.8.3 Timer/Counter Register – TCNT2

Bit 7 6 5 4 3 2 1 0
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
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.
15.8.4 Output Compare Register A – OCR2A

Bit 7 6 5 4 3 2 1 0
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.
15.8.5 Output Compare Register B – OCR2B

Bit 7 6 5 4 3 2 1 0
OCR2B[7:0] OCR2B
166 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
The Output Compare Register B contains an 8-bit value that is continuously compared with the
generate a waveform output on the OC2B pin.
15.9 Asynchronous operation of the Timer/Counter
15.9.1 Asynchronous Status Register – ASSR

Bit 7 6 5 4 3 2 1 0
– 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.
• 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.
167
7593D–AVR–07/06
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.
15.9.2 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
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
168 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.

Bit 7 6 5 4 3 2 1 0
– – – – – 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
169
7593D–AVR–07/06
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.

Bit 7 6 5 4 3 2 1 0
– – – – – 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.
170 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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. Apply-
ing an external clock source to TOSC1 is not recommended.
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.
15.10.1 General Timer/Counter Control Register – GTCCR

Bit 7 6 5 4 3 2 1 0
TSM – – – – – PSRA- PSRSY GTCCR
SY NC
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 “General Timer/Counter Control
Register – GTCCR” on page 101 for a description of the Timer/Counter Synchronization mode.
171
7593D–AVR–07/06
16. Output Compare Modulator (OCM1C0A)
16.1 Overview
The Output Compare Modulator (OCM) allows generation of waveforms modulated with a carrier
frequency. The modulator uses the outputs from the Output Compare Unit C of the 16-bit
Timer/Counter1 and the Output Compare Unit of the 8-bit Timer/Counter0. For more details
about these Timer/Counters see “Timer/Counter0, Timer/Counter1, and Timer/Counter3 Pres-
calers” on page 100 and “8-bit Timer/Counter2 with PWM and Asynchronous Operation” on
page 151.
Figure 16-1. Output Compare Modulator, Block Diagram
Timer/Counter 1 OC1C
Pin
OC1C /
Timer/Counter 0 OC0A OC0A / PB7
When the modulator is enabled, the two output compare channels are modulated together as
shown in the block diagram (Figure 16-1).
16.2 Description
The Output Compare unit 1C and Output Compare unit 2 shares the PB7 port pin for output. The
outputs of the Output Compare units (OC1C and OC0A) overrides the normal PORTB7 Register
when one of them is enabled (i.e., when COMnx1:0 is not equal to zero). When both OC1C and
OC0A are enabled at the same time, the modulator is automatically enabled.
The functional equivalent schematic of the modulator is shown on Figure 16-2. The schematic
includes part of the Timer/Counter units and the port B pin 7 output driver circuit.
Figure 16-2. Output Compare Modulator, Schematic

COMA01 Vcc
COMA00
COM1C1 Modulator
COM1C0
0
1
( From Waveform Generator ) D Q
1
OC1C
Pin
0
OC1C /
( From Waveform Generator ) D Q OC0A/ PB7
OC0A
D Q D Q
PORTB7 DDRB7
DATABUS
172 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
When the modulator is enabled the type of modulation (logical AND or OR) can be selected by
the PORTB7 Register. Note that the DDRB7 controls the direction of the port independent of the
COMnx1:0 bit setting.
16.2.1 Timing Example

Figure 16-3 illustrates the modulator in action. In this example the Timer/Counter1 is set to oper-
ate in fast PWM mode (non-inverted) and Timer/Counter0 uses CTC waveform mode with toggle
Compare Output mode (COMnx1:0 = 1).
Figure 16-3. Output Compare Modulator, Timing Diagram
clk I/O
OC1C
(FPWM Mode)
OC0A
(CTC Mode)
PB7
(PORTB7 = 0)
PB7
(PORTB7 = 1)
1 2 3
(Period)
In this example, Timer/Counter2 provides the carrier, while the modulating signal is generated
by the Output Compare unit C of the Timer/Counter1.
The resolution of the PWM signal (OC1C) is reduced by the modulation. The reduction factor is
equal to the number of system clock cycles of one period of the carrier (OC0A). In this example
the resolution is reduced by a factor of two. The reason for the reduction is illustrated in Figure
16-3 at the second and third period of the PB7 output when PORTB7 equals zero. The period 2
high time is one cycle longer than the period 3 high time, but the result on the PB7 output is
equal in both periods.
173
7593D–AVR–07/06
17. Serial Peripheral Interface – SPI
The Serial Peripheral Interface (SPI) allows high-speed synchronous data transfer between the
AT90USB64/128 and peripheral devices or between several AVR devices. The
AT90USB64/128 SPI includes the following 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
USART can also be used in Master SPI mode, see “USART in SPI Mode” on page 208.
The Power Reduction SPI bit, PRSPI, in “Power Reduction Register 0 - PRR0” on page 56 on
page 50 must be written to zero to enable SPI module.
Figure 17-1. SPI Block Diagram(1)
DIVIDER
/2/4/8/16/32/64/128
SPI2X
SPI2X
Note: 1. Refer to Figure 1-1 on page 3, and Table 10-6 on page 82 for SPI pin placement.
The interconnection between Master and Slave CPUs with SPI is shown in Figure 17-2. 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
174 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 17-2. 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 frequency of the SPI clock should never exceed fosc/4.
175
7593D–AVR–07/06
When the SPI is enabled, the data direction of the MOSI, MISO, SCK, and SS pins is overridden
according to Table 17-1. For more details on automatic port overrides, refer to “Alternate Port
Functions” on page 79.
Table 17-1. 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 82 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.
176 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128

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)))
;
}
177
7593D–AVR–07/06
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;
}
17.1 SS Pin Functionality
17.1.1 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
178 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
17.1.2 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.
17.1.3 SPI Control Register – SPCR

Bit 7 6 5 4 3 2 1 0
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,
179
7593D–AVR–07/06
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 17-3 and Figure 17-4 for an example. The CPOL functionality is sum-
marized below:
Table 17-2. 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 17-3 and Figure 17-4 for an example. The CPOL
functionality is summarized below:
Table 17-3. CPHA Functionality

CPHA Leading Edge Trailing Edge
0 Sample Setup
1 Setup Sample
• 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 17-4. 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
17.1.4 SPI Status Register – SPSR

Bit 7 6 5 4 3 2 1 0
SPIF WCOL – – – – – SPI2X SPSR
• Bit 7 – SPIF: SPI Interrupt Flag
180 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 17-4). 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 AT90USB64/128 is also used for program memory and EEPROM
downloading or uploading. See page 382 for serial programming and verification.
17.1.5 SPI Data Register – SPDR

Bit 7 6 5 4 3 2 1 0
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.
17.2 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
17-3 and Figure 17-4. Data bits are shifted out and latched in on opposite edges of the SCK sig-
nal, ensuring sufficient time for data signals to stabilize. This is clearly seen by summarizing
Table 17-2 and Table 17-3, as done below:
181
7593D–AVR–07/06
Table 17-5. CPOL Functionality
Leading Edge Trailing eDge SPI Mode
CPOL=0, CPHA=0 Sample (Rising) Setup (Falling) 0
CPOL=0, CPHA=1 Setup (Rising) Sample (Falling) 1
CPOL=1, CPHA=0 Sample (Falling) Setup (Rising) 2
CPOL=1, CPHA=1 Setup (Falling) Sample (Rising) 3
Figure 17-3. 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 17-4. 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
182 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
18. USART
The Universal Synchronous and Asynchronous serial Receiver and Transmitter (USART) is a
highly flexible serial communication device. The main features are:
• 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
.
18.1 Overview
A simplified block diagram of the USART Transmitter is shown in Figure 18-1 on page 184. CPU
accessible I/O Registers and I/O pins are shown in bold.
183
7593D–AVR–07/06
Figure 18-1. 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 3, Table 10-12 on page 86 and 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.
18.2 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
UCSRnA Register. When using synchronous mode (UMSELn = 1), the Data Direction Register
184 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 18-2 shows a block diagram of the clock generation logic.
Figure 18-2. 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).
18.2.1 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 18-2.
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.
185
7593D–AVR–07/06
Table 18-1 contains equations for calculating the baud rate (in bits per second) and for calculat-
ing the UBRRn value for each mode of operation using an internally generated clock source.
Table 18-1. Equations for Calculating Baud Rate Register Setting

Equation for Calculating Equation for Calculating
Operating Mode Baud Rate(1) UBRR Value
f OSC
UBRRn = ------------------------ – 1
16BAUD
Asynchronous Normal f OSC

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

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

BAUD = ---------------------------------------
mode 2 ( UBRRn + 1 )
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)
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 18-9 on
page 205.
18.2.2 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.
186 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
18.2.3 External Clock
External clocking is used by the synchronous slave modes of operation. The description in this
section refers to Figure 18-2 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-
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.
18.2.4 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 18-3. 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 18-3 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.
18.3 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
187
7593D–AVR–07/06
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 18-4 illustrates the possible combinations of the frame formats. Bits inside brackets are
optional.
Figure 18-4. 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.
18.3.1 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

odd
P 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.
188 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
18.4 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
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.
189
7593D–AVR–07/06
18.5 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.
18.5.1 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.
18.5.2 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
190 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
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.
18.5.3 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
transmission is used, the Data Register Empty interrupt routine must either write new data to
191
7593D–AVR–07/06
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.
18.5.4 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.
18.5.5 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.
18.6 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.
18.6.1 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
192 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.

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.
18.6.2 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.
193
7593D–AVR–07/06
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.
18.6.3 Receive Compete Flag and Interrupt

The USART Receiver has one flag that indicates the Receiver state.
194 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
18.6.4 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 188 and “Parity Checker” on page 195.
18.6.5 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
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.
195
7593D–AVR–07/06
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.
18.6.6 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
18.6.7 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;
}
18.7 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.
18.7.1 Asynchronous Clock Recovery

The clock recovery logic synchronizes internal clock to the incoming serial frames. Figure 18-5
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
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).
196 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 18-5. 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.
18.7.2 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 18-6 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 18-6. 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 18-7 shows the sampling of the stop bit and the earliest possible beginning of the start bit
of the next frame.
197
7593D–AVR–07/06
Figure 18-7. 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 18-7. 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.
18.7.3 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 18-2) 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.
Table 1.
( 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 18-2 and Table 18-3 list the maximum receiver baud rate error that can be tolerated. Note
that Normal Speed mode has higher toleration of baud rate variations.
198 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 18-2. 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 18-3. 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.
18.8 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
199
7593D–AVR–07/06
nine data bits, then the ninth bit (RXB8n) is used for identifying address and data frames. When
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.
18.8.1 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.
18.9 USART Register Description
18.9.1 USART I/O Data Register n– UDRn

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-
200 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
18.9.2 USART Control and Status Register A – UCSRnA

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
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
201
7593D–AVR–07/06
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 199.
18.9.3 USART Control and Status Register n B – UCSRnB

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.
• 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
202 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
18.9.4 USART Control and Status Register n C – UCSRnC

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 18-4..
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)(1)
Note: 1. See “USART in SPI Mode” on page 208 for full description of the Master SPI Mode (MSPIM)
operation
• 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 18-5. UPMn Bits Settings

UPMn1 UPMn0 Parity Mode
0 0 Disabled
0 1 Reserved
1 0 Enabled, Even Parity
1 1 Enabled, Odd Parity
203
7593D–AVR–07/06
• 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 18-6. 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 18-7. 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
• 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 18-8. UCPOLn Bit Settings

Transmitted Data Changed (Output Received Data Sampled (Input on
UCPOLn of TxDn Pin) RxDn Pin)
0 Rising XCKn Edge Falling XCKn Edge
1 Falling XCKn Edge Rising XCKn Edge
18.9.5 USART Baud Rate Registers – UBRRLn and UBRRHn

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
204 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
18.10 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 18-9 to Table 18-12.
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 198). The error values are calculated using the following equation:
BaudRate Closest Match

Error[%] = ⎛ -------------------------------------------------------- – 1⎞ • 100%
⎝ BaudRate ⎠
Table 18-9. 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%
205
7593D–AVR–07/06
Table 18-10. 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%
206 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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%
207
7593D–AVR–07/06
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%
19. USART in SPI Mode

The Universal Synchronous and Asynchronous serial Receiver and Transmitter (USART) can be
set to a master SPI compliant mode of operation. The Master SPI Mode (MSPIM) has the follow-
ing 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
19.1 Overview
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
208 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
19.2 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 19-1:
Table 19-1. Equations for Calculating Baud Rate Register Setting

Equation for Calculating Baud Equation for Calculating
Operating Mode Rate(1) UBRRn Value
Synchronous Master f OSC f OSC

BAUD = --------------------------------------- UBRRn = -------------------- – 1
mode 2 ( UBRRn + 1 ) 2BAUD
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)
fOSC System Oscillator clock frequency
UBRRn Contents of the UBRRnH and UBRRnL Registers, (0-4095)
19.3 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 19-1. 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-
209
7593D–AVR–07/06
ality is summarized in Table 19-2. Note that changing the setting of any of these bits will corrupt
all ongoing communication for both the Receiver and Transmitter.
Table 19-2. 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 19-1. 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)
19.4 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.
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.
19.4.1 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-
210 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
211
7593D–AVR–07/06
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;
}
19.5 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.
212 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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
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;
}
213
7593D–AVR–07/06
19.5.1 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.
19.5.2 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.
19.6 USART MSPIM Register Description

The following section describes the registers used for SPI operation using the USART.
19.6.1 USART MSPIM I/O Data Register - UDRn

The function and bit description of the USART data register (UDRn) in MSPI mode is identical to
normal USART operation. See “USART I/O Data Register n– UDRn” on page 200.
19.6.2 USART MSPIM Control and Status Register n A - UCSRnA
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.
214 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
19.6.3 USART MSPIM Control and Status Register n B - UCSRnB
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.
• 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.
19.6.4 USART MSPIM Control and Status Register n C - UCSRnC
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
215
7593D–AVR–07/06
These bits select the mode of operation of the USART as shown in Table 19-3. See “USART
Control and Status Register n C – UCSRnC” on page 203 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 19-3. UMSELn Bits Settings

UMSELn1 UMSELn0 Mode
0 0 Asynchronous USART
0 1 Synchronous USART
1 0 (Reserved)
1 1 Master SPI (MSPIM)

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.
19.6.5 USART MSPIM Baud Rate Registers - UBRRnL and UBRRnH

The function and bit description of the baud rate registers in MSPI mode is identical to normal
USART operation. See “USART Baud Rate Registers – UBRRLn and UBRRHn” on page 204.
19.7 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.
216 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
• 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 19-4 on page
217.
Table 19-4. 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
217
7593D–AVR–07/06
20. 2-wire Serial Interface
20.1 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
20.2 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 20-1. TWI Bus Interconnection

VCC
Device 1 Device 2 Device 3 ........ Device n R1 R2
SDA
SCL
20.2.1 TWI Terminology

The following definitions are frequently encountered in this section.
Table 20-1. TWI Terminology

Term Description
The device that initiates and terminates a transmission. The Master also
Master
generates the 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.
218 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
The Power Reduction TWI bit, PRTWI bit in “Power Reduction Register 0 - PRR0” on page 56
must be written to zero to enable the 2-wire Serial Interface.
20.2.2 Electrical Interconnection

As depicted in Figure 20-1, 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 405. 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.
20.3 Data Transfer and Frame Format
20.3.1 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 20-2. Data Validity
SDA
SCL
Data Stable Data Stable
Data Change
20.3.2 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
219
7593D–AVR–07/06
depicted below, START and STOP conditions are signalled by changing the level of the SDA
line when the SCL line is high.
Figure 20-3. START, REPEATED START and STOP conditions
SDA
SCL
START STOP START REPEATED START STOP
20.3.3 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 20-4. Address Packet Format
Addr MSB Addr LSB R/W ACK
SDA
SCL
1 2 7 8 9
START
220 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
20.3.4 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 20-5. 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
20.3.5 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 20-6 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 imple-
mented by the application software.
Figure 20-6. Typical Data Transmission
Addr MSB Addr LSB R/W ACK Data MSB Data LSB ACK
SDA
SCL
1 2 7 8 9 1 2 7 8 9
START SLA+R/W Data Byte STOP
221
7593D–AVR–07/06
20.4 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 20-7. 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
222 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
bits. If several masters are trying to address the same Slave, arbitration will continue into the
data packet.
Figure 20-8. 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.
20.5 Overview of the TWI Module

The TWI module is comprised of several submodules, as shown in Figure 20-9. All registers
drawn in a thick line are accessible through the AVR data bus.
223
7593D–AVR–07/06
Figure 20-9. 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
20.5.1 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.
20.5.2 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:
224 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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: TWBR should be 10 or higher if the TWI operates in Master mode. If TWBR is lower than 10, the
Master may produce an incorrect output on SDA and SCL for the reminder of the byte. The prob-
lem occurs when operating the TWI in Master mode, sending Start + SLA + R/W to a Slave (a
Slave does not need to be connected to the bus for the condition to happen).
20.5.3 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.
20.5.4 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.
20.5.5 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.
225
7593D–AVR–07/06
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.
20.6 TWI Register Description
20.6.1 TWI Bit Rate Register – TWBR

Bit 7 6 5 4 3 2 1 0
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 224 for calculating bit rates.
20.6.2 TWI Control Register – TWCR

Bit 7 6 5 4 3 2 1 0
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
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:
226 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
• 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.
20.6.3 TWI Status Register – TWSR

Bit 7 6 5 4 3 2 1 0
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-
227
7593D–AVR–07/06
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.

• Bits 1..0 – TWPS: TWI Prescaler Bits

These bits can be read and written, and control the bit rate prescaler.
Table 20-2. 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 224. The value of TWPS1..0 is
used in the equation.
20.6.4 TWI Data Register – TWDR

Bit 7 6 5 4 3 2 1 0
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
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.
20.6.5 TWI (Slave) Address Register – TWAR

Bit 7 6 5 4 3 2 1 0
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.
228 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
20.6.6 TWI (Slave) Address Mask Register – TWAMR

Bit 7 6 5 4 3 2 1 0
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 20-10 shows the address match logic in
detail.
Figure 20-10. TWI Address Match Logic, Block Diagram
TWAR0
Address
Address Match
Bit 0
TWAMR0
Address Bit Comparator 0
Address Bit Comparator 6..1

20.7 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
229
7593D–AVR–07/06
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 20-11 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 20-11. 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
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
application 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 success-
fully been sent. The status code will also reflect whether a Slave acknowledged the
packet or not.
230 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
the 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.
the 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 spe-
cial 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 condition. 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.
• 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.
231
7593D–AVR–07/06
Table 2.
Assembly Code Example C Example Comments
ldi r16, TWCR = (1<<TWINT)|(1<<TWSTA)|
(1<<TWINT)|(1<<TWSTA)| (1<<TWEN)
1 Send START condition
(1<<TWEN)
out TWCR, r16
wait1: while (!(TWCR & (1<<TWINT)))
in r16,TWCR ; Wait for TWINT Flag set. This
2 indicates that the START
sbrs r16,TWINT condition 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
brne ERROR ERROR
3 ldi r16, SLA_W TWDR = SLA_W;
out TWDR, r16 TWCR = (1<<TWINT) | Load SLA_W into TWDR
(1<<TWEN); Register. Clear TWINT bit in
ldi r16, (1<<TWINT) | TWCR to start transmission of
(1<<TWEN) address
out TWCR, r16
wait2: while (!(TWCR & (1<<TWINT))) Wait for TWINT Flag set. This
in r16,TWCR ; indicates that the SLA+W has
4
sbrs r16,TWINT been 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) | Load DATA into TWDR Register.
ldi r16, (1<<TWINT) | (1<<TWEN); Clear TWINT bit in TWCR to start
(1<<TWEN) transmission of data
out TWCR, r16
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
MT_DATA_ACK go to ERROR
brne ERROR
7
ldi r16, TWCR = (1<<TWINT)|(1<<TWEN)|
(1<<TWINT)|(1<<TWEN)| (1<<TWSTO);
Transmit STOP condition
(1<<TWSTO)
out TWCR, r16
232 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
20.8 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 20-13 to Figure 20-19, 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 continue 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 20-3 to Table 20-6. Note that the prescaler bits are masked to zero in
these tables.
20.8.1 Master Transmitter Mode

In the Master Transmitter mode, a number of data bytes are transmitted to a Slave Receiver
(see Figure 20-12). 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 trans-
mitted, MR mode is entered. All the status codes mentioned in this section assume that the
prescaler bits are zero or are masked to zero.
233
7593D–AVR–07/06
Figure 20-12. 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 20-3). 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 follow-
ing 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 20-3.
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
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
234 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 20-3. 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
are 0 Hardware STA STO TWIN TWE Next Action Taken by TWI Hardware
T A
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
No TWDR action or 1 0 1 X Repeated START will be transmitted
0x28 Data byte has been transmit- Load data byte or 0 0 1 X Data byte will be transmitted and ACK or NOT ACK will
ted; be received
ACK has been received No TWDR action or 1 0 1 X Repeated START will be transmitted
0x30 Data byte has been transmit- Load data byte or 0 0 1 X Data byte will be transmitted and ACK or NOT ACK will
ted; be received
NOT ACK has been received No TWDR action or 1 0 1 X Repeated START will be transmitted
0x38 Arbitration lost in SLA+W or No TWDR action or 0 0 1 X 2-wire Serial Bus will be released and not addressed
data bytes Slave mode entered
No TWDR action 1 0 1 X A START condition will be transmitted when the bus
becomes free
235
7593D–AVR–07/06
Figure 20-13. 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
20.8.2 Master Receiver Mode

In the Master Receiver mode, a number of data bytes are received from a Slave Transmitter
(Slave see Figure 20-14). In order to enter a Master mode, a START condition must be transmit-
ted. 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.
236 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 20-14. 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 20-3). 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 20-4. 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
237
7593D–AVR–07/06
the Master to switch between Slaves, Master Transmitter mode and Master Receiver mode with-
out losing control over the bus.
Table 20-4. Status codes for Master Receiver Mode

(TWSR) Status of the 2-wire Serial Bus To TWCR
Prescaler Bits and 2-wire Serial Interface To/from TWDR
T A
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 No TWDR action or 0 0 1 X 2-wire Serial Bus will be released and not addressed
NOT 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
238 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 20-15. 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


20.8.3 Slave Receiver Mode

In the Slave Receiver mode, a number of data bytes are received from a Master Transmitter
(see Figure 20-16). All the status codes mentioned in this section assume that the prescaler bits
are zero or are masked to zero.
Figure 20-16. 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
239
7593D–AVR–07/06
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 20-5.
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.
240 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 20-5. Status Codes for Slave Receiver Mode

Prescaler Bits and 2-wire Serial Interface Hard- To/from TWDR
are 0 ware STA STO TWIN TWE Next Action Taken by TWI Hardware
T A
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
241
7593D–AVR–07/06
Figure 20-17. 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


20.8.4 Slave Transmitter Mode

In the Slave Transmitter mode, a number of data bytes are transmitted to a Master Receiver
(see Figure 20-18). All the status codes mentioned in this section assume that the prescaler bits
are zero or are masked to zero.
Figure 20-18. Data Transfer in Slave Transmitter Mode

VCC
Device 1 Device 2
SLAVE MASTER Device 3 ........ Device n R1 R2
TRANSMITTER RECEIVER
SDA
SCL
242 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 20-6.
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.
Table 20-6. Status Codes for Slave Transmitter Mode

Prescaler and 2-wire Serial Interface Hard- To/from TWDR
Bits ware STA STO TWIN TWE Next Action Taken by TWI Hardware
are 0 T A
243
7593D–AVR–07/06
Table 20-6. Status Codes for Slave Transmitter Mode
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
Figure 20-19. 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


20.8.5 Miscellaneous States

There are two status codes that do not correspond to a defined TWI state, see Table 20-7.
244 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 20-7. Miscellaneous States

Prescaler Bits and 2-wire Serial Interface To/from TWDR
T A
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.
20.8.6 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.
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 20-20. 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
245
7593D–AVR–07/06
20.9 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 20-21. 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.
• 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 20-22. Possible status values are given in circles.
246 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 20-22. 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
247
7593D–AVR–07/06
21. USB controller
21.1 Features
• Support full-speed and low-speed.
• Support ping-pong mode (dual bank)
• 832 bytes of DPRAM.
• 1 endpoint 64 bytes max (default control endpoint),
• 1 endpoints of 256 bytes max, (one or two banks),
• 5 endpoints of 64 bytes max, (one or two banks)
21.2 Block Diagram

The USB controller provides the hardware to interface a USB link to a data flow stored in a dou-
ble port memory (DPRAM).
The USB controller requires a 48 MHz ±0.25% reference clock, which is the output of an internal
PLL. The PLL generates the internal high frequency (48 MHz) clock for USB interface, the PLL
input is generated from an external low-frequency (the crystal oscillator or external clock input
pin from XTAL1, to satisfy the USB frequency accuracy and jitter, only these sources clock allow
proper functionnality of the USB controller).
The 48MHz clock is used to generate a 12 MHz Full-speed (or 1.5 MHz Low-Speed bit clock
from the received USB differential data and to transmit data according to full or low speed USB
device tolerance. Clock recovery is done by a Digital Phase Locked Loop (DPLL) block, which is
compliant with the jitter specification of the USB bus.
To comply the USB DC characteristics, USB Pads (D+ or D-) should be powered within the 3.0
to 3.6V range. As AT90USB64/128 can be powered up to 5.5V, an internal regulator can insure
the USB pads power supply.
248 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 21-1. USB controller Block Diagram overview
UVCC AVCC XTAL1
clk
USB Regulator PLL 2MHz PLL clock
UCAP
24x Prescaler
clk
48MHz
D- CPU
DPLL
Clock
Recovery
D+
USB
Interface
VBUS
On-Chip
USB DPRAM
UID
21.3 Typical Application Implementation

Depending on the USB operating mode (Device only, Reduced Host or OTG mode) and target
application power supply, the AT90USB64/128 requires different hardware typical
implementations.
Figure 21-2. Operating modes versus frequency and power-supply

VCC (V) Internal Regulator Max
Functionnal operation Operating Frequency (MHz)
5.5
16 MHz
4.5
USB operationnal
Normal operation
with internal regulator
3.6 8 MHz
3.4 USB operationnal

internal regulator not needed
3.0
Extra power consumption
2.7
USB not operationnal
2 MHz
VCC min
249
7593D–AVR–07/06
21.3.1 Device mode
21.3.1.1 Bus Powered device
Figure 21-3. Typical Bus powered application with 5V I/O
UVCC AVCC VCC
UCAP
1µF
VBUS VBUS
UDP D+
Rs=22
UDM D-
Rs=22
UVSS UVSS
UID UID
XTAL1 XTAL2 AVSS DVSS
Figure 21-4. Typical Bus powered application with 3V I/O

External
3V Regulator
UVCC AVCC VCC
UCAP
1µF
VBUS VBUS
UDP D+
Rs=22
UDM D-
Rs=22
UVSS UVSS
UVSS
UID UID
250 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
21.3.1.2 Self Powered device
Figure 21-5. Typical Self powered application with 3.4V to 5.5V I/O
External 3.4V - 5.5V

Power Supply
UVCC AVCC VCC
UCAP
1µF
VBUS VBUS
UDP D+
Rs=22
UDM D-
Rs=22
UVSS UVSS
UID UID
Figure 21-6. Typical Self powered application with 3.0V to 3.4 I/O

Power Supply
UVCC AVCC VCC
UCAP
1µF
VBUS VBUS
UDP D+
Rs=22
UDM D-
Rs=22
UVSS UVSS
UID UID
251
7593D–AVR–07/06
21.3.2 Host / OTG mode
Figure 21-7. Host/OTG powered application with 3.0V to 3.4 I/O

Power Supply
5V UVCC AVCC VCC

5V DC/DC
generator
UCAP
1µF
UVCON
VBUS VBUS
UDM D+
Rs=22
UDP D-
Rs=22
UVSS UVSS
UID UID
Figure 21-8. Host/OTG powered application with 5V I/O
External 5.0V
5V Power Supply
UVCC AVCC VCC
UCAP
1µF
UVCON
VBUS VBUS
UDP D+
Rs=22
UDM D-
Rs=22
UVSS UVSS
UID UID
252 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
21.4 General Operating Modes
21.4.1 Introduction
After a hardware reset, the USB controller is disabled. When enabled, the USB controller has to
run the Device Controller or the Host Controller. This is performed using the ID detection.
• If the ID pin is not connected to ground, the ID bit is set by hardware (internal pull up on the
UID pad) and the USB Device controller is selected.
• The ID bit is cleared by hardware when a low level has been detected on the ID pin. The
Device controller is then disabled and the Host controller enabled.
The software anyway has to select the mode (Host, Device) in order to access to the Device
controller registers or to the Host controller registers, which are multiplexed. For example, even
if the USB controller has detected a Device mode (pin ID high), the software shall select the
device mode (bit HOST cleared), otherwise it will access to the host registers. This is also true
for the Host mode.
21.4.2 Power-on and reset

The next diagram explains the USB controller main states on power-on:
Figure 21-9. USB controller states after reset
Clockstopped <any other

FRZCLK=1 USBE=0
state>
Macrooff
Reset HW
RESET
USBE=1
ID=1 USBE=0
USBE=0 USBE=1
ID=0
Dev ice USBE=0
Host
USB Controller state after an hardware reset is ‘Reset’. In this state:

• USBE is not set
• the macro clock is stopped in order to minimize the power consumption (FRZCLK=1),
• the macro is disabled,
• the pad is in the suspend mode,
• the Host and Device USB controllers internal states are reset.
• The DPACC bit and the DPADD10:0 field can be set by software. The DPRAM is not cleared.
• The SPDCONF bits can be set by software.
After setting USBE, the USB Controller enters in the Host or in the Device state (according to the
IP pin). The selected controller is ‘Idle’.
The USB Controller can at any time be ‘stopped’ by clearing USBE. In fact, clearing USBE acts
as an hardware reset.
253
7593D–AVR–07/06
21.4.3 Interrupts
Two interrupts vectors are assigned to USB interface.
Figure 21-10. USB Interrupt System
USB General
& OTG Interrupt
USB Device USB General

Interrupt Interrupt Vector
USB Host
Interrupt
Endpoint
Interrupt
USB Endpoint/Pipe
Interrupt Vector
Pipe
Interrupt
See Section 22.18, page 279 and Section 23.15, page 299 for more details on the Host and
Device interrupts.
254 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 21-11. USB General interrupt vector sources
IDTI
USBINT.1
IDTE
USBCON.1
VBUSTI
USBINT.0
VBUSTE
USBCON.0
STOI
OTGINT.5
STOE
OTGIEN.5
HNPERRI
OTGINT.4
HNPERRE USB General
OTGIEN.4 Interrupt Vector
ROLEEXI
OTGINT.3
ROLEEXE
OTGIEN.3
BCERRI
OTGINT.2
BCERRE
OTGIEN.2
VBERRI
OTGINT.1
VBERRE
OTGIEN.1
SRPI
OTGINT.0
SRPE
OTGIEN.0
UPRSMI
UDINT.6
UPRSME
UDIEN.6
EORSMI
UDINT.5
EORSME
UDIEN.5
WAKEUPI
UDINT.4
WAKEUPE USB Device USB General
UDIEN.4 Interrupt Interrupt Vector
EORSTI
UDINT.3
EORSTE
UDIEN.3
SOFI
UDINT.2
SOFE
UDIEN.2
SUSPI
UDINT.0
SUSPE
UDIEN.0
HWUPI
UHINT.6
HWUPE
UHIEN.6
HSOFI
UHINT.5
HSOFE
UHIEN.5
RXRSMI
UHINT.4
RXRSME
UHIEN.4
USB Host
RSMEDI Interrupt
UHINT.3
RSMEDE
UHIEN.3
RSTI
UHINT.2
RSTE
UHIEN.2
DDISCI
UHINT.1
DDISCE Asynchronous Interrupt source
UHIEN.1 (allows the CPU to wake up from power down mode)
DCONNI
UHINT.0
DCONNE
UHIEN.0
255
7593D–AVR–07/06
Figure 21-12. USB Endpoint/Pipe Interrupt vector sources
Endpoint 6
Endpoint 5
Endpoint 4
Endpoint 3
Endpoint 2
Endpoint 1
Endpoint 0
OVERFI
UESTAX.6
UNDERFI
FLERRE
UESTAX.5
UEIENX.7
NAKINI
UEINTX.6
NAKINE
UEIENX.6
NAKOUTI
UEINTX.4
TXSTPE
Endpoint Interrupt
UEIENX.4
RXSTPI EPINT
UEINTX.3 UEINT.X
TXOUTE
UEIENX.3
RXOUTI
UEINTX.2
RXOUTE
UEIENX.2
STALLEDI
UEINTX.1
STALLEDE
UEIENX.1
TXINI
UEINTX.0
TXINE
UEIENX.0
USB Endpoint/P
Interrupt Vecto
PIPE 6
PIPE 5
PIPE 4
PIPE 3
PIPE 2
PIPE 1
OVERFI PIPE 0
UPSTAX.6
UNDERFI
FLERRE
UPSTAX.5
UPIEN.7
NAKEDI
UPINTX.6
NAKEDE
UPIEN.6
PERRI
UPINTX.4
PERRE
UPIEN.4
Pipe Interrupt
TXSTPI FLERRE
UPINTX.3 UPIEN.7
TXSTPE
UPIEN.3
TXOUTI
UPINTX.2
TXOUTE
UPIEN.2
RXSTALLI
UPINTX.1
RXSTALLE
UPIEN.1
RXINI
UPINTX.0
RXINE
UPIEN.0
256 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 21-13. USB General and OTG Controller Interrupt System
IDTI
USBINT.1
IDTE
USBCON.1
VBUSTI
USBINT.0
VBUSTE
USBCON.0
STOI
OTGINT.5
STOE
OTGIEN.5
HNPERRI
OTGINT.4
HNPERRE USB General & OTG
OTGIEN.4 Interrupt Vector
ROLEEXI
OTGINT.3
ROLEEXE
OTGIEN.3
BCERRI
OTGINT.2
BCERRE
OTGIEN.2
VBERRI
OTGINT.1
VBERRE Asynchronous Interrupt source
OTGIEN.1 (allows the CPU to wake up from power down mode)
SRPI
OTGINT.0
SRPE
OTGIEN.0
There are 2 kind of interrupts: processing (i.e. their generation are part of the normal processing)
and exception (errors).
Processing interrupts are generated when:
• ID Pad detection (insert, remove)(IDTI)
• VBUS plug-in detection (insert, remove) (VBUSTI)
• SRP detected(SRPI)
• Role Exchanged(ROLEEXI)
Exception Interrupts are generated when:
• Drop on VBus Detected(VBERRI)
• Error during the B-Connection(BCERRI)
• HNP Error(HNPERRI)
• Time-out detected during Suspend mode(STOII)
21.5 Power modes
21.5.1 Idle mode

In this mode, the CPU core is halted (CPU clock stopped). The Idle mode is taken whatever the
USB controller is running or not. The CPU “wakes up” on any USB interrupts.
21.5.2 Power down

In this mode, the oscillator is stopped and halts all the product (CPU and peripherals). The CPU
“wakes up” when:
• the WAKEUPI interrupt is triggered in the Peripheral mode (HOST cleared),
257
7593D–AVR–07/06
• the HWUPI interrupt is triggered in the Host mode (HOST set).
• the IDTI interrupt is triggered
• the VBUSTI interrupt is triggered
21.5.3 Freeze clock

The firmware has the ability to reduce the power consumption by setting the FRZCLK bit, which
freeze the clock of USB controller. When FRZCLK is set, it is still possible to access to the fol-
lowing registers:
• USBCON, USBSTA, USBINT
• DPRAM direct access (DPADD10:0, UxDATX)
• UDCON (detach, ...)
• UDINT
• UDIEN
• UHCON
• UHINT
• UHIEN
Moreover, when FRZCLK is set, only the following interrupts may be triggered:
• WAKEUPI
• IDTI
• VBUSTI
• HWUPI
21.6 Speed Control
21.6.1 Device mode

When the USB interface is configured in device mode, the speed selection (Full Speed or Low
Speed) depends on the UDP/UDM pull-up. UDSS register allows to select an internal pull up on
UDM (Low Speed mode) or UDP(Full Speed mode) data lines. UDSS should be configure
before attaching the device.
Figure 21-14. Device mode Speed Selection
UCAP USB
Regulator
DETACH
UDCON.0
LSM
UDCON.2
RPU
RPU
UDP
UDM
258 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
21.6.2 Host mode
When the USB interface is configured in device mode, internal Pull Down resistors are activated
on both UDP UDM lines and the interface detects the type of device connected.
21.7 Memory management

The controller does only support the following memory allocation management:
The reservation of a Pipe or an Endpoint can only be made in the growing order (Pipe/Endpoint
0 to the last Pipe/Endpoint). The firmware shall thus configure them in the same order.
The reservation of a Pipe or an Endpoint “ki” is done when its ALLOC bit is set. Then, the hard-
ware allocates the memory and insert it between the Pipe/Endpoints “ki-1” and “ki+1”. The “ki+1”
Pipe/Endpoint memory “slides” up and its data is lost. Note that the “ki+2” and upper Pipe/End-
point memory does not slide.
Clearing a Pipe enable (PEN) or an Endpoint enable (EPEN) does not clear neither its ALLOC
bit, nor its configuration (EPSIZE/PSIZE, EPBK/PBK). To free its memory, the firmware should
clear ALLOC. Then, the “ki+1” Pipe/Endpoint memory automatically “slides” down. Note that the
“ki+2” and upper Pipe/Endpoint memory does not slide.
The following figure illustrates the allocation and reorganization of the USB memory in a typical
example:
Table 21-1. Allocation and reorganization USB memory flow
Free memory Free memory Free memory Free memory
5 5 5 5
4 Conflict
4 4 Lost memory
3
EPEN=0
4 3 (bigger size)
(ALLOC=1)
2 2 2 2
1 1 1 1
0 0 0 0
EPEN=1
ALLOC=1
Pipe/Endpoints Pipe/Endpoint Free its memory Pipe/Endpoint

activation Disable (ALLOC=0) Activatation
• First, Pipe/Endpoint 0 to Pipe/Endpoint 5 are configured, in the growing order. The memory of
each is reserved in the DPRAM.
• Then, the Pipe/Endpoint 3 is disabled (EPEN=0), but its memory reservation is internally kept
by the controller.
• Its ALLOC bit is cleared: the Pipe/Endpoint 4 “slides” down, but the Pipe/Endpoint 5 does not
“slide”.
• Finally, if the firmware chooses to reconfigure the Pipe/Endpoint 3, with a bigger size. The
controller reserved the memory after the endpoint 2 memory and automatically “slide” the
Pipe/Endpoint 4. The Pipe/Endpoint 5 does not move and a memory conflict appear, in that
259
7593D–AVR–07/06
both Pipe/Endpoint 4 and 5 use a common area. The data of those endpoints are potentially
lost.
Note that:
• the data of Pipe/Endpoint 0 are never lost whatever the activation or deactivation of the
higher Pipe/Endpoint. Its data is lost if it is deactivated.
• Deactivate and reactivate the same Pipe/Endpoint with the same parameters does not lead
to a “slide” of the higher endpoints. For those endpoints, the data are preserved.
• CFGOK is set by hardware even in the case that there is a “conflict” in the memory allocation.
21.8 PAD suspend

The next figures illustrates the pad behaviour:
• In the “idle” mode, the pad is put in low power consumption mode.
• In the “active” mode, the pad is working.
Figure 21-15. Pad behaviour
USBE=1
& DETACH=0
Idle mode & suspend
USBE=0 Active mode

| DETACH=1
| suspend
The SUSPI flag indicated that a suspend state has been detected on the USB bus. This flag
automatically put the USB pad in Idle. The detection of a non-idle event sets the WAKEUPI flag
and wakes-up the USB pad.
260 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Suspend detected = USB pad power down

SUSPI Clear Suspend by software
WAKEUPI Clear Resume by software

Resume = USB pad wake-up
PAD status
Active Power Down Active
Moreover, the pad can also be put in the “idle” mode if the DETACH bit is set. It come back in
the active mode when the DETACH bit is cleared.
21.9 OTG timers customizing

It is possible to refine some OTG timers thanks to the OTGTCON and OTGCON registers
• PAGE=00b: AWaitVrise time-out. [OTG] chapter 6.6.5.1
• VALUE=00bTime-out is set to 20 ms
• VALUE=10bTime-out is set to 70ms
• PAGE=01b: VbBusPulsing. [OTG] chapter 5.3.4
• PAGE=10b: PdTmOutCnt. [OTG] chapter 5.3.2
• PAGE=11b: SRPDetTmOut. [OTG] chapter 5.3.3
• VALUE=00bTime-out is set to 10 us
• VALUE=01bTime-out is set to 100 us
21.10 Plug-in detection

The USB connection is detected by the VBUS pad, thanks to the following architecture:
261
7593D–AVR–07/06
Figure 21-16. Plug-in Detection Input Block Diagram
VDD
VBus_pulsing
RPU
Session_valid
UVBUS Logic VBUS VBUSTI

Va_Vbus_valid
USBSTA.0 USBINT.0
RPU
VBus_discharge
VSS
Pad logic
The control logic of the UVBUS pad outputs 2 signals:

• The “session_valid” signal is active high when the voltage on the UVBUS pad is higher or
equal to 1.4V.
• The “Va_Vbus_valid” signal is active high when the voltage on the UVBUS pad is higher or
equal to 4.4V.
In the Host mode, the VBUS flag follows the next hysteresis rule:
• VBUS is set when the voltage on the UVBUS pad is higher or equal to 4.4 V.
• VBUS is cleared when the voltage on the UVBUS pad is lower than 1.4 V.
In the Peripheral mode, the VBUS flag follows the next rule:
• VBUS is set when the voltage on the UVBUS pad is higher or equal to 1.4 V.
• VBUS is cleared when the voltage on the UVBUS pad is lower than 1.4 V.
The VBUSTI interrupt is triggered at each transition of the VBUS flag.
21.11 ID detection
The ID pin transition is detected thanks to the following architecture:
Figure 21-17. ID Detection Input Block Diagram
VDD
Internal Pull Up
RPU
1
UID
0 ID
UIMOD USBSTA.1
UHWCON.7
UIDE
UHWCON.6
The ID pin can be used to detect the USB mode (Peripheral or Host) or software selected. This
allows the UID pin to be used has general purpose I/O even when USB interface is enable.
When the UID pin is selected, by default, (no A-plug or B-plug), the macro is in the Peripheral
262 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
mode (internal pull-up). The IDTI interrupt is triggered when a A-plug (Host) is plugged or
unplugged. The interrupt is not triggered when a B-plug (Periph) is plugged or unplugged.
ID detection is independant of USB global interface enable.
21.12 Registers description
21.12.1 USB general registers
Bit 7 6 5 4 3 2 1 0
UIMOD UIDE UVCONE UVREGE UHWCON
Read/Write R/W R/W R R/W R R R R/W
• 7 – UIMOD: USB Mode Bit

This bit has no effect when the UIDE bit is set (external UID pin activated). Set to enable the
USB device mode. Clear to enable the USB host mode
• 6 – UIDE: UID pin Enable

Set to enable the USB mode selection (peripheral/host) through the UID pin. Clear to enable the
USB mode selection (peripheral/host) with UIMOD bit register.
UIDE should be modified only when the USB interface is disabled (USBE bit cleared).
• 5 – Reserved
The value read from this bit is always 0. Do not set this bit.
• 4 – UVCONE: UVCON pin Enable

Set to enable the UVCON pin control. Clear to disable the UVCON pin control. This bit should be
set only when the USB interface is enable.
• 3-1 – Reserved
The value read from these bits is always 0. Do not set these bits.
• 0 – UVREGE: USB pad regulator Enable

Set to enable the USB pad regulator. Clear to disable the USB pad regulator.
Bit 7 6 5 4 3 2 1 0
USBE HOST FRZCLK OTGPADE - - IDTE VBUSTE USBCON
• 7 – USBE: USB macro Enable Bit

Set to enable the USB controller. Clear to disable and reset the USB controller, to disable the
USB transceiver and to disable the USB controller clock inputs.
• 6 – HOST: HOST Bit

Set to enable the Host mode. Clear to enable the device mode.
• 5 – FRZCLK: Freeze USB Clock Bit
263
7593D–AVR–07/06
Set to disable the clock inputs (the ”Resume Detection” is still active). This reduces the power
consumption. Clear to enable the clock inputs.
• 4 – OTGPADE: OTG Pad Enable

Set to enable the OTG pad. Clear to disable the OTG pad.
Note that this bit can be set/cleared even if USBE=0 (this allows the VBUS detection even if the
USB macro is disable).
• 3-2 – Reserved
• 1 – IDTE: ID Transition Interrupt Enable Bit

Set this bit to enable the ID Transition interrupt generation. Clear this bit to disable the ID Transi-
tion interrupt generation.
• 0 – VBUSTE: VBUS Transition Interrupt Enable Bit

Set this bit to enable the VBUS Transition interrupt generation.
Clear this bit to disable the VBUS Transition interrupt generation.
Bit 7 6 5 4 3 2 1 0
- - - - SPEED ID VBUS USBSTA
Read/Write R R R R R R R R
• 7-4 - Reserved
• 3 – SPEED: Speed Status Flag

This should be read only when the USB controller operates in host mode, in device mode the
value read from this bit is underterminate.
Set by hardware when the controller is in FULL-SPEED mode. Cleared by hardware when the
controller is in LOW-SPEED mode.
• 2 – Reserved
• 1 – ID: IUD pin Flag

The value read from this bit indicates the state of the UID pin.
• 0 – VBUS: VBus Flag

The value read from this bit indicates the state of the UVBUS pin. This bit can be used in device
mode to monitor the USB bus connection state of the appication. See Section 21.10, page 261
for more details.
Bit 7 6 5 4 3 2 1 0
- - - - - - IDTI VBUSTI USBINT
264 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Bit 7 6 5 4 3 2 1 0
7-2 - Reserved
1 – IDTI: D Transition Interrupt Flag
Set by hardware when a transition (high to low, low to high) has been detected on the UID pin.
Shall be cleared by software.
• 0 – VBUSTI: IVBUS Transition Interrupt Flag

Set by hardware when a transition (high to low, low to high) has been detected on the VBUS
pad.
Shall be cleared by software.
Bit 7 6 5 4 3 2 1 0
- - HNPREQ SRPREQ SRPSEL VBUSHWC VBUSREQ VBUSRQC OTGCON
• 7-6 - Reserved
• 5 – HNPREQ: HNP Request Bit

Set to initiate the HNP when the controller is in the Device mode (B). Set to accept the HNP
when the controller is in the Host mode (A).
Clear otherwise.
• 4 – SRPREQ: SRP Request Bit

Set to initiate the SRP when the controller is in Device mode. Cleared by hardware when the
controller is initiating a SRP.
• 3 – SRPSEL: SRP Selection Bit

Set to choose VBUS pulsing as SRP method.
Clear to choose data line pulsing as SRP method.
• 2 – VBUSHWC: VBus Hardware Control Bit

Set to disable the hardware control over the UVCON pin.
Clear to enable the hardware control over the UVCON pin.
See for more details
• 1 – VBUSREQ: VBUS Request Bit

Set to assert the UVCON pin in order to enable the VBUS power supply generation. This bit
shall be used when the controller is in the Host mode.
Cleared by hardware when VBUSRQC is set.
• 0 – VBUSRQC: VBUS Request Clear Bit
265
7593D–AVR–07/06
Set to deassert the UVCON pin in order to enable the VBUS power supply generation. This bit
shall be used when the controller is in the Host mode.
Cleared by hardware immediately after the set.
Bit 7 6 5 4 3 2 1 0
1 PAGE - - VALUE OTGTCON
Read/Write R R/W R/W R R R/W R/W R/W
• 7 – Reserved
This bit is reserved and always set.
• 6-5 – PAGE: Timer page access Bit

Set/clear to access a special timer register. See Section 21.9, page 261 for more details.
• 4-3 - Reserved
• 2-0 – VALUE: Value Bit

Set to initialize the new value of the timer. See Section 21.9, page 261 for more details.
Bit 7 6 5 4 3 2 1 0
- - STOE HNPERRE ROLEEXE BCERRE VBERRE SRPE OTGIEN
• 7-6 - Reserved
• 5 – STOE: Suspend Time-out Error Interrupt Enable Bit

Set to enable the STOI interrupt. Clear to disable the STOI interrupt.
• 4 – HNPERRE: HNP Error Interrupt Enable Bit

Set to enable the HNPERRI interrupt. Clear to disable the HNPERRI interrupt.
• 3 – ROLEEXE: Role Exchange Interrupt Enable Bit

Set to enable the ROLEEXI interrupt. Clear to disable the ROLEEXI interrupt.
• 2 – BCERRE: B-Connection Error Interrupt Enable Bit

Set to enable the BCERRI interrupt. Clear to disable the BCERRI interrupt.
• 1 – VBERRE: VBus Error Interrupt Enable Bit

Set to enable the VBERRI interrupt. Clear to disable the VBERRI interrupt.
• 0 – SRPE: SRP Interrupt Enable Bit

Set to enable the SRPI interrupt. Clear to disable the SRPI interrupt.
266 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Bit 7 6 5 4 3 2 1 0
- - STOI HNPERRI ROLEEXI BCERRI VBERRI SRPI OTGINT
• 7-6 - Reserved
• 5 – STOI: Suspend Time-out Error Interrupt Flag

Set by hardware when a time-out error (more than 150 ms) has been detected after a suspend.
Shall be cleared by software. See for more details.
• 4 – HNPERRI: HNP Error Interrupt Flag

Set by hardware when an error has been detected during the protocol.
Shall be cleared by software. See for more details.
• 3 – ROLEEXI: Role Exchange Interrupt Flag

Set by hardware when the USB controller has successfully swapped its mode, due to an HNP
negotiation: Host to Device or Device to Host. Shall be cleared by software. See for more
details.
• 2 – BCERRI: B-Connection Error Interrupt Flag

Set by hardware when an error occur during the B-Connection. Shall be cleared by software.
• 1 – VBERRI: V-Bus Error Interrupt Flag

Set by hardware when a drop on VBus has been detected. Shall be cleared by software.
• 0 – SRPI: SRP Interrupt Flag

Set by hardware when a SRP has been detected. Shall be used in the Host mode only Shall be
cleared by software.
21.13 USB Software Operating modes

Depending on the USB operating mode, the software should perform some the following
operations:
Power On the USB interface
• Power-On USB pads regulator
• Wait USB pads regulator ready state
• Configure PLL interface
• Enable PLL
• Check PLL lock
• Enable USB interface
• Configure USB interface (USB speed, Endpoints configuration...)
• Wait for USB VBUS information connection
267
7593D–AVR–07/06
• Attach USB device
Power Off the USB interface

• Detach USB interface
• Disable USB interface
• Disable PLL
• Disable USB pad regulator
Suspending the USB interface

• Clear Suspend Bit
• Set USB suspend clock
• Disable PLL
• Be sure to have interrupts enable to exit sleep mode
• Make the MCU enter sleep mode
Resuming the USB interface

• Enable PLL
• Wait PLL lock
• Clear USB suspend clock
• Clear Resume information
268 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
22. USB Device Operating modes
22.1 Introduction
The USB device controller supports full speed and low speed data transfers. In addition to the
default control endpoint, it provides six other endpoints, which can be configured in control, bulk,
interrupt or isochronous modes:
• Endpoint 0:programmable size FIFO up to 64 bytes, default control endpoint
• Endpoints 1 programmable size FIFO up to 256 bytes in ping-pong mode.
• Endpoints 2 to 6: programmable size FIFO up to 64 bytes in ping-pong mode.
The controller starts in the “idle” mode. In this mode, the pad consumption is reduced to the
minimum.
22.2 Power-on and reset

The next diagram explains the USB device controller main states on power-on:
Figure 22-1. USB device controller states after reset
<any
other
USBE=0 state>
USBE=0 Idle
Reset USBE=1
UID=1
HW
RESET
The reset state of the Device controller is:

• the macro clock is stopped in order to minimize the power consumption (FRZCLK set),
• the USB device controller internal state is reset (all the registers are reset to their default
value. Note that DETACH is set.)
• the endpoint banks are reset
• the D+ or D- pull up are not activated (mode Detach)
The D+ or D- pull-up will be activated as soon as the DETACH bit is cleared and VBUS is
present.
The macro is in the ‘Idle’ state after reset with a minimum power consumption and does not
need to have the PLL activated to enter in this state.
The USB device controller can at any time be reset by clearing USBE (disable USB interface).
22.3 Speed identification on startup

The usb bus reset is managed by the hardware. At the connection, the host makes a reset that
can be:
At the end of the reset process (full speed or low speed mode), the end of reset interrupt
(EORSTI) is generated. Then the CPU can read the SPEED1 bit to know the speed mode of the
device.
269
7593D–AVR–07/06
22.4 Endpoint reset
An endpoint can be reset at any time by setting in the UERST register the bit corresponding to
the endpoint (EPRSTx). This resets:
• the internal state machine on that endpoint,
• the Rx and Tx banks are cleared and their internal pointers are restored,
• the UEINTX, UESTA0X and UESTA1X are restored to their reset value.
The data toggle field remains unchanged.
The other registers remain unchanged.
The endpoint configuration remains active and the endpoint is still enabled.
The endpoint reset may be associated with a clear of the data toggle command (RSTDT bit) as
an answer to the CLEAR_FEATURE USB command.
22.5 USB reset

When an USB reset is detected on the USB line, the next operations are performed by the
controller:
• all the endpoints are disabled, except the default control endpoint,
• the default control endpoint is reset (see Section 22.4, page 270 for more details).
• The data toggle of the default control endpoint is cleared.
22.6 Endpoint selection

Prior to any operation performed by the CPU, the endpoint must first be selected. This is done
by:
• Clearing EPNUMS.
• Setting EPNUM with the endpoint number which will be managed by the CPU.
The CPU can then access to the various endpoint registers and data.
22.7 Endpoint activation

The endpoint is maintained under reset as long as the EPEN bit is not set.
The following flow must be respected in order to activate an endpoint:
270 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 22-2. Endpoint activation flow:
Endpoint
Activation
UENUM Select the endpoint

EPNUM=x
EPEN=1 Activate the endpoint
UECFG0X Configure:
EPDIR - the endpoint direction
EPTYPE - the endpoint type
... - the Not Yet Disable feature
Configure:
UECFG1X - the endpoint size
ALLOC - the bank parametrization
EPSIZE Allocation and reorganization of
EPBK the memory is made on-the-fly
Test the correct endpoint

CFGOK=1 configuration
No
Yes
Endpoint activated ERROR
As long as the endpoint is not correctly configured (CFGOK cleared), the hardware does not
acknowledge the packets sent by the host.
CFGOK is will not be sent if the Endpoint size parameter is bigger than the DPRAM size.
A clear of EPEN acts as an endpoint reset (see Section 22.4, page 270 for more details). It also
performs the next operation:
• The configuration of the endpoint is kept (EPSIZE, EPBK, ALLOC kept)
• It resets the data toggle field.
• The DPRAM memory associated to the endpoint is still reserved.
See Section 21.7, page 259 for more details about the memory allocation/reorganization.
22.8 Address Setup

The USB device address is set up according to the USB protocol:
• the USB device, after power-up, responds at address 0
• the host sends a SETUP command (SET_ADDRESS(addr)),
• the firmware records that address in UADD, but keep ADDEN cleared,
• the USB device sends an IN command of 0 bytes (IN 0 Zero Length Packet),
• then, the firmware can enable the USB device address by setting ADDEN. The only accepted
address by the controller is the one stored in UADD.
ADDEN and UADD shall not be written at the same time.
UADD contains the default address 00h after a power-up or USB reset.
271
7593D–AVR–07/06
ADDEN is cleared by hardware:
• after a power-up reset,
• when an USB reset is received,
• or when the macro is disabled (USBE cleared)
When this bit is cleared, the default device address 00h is used.
22.9 Suspend, Wake-up and Resume

After a period of 3 ms during which the USB line was inactive, the controller switches to the full-
speed mode and triggers (if enabled) the SUSPI (suspend) interrupt. The firmware may then set
the FRZCLK bit.
The CPU can also, depending on software architecture, enter in the idle mode to lower again the
power consumption.
There are two ways to recover from the “Suspend” mode:
• First one is to clear the FRZCLK bit. This is possible if the CPU is not in the Idle mode.
• Second way, if the CPU is “idle”, is to enable the WAKEUPI interrupt (WAKEUPE set). Then,
as soon as an non-idle signal is seen by the controller, the WAKEUPI interrupt is triggered.
The firmware shall then clear the FRZCLK bit to restart the transfer.
There are no relationship between the SUSPI interrupt and the WAKEUPI interrupt: the WAKE-
UPI interrupt is triggered as soon as there are non-idle patterns on the data lines. Thus, the
WAKEUPI interrupt can occurs even if the controller is not in the “suspend” mode.
When the WAKEUPI interrupt is triggered, if the SUSPI interrupt bit was already set, it is cleared
by hardware.
When the SUSPI interrupt is triggered, if the WAKEUPI interrupt bit was already set, it is cleared
by hardware.
22.10 Detach
The reset value of the DETACH bit is 1.
It is possible to re-enumerate a device, simply by setting and clearing the DETACH bit.
• If the USB device controller is in full-speed mode, setting DETACH will disconnect the pull-up
on the D+ or D- pad (depending on full or low speed mode selected). Then, clearing DETACH
will connect the pull-up on the D+ or D- pad.
Figure 22-3. Detach a device in Full-speed:

UVREF UVREF
D+ D+
D- D-
Detach, then
EN=1 Attach EN=1
272 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
22.11 Remote Wake-up
The “Remote Wake-up” (or “upstream resume”) request is the only operation allowed to be sent
by the device on its own initiative. Anyway, to do that, the device should first have received a
DEVICE_REMOTE_WAKEUP request from the host.
• First, the USB controller must have detected the “suspend” state of the line: the remote
wake-up can only be sent after a SUSPI interrupt has been triggered.
• The firmware has then the ability to set RMWKUP to send the “upstream resume” stream.
This will automatically be done by the controller after 5ms of inactivity on the USB line.
• When the controller starts to send the “upstream resume”, the UPRSMI interrupt is triggered
(if enabled). If SUSPI was set, SUSPI is cleared by hardware.
• RMWKUP is cleared by hardware at the end of the “upstream resume”.
• If the controller detects a good “End Of Resume” signal from the host, an EORSMI interrupt
is triggered (if enabled).
22.12 STALL request

For each endpoint, the STALL management is performed using 2 bits:
– STALLRQ (enable stall request)
– STALLRQC (disable stall request)
– STALLEDI (stall sent interrupt)
To send a STALL handshake at the next request, the STALLRQ request bit has to be set. All fol-
lowing requests will be handshak’ed with a STALL until the STALLRQC bit is set.
Setting STALLRQC automatically clears the STALLRQ bit. The STALLRQC bit is also immedi-
ately cleared by hardware after being set by software. Thus, the firmware will never read this bit
as set.
Each time the STALL handshake is sent, the STALLEDI flag is set by the USB controller and the
EPINTx interrupt will be triggered (if enabled).
The incoming packets will be discarded (RXOUTI and RWAL will not be set).
The host will then send a command to reset the STALL: the firmware just has to set the STALL-
RQC bit and to reset the endpoint.
22.12.1 Special consideration for Control Endpoints

A SETUP request is always ACK’ed.
If a STALL request is set for a Control Endpoint and if a SETUP request occurs, the SETUP
request has to be ACK’ed and the STALLRQ request and STALLEDI sent flags are automati-
cally reset (RXSETUPI set, TXIN cleared, STALLED cleared, TXINI cleared...).
This management simplifies the enumeration process management. If a command is not sup-
ported or contains an error, the firmware set the STALL request flag and can return to the main
task, waiting for the next SETUP request.
This function is compliant with the Chapter 8 test from PMTC that send extra status for a
GET_DESCRIPTOR. The firmware sets the STALL request just after receiving the status. All
extra status will be automatically STALL’ed until the next SETUP request.
273
7593D–AVR–07/06
22.12.2 STALL handshake and Retry mechanism
The Retry mechanism has priority over the STALL handshake. A STALL handshake is sent if the
STALLRQ request bit is set and if there is no retry required.
22.13 CONTROL endpoint management

A SETUP request is always ACK’ed. When a new setup packet is received, the RXSTPI inter-
rupt is triggered (if enabled). The RXOUTI interrupt is not triggered.
The FIFOCON and RWAL fields are irrelevant with CONTROL endpoints. The firmware shall
thus never use them on that endpoints. When read, their value is always 0.
CONTROL endpoints are managed by the following bits:
• RXSTPI is set when a new SETUP is received. It shall be cleared by firmware to
acknowledge the packet and to clear the endpoint bank.
• RXOUTI is set when a new OUT data is received. It shall be cleared by firmware to
acknowledge the packet and to clear the endpoint bank.
• TXINI is set when the bank is ready to accept a new IN packet. It shall be cleared by firmware
to send the packet and to clear the endpoint bank.
CONTROL endpoints should not be managed by interrupts, but only by polling the status bits.
22.13.1 Control Write

The next figure shows a control write transaction. During the status stage, the controller will not
necessary send a NAK at the first IN token:
• If the firmware knows the exact number of descriptor bytes that must be read, it can then
anticipate on the status stage and send a ZLP for the next IN token,
• or it can read the bytes and poll NAKINI, which tells that all the bytes have been sent by the
host, and the transaction is now in the status stage.
SETUP DATA STATUS
USB line SETUP OUT OUT IN IN

NAK
RXSTPI HW SW
RXOUTI HW SW HW SW
TXINI SW
274 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
22.13.2 Control Read
The next figure shows a control read transaction. The USB controller has to manage the simulta-
neous write requests from the CPU and the USB host:
SETUP DATA STATUS
USB line SETUP IN IN OUT OUT

NAK
RXSTPI HW SW
RXOUTI HW SW
TXINI SW HW SW
Wr Enable
HOST
Wr Enable
CPU
A NAK handshake is always generated at the first status stage command.

When the controller detect the status stage, all the data writen by the CPU are erased, and
clearing TXINI has no effects.
The firmware checks if the transmission is complete or if the reception is complete.
The OUT retry is always ack’ed. This reception:
- set the RXOUTI flag (received OUT data)
- set the TXINI flag (data sent, ready to accept new data)
software algorithm:
set transmit ready

wait (transmit complete OR Receive complete)
if receive complete, clear flag and return
if transmit complete, continue
Once the OUT status stage has been received, the USB controller waits for a SETUP request.
The SETUP request have priority over any other request and has to be ACK’ed. This means that
any other flag should be cleared and the fifo reset when a SETUP is received.
WARNING: the byte counter is reset when the OUT Zero Length Packet is received. The firm-
ware has to take care of this.
22.14 OUT endpoint management

OUT packets are sent by the host. All the data can be read by the CPU, which acknowledges or
not the bank when it is empty.
22.14.1 Overview
The Endpoint must be configured first.
Each time the current bank is full, the RXOUTI and the FIFOCON bits are set. This triggers an
interrupt if the RXOUTE bit is set. The firmware can acknowledge the USB interrupt by clearing
the RXOUTI bit. The Firmware read the data and clear the FIFOCON bit in order to free the cur-
rent bank. If the OUT Endpoint is composed of multiple banks, clearing the FIFOCON bit will
275
7593D–AVR–07/06
switch to the next bank. The RXOUTI and FIFOCON bits are then updated by hardware in
accordance with the status of the new bank.
RXOUTI shall always be cleared before clearing FIFOCON.
The RWAL bit always reflects the state of the current bank. This bit is set if the firmware can
read data from the bank, and cleared by hardware when the bank is empty.
Example with 1 OUT data bank
DATA NAK DATA

OUT ACK OUT ACK
(to bank 0) (to bank 0)
HW HW
RXOUTI SW SW
FIFOCON read data from CPU SW

BANK 0 read data from CPU
BANK 0
Example with 2 OUT data banks
DATA DATA
OUT ACK OUT ACK
HW
HW
RXOUTI SW SW
FIFOCON read data from CPU SW

BANK 0 read data from CPU
BANK 1
22.14.2 Detailed description
22.14.2.1
The data are read by the CPU, following the next flow:
• When the bank is filled by the host, an endpoint interrupt (EPINTx) is triggered, if enabled
(RXOUTE set) and RXOUTI is set. The CPU can also poll RXOUTI or FIFOCON, depending
on the software architecture,
• The CPU acknowledges the interrupt by clearing RXOUTI,
• The CPU can read the number of byte (N) in the current bank (N=BYCT),
• The CPU can read the data from the current bank (“N” read of UEDATX),
• The CPU can free the bank by clearing FIFOCON when all the data is read, that is:
• after “N” read of UEDATX,
• as soon as RWAL is cleared by hardware.
If the endpoint uses 2 banks, the second one can be filled by the HOST while the current one is
being read by the CPU. Then, when the CPU clear FIFOCON, the next bank may be already
ready and RXOUTI is set immediately.
276 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
22.15 IN endpoint management
IN packets are sent by the USB device controller, upon an IN request from the host. All the data
can be written by the CPU, which acknowledge or not the bank when it is full.Overview
The Endpoint must be configured first.
The TXINI bit is set by hardware when the current bank becomes free. This triggers an interrupt
if the TXINE bit is set. The FIFOCON bit is set at the same time. The CPU writes into the FIFO
and clears the FIFOCON bit to allow the USB controller to send the data. If the IN Endpoint is
composed of multiple banks, this also switches to the next data bank. The TXINI and FIFOCON
bits are automatically updated by hardware regarding the status of the next bank.
TXINI shall always be cleared before clearing FIFOCON.
The RWAL bit always reflects the state of the current bank. This bit is set if the firmware can
write data to the bank, and cleared by hardware when the bank is full.
Example with 1 IN data bank
NAK DATA
IN ACK IN
(bank 0)
HW
TXINI SW SW
FIFOCON write data from CPU SW SW

BANK 0 write data from CPU
BANK 0
Example with 2 IN data banks
DATA DATA
IN ACK IN ACK
(bank 0) (bank 1)
HW
TXINI SW SW SW
FIFOCON write data from CPU SW write data from CPU SW write data from CPU
BANK 0 BANK 1 BANK0
22.15.1 Detailed description
The data are written by the CPU, following the next flow:
• When the bank is empty, an endpoint interrupt (EPINTx) is triggered, if enabled (TXINE set)
and TXINI is set. The CPU can also poll TXINI or FIFOCON, depending the software
architecture choice,
• The CPU acknowledges the interrupt by clearing TXINI,
• The CPU can write the data into the current bank (write in UEDATX),
• The CPU can free the bank by clearing FIFOCON when all the data are written, that is:
277
7593D–AVR–07/06
• after “N” write into UEDATX
• as soon as RWAL is cleared by hardware.
If the endpoint uses 2 banks, the second one can be read by the HOST while the current is
being written by the CPU. Then, when the CPU clears FIFOCON, the next bank may be already
ready (free) and TXINI is set immediately.
22.15.1.1 Abort
An “abort” stage can be produced by the host in some situations:
• In a control transaction: ZLP data OUT received during a IN stage,
• In an isochronous IN transaction: ZLP data OUT received on the OUT endpoint during a IN
stage on the IN endpoint
• ...
The KILLBK bit is used to kill the last “written” bank. The best way to manage this abort is to per-
form the following operations:
Table 22-1. Abort flow
Endpoint
Abort
Clear Disable the TXINI interrupt.

UEIENX.
TXINE
Abort is based on the fact

NBUSYBK No that no banks are busy,
=0 meaning that nothing has to
be sent.
Yes
Endpoint Kill the last written
KILLBK=1 bank.
reset
Wait for the end of the

Yes KILLBK=1 procedure.
No
Abort done
22.16 Isochronous mode

For Isochronous IN endpoints, it is possible to automatically switch the banks on each start of
frame (SOF). This is done by setting ISOSW. The CPU has to fill the bank of the endpoint; the
bank switching will be automatic as soon as a SOF is seen by the hardware.
A clear of FIFOCON does not have any effects in this mode.
In the case that a SOF is missing (noise on USB pad, ...), the controller will automatically build
internally a “pseudo” start of frame and the bank switching is made. The SOFI interrupt is trig-
gered and the frame number FNUM10:0 is increased.
22.16.1 Underflow
An underflow can occur during IN stage if the host attempts to read a bank which is empty. In
this situation, the UNDERFI interrupt is triggered.
278 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
An underflow can also occur during OUT stage if the host send a packet while the banks are
already full. Typically, he CPU is not fast enough. The packet is lost.
It is not possible to have underflow error during OUT stage, in the CPU side, since the CPU
should read only if the bank is ready to give data (RXOUTI=1 or RWAL=1)
22.16.2 CRC Error

A CRC error can occur during OUT stage if the USB controller detects a bad received packet. In
this situation, the STALLEDI interrupt is triggered. This does not prevent the RXOUTI interrupt
from being triggered.
22.17 Overflow
In Control, Isochronous, Bulk or Interrupt Endpoint, an overflow can occur during OUT stage, if
the host attempts to write in a bank that is too small for the packet. In this situation, the OVERFI
interrupt is triggered (if enabled). The packet is hacknowledged and the RXOUTI interrupt is also
triggered (if enabled). The bank is filled with the first bytes of the packet.
It is not possible to have overflow error during IN stage, in the CPU side, since the CPU should
write only if the bank is ready to access data (TXINI=1 or RWAL=1).
22.18 Interrupts
The next figure shows all the interrupts sources:
Figure 22-4. USB Device Controller Interrupt System

UPRSMI
UDINT.6
UPRSME
UDIEN.6
EORSMI
UDINT.5
EORSME
UDIEN.5
WAKEUPI
UDINT.4
WAKEUPE USB Device
UDIEN.4 Interrupt
EORSTI
UDINT.3
EORSTE
UDIEN.3
SOFI
UDINT.2
SOFE
UDIEN.2
SUSPI
UDINT.0
SUSPE
UDIEN.0
There are 2 kind of interrupts: processing (i.e. their generation are part of the normal processing)
and exception (errors).
• VBUS plug-in detection (insert, remove)(VBUSTI)
• Upstream resume(UPRSMI)
• End of resume(EORSMI)
• Wake up(WAKEUPI)
• End of reset (Speed Initialization)(EORSTI)
279
7593D–AVR–07/06
• Start of frame(SOFI, if FNCERR=0)
• Suspend detected after 3 ms of inactivity(SUSPI)
• CRC error in frame number of SOF(SOFI, FNCERR=1)
Figure 22-5. USB Device Controller Endpoint Interrupt System

Endpoint 6
Endpoint 5
Endpoint 4
Endpoint 3
Endpoint 2
Endpoint 1
Endpoint 0
OVERFI
UESTAX.6
UNDERFI
FLERRE
UESTAX.5
UEIENX.7
NAKINI
UEINTX.6
NAKINE
UEIENX.6
NAKOUTI
UEINTX.4
TXSTPE
Endpoint Interrupt
UEIENX.4
RXSTPI EPINT
UEINTX.3 UEINT.X
TXOUTE
UEIENX.3
RXOUTI
UEINTX.2
RXOUTE
UEIENX.2
STALLEDI
UEINTX.1
STALLEDE
UEIENX.1
TXINI
UEINTX.0
TXINE
UEIENX.0

• Ready to accept IN data(EPINTx, TXINI=1)
• Received OUT data(EPINTx, RXOUTI=1)
• Received SETUP(EPINTx, RXSTPI=1)
• Stalled packet(EPINTx, STALLEDI=1)
• CRC error on OUT in isochronous mode(EPINTx, STALLEDI=1)
• Overflow in isochronous mode(EPINTx, OVERFI=1)
• Underflow in isochronous mode(EPINTx, UNDERFI=1)
• NAK IN sent(EPINTx, NAKINI=1)
• NAK OUT sent(EPINTx, NAKOUTI=1)
280 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
22.19 Registers
22.19.1 USB device general registers
Bit 7 6 5 4 3 2 1 0
- - - - - LSM RMWKUP DETACH UDCON
• 7-3 - Reserved
• 2 - LSM - USB Device Low Speed Mode Selection

When configured USB is configured in device mode, this bit allows to select the USB the USB
Low Speed or Full Speed Mod.
Clear to select full speed mode (D+ internal pull-up will be activate with the ATTACH bit will be
set) .
Set to select low speed mode (D- internal pull-up will be activate with the ATTACH bit will be
set). This bit has no effect when the USB interface is configured in HOST mode.
• 1- RMWKUP - Remote Wake-up Bit

Set to send an “upstream-resume” to the host for a remote wake-up.
Cleared by hardware. Clearing by software has no effect.
See Section 22.11, page 273 for more details.
• 0 - DETACH - Detach Bit

Set to physically detach de device (disconnect internal pull-up on D+ or D-).
Clear to reconnect the device. See Section 22.10, page 272 for more details.
Bit 7 6 5 4 3 2 1 0
- UPRSMI EORSMI WAKEUPI EORSTI SOFI - SUSPI UDINT
Read/Write
• 7 - Reserved
The value read from this bits is always 0. Do not set this bit.
• 6 - UPRSMI - Upstream Resume Interrupt Flag

Set by hardware when the USB controller is sending a resume signal called “Upstream
Resume”. This triggers an USB interrupt if UPRSME is set.
Shall be cleared by software (USB clocks must be enabled before). Setting by software has no
effect.
• 5 - EORSMI - End Of Resume Interrupt Flag

Set by hardware when the USB controller detects a good “End Of Resume” signal initiated by
the host. This triggers an USB interrupt if EORSME is set.
281
7593D–AVR–07/06
Shall be cleared by software. Setting by software has no effect.
• 4 - WAKEUPI - Wake-up CPU Interrupt Flag

Set by hardware when the USB controller is re-activated by a filtered non-idle signal from the
lines (not by an upstream resume). This triggers an interrupt if WAKEUPE is set.
Shall be cleared by software (USB clock inputs must be enabled before). Setting by software
has no effect.
• 3 - EORSTI - End Of Reset Interrupt Flag

Set by hardware when an “End Of Reset” has been detected by the USB controller. This triggers
an USB interrupt if EORSTE is set.
• 2 - SOFI - Start Of Frame Interrupt Flag

Set by hardware when an USB “Start Of Frame” PID (SOF) has been detected (every 1 ms).
This triggers an USB interrupt if SOFE is set..
• 1 - Reserved
The value read from this bits is always 0. Do not set this bit
• 0 - SUSPI - Suspend Interrupt Flag

Set by hardware when an USB “Suspend” ‘idle bus for 3 frame periods: a J state for 3 ms) is detected. This
triggers an USB interrupt if SUSPE is set.
The interrupt bits are set even if their corresponding ‘Enable’ bits is not set.
Bit 7 6 5 4 3 2 1 0
- UPRSME EORSME WAKEUPE EORSTE SOFE - SUSPE UDIEN
Read/Write
• 7 - Reserved
• 6 - UPRSME - Upstream Resume Interrupt Enable Bit

Set to enable the UPRSMI interrupt.
Clear to disable the UPRSMI interrupt.
• 5 - EORSME - End Of Resume Interrupt Enable Bit

Set to enable the EORSMI interrupt.
Clear to disable the EORSMI interrupt.
• 4 - WAKEUPE - Wake-up CPU Interrupt Enable Bit

Set to enable the WAKEUPI interrupt.
282 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Clear to disable the WAKEUPI interrupt.
• 3 - EORSTE - End Of Reset Interrupt Enable Bit

Set to enable the EORSTI interrupt. This bit is set after a reset.
Clear to disable the EORSTI interrupt.
• 2 - SOFE - Start Of Frame Interrupt Enable Bit

Set to enable the SOFI interrupt.
Clear to disable the SOFI interrupt.
• 1 - Reserved
The value read from this bits is always 0. Do not set this bit
• 0 - SUSPE - Suspend Interrupt Enable Bit

Set to enable the SUSPI interrupt.
Clear to disable the SUSPI interrupt.
Bit 7 6 5 4 3 2 1 0
ADDEN UADD6:0 UDADDR
Read/Write W R/W R/W R/W R/W R/W R/W R/W
Initial Val- 0 0 0 0 0 0 0 0
ue
• 7 - ADDEN - Address Enable Bit

Set to activate the UADD (USB address).
Cleared by hardware. Clearing by software has no effect.
• 6-0 - UADD6:0 - USB Address Bits

Load by software to configure the device address.
.
Bit 7 6 5 4 3 2 1 0
- - - - - FNUM10:8 UDFNUMH
• 7-3 - Reserved
• 2-0 - FNUM10:8 - Frame Number Upper Flag

Set by hardware. These bits are the 3 MSB of the 11-bits Frame Number information. They are
provided in the last received SOF packet. FNUM is updated if a corrupted SOF is received.
283
7593D–AVR–07/06
Bit 7 6 5 4 3 2 1 0
FNUM7:0 UDFNUML
• Frame Number Lower Flag

Set by hardware. These bits are the 8 LSB of the 11-bits Frame Number information.
Bit 7 6 5 4 3 2 1 0
- - - FNCERR - - - - UDMFN
Read/W R
rite
Initial 0 0 0 0 0 0 0 0
Value
• 7-5 - Reserved
• 4 - FNCERR -Frame Number CRC Error Flag

Set by hardware when a corrupted Frame Number in start of frame packet is received.
This bit and the SOFI interrupt are updated at the same time.
• 3-0 - Reserved
22.19.2 USB device endpoint registers
Bit 7 6 5 4 3 2 1 0
- - - - - EPNUM2:0 UENUM
• 7-3 - Reserved
• 2-0 - EPNUM2:0 Endpoint Number Bits

Load by software to select the number of the endpoint which shall be accessed by the CPU. See
Section 22.6, page 270 for more details.
EPNUM = 111b is forbidden.
Bit 7 6 5 4 3 2 1 0
- EPRST6 EPRST5 EPRST4 EPRST3 EPRST2 EPRST1 EPRST0 UERST
Read/Write R R/W R/W R/W R/W R/W R/W R/W
284 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
• 7 - Reserved
• 6-0 - EPRST6:0 - Endpoint FIFO Reset Bits

Set to reset the selected endpoint FIFO prior to any other operation, upon hardware reset or
when an USB bus reset has been received. See Section 22.4, page 270 for more information
Then, cleared by software to complete the reset operation and start using the endpoint.
Bit 7 6 5 4 3 2 1 0
- - STALLRQ STALLRQC RSTDT - - EPEN UECONX
Read/Write R R W W W R R R/W
• 7-6 - Reserved
• 5 - STALLRQ - STALL Request Handshake Bit

Set to request a STALL answer to the host for the next handshake.
Cleared by hardware when a new SETUP is received. Clearing by software has no effect.
• 4 - STALLRQC - STALL Request Clear Handshake Bit

Set to disable the STALL handshake mechanism.
Cleared by hardware immediately after the set. Clearing by software has no effect.
3
• RSTDT - Reset Data Toggle Bit

Set to automatically clear the data toggle sequence:
For OUT endpoint: the next received packet will have the data toggle 0.
For IN endpoint: the next packet to be sent will have the data toggle 0.
Cleared by hardware instantaneously. The firmware does not have to wait that the bit is cleared.
Clearing by software has no effect.
• 2 - Reserved
• 1 - Reserved
• 0 - EPEN - Endpoint Enable Bit

Set to enable the endpoint according to the device configuration. Endpoint 0 shall always be
enabled after a hardware or USB reset and participate in the device configuration.
Clear this bit to disable the endpoint. See Section 22.7, page 270 for more details.
285
7593D–AVR–07/06
Bit 7 6 5 4 3 2 1 0
EPTYPE1:0 - - - - - EPDIR UECFG0X
Read/Write R/W R/W R R R R R R/W
• 7-6 - EPTYPE1:0 - Endpoint Type Bits

Set this bit according to the endpoint configuration:
00b: Control10b: Bulk
01b: Isochronous11b: Interrupt
• 5-4 - Reserved
• 3-2 - Reserved for test purpose

• 1 - Reserved
• 0 - EPDIR - Endpoint Direction Bit

Set to configure an IN direction for bulk, interrupt or isochronous endpoints.
Clear to configure an OUT direction for bulk, interrupt, isochronous or control endpoints.
Bit 7 6 5 4 3 2 1 0
- EPSIZE2:0 EPBK1:0 ALLOC - UECFG1X
Read/Write R R/W R/W R/W R/W R/W R/W R
• 7 - Reserved
• 6-4 - EPSIZE2:0 - Endpoint Size Bits

Set this bit according to the endpoint size:
000b: 8 bytes100b: 128 bytes
011b: 64 bytes111b: Reserved. Do not use this configuration.
• 3-2 - EPBK1:0 - Endpoint Bank Bits

Set this field according to the endpoint size:
00b: One bank
01b: Double bank
1xb: Reserved. Do not use this configuration.
286 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
• 1 - ALLOC - Endpoint Allocation Bit
Set this bit to allocate the endpoint memory.
Clear to free the endpoint memory.
• 0 - Reserved
Bit 7 6 5 4 3 2 1 0
CFGOK OVERFI UNDERFI ZLPSEEN DTSEQ1:0 NBUSYBK1:0 UESTA0X
Read/Write R R/W R/W R/W R R R R
• 7 - CFGOK - Configuration Status Flag

Set by hardware when the endpoint X size parameter (EPSIZE) and the bank parametrization
(EPBK) are correct compared to the max FIFO capacity and the max number of allowed bank.
This bit is updated when the bit ALLOC is set.
If this bit is cleared, the user should reprogram the UECFG1X register with correct EPSIZE and
EPBK values.
• 6 - OVERFI - Overflow Error Interrupt Flag

Set by hardware when an overflow error occurs in an isochronous endpoint. An interrupt
(EPINTx) is triggered (if enabled).
• 5 - UNDERFI - Flow Error Interrupt Flag

Set by hardware when an underflow error occurs in an isochronous endpoint. An interrupt
(EPINTx) is triggered (if enabled).
• 4 - ZLPSEEN - Zero Length Packet Seen (bit / Flag)

Set by hardware, as soon as a ZLP has been filtered during a transfer.
Shall be cleared by the software. Setting by software has no effect.
• 3-2 - DTSEQ1:0 - Data Toggle Sequencing Flag

Set by hardware to indicate the PID data of the current bank:
00b Data0
01b Data1
1xb Reserved.
For OUT transfer, this value indicates the last data toggle received on the current bank.
287
7593D–AVR–07/06
For IN transfer, it indicates the Toggle that will be used for the next packet to be sent. This is not
relative to the current bank.
• 1-0 - NBUSYBK1:0 - Busy Bank Flag

Set by hardware to indicate the number of busy bank.
For IN endpoint, it indicates the number of busy bank(s), filled by the user, ready for IN transfer.
For OUT endpoint, it indicates the number of busy bank(s) filled by OUT transaction from the
host.
00b All banks are free
01b 1 busy bank
10b 2 busy banks
11b Reserved.
Bit 7 6 5 4 3 2 1 0
- - - - - CTRLDIR CURRBK1:0 UESTA1X
• 7-3 - Reserved
• 2 - CTRLDIR - Control Direction (Flag, and bit for debug purpose)

Set by hardware after a SETUP packet, and gives the direction of the following packet:
- 1 for IN endpoint
- 0 for OUT endpoint.
Can not be set or cleared by software.
• 1-0 - CURRBK1:0 - Current Bank (all endpoints except Control endpoint) Flag
Set by hardware to indicate the number of the current bank:
00b Bank0
01b Bank1
1xb Reserved.
Can not be set or cleared by software.
Bit 7 6 5 4 3 2 1 0
FIFOCON NAKINI RWAL NAKOUTI RXSTPI RXOUTI STALLEDI TXINI UEINTX
• 7 - FIFOCON - FIFO Control Bit

For OUT and SETUP Endpoint:
288 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Set by hardware when a new OUT message is stored in the current bank, at the same time than
RXOUT or RXSTP.
Clear to free the current bank and to switch to the following bank. Setting by software has no
effect.
For IN Endpoint:
Set by hardware when the current bank is free, at the same time than TXIN.
Clear to send the FIFO data and to switch the bank. Setting by software has no effect.
• 6 - NAKINI - NAK IN Received Interrupt Flag

Set by hardware when a NAK handshake has been sent in response of a IN request from the
host. This triggers an USB interrupt if NAKINE is sent.
• 5 - RWAL - Read/Write Allowed Flag

Set by hardware to signal:
- for an IN endpoint: the current bank is not full i.e. the firmware can push data into the FIFO,
- for an OUT endpoint: the current bank is not empty, i.e. the firmware can read data from the
FIFO.
The bit is never set if STALLRQ is set, or in case of error.
Cleared by hardware otherwise.
This bit shall not be used for the control endpoint.
• 4 - NAKOUTI - NAK OUT Received Interrupt Flag

Set by hardware when a NAK handshake has been sent in response of a OUT/PING request
from the host. This triggers an USB interrupt if NAKOUTE is sent.
• 3 - RXSTPI - Received SETUP Interrupt Flag

Set by hardware to signal that the current bank contains a new valid SETUP packet. An inter-
rupt (EPINTx) is triggered (if enabled).
Shall be cleared by software to handshake the interrupt. Setting by software has no effect.
This bit is inactive (cleared) if the endpoint is an IN endpoint.
• 2 - RXOUTI / KILLBK - Received OUT Data Interrupt Flag

Set by hardware to signal that the current bank contains a new packet. An interrupt (EPINTx) is
triggered (if enabled).
Kill Bank IN Bit
Set this bit to kill the last written bank.
Cleared by hardware when the bank is killed. Clearing by software has no effect.
See page 278 for more details on the Abort.
289
7593D–AVR–07/06
• 1 - STALLEDI - STALLEDI Interrupt Flag
Set by hardware to signal that a STALL handshake has been sent, or that a CRC error has been
detected in a OUT isochronous endpoint.
• 0 - TXINI - Transmitter Ready Interrupt Flag

Set by hardware to signal that the current bank is free and can be filled. An interrupt (EPINTx) is
triggered (if enabled).
This bit is inactive (cleared) if the endpoint is an OUT endpoint.
Bit 7 6 5 4 3 2 1 0
FLERRE NAKINE - NAKOUTE RXSTPE RXOUTE STALLEDE TXINE UEIENX
• 7 - FLERRE - Flow Error Interrupt Enable Flag

Set to enable an endpoint interrupt (EPINTx) when OVERFI or UNDERFI are sent.
Clear to disable an endpoint interrupt (EPINTx) when OVERFI or UNDERFI are sent.
• 6 - NAKINE - NAK IN Interrupt Enable Bit

Set to enable an endpoint interrupt (EPINTx) when NAKINI is set.
Clear to disable an endpoint interrupt (EPINTx) when NAKINI is set.
• 5 - Reserved
• 4 - NAKOUTE - NAK OUT Interrupt Enable Bit

Set to enable an endpoint interrupt (EPINTx) when NAKOUTI is set.
Clear to disable an endpoint interrupt (EPINTx) when NAKOUTI is set.
• 3 - RXSTPE - Received SETUP Interrupt Enable Flag

Set to enable an endpoint interrupt (EPINTx) when RXSTPI is sent.
Clear to disable an endpoint interrupt (EPINTx) when RXSTPI is sent.
• 2 - RXOUTE - Received OUT Data Interrupt Enable Flag

Set to enable an endpoint interrupt (EPINTx) when RXOUTI is sent.
Clear to disable an endpoint interrupt (EPINTx) when RXOUTI is sent.
• 1 - STALLEDE - Stalled Interrupt Enable Flag

Set to enable an endpoint interrupt (EPINTx) when STALLEDI is sent.
Clear to disable an endpoint interrupt (EPINTx) when STALLEDI is sent.
• 0 - TXINE - Transmitter Ready Interrupt Enable Flag

Set to enable an endpoint interrupt (EPINTx) when TXINI is sent.
290 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Clear to disable an endpoint interrupt (EPINTx) when TXINI is sent.
Bit 7 6 5 4 3 2 1 0
DAT D7 DAT D6 DAT D5 DAT D4 DAT D3 DAT D2 DAT D1 DAT D0 UEDATX
• 7-0 - DAT7:0 -Data Bits

Set by the software to read/write a byte from/to the endpoint FIFO selected by EPNUM.
Bit 7 6 5 4 3 2 1 0
- - - - - BYCT D10 BYCT D9 BYCT D8 UEBCHX
• 7-3 - Reserved
• 2-0 - BYCT10:8 - Byte count (high) Bits

Set by hardware. This field is the MSB of the byte count of the FIFO endpoint. The LSB part is
provided by the UEBCLX register.
Bit 7 6 5 4 3 2 1 0
BYCT D7 BYCT D6 BYCT D5 BYCT D4 BYCT D3 BYCT D2 BYCT D1 BYCT D0 UEBCLX
• 7-0 - BYCT7:0 - Byte Count (low) Bits

Set by the hardware. BYCT10:0 is:
- (for IN endpoint) increased after each writing into the endpoint and decremented after each
byte sent,
- (for OUT endpoint) increased after each byte sent by the host, and decremented after each
byte read by the software.
Bit 7 6 5 4 3 2 1 0
- EPINT D6 EPINT D5 EPINT D4 EPINT D3 EPINT D2 EPINT D1 EPINT D0 UEINT
• 7 - Reserved
• 6-0 - EPINT6:0 - Endpoint Interrupts Bits

Set by hardware when an interrupt is triggered by the UEINTX register and if the corresponding
endpoint interrupt enable bit is set.
291
7593D–AVR–07/06
Cleared by hardware when the interrupt source is served.
292 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
23. USB Host Operating Modes
23.1 Pipe description

For the USB Host controller, the term of Pipe is used instead of Endpoint for the USB Device
controller. A Host Pipe corresponds to a Device Endpoint, as described in the USB specification:
Figure 23-1. Pipes and Endpoints in a USB system
In the USB Host controller, a Pipe will be associated to a Device Endpoint, considering the
Device Configuration Descriptors.
23.2 Detach
The reset value of the DETACH bit is 1. Thus, the firmware has the responsibility of clearing this
bit before switching to the Host mode (HOST set).
23.3 Power-on and Reset

The next diagram explains the USB host controller main states on power-on:
Figure 23-2. USB host controller states after reset

Device <any
Clock stopped other
Macro off disconnection
state>
Host
Idle
Device
connection
Device
disconnection
Host
Ready
SOFE=0
SOFE=1 Host
Suspend
293
7593D–AVR–07/06
USB host controller state after an hardware reset is ‘Reset’. When the USB controller is enabled
and the USB Host controller is selected, the USB controller is in ‘Idle’ state. In this state, the
USB Host controller waits for the Device connection, with a minimum power consumption.
The USB Pad should be in Idle mode. The macro does not need to have the PLL activated to
enter in ‘Host Ready’ state.
The Host controller enters in Suspend state when the USB bus is in Suspend state, i.e. when the
Host controller doesn’t generate the Start of Frame. In this state, the USB consumption is mini-
mum. The Host controller exits to the Suspend state when starting to generate the SOF over the
USB line.
23.4 Device Detection

A Device is detected by the USB controller when the USB bus if different from D+ and D- low. In
other words, when the USB Host Controller detects the Device pull-up on the D+ line. To enable
this detection, the Host Controller has to provide the Vbus power supply to the Device.
The Device Disconnection is detected by the USB Host controller when the USB Idle correspond
to D+ and D- low on the USB line.
23.5 Pipe Selection

Prior to any operation performed by the CPU, the Pipe must first be selected. This is done by:
• Clearing PNUMS.
• Setting PNUM with the Pipe number which will be managed by the CPU.
The CPU can then access to the various Pipe registers and data.
23.6 Pipe Configuration

The following flow must be respected in order to activate a Pipe:
294 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 23-3. Pipe activation flow:
Pipe
Activ ation
UPCONX Enablethepipe
PENABLE=1
UPCFG0X SelectthePipetype:
PTYPE * Type(Control,Bulk,Interrupt)
PTOKEN * Token(IN,OUT ,SET UP)
PEPNUM * Endpointnumber
UPCFG1X ConfigurethePipememory:
PSIZE * Pipesize
PBK * Numberofbanks
CFGMEM
CFGOK=1 No
Y es
ERROR
UPCFG2X
INT FRQ
(interruptonly) Configurethepollinginterval
forInterruptpipe
Pipeactiv ated
and f reezed
Once the Pipe is activated (EPEN set) and, the hardware is ready to send requests to the
Device.
When configured (CFGOK = 1), only the Pipe Token (PTOKEN) and the polling interval for Inter-
rupt pipe can be modified.
A Control type pipe supports only 1 bank. Any other value will lead to a configuration error
(CFGOK = 0).
A clear of PEN will reset the configuration of the Pipe. All the corresponding Pipe registers are
reset to there reset values. Please refers to the Memory Management chapter for more details.
Note: The firmware has to configure the Default Control Pipe with the following parameters:
• Type: Control
• Token: SETUP
• Data bank: 1
• Size: 64 Bytes
The firmware asks for 8 bytes of the Device Descriptor sending a GET_DESCRIPTOR request.
These bytes contains the MaxPacketSize of the Device default control endpoint and the firm-
ware re-configures the size of the Default Control Pipe with this size parameter.
295
7593D–AVR–07/06
23.7 USB Reset
The USB controller sends a USB Reset when the firmware set the RESET bit. The RSTI bit is
set by hardware when the USB Reset has been sent. This triggers an interrupt if the RSTE has
been set.
When a USB Reset has been sent, all the Pipe configuration and the memory allocation are
reset. The General Host interrupt enable register is left unchanged.
If the bus was previously in suspend mode (SOFEN = 0), the USB controller automatically
switches to the resume mode (HWUPI is set) and the SOFEN bit is set by hardware in order to
generate SOF immediately after the USB Reset.
23.8 Address Setup

Once the Device has answer to the first Host requests with the default address (0), the Host
assigns a new address to the device. The Host controller has to send a USB reset to the device
and perform a SET ADDRESS control request, with the new address to be used by the Device.
This control request ended, the firmware write the new address into the UHADDR register. All
following requests, on every Pipes, will be performed using this new address.
When the Host controller send a USB reset, the UHADDR register is reset by hardware and the
following Host requests will be performed using the default address (0).
23.9 Remote Wake-Up detection

The Host Controller enters in Suspend mode when clearing the SOFEN bit. No more Start Of
Frame is sent on the USB bus and the USB Device enters in Suspend mode 3ms later.
The Device awakes the Host Controller by sending an Upstream Resume (Remote Wake-Up
feature). The Host Controller detects a non-idle state on the USB bus and set the HWUPI bit. If
the non-Idle correspond to an Upstream Resume (K state), the RXRSMI bit is set by hardware.
The firmware has to generate a downstream resume within 1ms and for at least 20ms by setting
the RESUME bit.
Once the downstream Resume has been generated, the SOFEN bit is automatically set by hard-
ware in order to generate SOF immediately after the USB resume.
Host
Ready
SOFE=0
SOFE=1 Host
or HWUP=1 Suspend
23.10 USB Pipe Reset

The firmware can reset a Pipe using the pipe reset register. The configuration of the pipe and
the data toggle remains unchanged. Only the bank management and the status bits are reset to
their initial values.
To completely reset a Pipe, the firmware has to disable and then enable the pipe.
23.11 Pipe Data Access

In order to read or to write into the Pipe Fifo, the CPU selects the Pipe number with the UPNUM
register and performs read or write action on the UPDATX register.
296 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
23.12 Control Pipe management
A Control transaction is composed of 3 phases:
• SETUP
• Data (IN or OUT)
• Status (OUT or IN)
The firmware has to change the Token for each phase.
The initial data toggle is set for the corresponding token (ONLY for Control Pipe):
• SETUP: Data0
• OUT: Data1
• IN: Data1 (expected data toggle)
23.13 OUT Pipe management

The Pipe must be configured and not frozen first.
Note: if the firmware decides to switch to suspend mode (clear SOFEN) even if a bank is ready
to be sent, the USB controller will automatically exit from Suspend mode and the bank will be
sent.
The TXOUT bit is set by hardware when the current bank becomes free. This triggers an inter-
rupt if the TXOUTE bit is set. The FIFOCON bit is set at the same time. The CPU writes into the
FIFO and clears the FIFOCON bit to allow the USB controller to send the data.
297
7593D–AVR–07/06
If the OUT Pipe is composed of multiple banks, this also switches to the next data bank. The
TXOUT and FIFOCON bits are automatically updated by hardware regarding the status of the
next bank.
Example with 1 OUT data bank
DATA
OUT ACK OUT
(bank 0)
HW
TXOUT SW SW
FIFOCON SW SW
write data from CPU write data from CPU
BANK 0 BANK 0
DATA DATA
OUT ACK OUT ACK
(bank 0) (bank 1)
HW
TXOUT SW SW SW
FIFOCON SW SW
write data from CPU write data from CPU write data from CPU
BANK 0 BANK 1 BANK0
DATA DATA
OUT ACK OUT ACK
(bank 0) (bank 1)
HW
TXOUT SW SW SW
FIFOCON SW
write data from CPU write data from CPU SW write data from CPU
BANK 0 BANK 1 BANK0
23.14 IN Pipe management

The Pipe must be configured first.
When the Host requires data from the device, the firmware has to determine first the IN mode to
use using the INMODE bit:
• INMODE = 0. The INRQX register is taken in account. The Host controller will perform
(INRQX+1) IN requests on the selected Pipe before freezing the Pipe. This mode avoids to
have extra IN requests on a Pipe.
• INMODE = 1. The USB controller will perform infinite IN request until the firmware freezes the
Pipe.
The IN request generation will start when the firmware clear the PFREEZE bit.
298 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Each time the current bank is full, the RXIN and the FIFOCON bits are set. This triggers an inter-
rupt if the RXINE bit is set. The firmware can acknowledge the USB interrupt by clearing the
RXIN bit. The Firmware read the data and clear the FIFOCON bit in order to free the current
bank. If the IN Pipe is composed of multiple banks, clearing the FIFOCON bit will switch to the
next bank. The RXIN and FIFOCON bits are then updated by hardware in accordance with the
status of the new bank.
Example with 1 IN data bank
DATA DATA
IN ACK IN ACK
HW HW
RXIN SW SW
FIFOCON SW
read data from CPU read data from CPU
BANK 0 BANK 0
Example with 2 IN data banks
DATA DATA
IN ACK IN ACK
HW HW
RXIN SW SW
FIFOCON SW
read data from CPU read data from CPU
BANK 0 BANK 1
23.14.1 CRC Error (isochronous only)

A CRC error can occur during IN stage if the USB controller detects a bad received packet. In
this situation, the STALLEDI/CRCERRI interrupt is triggered. This does not prevent the RXINI
interrupt from being triggered.
23.15 Interrupt system
299
7593D–AVR–07/06
Figure 23-4. USB Host Controller Interrupt System
HWUPI
UHINT.6
HWUPE
UHIEN.6
HSOFI
UHINT.5
HSOFE
UHIEN.5
RXRSMI
UHINT.4
RXRSME
UHIEN.4
USB Host
RSMEDI Interrupt
UHINT.3
RSMEDE
UHIEN.3
RSTI
UHINT.2
RSTE
UHIEN.2
DDISCI
UHINT.1
DDISCE
UHIEN.1
DCONNI
UHINT.0
DCONNE
UHIEN.0
Figure 23-5. USB Device Controller Pipe Interrupt System

PIPE 6
PIPE 5
PIPE 4
PIPE 3
PIPE 2
PIPE 1
OVERFI PIPE 0
UPSTAX.6
UNDERFI
FLERRE
UPSTAX.5
UPIEN.7
NAKEDI
UPINTX.6
NAKEDE
UPIEN.6
PERRI
UPINTX.4
PERRE Pipe Interrupt
UPIEN.4
TXSTPI FLERRE
UPINTX.3 UPIEN.7
TXSTPE
UPIEN.3
TXOUTI
UPINTX.2
TXOUTE
UPIEN.2
RXSTALLI
UPINTX.1
RXSTALLE
UPIEN.1
RXINI
UPINTX.0
RXINE
UPIEN.0
23.16 Registers
300 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
23.16.1 General USB Host registers
Bit 7 6 5 4 3 2 1 0
- - - - - RESUME RESET SOFEN UHCON
• 7-3 - Reserved
• 2 - RESUME - Send USB Resume

Set this bit to generate a USB Resume on the USB bus.
Cleared by hardware when the USB Resume has been sent. Clearing by software has no effect.
This bit should be set only when the start of frame generation is enable (SOFEN bit set).
• 1 - RESET - Send USB Reset

Set this bit to generate a USB Reset on the USB bus.
Cleared by hardware when the USB Reset has been sent. Clearing by software has no effect.
Refer to the USB reset section for more details.
• 0 - SOFEN - Start Of Frame Generation Enable

Set this bit to generate SOF on the USB bus in full speed mode and keep-alive in low speed
mode.
Clear this bit to disable the SOF generation and to leave the USB bus in Idle state.
Bit 7 6 5 4 3 2 1 0
- HWUP HSOF RXRSMI RSMEDI RSTI DDISCI DCONNI UHINT
• 7 - Reserved
• 6 - HWUP
Host Wake-Up Interrupt
Set by hardware when a non-idle state is detected on the USB bus.
Shall be clear by software to acknowledge the interrupt. Setting by software has no effect.
• 5 - HSOFI - Host Start Of Frame Interrupt

Set by hardware when a SOF is issued by the Host controller. This triggers a USB interrupt
when HSOFE is set. When using the host controller in low speed mode, this bit is also set when
a keep-alive is sent.
Shall be cleared by software to acknowledge the interrupt. Setting by software has no effect.
• 4 - RXRSMI - Upstream Resume Received Interrupt
301
7593D–AVR–07/06
Set by hardware when an Upstream Resume has been received from the Device.
• 3 - RSMEDI - Downstream Resume Sent Interrupt

Set by hardware when a Downstream Resume has been sent to the Device.
• 2 - RSTI - USB Reset Sent Interrupt

Set by hardware when a USB Reset has been sent to the Device.
• 1 - DDISCI Device Disconnection Interrupt

Set by hardware when the device has been removed from the USB bus.
• 0 - DCONNI - Device Connection Interrupt

Set by hardware when a new device has been connected to the USB bus.
Bit 7 6 5 4 3 2 1 0
HWUPE HSOFE RXRSME RSMEDE RSTE DDISCE DCONNE UHIEN
• 7 - Reserved
• 6 - HWUPE - Host Wake-Up Interrupt Enable

Set this bit to enable HWUP interrupt.
Clear this bit to disable HWUP interrupt.
• 5 - HSOFE - Host Start Of frame Interrupt Enable

Set this bit to enable HSOF interrupt.
Clear this bit to disable HSOF interrupt.
• 4 - RXRSME -Upstream Resume Received Interrupt Enable

Set this bit to enable the RXRSMI interrupt.
Clear this bit to disable the RXRSMI interrupt.
• 3 - RSMEDE - Downstream Resume Sent Interrupt Enable

Set this bit to enable the RSMEDI interrupt.
Clear this bit to disable the RSMEDI interrupt.
• 2 - RSTE - USB Reset Sent Interrupt Enable
302 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Set this bit to enable the RSTI interrupt.
Clear this bit to disable the RSTI interrupt.
• 1 - DDISCE - Device Disconnection Interrupt Enable

Set this bit to enable the DDISCI interrupt.
Clear this bit to disable the DDISCI interrupt.
• 0 - DCONNE - Device Connection Interrupt Enable

Set this bit to enable the DCONNI interrupt.
Clear this bit to disable the DCONNI interrupt.
Bit 7 6 5 4 3 2 1 0
HADDR6 HADDR5 HADDR4 HADDR3 HADDR2 HADDR1 HADDR0 HADDR6 UHADDR
• 7 - Reserved
• 6-0 - HADDR6:0 - USB Host Address

These bits contain the address of the USB Device.
Bit 7 6 5 4 3 2 1 0
- - - - - FNUM10 FNUM9 FNUM8 UHFNUMH
• 7-4 - Reserved
• 3-0 - FNUM10:8 - Frame Number

The value contained in this register is the current SOF number.
This value can be modified by software.
Bit 7 6 5 4 3 2 1 0
FNUM7 FNUM6 FNUM5 FNUM4 FNUM3 FNUM2 FNUM1 FNUM0 UHFNUML
• 7-0 - FNUM7:0 - Frame Number

The value contained in this register is the current SOF number.
This value can be modified by software.
303
7593D–AVR–07/06
Bit 7 6 5 4 3 2 1 0
FLEN7 FLEN6 FLEN5 FLEN4 FLEN3 FLEN2 FLEN1 FLEN0 UHFLEN
• 7-0 - FLEN7:0 - Frame Length

The value contained the data frame length transmited.
23.16.2 USB Host Pipe registers
Bit 7 6 5 4 3 2 1 0
PNUM2 PNUM1 PNUM0 UPNUM
Read/Write RW RW RW
• 7-3 - Reserved
• 2-0 - PNUM2:0 - Pipe Number

Select the pipe using this register. The USB Host registers ended by a X correspond then to this
number.
This number is used for the USB controller following the value of the PNUMD bit.
Bit 7 6 5 4 3 2 1 0
- P6RST P5RST P4RST P3RST P2RST P1RST P0RST UPRST
Read/Write RW RW RW RW RW RW RW
• 7 - Reserved
• 6 - P6RST - Pipe 6 Reset

Set this bit to 1 and reset this bit to 0 to reset the Pipe 6.




304 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128

Bit 7 6 5 4 3 2 1 0
- PFREEZE INMODE - RSTDT - - PEN UPCONX
Read/Write RW RW RW RW
• 7 - Reserved
• 6 - PFREEZE - Pipe Freeze

Set this bit to Freeze the Pipe requests generation.
Clear this bit to enable the Pipe request generation.
This bit is set by hardware when:
- the pipe is not configured
- a STALL handshake has been received on this Pipe
- An error occurs on the Pipe (PERR = 1)
- (INRQ+1) In requests have been processed
This bit is set at 1 by hardware after a Pipe reset or a Pipe enable.
• 5 - INMODE - IN Request mode

Set this bit to allow the USB controller to perform infinite IN requests when the Pipe is not frozen.
Clear this bit to perform a pre-defined number of IN requests. This number is stored in the UIN-
RQX register.
• 4 - Reserved
• 3 - RSTDT - Reset Data Toggle

Set this bit to reset the Data Toggle to its initial value for the current Pipe.
Cleared by hardware when proceed. Clearing by software has no effect.
• 2 - Reserved
• 1 - Reserved
• 0 - PEN - Pipe Enable

Set to enable the Pipe.
305
7593D–AVR–07/06
Clear to disable and reset the Pipe.
Bit 7 6 5 4 3 2 1 0
PTYPE1 PTYPE0 PTOKEN1 PTOKEN0 PEPNUM3 PEPNUM2 PEPNUM1 PEPNUM0 UPCFG0X
Read/Write RW RW RW RW RW RW RW RW
• 7-6 - PTYPE1:0 - Pipe Type

Select the type of the Pipe:
- 00: Control
- 01: Isochronous
- 10: Bulk
- 11: Interrupt
• 5-4 - PTOKEN1:0 - Pipe Token

Select the Token to associate to the Pipe
- 00: SETUP
- 01: IN
- 10: OUT
- 11: reserved
• 3-0 - PEPNUM3:0 - Pipe Endpoint Number

Set this field according to the Pipe configuration. Set the number of the Endpoint targeted by the
Pipe. This value is from 0 and 15.
Bit 7 6 5 4 3 2 1 0
- PSIZE2:0 PBK1:0 ALLOC - UPCFG1X
Read/Write R RW RW RW RW RW RW
• 7 - Reserved
• 6-4 - PSIZE2:0 - Pipe Size

Select the size of the Pipe:
- 000: 8
- 001: 16
- 010: 32
- 011: 64
- 100: 128
- 101: 256
306 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
- 110: 512
- 111: 1024
• 3-2 - PBK1:0 - Pipe Bank

Select the number of bank to declare for the current Pipe.
- 00: 1 bank
- 01: 2 banks
- 10: invalid
- 11: invalid
• ALLOC
Configure Pipe Memory
Set to configure the pipe memory with the characteristics.
Clear to update the memory allocation. Refer to the Memory Management chapter for more
details.
7 - Reserved
Bit 7 6 5 4 3 2 1 0
INTFRQ7 INTFRQ6 INTFRQ5 INTFRQ4 INTFRQ3 INTFRQ2 INTFRQ1 INTFRQ0 UPCFG2X
• 7 - INTFRQ7:0 - Interrupt Pipe Request Frequency

These bits are the maximum value in millisecond of the polling period for an Interrupt Pipe.
This value has no effect for a non-Interrupt Pipe.
Bit 7 6 5 4 3 2 1 0
CFGOK OVERFI UNDERFI - DTSEQ1:0 NBUSYBK UPSTAX
Read/Write R RW RW R R R R
• 7 - CFGOK - Configure Pipe Memory OK

Set by hardware if the required memory configuration has been successfully performed.
Cleared by hardware when the pipe is disabled. The USB reset and the reset pipe have no effect
on the configuration of the pipe.
• 6 - OVERFI - Overflow
Set by hardware when a the current Pipe has received more data than the maximum length of
the current Pipe. An interrupt is triggered if the FLERRE bit is set.
• 5 - UNDERFI - Underflow
307
7593D–AVR–07/06
Set by hardware when a transaction underflow occurs in the current isochronous or interrupt
Pipe. The Pipe can’t send the data flow required by the device. A ZLP will be sent instead. An
interrupt is triggered if the FLERRE bit is set.
Note: the Host controller has to send a OUT packet, but the bank is empty. A ZLP will be sent
and the UNDERFI bit is set
underflow for interrupt Pipe:
• 4 - Reserved
• 3-2 - DTSEQ1:0 - Toggle Sequencing Flag

Set by hardware to indicate the PID data of the current bank:
00b Data0
01b Data1
1xb Reserved.
For OUT Pipe, this value indicates the next data toggle that will be sent. This is not relative to the
current bank.
For IN Pipe, this value indicates the last data toggle received on the current bank.
• 1-0 - NBUSYBK1:0 - Busy Bank Flag

Set by hardware to indicate the number of busy bank.
For OUT Pipe, it indicates the number of busy bank(s), filled by the user, ready for OUT transfer.
For IN Pipe, it indicates the number of busy bank(s) filled by IN transaction from the Device.
00b All banks are free
01b 1 busy bank
10b 2 busy banks
11b Reserved.
Bit 7 6 5 4 3 2 1 0
INRQ7 INRQ6 INRQ5 INRQ4 INRQ3 INRQ2 INRQ1 INRQ0 UPINRQX
• 7-0 - INRQ7:0 - IN Request Number Before Freeze

Enter the number of IN transactions before the USB controller freezes the pipe. The USB con-
troller will perform (INRQ+1) IN requests before to freeze the Pipe. This counter is automatically
decreased by 1 each time a IN request has been successfully performed.
This register has no effect when the INMODE bit is set (infinite IN requests generation till the
pipe is not frozen).
308 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Bit 7 6 5 4 3 2 1 0
- COUNTER1:0 CRC16 TIMEOUT PID DATAPID DATATGL UPERRX
• 7-6 - Reserved
• 5 - COUNTER1:0 - Error counter

This counter is increased by the USB controller each time an error occurs on the Pipe. When this
value reaches 3, the Pipe is automatically frozen.
Clear these bits by software.
• 4 - CRC16 - CRC16 Error

Set by hardware when a CRC16 error has been detected.
• 3 - TIMEOUT - Time-out Error

Set by hardware when a time-out error has been detected.
• 2 - PID - PID Error

Set by hardware when a PID error has been detected.
• 1 - DATAPID - Data PID Error

Set by hardware when a data PID error has been detected.
• 0 - DATATGL - Bad Data Toggle

Set by hardware when a data toggle error has been detected.
Bit 7 6 5 4 3 2 1 0
FIFOCON NAKEDI RWAL PERRI TXSTPI TXOUTI RXSTALLI RXINI UPINTX
• 7 - FIFOCON - FIFO Control

For OUT and SETUP Pipe:
Set by hardware when the current bank is free, at the same time than TXOUT or TXSTP.
Clear to send the FIFO data and to switch the bank. Setting by software has no effect.
309
7593D–AVR–07/06
For IN Pipe:
Set by hardware when a new IN message is stored in the current bank, at the same time than
RXIN.
Clear to free the current bank and to switch to the following bank. Setting by software has no
effect.
• 6 - NAKEDI - NAK Handshake received

Set by hardware when a NAK has been received on the current bank of the Pipe. This triggers
an interrupt if the NAKEDE bit is set in the UPIENX register.
Shall be clear to handshake the interrupt. Setting by software has no effect.
• 5 - RWAL - Read/Write Allowed

OUT Pipe:
Set by hardware when the firmware can write a new data into the Pipe FIFO.
Cleared by hardware when the current Pipe FIFO is full.
IN Pipe:
Set by hardware when the firmware can read a new data into the Pipe FIFO.
Cleared by hardware when the current Pipe FIFO is empty.
This bit is also cleared by hardware when the RXSTALL or the PERR bit is set
• 4 - PERRI -PIPE Error

Set by hardware when an error occurs on the current bank of the Pipe. This triggers an interrupt
if the PERRE bit is set in the UPIENX register. Refers to the UPERRX register to determine the
source of the error.
Automatically cleared by hardware when the error source bit is cleared.
• 3 - TXSTPI - SETUP Bank ready

Set by hardware when the current SETUP bank is free and can be filled. This triggers an inter-
rupt if the TXSTPE bit is set in the UPIENX register.
Shall be cleared to handshake the interrupt. Setting by software has no effect.
• 2 - TXOUTI -OUT Bank ready

Set by hardware when the current OUT bank is free and can be filled. This triggers an interrupt if
the TXOUTE bit is set in the UPIENX register.
• 1 - RXSTALLI / CRCERR - STALL Received / Isochronous CRC Error

Set by hardware when a STALL handshake has been received on the current bank of the Pipe.
The Pipe is automatically frozen. This triggers an interrupt if the RXSTALLE bit is set in the UPI-
ENX register.
For Isochronous Pipe:
Set by hardware when a CRC error occurs on the current bank of the Pipe. This triggers an inter-
rupt if the TXSTPE bit is set in the UPIENX register.
310 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
• 0 - RXINI - IN Data received

Set by hardware when a new USB message is stored in the current bank of the Pipe. This trig-
gers an interrupt if the RXINE bit is set in the UPIENX register.
Bit 7 6 5 4 3 2 1 0
FLERRE NAKEDE - PERRE TXSTPE TXOUTE RXSTALLE RXINE UPIENX
• 7 - FLERRE - Flow Error Interrupt enable

Set to enable the OVERFI and UNDERFI interrupts.
Clear to disable the OVERFI and UNDERFI interrupts.
• 6 - NAKEDE -NAK Handshake Received Interrupt Enable

Set to enable the NAKEDI interrupt.
Clear to disable the NAKEDI interrupt.
• 5 - Reserved
• 4 - PERRE -PIPE Error Interrupt Enable

Set to enable the PERRI interrupt.
Clear to disable the PERRI interrupt.
• 3 - TXSTPE - SETUP Bank ready Interrupt Enable

Set to enable the TXSTPI interrupt.
Clear to disable the TXSTPI interrupt.
• 2 - TXOUTE - OUT Bank ready Interrupt Enable

Set to enable the TXOUTI interrupt.
Clear to disable the TXOUTI interrupt.
• 1 - RXSTALLE - STALL Received Interrupt Enable

Set to enable the RXSTALLI interrupt.
Clear to disable the RXSTALLI interrupt.
• 0 - RXINE - IN Data received Interrupt Enable

Set to enable the RXINI interrupt.
Clear to disable the RXINI interrupt.
Bit 7 6 5 4 3 2 1 0
PDAT7 PDAT6 PDAT5 PDAT4 PDAT3 PDAT2 PDAT1 PDAT0 UPDATX
311
7593D–AVR–07/06
Bit 7 6 5 4 3 2 1 0
• 7-0 - PDAT7:0 - Pipe Data Bits

Set by the software to read/write a byte from/to the Pipe FIFO selected by PNUM.
Bit 7 6 5 4 3 2 1 0
- - - - - PBYCT10 PBYCT9 PBYCT8 UPBCHX
Read/Write R R R
• 7-3 - Reserved
• 2-0 - PBYCT10:8 - Byte count (high) Bits

Set by hardware. This field is the MSB of the byte count of the FIFO endpoint. The LSB part is
provided by the UPBCLX register.
Bit 7 6 5 4 3 2 1 0
PBYCT7 PBYCT6 PBYCT5 PBYCT4 PBYCT3 PBYCT2 PBYCT1 PBYCT0 UPBCLX
• 7-0 - PBYCT7:0 - Byte Count (low) Bits

Set by the hardware. PBYCT10:0 is:
- (for OUT Pipe) increased after each writing into the Pipe and decremented after each byte
sent,
- (for IN Pipe) increased after each byte received by the host, and decremented after each byte
read by the software.
Bit 7 6 5 4 3 2 1 0
- PINT6 PINT5 PINT4 PINT3 PINT2 PINT1 PINT0 UPINT
Read/Write
• 7 - Reserved
• 6-0 - PINT6:0 - Pipe Interrupts Bits

Set by hardware when an interrupt is triggered by the UPINTX register and if the corresponding
endpoint interrupt enable bit is set.
Cleared by hardware when the interrupt source is served.
312 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
24. Analog Comparator
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 24-1.
The Power Reduction ADC bit, PRADC, in “Power Reduction Register 0 - PRR0” on page 56
must be disabled by writing a logical zero to be able to use the ADC input MUX.
Figure 24-1. Analog Comparator Block Diagram(2)

BANDGAP
REFERENCE
ACBG
ACME
ADEN
ADC MULTIPLEXER
OUTPUT (1)
Notes: 1. See Table 24-2 on page 315.

2. Refer to Figure 1-1 on page 3 and Table 10-6 on page 82 for Analog Comparator pin
placement.
24.0.1 ADC Control and Status Register B – ADCSRB

Bit 7 6 5 4 3 2 1 0
– ACME – – - ADTS2 ADTS1 ADTS0 ADCSRB
Read/Write R R/W R R R 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 315.
24.0.2 Analog Comparator Control and Status Register – ACSR

Bit 7 6 5 4 3 2 1 0
ACD ACBG ACO ACI ACIE ACIC ACIS1 ACIS0 ACSR
Initial Value 0 0 N/A 0 0 0 0 0
• Bit 7 – ACD: Analog Comparator Disable
313
7593D–AVR–07/06
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. See “Internal Voltage Reference” on page 64.
• 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.
• 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 24-1.
Table 24-1. 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.
314 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
24.1 Analog Comparator Multiplexed Input
It is possible to select any of the ADC15..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), and MUX2..0 in
ADMUX select the input pin to replace the negative input to the Analog Comparator, as shown in
Table 24-2. If ACME is cleared or ADEN is set, AIN1 is applied to the negative input to the Ana-
log Comparator.
Table 24-2. Analog Comparator Mulitiplexed Input

ACME ADEN MUX2..0 Analog Comparator Negative Input
0 x xxx AIN1
1 1 xxx AIN1
1 0 000 ADC0
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
24.1.1 Digital Input Disable Register 1 – DIDR1

Bit 7 6 5 4 3 2 1 0
– – – – – – AIN1D AIN0D DIDR1
• 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.
315
7593D–AVR–07/06
25. Analog to Digital Converter - ADC
25.1 Features
• 10-bit Resolution
• 0.5 LSB Integral Non-linearity
• ± 2 LSB Absolute Accuracy
• 65 - 260 µs Conversion Time
• Up to 15 kSPS at Maximum Resolution
• Eight Multiplexed Single Ended Input Channels
• Seven Differential input channels
• Optional Left Adjustment for ADC Result Readout
• 0 - VCC ADC Input Voltage Range
• Selectable 2.56 V 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
The AT90USB64/128 features a 10-bit successive approximation ADC. The ADC is connected
to an 8-channel Analog Multiplexer which allows eight single-ended voltage inputs constructed
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, providing
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, 7-bit resolution can be
expected.
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 25-1.
The ADC has a separate analog supply voltage pin, AVCC. AVCC must not differ more than ±
0.3V from VCC. See the paragraph “ADC Noise Canceler” on page 323 on how to connect this
pin.
Internal reference voltages of nominally 2.56V or AVCC are provided On-chip. The voltage refer-
ence may be externally decoupled at the AREF pin by a capacitor for better noise performance.
316 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 25-1. Analog to Digital Converter Block Schematic
ADC CONVERSION
COMPLETE IRQ
INTERRUPT
FLAGS
ADTS[2:0]
8-BIT DATA BUS
ADIE
ADIF
15 0
ADC MULTIPLEXER ADC CTRL. & STATUS ADC DATA REGISTER
SELECT (ADMUX) REGISTER (ADCSRA) (ADCH/ADCL)
REFS1
REFS0
ADLAR
MUX4
MUX3
MUX2
MUX1
MUX0
ADEN
ADSC
ADATE
ADPS2
ADPS1
ADPS0
ADIF
ADC[9:0]
TRIGGER
SELECT
MUX DECODER
PRESCALER
CHANNEL SELECTION
START
GAIN SELECTION CONVERSION LOGIC

AVCC
INTERNAL
REFERENCE SAMPLE & HOLD
COMPARATOR
AREF 10-BIT DAC -
+
ADHSM
GND
BANDGAP
REFERENCE
ADC7
SINGLE ENDED / DIFFERENTIAL SELECTION
ADC6
POS. ADC MULTIPLEXER
ADC5 INPUT OUTPUT
MUX
ADC4
ADC3 DIFFERENTIAL
AMPLIFIER
+
ADC2
-
ADC1
ADC0
NEG.
INPUT
MUX
25.2 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
317
7593D–AVR–07/06
the AREF pin minus 1 LSB. Optionally, AVCC or an internal 2.56V reference voltage may be con-
nected 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 amplifier.
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. The ADC
access to the Data Registers is prohibited between reading of ADCH and ADCL, the interrupt
will trigger even if the result is lost.
25.3 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 is still 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.
318 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 25-2. 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.
25.4 Prescaling and Conversion Timing
Figure 25-3. 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. Alter-
natively, setting the ADHSM bit in ADCSRB allows an increased ADC clock frequency at the
expense of higher power consumption.
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
319
7593D–AVR–07/06
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 Channels” on page
321 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.
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 an 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 two ADC clock cycles after the rising edge on the trigger source signal. Three addi-
tional CPU clock cycles are used for synchronization logic.
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 25-1.
Figure 25-4. 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 Sign and MSB of Result
ADCL LSB of Result
MUX
MUX and REFS Conversion and REFS
Update Sample & Hold Complete Update
Figure 25-5. 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
ADCL LSB of Result
Sample & Hold

Conversion MUX and REFS
MUX and REFS Complete Update
Update
320 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 25-6. 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
ADCL LSB of Result
Sample & Conversion Prescaler

Prescaler Hold Reset
Complete
Reset
MUX and REFS
Update
Figure 25-7. ADC Timing Diagram, Free Running Conversion

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

Complete MUX and REFS
Update
Table 25-1. ADC Conversion Time

Normal
First Conversion, Auto Triggered
Condition Conversion Single Ended Convertion
Sample & Hold
14.5 1.5 2
(Cycles from Start of Convertion)
Conversion Time
25 13 13.5
(Cycles)
25.4.1 Differential Channels

When using differential channels, certain aspects of the conversion need to be taken into
consideration.
Differential conversions are synchronized to the internal clock CKADC2 equal to half the ADC
clock frequency. 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
321
7593D–AVR–07/06
clock cycles due to the synchronization mechanism. In Free Running mode, a new conversion is
initiated 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.
If differential 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 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 conversions are
performed. The result from the extended conversions will be valid. See “Prescaling and Conver-
sion Timing” on page 319 for timing details.
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. E.g. the ADC clock period
may be 6 µs, allowing a channel to be sampled at 12 kSPS, regardless of the bandwidth of this
channel.
25.5 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.
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:
a. When ADATE or ADEN is cleared.
b. During conversion, minimum one ADC clock cycle after the trigger event.
c. 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 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 chan-
nel. 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).
322 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
The settling time and gain stage bandwidth is independent of the ADHSM bit setting.
25.5.1 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
channel 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
channel 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.
25.5.2 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.
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 indi-
cated in Table 30-5 on page 407.
25.6 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:
323
7593D–AVR–07/06
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 inter-
rupt 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 com-
plete, 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 conver-
sions, 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.
25.6.1 Analog Input Circuitry

The analog input circuitry for single ended channels is illustrated in Figure 25-8. 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.
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 25-8. Analog Input Circuitry
IIH
ADCn
1..100 kΩ
CS/H= 14 pF
IIL
VCC/2
324 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
25.6.2 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 digi-
tal 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 25-9.
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 25-9. ADC Power Connections
(AD0) PA0 51
VCC
52
GND 53
(ADC7) PF7 54
(ADC6) PF6 55
(ADC5) PF5 56
(ADC4) PF4 57
(ADC3) PF3 58
(ADC2) PF2 59
(ADC1) PF1 60
(ADC0) PF0 61
10μH
AREF 62
GND
63
AVCC
64
100nF
1
Analog Ground Plane
NC
25.6.3 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.
25.6.4 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:
325
7593D–AVR–07/06
• Offset: The deviation of the first transition (0x000 to 0x001) compared to the ideal transition
(at 0.5 LSB). Ideal value: 0 LSB.
Figure 25-10. 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 25-11. 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.
326 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 25-12. 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 25-13. 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.
25.7 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:
327
7593D–AVR–07/06
V IN ⋅ 1023
ADC = --------------------------
V REF
where VIN is the voltage on the selected input pin and VREF the selected voltage reference (see
Table 25-3 on page 330 and Table 25-4 on page 330). 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 result, it is sufficient to read the MSB of the result
(ADC9 in ADCH). If the bit is one, the result is negative, and if this bit is zero, the result is posi-
tive. Figure 25-14 shows the decoding of the differential input range.
Table 82 shows the resulting output codes if the differential input channel pair (ADCn - ADCm) is
selected with a reference voltage of VREF.
Figure 25-14. Differential Measurement Range
Output Code
0x1FF
0x000
- VREF 0 VREF Differential Input

0x3FF
Voltage (Volts)
0x200
328 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 25-2. 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
Example 1:
– 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.
Example 2:
– ADMUX = 0xFB (ADC3 - ADC2, 1x gain, 2.56V reference, left adjusted result)
– Voltage on ADC3 is 300 mV, voltage on ADC2 is 500 mV.
– ADCR = 512 * 1 * (300 - 500) / 2560 = -41 = 0x029.
– ADCL will thus read 0x40, and ADCH will read 0x0A.
Writing zero to ADLAR right adjusts the result: ADCL = 0x00, ADCH = 0x29.
25.8 ADC Register Description
25.8.1 ADC Multiplexer Selection Register – ADMUX

Bit 7 6 5 4 3 2 1 0
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 25-3. If these bits are
changed during a conversion, the change will not go in effect until this conversion is complete
329
7593D–AVR–07/06
(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 25-3. Voltage Reference Selections for ADC

REFS1 REFS0 Voltage Reference Selection
0 0 AREF, Internal Vref turned off
0 1 AVCC with external capacitor on AREF pin
1 0 Reserved
1 1 Internal 2.56V Voltage Reference with external capacitor on AREF pin
• 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 “The ADC Data Register – ADCL and ADCH” on
page 332.
• Bits 4:0 – MUX4:0: Analog Channel 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 25-4 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).
Table 25-4. 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
330 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 25-4. Input Channel and Gain Selections (Continued)
Single Ended Positive Differential Negative Differential
MUX4..0 Input Input Input Gain
01000 (ADC0 / ADC0 / 10x)
01001 ADC1 ADC0 10x
01010 (ADC0 / ADC0 / 200x)
01011 ADC1 ADC0 200x
01100 (Reserved - ADC2 / ADC2 / 10x)
01101 ADC3 ADC2 10x
01110 (ADC2 / ADC2 / 200x)
01111 ADC3 ADC2 200x
10000 ADC0 ADC1 1x
10001 (ADC1 / ADC1 / 1x)
10010 ADC2 ADC1 1x
N/A
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 (VBand Gap)
N/A
11111 0V (GND)
25.8.2 ADC Control and Status Register A – ADCSRA

Bit 7 6 5 4 3 2 1 0
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,
331
7593D–AVR–07/06
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 25-5. ADC Prescaler Selections

ADPS2 ADPS1 ADPS0 Division Factor
0 0 0 2
0 0 1 2
0 1 0 4
0 1 1 8
1 0 0 16
1 0 1 32
1 1 0 64
1 1 1 128
25.8.3 The ADC Data Register – ADCL and ADCH
25.8.3.1 ADLAR = 0
Bit 15 14 13 12 11 10 9 8
– – – – – – ADC9 ADC8 ADCH
ADC7 ADC6 ADC5 ADC4 ADC3 ADC2 ADC1 ADC0 ADCL
Bit 7 6 5 4 3 2 1 0
R R R R R R R R
0 0 0 0 0 0 0 0
332 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
25.8.3.2 ADLAR = 1
Bit 15 14 13 12 11 10 9 8
ADC9 ADC8 ADC7 ADC6 ADC5 ADC4 ADC3 ADC2 ADCH
ADC1 ADC0 – – – – – – ADCL
Bit 7 6 5 4 3 2 1 0
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 (7 bit + sign bit for differential input
channels) 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 327.
25.8.4 ADC Control and Status Register B – ADCSRB

Bit 7 6 5 4 3 2 1 0
ADHSM ACME – – – ADTS2 ADTS1 ADTS0 ADCSRB
Read/Write R/W R/W R R R R/W R/W R/W
• Bit 7 – ADHSM: ADC High Speed Mode

Writing this bit to one enables the ADC High Speed mode. This mode enables higher conversion
rate at the expense of higher power consumption.
• 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 ADTS2:0 settings will have no effect. A conversion
will be triggered by the rising edge of the selected interrupt flag. Note that switching from a trig-
ger 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 25-6. 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
333
7593D–AVR–07/06
Table 25-6. ADC Auto Trigger Source Selections
ADTS2 ADTS1 ADTS0 Trigger Source
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
25.8.5 Digital Input Disable Register 0 – DIDR0

Bit 7 6 5 4 3 2 1 0
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.
334 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
26. JTAG Interface and On-chip Debug System
26.0.1 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®
26.1 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 387 and “IEEE 1149.1 (JTAG) Boundary-scan” on page
341, respectively. The On-chip Debug support is considered being private JTAG instructions,
and distributed within ATMEL and to selected third party vendors only.
Figure 26-1 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.
26.2 Test Access Port – TAP

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.
• TCK: Test Clock. JTAG operation is synchronous to TCK.
335
7593D–AVR–07/06
• 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 26-1. 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
336 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 26-2. 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
26.3 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 26-2 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.
337
7593D–AVR–07/06
• 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 340.
26.4 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 341.
26.5 Using the On-chip Debug System

As shown in Figure 26-1, 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”).
338 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
• 2 single Program Memory Break Points + 1 Data Memory Break Point with mask (“range
Break Point”).
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 339.
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.
26.6 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.
26.6.1 PRIVATE0; 0x8

Private JTAG instruction for accessing On-chip debug system.
26.6.2 PRIVATE1; 0x9

26.6.3 PRIVATE2; 0xA

26.6.4 PRIVATE3; 0xB

339
7593D–AVR–07/06
26.7 On-chip Debug Related Register in I/O Memory
26.7.1 On-chip Debug Register – OCDR

Bit 7 6 5 4 3 2 1 0
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.
26.8 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.
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 387.
26.9 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.
340 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
27. IEEE 1149.1 (JTAG) Boundary-scan
27.1 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
27.2 System 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.
27.3 Data Registers

The Data Registers relevant for Boundary-scan operations are:
• Bypass Register
• Device Identification Register
• Reset Register
• Boundary-scan Chain
341
7593D–AVR–07/06
27.3.1 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.
27.3.2 Device Identification Register

Figure 27-1 shows the structure of the Device Identification Register.
Figure 27-1. 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
27.3.2.1 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.
27.3.2.2 Part Number

The part number is a 16-bit code identifying the component. The JTAG Part Number for
AT90USB64/128 is listed in Table 27-1.
Table 27-1. AVR JTAG Part Number

Part Number JTAG Part Number (Hex)
AVR USB 0x9782
27.3.2.3 Manufacturer ID
The Manufacturer ID is a 11-bit code identifying the manufacturer. The JTAG manufacturer ID
for ATMEL is listed in Table 27-2.
Table 27-2. Manufacturer ID

Manufacturer JTAG Manufactor ID (Hex)
ATMEL 0x01F
27.3.3 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 40) 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 27-2.
342 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 27-2. Reset Register
To
TDO
From Other Internal and

External Reset Sources
From Internal reset

D Q
TDI
ClockDR · AVR_RESET
27.3.4 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 345 for a complete description.
27.4 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.
27.4.1 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.
27.4.2 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.
343
7593D–AVR–07/06
• 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.
27.4.3 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.
27.4.4 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.
27.4.5 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.
27.5 Boundary-scan Related Register in I/O Memory

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.
344 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
27.5.2 MCU Status Register – MCUSR
The MCU Status Register provides information on which reset source caused an MCU reset.
Bit 7 6 5 4 3 2 1 0
– – – JTRF WDRF BORF EXTRF PORF MCUSR
Read/Write R R R R/W R/W R/W R/W R/W
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.
27.6 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.
27.6.1 Scanning the Digital Port Pins

Figure 27-3 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
The Boundary-scan logic is not included in the figures in the datasheet. Figure 27-4 shows a
simple digital port pin as described in the section “I/O-Ports” on page 74. The Boundary-scan
details from Figure 27-3 replaces the dashed box in Figure 27-4.
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 27-4 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.
345
7593D–AVR–07/06
Figure 27-3. 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
346 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 27-4. 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
27.6.2 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 27-5 is
inserted for the 5V reset signal.
Figure 27-5. Observe-only Cell

To
Next
ShiftDR Cell
From System Pin To System Logic
FF1
0
D Q
1
From ClockDR
Previous
Cell
347
7593D–AVR–07/06
27.7 AT90USB64/128 Boundary-scan Order
Table 27-3 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 Fis
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 27-3, 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. The USB pads are not included in
the boundary-scan.
Table 27-3. AT90USB64/128 Boundary-scan Order

Bit Number Signal Name Module
88 PE6.Data
87 PE6.Control
86 PE7.Data
Port E
85 PE7.Control
84 PE3.Data
83 PE3.Control
82 PB0.Data
81 PB0.Control
80 PB1.Data
79 PB1.Control
78 PB2.Data
77 PB2.Control
76 PB3.Data
75 PB3.Control
Port B
74 PB4.Data
73 PB4.Control
72 PB5.Data
71 PB5.Control
70 PB6.Data
69 PB6.Control
68 PB7.Data
67 PB7.Control
66 PE4.Data
65 PE4.Control
PORTE
64 PE5.Data
63 PE5.Control
348 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 27-3. AT90USB64/128 Boundary-scan Order (Continued)
62 RSTT Reset Logic (Observe Only)
61 PD0.Data
60 PD0.Control
59 PD1.Data
58 PD1.Control
57 PD2.Data
56 PD2.Control
55 PD3.Data
54 PD3.Control
Port D
53 PD4.Data
52 PD4.Control
51 PD5.Data
50 PD5.Control
49 PD6.Data
48 PD6.Control
47 PD7.Data
46 PD7.Control
45 PE0.Data
44 PE0.Control
Port E
43 PE1.Data
42 PE1.Control
349
7593D–AVR–07/06
41 PC0.Data
40 PC0.Control
39 PC1.Data
38 PC1.Control
37 PC2.Data
36 PC2.Control
35 PC3.Data
34 PC3.Control
Port C
33 PC4.Data
32 PC4.Control
31 PC5.Data
30 PC5.Control
29 PC6.Data
28 PC6.Control
27 PC7.Data
26 PC7.Control
25 PE2.Data
Port E
24 PE2.Control
23 PA7.Data
22 PA7.Control
21 PA6.Data
20 PA6.Control
19 PA5.Data
18 PA5.Control
17 PA4.Data
16 PA4.Control
Port A
15 PA3.Data
14 PA3.Control
13 PA2.Data
12 PA2.Control
11 PA1.Data
10 PA1.Control
9 PA0.Data
8 PA0.Control
350 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
7 PF3.Data
6 PF3.Control
5 PF2.Data
4 PF2.Control
Port F
3 PF1.Data
2 PF1.Control
1 PF0.Data
0 PF0.Control
27.8 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
AT90USB64/128.
351
7593D–AVR–07/06
28. Boot Loader Support – Read-While-Write Self-Programming
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. General information on SPM and ELPM is provided in See “AVR CPU
Core” on page 10.
28.1 Boot Loader 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 29-11 on page 373)
used during programming. The page organization does not affect normal operation.
28.2 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 28-2). The size of the different sections is configured by the
BOOTSZ Fuses as shown in Table 28-8 on page 366 and Figure 28-2. These two sections can
have different level of protection since they have different sets of Lock bits.
28.2.1 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 28-2 on page 356. The Application section can never store any
Boot Loader code since the SPM instruction is disabled when executed from the Application
section.
28.2.2 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 28-3 on page 356.
28.3 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
352 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 28-
1 and Figure 28-1 on page 354. 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.
28.3.1 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 “Store Program Memory Control and Status Register –
SPMCSR” on page 358. for details on how to clear RWWSB.
28.3.2 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 28-1. Read-While-Write Features

Which Section does the Z- Which Section Can
pointer Address During the be Read During Is the CPU Read-While-Write
Programming? Programming? Halted? Supported?
RWW Section NRWW Section No Yes
NRWW Section None Yes No
353
7593D–AVR–07/06
Figure 28-1. 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
354 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 28-2. 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 28-8 on page 366.
28.4 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 28-2 and Table 28-3 for further details. The Boot Lock bits can be set by software and
in Serial or in Parallel Programming mode. They can only be cleared by a Chip Erase command
only. The general Write Lock (Lock Bit mode 2) does not control the programming 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.
355
7593D–AVR–07/06
Table 28-2. Boot Lock Bit0 Protection Modes (Application Section)(1)
BLB0 Mode BLB02 BLB01 Protection
No restrictions for SPM or (E)LPM accessing the
1 1 1
Application 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
3 0 0 allowed to read from the Application section. If Interrupt
Vectors are placed in the Boot Loader section, interrupts
are disabled while executing from the Application section.
allowed to read from the Application section. If Interrupt
4 0 1
Note: 1. “1” means unprogrammed, “0” means programmed
Table 28-3. Boot Lock Bit1 Protection Modes (Boot Loader Section)(1)
BLB1 Mode BLB12 BLB11 Protection
No restrictions for SPM or (E)LPM accessing the Boot
1 1 1
Loader 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 read from the Boot Loader section. If Interrupt
3 0 0
Vectors are placed in the Application section, interrupts
are disabled while executing from the Boot Loader
section.
(E)LPM executing from the Application section is not
4 0 1 Vectors are placed in the Application section, interrupts
section.
28.5 Entering the Boot Loader Program

The bootloader can be executed with three different conditions:
28.5.1 Regular application conditions.

A jump or call from the application program. This may be initiated by a trigger such as a com-
mand received via USART, SPI or USB.
28.5.2 Boot Reset Fuse

The Boot Reset Fuse (BOOTRST) 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 application 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
356 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
is programmed, 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 28-4. Boot Reset Fuse(1)

BOOTRST Reset Address
1 Reset Vector = Application Reset (address 0x0000)
0 Reset Vector = Boot Loader Reset (see Table 28-8 on page 366)
28.5.3 External Hardware conditions

The Hardware Boot Enable Fuse (HWBE) can be programmed (See Table 28-5) so that upon
special hardware conditions under reset, the bootloader execution is forced after reset.
Table 28-5. Hardware Boot Enable Fuse(1)

HWBE Reset Address
1 ALE/HWB pin can not be used to force Boot Loader execution after reset
0 ALE/HWB pin is used during reset to force bootloader execution after reset

When the HWBE fuse is enable the ALE/HWB pin is configured as input during reset and sam-
pled during reset rising edge. When ALE/HWB pin is ‘0’ during reset rising edge, the reset vector
will be set as the Boot Loader Reset address and the Boot Loader will be executed (See Figures
28-3).
357
7593D–AVR–07/06
Figure 28-3. Boot Process Description
RESET
tSHRH tHHRH
ALE/HWB
HWBE ?
Ext. Hardware
Conditions ?
BOOTRST ?
Reset Vector = Application Reset Reset Vector =Boot Lhoader Reset
28.5.4 Store Program Memory Control and Status Register – SPMCSR

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
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 363 for details. An SPM instruction within four cycles
358 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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 363 for
details.
• 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.
359
7593D–AVR–07/06
Note: Only one SPM instruction should be active at any time.
28.6 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 29-11 on page 373), 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 28-4. 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 28-4. Addressing the Flash During SPM(1)

BIT 23 ZPCMSB ZPAGEMSB 1 0
0 Z - POINTER
PCMSB PAGEMSB
PROGRAM COUNTER PCPAGE PCWORD
PAGE ADDRESS WORD ADDRESS

WITHIN THE FLASH WITHIN A PAGE
PROGRAM MEMORY PAGE PCWORD[PAGEMSB:0]:
PAGE INSTRUCTION WORD 00
01
02
PAGEEND
360 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Note: 1. The different variables used in Figure 28-4 are listed in Table 28-10 on page 367.
28.7 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 364 for an
assembly code example.
28.7.1 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.
28.7.2 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.
28.7.3 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.
361
7593D–AVR–07/06
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.
28.7.4 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 70.
28.7.5 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.
28.7.6 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 70, 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 364 for an example.
28.7.7 Setting the Boot Loader Lock Bits by SPM

To set the Boot Loader Lock bits, write the desired data to R0, write “X0001001” to SPMCSR
and execute SPM within four clock cycles after writing SPMCSR. The only accessible Lock bits
are the Boot Lock bits that may prevent the Application and Boot Loader section from any soft-
ware update by the MCU.
Bit 7 6 5 4 3 2 1 0
R0 1 1 BLB12 BLB11 BLB02 BLB01 1 1
See Table 28-2 and Table 28-3 for how the different settings of the Boot Loader bits affect the
Flash access.
If bits 5..2 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, 6, 1, and 0 in R0 to “1” when writing the Lock bits. When pro-
gramming the Lock bits the entire Flash can be read during the operation.
28.7.8 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
362 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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.
28.7.9 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 29-5 on page 370 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 29-4 on page 370 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 29-3 on page 369 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.
28.7.10 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 28-6 on page 364 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
363
7593D–AVR–07/06
instruction is executed within three CPU cycles. When SIGRD and SPMEN are cleared, LPM will
work as described in the Instruction set Manual.
Table 28-6. 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.
28.7.11 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
voltage 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.
28.7.12 Programming Time for Flash when Using SPM

The calibrated RC Oscillator is used to time Flash accesses. Table 28-7 shows the typical pro-
gramming time for Flash accesses from the CPU.
Table 28-7. SPM Programming Time

Symbol Min Programming Time Max Programming Time
Flash write (Page Erase, Page Write,
3.7 ms 4.5 ms
and write Lock bits by SPM)
28.7.13 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
364 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
; (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
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
365
7593D–AVR–07/06
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
28.7.14 AT90USB64/128 Boot Loader Parameters

In Table 28-8 through Table 28-10, the parameters used in the description of the Self-Program-
ming are given.
Table 28-8. Boot Size Configuration (Word Addresses)(1)
Reset Address
Flash Section
Flash Section
Loader Section)
BOOTSZ1
BOOTSZ0
Boot Size
Boot Loader
Application
Application
Device
(Start Boot
Section
Pages
Boot
End
1 1 512 words 4 0x0000 - 0x7DFF 0x7E00 - 0x7FFF 0x7DFF 0x7E00

AT90USB64
1 0 1024 words 8 0x0000 - 0x7BFF 0x7C00 - 0x7FFF 0x7BFF 0x7C00

0 1 2048 words 16 0x0000 - 0x77FF 0x7800 - 0x7FFF 0x77FF 0x7800
0 0 4096 words 32 0x0000 - 0x6FFF 0x7000 - 0x7FFF 0x6FFF 0x7000
1 1 512 words 4 0x0000 - 0xFDFF 0xFE00 - 0xFFFF 0xFDFF 0xFE00
AT90USB128
1 0 1024 words 8 0x0000 - 0xFBFF 0xFC00 - 0xFFFF 0xFBFF 0xFC00

0 1 2048 words 16 0x0000 - 0xF7FF 0xF800 - 0xFFFF 0xF7FF 0xF800
0 0 4096 words 32 0x0000 - 0xEFFF 0xF000 - 0xFFFF 0xEFFF 0xF000
Note: 1. The different BOOTSZ Fuse configurations are shown in Figure 28-2
366 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 28-9. Read-While-Write Limit (Word Addresses)(1)

Device Section Pages Address
Read-While-Write section (RWW) 224 0x0000 - 0x6FFF
AT90USB64
No Read-While-Write section (NRWW) 32 0x7000 - 0x7FFF
Read-While-Write section (RWW) 480 0x0000 - 0xEFFF

AT90USB28
No Read-While-Write section (NRWW) 32 0xF000 - 0xFFFF
Note: 1. For details about these two section, see “NRWW – No Read-While-Write Section” on page
353 and “RWW – Read-While-Write Section” on page 353.
Table 28-10.
Explanation of different variables used in Figure 28-4 and the mapping to the Z-
pointer
Corresponding
Variable Z-value Description(1)
Most significant bit in the Program Counter. (The
PCMSB 16
Program Counter is 17 bits PC[16:0])
Most significant bit which is used to address the
PAGEMSB 6 words within one page (128 words in a page requires
seven bits PC [6:0]).
Bit in Z-pointer that is mapped to PCMSB. Because
ZPCMSB Z17
Z0 is not used, the ZPCMSB equals PCMSB + 1.
Bit in Z-pointer that is mapped to PCMSB. Because
ZPAGEMSB Z7 Z0 is not used, the ZPAGEMSB equals PAGEMSB +
1.
Program Counter page address: Page select, for
PCPAGE PC[16:7] Z17:Z8
Page Erase and Page Write
Program Counter word address: Word select, for
PCWORD PC[6:0] Z7:Z1 filling temporary buffer (must be zero during Page
Write operation)
Most significant bit in the program counter. (The program
PCMSB 15
counter is 16 bits PC[15:0])
Most significant bit which is used to address the words

PAGEMSB 6 within one page (128 words in a page requires 7 bits PC
[6:0]).
Bit in Z-register that is mapped to PCMSB. Because Z0 is

ZPCMSB Z16
not used, the ZPCMSB equals PCMSB + 1.
Bit in Z-register that is mapped to PAGEMSB. Because Z0

ZPAGEMSB Z7
is not used, the ZPAGEMSB equals PAGEMSB + 1.
Program counter page address: Page select, for Page

PCPAGE PC[15:7] Z16:Z7
Erase and Page Write.
Program counter word address: Word select, for filling

PCWORD PC[6:0] Z7:Z1 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.
See “Addressing the Flash During Self-Programming” on page 360 for details about the use of
Z-pointer during Self-Programming.
367
7593D–AVR–07/06
29. Memory Programming
29.1 Program And Data Memory Lock Bits

The AT90USB64/128 provides six Lock bits which can be left unprogrammed (“1”) or can be pro-
grammed (“0”) to obtain the additional features listed in Table 29-2. The Lock bits can only be
erased to “1” with the Chip Erase command.
Table 29-1. 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)
BLB11 4 Boot Lock bit 0 (programmed)
LB2 1 Lock bit 0 (programmed)
LB1 0 Lock bit 0 (programmed)
Table 29-2. 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 Parallel and Serial Programming mode. The
2 1 0
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
3 0 0
mode. The Boot 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
1 1 1
Application 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 allowed to read from the Application section. If Interrupt
allowed to read from the Application section. If Interrupt
4 0 1
368 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 29-2. 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
1 1 1
Loader 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
3 0 0
Vectors are placed in the Application section, interrupts
section.
(E)LPM executing from the Application section is not
4 0 1 Vectors are placed in the Application section, interrupts
section.
Notes: 1. Program the Fuse bits and Boot Lock bits before programming the LB1 and LB2.
2. “1” means unprogrammed, “0” means programmed
29.2 Fuse Bits

The AT90USB64/128 has four Fuse bytes. Table 29-3 - Table 29-5 describe briefly the function-
ality 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 29-3. Extended Fuse Byte

Fuse Low Byte Bit No Description Default Value
– 7 – 1
– 6 – 1
– 5 – 1
– 4 – 1
HWBE 3 Hardware Boot Enable 0 (programmed)
BODLEVEL2(1) 2 Brown-out Detector trigger level 0 (programmed)
BODLEVEL1(1) 1 Brown-out Detector trigger level 1 (unprogrammed)
(1)
BODLEVEL0 0 Brown-out Detector trigger level 1 (unprogrammed)
Note: 1. See Table 8-2 on page 62 for BODLEVEL Fuse decoding.
369
7593D–AVR–07/06
Table 29-4. Fuse High Byte
Fuse High Byte Bit No Description Default Value
1 (unprogrammed, OCD
OCDEN(4) 7 Enable OCD
disabled)
0 (programmed, JTAG
JTAGEN 6 Enable 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,
EESAVE 3
through the Chip Erase EEPROM not preserved)
Select Boot Size (see Table 29-6
BOOTSZ1 2 0 (programmed)(2)
for details)
Select Boot Size (see Table 29-6
BOOTSZ0 1 0 (programmed)(2)
for 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 28-8 on page 366
for details.
3. See “Watchdog Timer Control Register - WDTCSR” on page 67 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 29-5. Fuse Low Byte
Fuse Low Byte Bit No Description Default Value
(4)
CKDIV8 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 Table 8-1 on page 60 for details.
2. The default setting of CKSEL3..0 results in internal RC Oscillator @ 8 MHz. See Table 6-1 on
page 40 for details.
3. The CKOUT Fuse allow the system clock to be output on PORTC7. See “Clock Output Buffer”
on page 48 for details.
4. See “System Clock Prescaler” on page 48 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.
370 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
29.2.1 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.
29.3 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.
AT90USB64/128 Signature Bytes:

1. 0x000: 0x1E (indicates manufactured by Atmel).
2. 0x001: 0x97 (indicates 128KB Flash memory).
3. 0x002: 0x82 (indicates AT90USBxxxdevice).
29.4 Calibration Byte

The AT90USB64/128 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 calibrated
RC Oscillator.
29.5 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 AT90USB64/128. Pulses are assumed to
be at least 250 ns unless otherwise noted.
29.5.1 Signal Names

In this section, some pins of the AT90USB64/128 are referenced by signal names describing
their functionality during parallel programming, see Figure 29-1 and Table 29-6. 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 29-9.
When pulsing WR or OE, the command loaded determines the action executed. The different
commands are shown in Table 29-10.
371
7593D–AVR–07/06
Figure 29-1. 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 29-6. Pin Name Mapping

Signal Name in
Programming Mode Pin Name I/O Function
0: Device is busy programming, 1: Device is
RDY/BSY PD1 O
ready for 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
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 29-7. BS2 and BS1 Encoding

Flash / Flash Data
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
372 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
,
Table 29-8. 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 29-9. XA1 and XA0 Enoding

XA1 XA0 Action when XTAL1 is Pulsed
Load Flash or EEPROM Address (High or low address byte
0 0
determined 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 29-10. 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
0000 1000 Read Signature Bytes and Calibration byte
0000 0100 Read Fuse and Lock bits
0000 0010 Read Flash
0000 0011 Read EEPROM
Table 29-11. No. of Words in a Page and No. of Pages in the Flash
No. of
Flash Size Page Size PCWORD Pages PCPAGE PCMSB
128K words (256K bytes) 128 words PC[6:0] 1024 PC[16:7] 16
373
7593D–AVR–07/06
Table 29-12. No. of Words in a Page and No. of Pages in the EEPROM
No. of
EEPROM Size Page Size PCWORD Pages PCPAGE EEAMSB
4K bytes 8 bytes EEA[2:0] 512 EEA[11:3] 11
29.6 Parallel Programming
29.6.1 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 29-8 on page 373 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.
29.6.2 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.
29.6.3 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.
374 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
29.6.4 Programming the Flash
The Flash is organized in pages, see Table 29-11 on page 373. 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 29-3 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 29-2 on page 376. 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.
375
7593D–AVR–07/06
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 29-3 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 29-2. 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 29-11 on page 373.
376 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 29-3. 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.
29.6.5 Programming the EEPROM

The EEPROM is organized in pages, see Table 29-12 on page 374. 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 375 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 29-4
for signal waveforms).
377
7593D–AVR–07/06
Figure 29-4. 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
29.6.6 Reading the Flash

The algorithm for reading the Flash memory is as follows (refer to “Programming the Flash” on
page 375 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”.
29.6.7 Reading the EEPROM

The algorithm for reading the EEPROM memory is as follows (refer to “Programming the Flash”
on page 375 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.
29.6.8 Programming the Fuse Low Bits

The algorithm for programming the Fuse Low bits is as follows (refer to “Programming the Flash”
on page 375 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.
378 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
29.6.9 Programming the Fuse High Bits
The algorithm for programming the Fuse High bits is as follows (refer to “Programming the
Flash” on page 375 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.
29.6.10 Programming the Extended Fuse Bits

The algorithm for programming the Extended Fuse bits is as follows (refer to “Programming the
Flash” on page 375 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 29-5. 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
29.6.11 Programming the Lock Bits

The algorithm for programming the Lock bits is as follows (refer to “Programming the Flash” on
page 375 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.
379
7593D–AVR–07/06
29.6.12 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 375 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 29-6. 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
29.6.13 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.
29.6.14 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.
380 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
29.6.15 Parallel Programming Characteristics
Figure 29-7. 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 29-8. 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 29-7 (i.e., tDVXH, tXHXL, and tXLDX) also apply to load-
ing operation.
Figure 29-9. 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
381
7593D–AVR–07/06
Note: 1. The timing requirements shown in Figure 29-7 (i.e., tDVXH, tXHXL, and tXLDX) also apply to read-
ing operation.
Table 29-13. 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
(1)
tWLRH WR Low to RDY/BSY High 3.7 4.5 ms
(2)
tWLRH_CE WR Low to RDY/BSY High for Chip Erase 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.
29.7 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, PDI
(input) and PDO (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 29-14 on
page 383, the pin mapping for serial programming is listed. Not all packages use the SPI pins
dedicated for the internal Serial Peripheral Interface - SPI.
382 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
29.8 Serial Programming Pin Mapping
Table 29-14. Pin Mapping Serial Programming

Pins
Symbol (TQFP-64) I/O Description
PDI PB2 I Serial Data in
PDO PB3 O Serial Data out
SCK PB1 I Serial Clock
Figure 29-10. Serial Programming and Verify(1)
+1.8 - 5.5V
VCC
+1.8 - 5.5V(2)
PDI
AVCC
PDO
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
29.8.1 Serial Programming Algorithm

When writing serial data to the AT90USB64/128, data is clocked on the rising edge of SCK.
When reading data from the AT90USB64/128, data is clocked on the falling edge of SCK. See
Figure 29-11 for timing details.
To program and verify the AT90USB64/128 in the serial programming mode, the following
sequence is recommended (See four byte instruction formats in Table 29-16):
383
7593D–AVR–07/06
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 PDI.
3. The serial programming instructions will not work if the communication is out of syn-
chronization. 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 29-
15.) 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 29-15.) 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
content at the selected address at serial output PDO. 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.
Table 29-15. 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
384 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 29-11. Serial Programming Waveforms
SERIAL DATA INPUT MSB LSB
(MOSI)
SERIAL DATA OUTPUT MSB LSB

(MISO)
SERIAL CLOCK INPUT

(SCK)
SAMPLE
385
7593D–AVR–07/06
Table 29-16. Serial Programming Instruction Set
Instruction Format
Instruction Byte 1 Byte 2 Byte 3 Byte4 Operation
1010 1100 0101 0011 xxxx xxxx xxxx xxxx Enable Serial Programming after
Programming Enable
RESET goes low.
Chip Erase 1010 1100 100x xxxx xxxx xxxx xxxx xxxx Chip Erase EEPROM and Flash.
0100 1101 0000 0000 cccc cccc xxxx xxxx Defines Extended Address Byte for
Load Extended Address Byte Read Program Memory and Write
Program Memory Page.
0010 H000 aaaa aaaa bbbb bbbb oooo oooo Read H (high or low) data o from
Read Program Memory Program memory at word address
c:a:b.
0100 H000 xxxx xxxx xxbb bbbb iiii iiii Write H (high or low) data i to Program
Memory page at word address b. Data
Load Program Memory Page low byte must be loaded before Data
high byte is applied within the same
address.
0100 1100 aaaa aaaa bbxx xxxx xxxx xxxx Write Program Memory Page at
Write Program Memory Page
address c:a:b.
1010 0000 0000 aaaa bbbb bbbb oooo oooo Read data o from EEPROM memory at
Read EEPROM Memory
address a:b.
1100 0000 0000 aaaa bbbb bbbb iiii iiii Write data i to EEPROM memory at
Write EEPROM Memory
address a:b.
1100 0001 0000 0000 0000 00bb iiii iiii Load data i to EEPROM memory page
Load EEPROM Memory
buffer. After data is loaded, program
Page (page access)
EEPROM page.
Write EEPROM Memory 1100 0010 0000 aaaa bbbb bb00 xxxx xxxx
Write EEPROM page at address a:b.
Page (page access)
0101 1000 0000 0000 xxxx xxxx xxoo oooo Read Lock bits. “0” = programmed, “1”
Read Lock bits = unprogrammed. See Table 29-1 on
1010 1100 111x xxxx xxxx xxxx 11ii iiii Write Lock bits. Set bits = “0” to
Write Lock bits program Lock bits. See Table 29-1 on
Read Signature Byte 0011 0000 000x xxxx xxxx xxbb oooo oooo Read Signature Byte o at address b.
1010 1100 1010 0000 xxxx xxxx iiii iiii Set bits = “0” to program, “1” to
Write Fuse bits
unprogram.
Write Fuse High bits
unprogram.
Write Extended Fuse Bits unprogram. See Table 29-3 on page
369 for details.
0101 0000 0000 0000 xxxx xxxx oooo oooo Read Fuse bits. “0” = programmed, “1”
Read Fuse bits
= unprogrammed.
0101 1000 0000 1000 xxxx xxxx oooo oooo Read Fuse High bits. “0” = pro-
Read Fuse High bits
grammed, “1” = unprogrammed.
386 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 29-16. Serial Programming Instruction Set (Continued)
Instruction Format
Instruction Byte 1 Byte 2 Byte 3 Byte4 Operation
0101 0000 0000 1000 xxxx xxxx oooo oooo Read Extended Fuse bits. “0” = pro-
Read Extended Fuse Bits grammed, “1” = unprogrammed. See
Table 29-3 on page 369 for details.
Read Calibration Byte 0011 1000 000x xxxx 0000 0000 oooo oooo Read Calibration Byte
1111 0000 0000 0000 xxxx xxxx xxxx xxxo If o = “1”, a programming operation is
Poll RDY/BSY still busy. Wait until this bit returns to
“0” before applying another command.
Note: 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
29.8.2 Serial Programming Characteristics

For characteristics of the Serial Programming module see “SPI Timing Characteristics” on page
405.
29.9 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.
29.9.1 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 29-12.
387
7593D–AVR–07/06
Figure 29-12. 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
29.9.2 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.
29.9.3 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.
388 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
29.9.4 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
29.9.5 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.
29.9.6 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.
29.9.7 Data Registers

The Data Registers are selected by the JTAG instruction registers described in section “Pro-
gramming Specific JTAG Instructions” on page 387. The Data Registers relevant for
programming operations are:
• Reset Register
• Programming Enable Register
• Programming Command Register
• Flash Data Byte Register
389
7593D–AVR–07/06
29.9.8 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 40) 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 8-1 on page 60.
29.9.9 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 29-13. Programming Enable Register

TDI
0xA370
D
A
T
= D Q Programming Enable
ClockDR & PROG_ENABLE
TDO
29.9.10 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 29-17. The state sequence when shifting
in the programming commands is illustrated in Figure 29-15.
390 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 29-14. 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
391
7593D–AVR–07/06
Table 29-17. 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)
392 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 29-17. 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
4g. Poll for Page Write Complete 0110011_00000000 xxxxxox_xxxxxxxx (2)
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
7b. Load Data Byte(9) 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)
393
7593D–AVR–07/06
Table 29-17. 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
8d. Read Fuse Low Byte(8)
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 29-3 on page 369
7. The bit mapping for Fuses High byte is listed in Table 29-4 on page 370
8. The bit mapping for Fuses Low byte is listed in Table 29-5 on page 370
9. The bit mapping for Lock bits byte is listed in Table 29-1 on page 368
10. Address bits exceeding PCMSB and EEAMSB (Table 29-11 and Table 29-12) are don’t care
11. All TDI and TDO sequences are represented by binary digits (0b...).
394 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 29-15. 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
29.9.11 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-
395
7593D–AVR–07/06
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 29-16. 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.
29.9.12 Programming Algorithm

All references below of type “1a”, “1b”, and so on, refer to Table 29-17.
29.9.13 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.
29.9.14 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.
396 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
29.9.15 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 29-13 on page 382).
29.9.16 Programming the Flash

Before programming the Flash a Chip Erase must be performed, see “Performing Chip Erase”
on page 397.
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 29-13 on page 382).
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 29-11 on page 373) 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, start-
ing 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
Register 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
29.9.17 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:
397
7593D–AVR–07/06
2. Enable Flash read using programming instruction 3a.
3. Load the page address using programming instructions 3b, 3c and 3d. PCWORD (refer
to Table 29-11 on page 373) 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.
29.9.18 Programming the EEPROM

Before programming the EEPROM a Chip Erase must be performed, see “Performing Chip
Erase” on page 397.
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
Note that the PROG_PAGELOAD instruction can not be used when programming the EEPROM.
29.9.19 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.
29.9.20 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
398 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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
29.9.21 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
29.9.22 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.
29.9.23 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.
29.9.24 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.
399
7593D–AVR–07/06
30. Electrical Characteristics
30.1 Absolute Maximum Ratings*

Operating Temperature.................................... -40°C to +85°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
30.2 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 = 10mA, VCC = 5V 0.7
VOL Output Low Voltage(3), V
IOL = 5mA, VCC = 3V 0.5
IOH = -20mA, VCC = 5V 4.2
VOH Output High Voltage(4), V
IOH = -10mA, 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
400 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
TA = -40°C to 85°C, VCC = 1.8V to 5.5V (unless otherwise noted) (Continued)
Symbol Parameter Condition Min.(5) Typ. Max.(5) Units
Active 1MHz, VCC = 2V
0.8 mA
(AT90USB64/128)
5 mA
(AT90USB64/128)
18 mA
(AT90USB64/128)
(6)
Power Supply Current
Idle 1MHz, VCC = 2V
ICC 0.4 0.75 mA
(AT90USB64/128)
Idle 4MHz, VCC = 3V
2.2 mA
(AT90USB64/128)
Idle 8MHz, VCC = 5V
8 mA
(AT90USB64/128)
WDT enabled, VCC = 3V <10 20 µA
Power-down mode
WDT disabled, VCC = 3V <1 3 µA
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:
AT90USB64/128:
1.)The sum of all IOL, for ports A0-A7, G2, C4-C7 should not exceed 100 mA.
2.)The sum of all IOL, for ports C0-C3, G0-G1, D0-D7 should not exceed 100 mA.
3.)The sum of all IOL, for ports G3-G5, B0-B7, E0-E7 should not exceed 100 mA.
4.)The sum of all IOL, for ports F0-F7 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:
AT90USB64/128:
1)The sum of all IOH, for ports A0-A7, G2, C4-C7 should not exceed 100 mA.
2)The sum of all IOH, for ports C0-C3, G0-G1, D0-D7 should not exceed 100 mA.
3)The sum of all IOH, for ports G3-G5, B0-B7, E0-E7 should not exceed 100 mA.
4)The sum of all IOH, for ports F0-F7 should not exceed 100 mA.
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
6. Values with “Power Reduction Register 1 - PRR1” disabled (0x00).
401
7593D–AVR–07/06
30.3 External Clock Drive Waveforms
Figure 30-1. External Clock Drive Waveforms
V IH1
V IL1
30.4 External Clock Drive
Table 30-1. 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.
30.5 Maximum speed vs. VCC

Maximum frequency is depending on VCC. As shown in Figure 30-2, the Maximum Frequency vs.
VCC curve is linear between 2.7V < VCC < 4.5V.
402 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 30-2. Maximum Frequency vs. VCC, AT90USB64/128
30.6 2-wire Serial Interface Characteristics

Table 30-2 describes the requirements for devices connected to the 2-wire Serial Bus. The AT90USB64/128 2-wire Serial
Interface meets or exceeds these requirements under the noted conditions.
Timing symbols refer to Figure 30-3.
Table 30-2. 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

(1) (2)
Vhys Hysteresis of Schmitt Trigger Inputs 0.05 VCC – V
(1)
VOL Output Low-voltage 3 mA sink current 0 0.4 V
(1) (3)(2)
tr Rise Time for both SDA and SCL 20 + 0.1Cb 300 ns
(1)
tof Output Fall Time from VIHmin to VILmax 10 pF < Cb < 400 pF(3) 20 + 0.1Cb(3)(2) 250 ns
(1) (2)
tSP Spikes Suppressed by Input Filter 0 50 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
fSCL > 100 kHz(7) 1.3 – µs
403
7593D–AVR–07/06
Table 30-2. 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 AT90USB64/128, 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 AT90USB64/128 2-wire Serial Interface operation. Other devices connected to the 2-wire
Serial Bus need only obey the general fSCL requirement.
6. The actual low period generated by the AT90USB64/128 2-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 AT90USB64/128 2-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, AT90USB64/128 devices connected to the bus may
communicate at full speed (400 kHz) with other AT90USB64/128 devices, as well as any other device with a proper tLOW
acceptance margin.
Figure 30-3. 2-wire Serial Bus Timing

tof tHIGH tr
tLOW tLOW
SCL
tSU;STA tHD;STA tHD;DAT tSU;DAT
tSU;STO
SDA
tBUF
404 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
30.7 SPI Timing Characteristics
See Figure 30-4 and Figure 30-5 for details.
Table 30-3. SPI Timing Parameters

Description Mode Min Typ Max
1 SCK period Master See Table 17-4
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 30-4. 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)
405
7593D–AVR–07/06
Figure 30-5. 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)
30.8 Hardware Boot EntranceTiming Characteristics
Figure 30-6. Hardware Boot Timing Requirements
RESET
tSHRH tHHRH
ALE/HWB
Table 30-4. Hardware Boot Timings

Symbol Parameter Min Max
tSHRH HWB low Setup before Reset High 0

StartUpTime(S
UT) + Time
tHHRH HWB low Hold after Reset High
Out
Delay(TOUT)
406 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
30.9 ADC Characteristics – Preliminary Data
Table 30-5. ADC Characteristics
Symbol Parameter Condition Min(1) Typ(1) Max(1) Units
Single Ended Conversion 10 Bits
Differential Conversion
8 Bits
Resolution Gain = 1x or 20x
Differential Conversion
7 Bits
Gain = 200x
Single Ended Conversion
VREF = 4V, VCC = 4V, 2 2.5 LSB
ADC clock = 200 kHz
VREF = 4V, VCC = 4V, 4.5 LSB
ADC clock = 1 MHz
Absolute accuracy (Including
INL, DNL, quantization error, Single Ended Conversion
gain and offset error) VREF = 4V, VCC = 4V,
2 LSB
ADC clock = 200 kHz
Noise Reduction Mode
VREF = 4V, VCC = 4V,
4.5 LSB
ADC clock = 1 MHz
Noise Reduction Mode
Integral Non-Linearity (INL) VREF = 4V, VCC = 4V, 0.5 LSB
ADC clock = 200 kHz
Differential Non-Linearity (DNL) VREF = 4V, VCC = 4V, 0.25 LSB
ADC clock = 200 kHz
Gain Error VREF = 4V, VCC = 4V, 2 LSB
ADC clock = 200 kHz
Offset Error VREF = 4V, VCC = 4V, 2 LSB
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
Single Ended Conversion 1.0 AVCC V
VREF Reference Voltage
Differential Conversion 1.0 AVCC - 0.5 V
Single ended channels GND VREF V
VIN Input Voltage
Differential Conversion 0 AVCC V
Single Ended Channels 38,5 kHz
Input Bandwidth
Differential Channels 4 kHz
VINT1 Internal Voltage Reference 1.1V 1.0 1.1 1.2 V
407
7593D–AVR–07/06
Table 30-5. ADC Characteristics (Continued)
Symbol Parameter Condition Min(1) Typ(1) Max(1) Units
VINT2 Internal Voltage Reference 2.56V 2.4 2.56 2.8 V
RREF Reference Input Resistance 32 kΩ
RAIN Analog Input Resistance 100 MΩ
Notes: 1. Values are guidelines only. Actual values are TBD
30.10 External Data Memory Timing

Table 30-6. External Data Memory Characteristics, 4.5 - 5.5 Volts, No Wait-state
8 MHz Oscillator Variable Oscillator
Symbol Parameter Min Max Min Max Unit
0 1/tCLCL Oscillator Frequency 0.0 16 MHz
1 tLHLL ALE Pulse Width 115 1.0tCLCL-10 ns
2 tAVLL Address Valid A to ALE Low 57.5 0.5tCLCL-5(1) ns
Address Hold After ALE Low,
3a tLLAX_ST 5 5 ns
write access
Address Hold after ALE Low,
3b tLLAX_LD 5 5 ns
read access
4 tAVLLC Address Valid C to ALE Low 57.5 0.5tCLCL-5(1) ns
5 tAVRL Address Valid to RD Low 115 1.0tCLCL-10 ns
6 tAVWL Address Valid to WR Low 115 1.0tCLCL-10 ns
7 tLLWL ALE Low to WR Low 47.5 67.5 0.5tCLCL-15(2) 0.5tCLCL+5(2) ns
(2) (2)
8 tLLRL ALE Low to RD Low 47.5 67.5 0.5tCLCL-15 0.5tCLCL+5 ns
9 tDVRH Data Setup to RD High 40 40 ns
10 tRLDV Read Low to Data Valid 75 1.0tCLCL-50 ns
11 tRHDX Data Hold After RD High 0 0 ns
12 tRLRH RD Pulse Width 115 1.0tCLCL-10 ns
(1)
13 tDVWL Data Setup to WR Low 42.5 0.5tCLCL-20 ns
14 tWHDX Data Hold After WR High 115 1.0tCLCL-10 ns
15 tDVWH Data Valid to WR High 125 1.0tCLCL ns
16 tWLWH WR Pulse Width 115 1.0tCLCL-10 ns
Notes: 1. This assumes 50% clock duty cycle. The half period is actually the high time of the external clock, XTAL1.
2. This assumes 50% clock duty cycle. The half period is actually the low time of the external clock, XTAL1.
408 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Table 30-7. External Data Memory Characteristics, 4.5 - 5.5 Volts, 1 Cycle Wait-state
Table 30-8. External Data Memory Characteristics, 4.5 - 5.5 Volts, SRWn1 = 1, SRWn0 = 0
Table 30-10. External Data Memory Characteristics, 2.7 - 5.5 Volts, No Wait-state
1 tLHLL ALE Pulse Width 235 tCLCL-15 ns
(1)
2 tAVLL Address Valid A to ALE Low 115 0.5tCLCL-10 ns
Address Hold After ALE Low,
3a tLLAX_ST 5 5 ns
write access
409
7593D–AVR–07/06
Table 30-10. External Data Memory Characteristics, 2.7 - 5.5 Volts, No Wait-state (Continued)
Address Hold after ALE Low,
3b tLLAX_LD 5 5 ns
read access
4 tAVLLC Address Valid C to ALE Low 115 0.5tCLCL-10(1) ns
5 tAVRL Address Valid to RD Low 235 1.0tCLCL-15 ns
6 tAVWL Address Valid to WR Low 235 1.0tCLCL-15 ns
(2) (2)
7 tLLWL ALE Low to WR Low 115 130 0.5tCLCL-10 0.5tCLCL+5 ns
(2) (2)
8 tLLRL ALE Low to RD Low 115 130 0.5tCLCL-10 0.5tCLCL+5 ns
9 tDVRH Data Setup to RD High 45 45 ns
11 tRHDX Data Hold After RD High 0 0 ns
(1)
13 tDVWL Data Setup to WR Low 105 0.5tCLCL-20 ns
Notes: 1. This assumes 50% clock duty cycle. The half period is actually the high time of the external clock, XTAL1.
2. This assumes 50% clock duty cycle. The half period is actually the low time of the external clock, XTAL1.
410 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 30-7. External Memory Timing (SRWn1 = 0, SRWn0 = 0

T1 T2 T3 T4
ALE
4 7

15
2 3a 13
Write
14
6 16
WR
3b 9 11
DA7:0 (XMBK = 0) Address Data
Read
5 10
8 12
RD
411
7593D–AVR–07/06
Figure 30-8. External Memory Timing (SRWn1 = 0, SRWn0 = 1)
T1 T2 T3 T4 T5
ALE
4 7

15
2 3a 13
Write
14
6 16
WR
3b 9 11
Read
5 10
8 12
RD
Figure 30-9. External Memory Timing (SRWn1 = 1, SRWn0 = 0)

T1 T2 T3 T4 T5 T6
ALE
4 7
15
2 3a 13
Write
14
6 16
WR
3b 9 11
Read
5 10
8 12
RD
412 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
Figure 30-10. External Memory Timing (SRWn1 = 1, SRWn0 = 1)()
T1 T2 T3 T4 T5 T6 T7

1
ALE
4 7
15
2 3a 13
Write
14
6 16
WR
3b 9 11
Read
5 10
8 12
RD
The ALE pulse in the last period (T4-T7) is only present if the next instruction accesses the RAM (internal
or external).
413
7593D–AVR–07/06
31. 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) OTGTCON 1 PAGE VALUE
(0xF8) UPINT PINT7:0
(0xF7) UPBCHX - - - - - PBYCT10:8
(0xF6) UPBCLX PBYCT7:0
(0xF5) UPERRX - COUNTER1:0 CRC16 TIMEOUT PID DATAPID DATATGL
(0xF4) UEINT EPINT6:0
(0xF3) UEBCHX - - - - - BYCT10:8
(0xF2) UEBCLX BYCT7:0
(0xF1) UEDATX DAT7:0
(0xF0) UEIENX FLERRE NAKINE - NAKOUTE RXSTPE RXOUTE STALLEDE TXINE
(0xEF) UESTA1X - - - - - CTRLDIR CURRBK1:0
(0xEE) UESTA0X CFGOK OVERFI UNDERFI ZLPSEEN DTSEQ1:0 NBUSYBK1:0
(0xED) UECFG1X EPSIZE2:0 EPBK1:0 ALLOC
(0xEC) UECFG0X EPTYPE1:0 ISOSW AUTOSW NYETSDIS EPDIR
(0xEB) UECONX STALLRQ STALLRQC RSTDT EPEN
(0xEA) UERST EPRST6:0
(0xE9) UENUM EPNUM2:0
(0xE8) UEINTX FIFOCON NAKINI RWAL NAKOUTI RXSTPI RXOUTI STALLEDI TXINI
(0xE7) UDTST OPMODE2 TSTPCKT TSTK TSTJ
(0xE6) UDMFN FNCERR
(0xE5) UDFNUMH FNUM10:8
(0xE4) UDFNUML FNUM7:0
(0xE3) UDADDR ADDEN UADD6:0
(0xE2) UDIEN UPRSME EORSME WAKEUPE EORSTE SOFE MSOFE SUSPE
(0xE1) UDINT UPRSMI EORSMI WAKEUPI EORSTI SOFI MSOFI SUSPI
(0xE0) UDCON LSM RMWKUP DETACH
(0xDF) OTGINT STOI HNPERRI ROLEEXI BCERRI VBERRI SRPI
(0xDE) OTGIEN STOE HNPERRE ROLEEXE BCERRE VBERRE SRPE
(0xDD) OTGCON 0 HNPREQ SRPREQ SRPSEL VBUSHWC VBUSREQ VBUSRQC
(0xDC) UDPADDH DPACC DPADD10:8
(0xDB) UDPADDL DPADD7:0
(0xDA) USBINT IDTI VBUSTI
(0xD9) USBSTA SPEED ID VBUS
(0xD8) USBCON USBE HOST FRZCLK OTGPADE IDTE VBUSTE
(0xD7) UHWCON UIMOD UIDE UVCONE UVREGE
(0xD6) Reserved
(0xD5) Reserved
(0xD4) Reserved
(0xD3) Reserved
(0xD2) Reserved - - - - - - - -
(0xD1) Reserved - - - - - - - -
(0xD0) Reserved - - - - - - - -
(0xCF) Reserved - - - - - - - -
(0xCE) UDR1 USART1 I/O Data Register
(0xCD) UBRR1H - - - - USART1 Baud Rate Register High Byte
(0xCC) UBRR1L USART1 Baud Rate Register Low Byte
(0xCB) Reserved - - - - - - - -
(0xCA) UCSR1C UMSEL11 UMSEL10 UPM11 UPM10 USBS1 UCSZ11 UCSZ10 UCPOL1
(0xC9) UCSR1B RXCIE1 TXCIE1 UDRIE1 RXEN1 TXEN1 UCSZ12 RXB81 TXB81
(0xC8) UCSR1A RXC1 TXC1 UDRE1 FE1 DOR1 PE1 U2X1 MPCM1
(0xC7) Reserved - - - - - - - -
(0xC6) Reserved - - - - - - - -
(0xC5) Reserved - - - - - - - -
(0xC4) Reserved - - - - - - - -
(0xC3) Reserved - - - - - - - -
(0xC2) Reserved - - - - - - - -
(0xC1) Reserved - - - - - - - -
(0xC0) Reserved - - - - - - - -
(0xBF) Reserved - - - - - - - -
414 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
(0xBE) Reserved - - - - - - - -
(0xBD) TWAMR TWAM6 TWAM5 TWAM4 TWAM3 TWAM2 TWAM1 TWAM0 -
(0xBC) TWCR TWINT TWEA TWSTA TWSTO TWWC TWEN - TWIE
(0xBB) TWDR 2-wire Serial Interface Data Register
(0xBA) TWAR TWA6 TWA5 TWA4 TWA3 TWA2 TWA1 TWA0 TWGCE
(0xB9) TWSR TWS7 TWS6 TWS5 TWS4 TWS3 - TWPS1 TWPS0
(0xB8) TWBR 2-wire Serial Interface Bit Rate Register
(0xB7) Reserved - - - - - - - -
(0xB6) ASSR - EXCLK AS2 TCN2UB OCR2AUB OCR2BUB TCR2AUB TCR2BUB
(0xB5) Reserved - - - - - - - -
(0xB4) OCR2B Timer/Counter2 Output Compare Register B
(0xB3) OCR2A Timer/Counter2 Output Compare Register A
(0xB2) TCNT2 Timer/Counter2 (8 Bit)
(0xB1) TCCR2B FOC2A FOC2B - - WGM22 CS22 CS21 CS20
(0xB0) TCCR2A COM2A1 COM2A0 COM2B1 COM2B0 - - WGM21 WGM20
(0xAF) UPDATX PDAT7:0
(0xAE) UPIENX FLERRE NAKEDE - PERRE TXSTPE TXOUTE RXSTALLE RXINE
(0xAD) UPCFG2X INTFRQ7:0
(0xAC) UPSTAX CFGOK OVERFI UNDERFI DTSEQ1:0 NBUSYBK1:0
(0xAB) UPCFG1X PSIZE2:0 PBK1:0 ALLOC
(0xAA) UPCFG0X PTYPE1:0 PTOKEN1:0 PEPNUM3:0
(0xA9) UPCONX PFREEZE INMODE AUTOSW RSTDT PEN
(0xA8) UPRST PRST6:0
(0xA7) UPNUM PNUM2:0
(0xA6) UPINTX FIFOCON NAKEDI RWAL PERRI TXSTPI TXOUTI RXSTALLI RXINI
(0xA5) UPINRQX INRQ7:0
(0xA4) UHFLEN FLEN7:0
(0xA3) UHFNUMH FNUM10:8
(0xA2) UHFNUML FNUM7:0
(0xA1) UHADDR HADD6:0
(0xA0) UHIEN UPRSME EORSME WAKEUPE EORSTE SOFE MSOFE SUSPE
(0x9F) UHINT HWUPI HSOFI RXRSMI RSMEDI RSTI DDISCI DCONNI
(0x9E) UHCON RESUME RESET SOFEN
(0x9D) OCR3CH Timer/Counter3 - Output Compare Register C High Byte
(0x9C) OCR3CL Timer/Counter3 - Output Compare Register C Low Byte
(0x9B) OCR3BH Timer/Counter3 - Output Compare Register B High Byte
(0x9A) OCR3BL Timer/Counter3 - Output Compare Register B Low Byte
(0x99) OCR3AH Timer/Counter3 - Output Compare Register A High Byte
(0x98) OCR3AL Timer/Counter3 - Output Compare Register A Low Byte
(0x97) ICR3H Timer/Counter3 - Input Capture Register High Byte
(0x96) ICR3L Timer/Counter3 - Input Capture Register Low Byte
(0x95) TCNT3H Timer/Counter3 - Counter Register High Byte
(0x94) TCNT3L Timer/Counter3 - Counter Register Low Byte
(0x93) Reserved - - - - - - - -
(0x92) TCCR3C FOC3A FOC3B FOC3C - - - - -
(0x91) TCCR3B ICNC3 ICES3 - WGM33 WGM32 CS32 CS31 CS30
(0x90) TCCR3A COM3A1 COM3A0 COM3B1 COM3B0 COM3C1 COM3C0 WGM31 WGM30
(0x8F) Reserved - - - - - - - -
(0x8E) Reserved - - - - - - - -
(0x8D) OCR1CH Timer/Counter1 - Output Compare Register C High Byte
(0x8C) OCR1CL Timer/Counter1 - Output Compare Register C Low Byte
(0x8B) OCR1BH Timer/Counter1 - Output Compare Register B High Byte
(0x8A) OCR1BL Timer/Counter1 - Output Compare Register B Low Byte
(0x89) OCR1AH Timer/Counter1 - Output Compare Register A High Byte
(0x88) OCR1AL Timer/Counter1 - Output Compare Register A Low Byte
(0x87) ICR1H Timer/Counter1 - Input Capture Register High Byte
(0x86) ICR1L Timer/Counter1 - Input Capture Register Low Byte
(0x85) TCNT1H Timer/Counter1 - Counter Register High Byte
(0x84) TCNT1L Timer/Counter1 - Counter Register Low Byte
(0x83) Reserved - - - - - - - -
(0x82) TCCR1C FOC1A FOC1B FOC1C - - - - -
(0x81) TCCR1B ICNC1 ICES1 - WGM13 WGM12 CS12 CS11 CS10
(0x80) TCCR1A COM1A1 COM1A0 COM1B1 COM1B0 COM1C1 COM1C0 WGM11 WGM10
(0x7F) DIDR1 - - - - - - AIN1D AIN0D
(0x7E) DIDR0 ADC7D ADC6D ADC5D ADC4D ADC3D ADC2D ADC1D ADC0D
(0x7D) - - - - - - - - -
415
7593D–AVR–07/06
(0x7C) ADMUX REFS1 REFS0 ADLAR MUX4 MUX3 MUX2 MUX1 MUX0
(0x7B) ADCSRB - ACME - - MUX5 ADTS2 ADTS1 ADTS0
(0x7A) ADCSRA ADEN ADSC ADATE ADIF ADIE ADPS2 ADPS1 ADPS0
(0x79) ADCH ADC Data Register High byte
(0x78) ADCL ADC Data Register Low byte
(0x77) Reserved - - - - - - - -
(0x76) Reserved - - - - - - - -
(0x75) XMCRB XMBK - - - - XMM2 XMM1 XMM0
(0x74) XMCRA SRE SRL2 SRL1 SRL0 SRW11 SRW10 SRW01 SRW00
(0x73) Reserved - - - - - - - -
(0x72) Reserved - - - - - - - -
(0x71) TIMSK3 - - ICIE3 - OCIE3C OCIE3B OCIE3A TOIE3
(0x70) TIMSK2 - - - - - OCIE2B OCIE2A TOIE2
(0x6F) TIMSK1 - - ICIE1 - OCIE1C OCIE1B OCIE1A TOIE1
(0x6E) TIMSK0 - - - - - OCIE0B OCIE0A TOIE0
(0x6D) Reserved - - - - - - - -
(0x6C) Reserved - - - - - - - -
(0x6B) PCMSK0 PCINT7 PCINT6 PCINT5 PCINT4 PCINT3 PCINT2 PCINT1 PCINT0
(0x6A) EICRB ISC71 ISC70 ISC61 ISC60 ISC51 ISC50 ISC41 ISC40
(0x69) EICRA ISC31 ISC30 ISC21 ISC20 ISC11 ISC10 ISC01 ISC00
(0x68) PCICR - - - - - - - PCIE0
(0x67) Reserved - - - - - - - -
(0x66) OSCCAL Oscillator Calibration Register
(0x65) PRR1 PRUSB - - - PRTIM3 - - PRUSART1
(0x64) PRR0 PRTWI PRTIM2 PRTIM0 - PRTIM1 PRSPI - PRADC
(0x63) Reserved - - - - - - - -
(0x62) Reserved - - - - - - - -
(0x61) CLKPR CLKPCE - - - CLKPS3 CLKPS2 CLKPS1 CLKPS0
(0x60) WDTCSR WDIF WDIE WDP3 WDCE WDE WDP2 WDP1 WDP0
0x3F (0x5F) SREG I T H S V N Z C
0x3E (0x5E) SPH SP15 SP14 SP13 SP12 SP11 SP10 SP9 SP8
0x3D (0x5D) SPL SP7 SP6 SP5 SP4 SP3 SP2 SP1 SP0
0x3C (0x5C) Reserved - - - - - - - -
0x3B (0x5B) RAMPZ - - - - - - RAMPZ1 RAMPZ0
0x3A (0x5A) Reserved - - - - - - - -
0x39 (0x59) Reserved - - - - - - - -
0x38 (0x58) Reserved - - - - - - - -
0x37 (0x57) SPMCSR SPMIE RWWSB SIGRD RWWSRE BLBSET PGWRT PGERS SPMEN
0x36 (0x56) Reserved - - - - - - - -
0x35 (0x55) MCUCR JTD - - PUD - - IVSEL IVCE
0x34 (0x54) MCUSR - - - JTRF WDRF BORF EXTRF PORF
0x33 (0x53) SMCR - - - - SM2 SM1 SM0 SE
0x32 (0x52) Reserved - - - - - - - -
OCDR/ OCDR7 OCDR6 OCDR5 OCDR4 OCDR3 OCDR2 OCDR1 OCDR0
0x31 (0x51) MONDR Monitor Data Register
0x30 (0x50) ACSR ACD ACBG ACO ACI ACIE ACIC ACIS1 ACIS0
0x2F (0x4F) Reserved - - - - - - - -
0x2E (0x4E) SPDR SPI Data Register
0x2D (0x4D) SPSR SPIF WCOL - - - - - SPI2X
0x2C (0x4C) SPCR SPIE SPE DORD MSTR CPOL CPHA SPR1 SPR0
0x2B (0x4B) GPIOR2 General Purpose I/O Register 2
0x2A (0x4A) GPIOR1 General Purpose I/O Register 1
0x29 (0x49) PLLCSR - - - PLLP2 PLLP1 PLLP0 PLLE PLOCK
0x28 (0x48) OCR0B Timer/Counter0 Output Compare Register B
0x27 (0x47) OCR0A Timer/Counter0 Output Compare Register A
0x26 (0x46) TCNT0 Timer/Counter0 (8 Bit)
0x25 (0x45) TCCR0B FOC0A FOC0B - - WGM02 CS02 CS01 CS00
0x24 (0x44) TCCR0A COM0A1 COM0A0 COM0B1 COM0B0 - - WGM01 WGM00
0x23 (0x43) GTCCR TSM - - - - - PSRASY PSRSYNC
0x22 (0x42) EEARH - - - - EEPROM Address Register High Byte
0x21 (0x41) EEARL EEPROM Address Register Low Byte
0x20 (0x40) EEDR EEPROM Data Register
0x1F (0x3F) EECR - - EEPM1 EEPM0 EERIE EEMPE EEPE EERE
0x1E (0x3E) GPIOR0 General Purpose I/O Register 0
0x1D (0x3D) EIMSK INT7 INT6 INT5 INT4 INT3 INT2 INT1 INT0
0x1C (0x3C) EIFR INTF7 INTF6 INTF5 INTF4 INTF3 INTF2 INTF1 INTF0
416 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
0x1B (0x3B) PCIFR - - - - - - - PCIF0
0x1A (0x3A) Reserved - - - - - - - -
0x19 (0x39) Reserved - - - - - - - -
0x18 (0x38) TIFR3 - - ICF3 - OCF3C OCF3B OCF3A TOV3
0x17 (0x37) TIFR2 - - - - - OCF2B OCF2A TOV2
0x16 (0x36) TIFR1 - - ICF1 - OCF1C OCF1B OCF1A TOV1
0x15 (0x35) TIFR0 - - - - - OCF0B OCF0A TOV0
0x14 (0x34) Reserved - - - - - - - -
0x13 (0x33) Reserved - - - - - - - -
0x12 (0x32) Reserved - - - - - - - -
0x11 (0x31) PORTF PORTF7 PORTF6 PORTF5 PORTF4 PORTF3 PORTF2 PORTF1 PORTF0
0x10 (0x30) DDRF DDF7 DDF6 DDF5 DDF4 DDF3 DDF2 DDF1 DDF0
0x0F (0x2F) PINF PINF7 PINF6 PINF5 PINF4 PINF3 PINF2 PINF1 PINF0
0x0E (0x2E) PORTE PORTE7 PORTE6 PORTE5 PORTE4 PORTE3 PORTE2 PORTE1 PORTE0
0x0D (0x2D) DDRE DDE7 DDE6 DDE5 DDE4 DDE3 DDE2 DDE1 DDE0
0x0C (0x2C) PINE PINE7 PINE6 PINE5 PINE4 PINE3 PINE2 PINE1 PINE0
0x0B (0x2B) PORTD PORTD7 PORTD6 PORTD5 PORTD4 PORTD3 PORTD2 PORTD1 PORTD0
0x0A (0x2A) DDRD DDD7 DDD6 DDD5 DDD4 DDD3 DDD2 DDD1 DDD0
0x09 (0x29) PIND PIND7 PIND6 PIND5 PIND4 PIND3 PIND2 PIND1 PIND0
0x08 (0x28) PORTC PORTC7 PORTC6 PORTC5 PORTC4 PORTC3 PORTC2 PORTC1 PORTC0
0x07 (0x27) DDRC DDC7 DDC6 DDC5 DDC4 DDC3 DDC2 DDC1 DDC0
0x06 (0x26) PINC PINC7 PINC6 PINC5 PINC4 PINC3 PINC2 PINC1 PINC0
0x05 (0x25) PORTB PORTB7 PORTB6 PORTB5 PORTB4 PORTB3 PORTB2 PORTB1 PORTB0
0x04 (0x24) DDRB DDB7 DDB6 DDB5 DDB4 DDB3 DDB2 DDB1 DDB0
0x03 (0x23) PINB PINB7 PINB6 PINB5 PINB4 PINB3 PINB2 PINB1 PINB0
0x02 (0x22) PORTA PORTA7 PORTA6 PORTA5 PORTA4 PORTA3 PORTA2 PORTA1 PORTA0
0x01 (0x21) DDRA DDA7 DDA6 DDA5 DDA4 DDA3 DDA2 DDA1 DDA0
0x00 (0x20) PINA PINA7 PINA6 PINA5 PINA4 PINA3 PINA2 PINA1 PINA0
Note: 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 reg-
isters as data space using LD and ST instructions, $20 must be added to these addresses. The AT90USB64/128 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 - $1FF in SRAM, only the ST/STS/STD and LD/LDS/LDD
instructions can be used.
417
7593D–AVR–07/06
32. 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
EIJMP Extended Indirect Jump to (Z) PC ←(EIND: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
EICALL Extended Indirect Call to (Z) PC ←(EIND: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
418 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
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
419
7593D–AVR–07/06
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
420 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
33. Ordering Information
Table 33-1. Possible Order Entries
USB Speed Power Supply
Ordering Code interface (MHz) (V) Package Operation Range Product Marking
Industrial (-40° to +85°C)

AT90USB1287-16AU OTG 8-16 2.7 - 5.5 TQFP64 90USB1287-16AU
Green
AT90USB1287-16MU OTG 8-16 2.7 - 5.5 QFN64 90USB1287-16MU
Green
Device Industrial (-40° to +85°C)
AT90USB1286-16MU 8-16 2.7 - 5.5 QFN64 90USB1286-16MU
only Green
AT90USB647-16AU OTG 8-16 2.7 - 5.5 TQFP64 90USB1287-16AU
Green
AT90USB647-16MU OTG 8-16 2.7 - 5.5 QFN64 90USB1287-16MU
Green
Device Industrial (-40° to +85°C)
AT90USB646-16MU 8-16 2.7 - 5.5 QFN64 90USB1286-16MU
only Green
421
7593D–AVR–07/06
33.1 TQFP64
422 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
33.2 QFN64
423
7593D–AVR–07/06
34. Errata
The revision letter in this section refers to the revision of the AT90USB64/128 device.
34.1 AT90USB1287/6 Rev B

• USB signal rate
• Spike on TWI pins when TWI is enabled
• High current consumption in sleep mode
• Async timer interrupt wake up from sleep generate multiple interrupts
4. UBS signal rate
The average USB signal rate may sometime be measured out of the USB specifications
(12MHz ±30kHz) with short frames. When measured on a long period, the average signal
rate value complies with the specifications. This bit rate deviation does not generates com-
munication or functional errors.
Problem fix/workaround
None.
3. Spike on TWI pins when TWI is enabled

100 ns negative spike occurs on SDA and SCL pins when TWI is enabled.
Problem Fix/workaround
No known workaround, enable AT90USB64/128 TWI first versus the others nodes of the
TWI network.
2. High current consumption in sleep mode

If a pending interrupt cannot wake the part up from the selected mode, the current consump-
tion will increase during sleep when executing the SLEEP instruction directly after a SEI
instruction.
Before entering sleep, interrupts not used to wake up the part from the sleep mode should
be disabled.
1. Asynchronous timer interrupt wake up from sleep generates multiple interrupts

If the CPU core is in sleep and wakes-up from an asynchronous timer interrupt and then go
back in sleep mode again it may wake up several times.
A software workaround is to wait with performing the sleep instruction until
TCNT2>OCR2+1.
424 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
34.2 AT90USB1287/6 Rev A
• USB signal rate
• VBUS residual level
• Spike on TWI pins when TWI is enabled
• High current consumption in sleep mode
• Async timer interrupt wake up from sleep generate multiple interrupts
5. UBS signal rate
The average USB signal rate may sometime be measured out of the USB specifications
(12MHz ±30kHz) with short frames. When measured on a long period, the average signal
rate value complies with the specifications. This bit rate deviation does not generates com-
munication or functional errors.
None.
4. VBUS residual level
In USB device and host mode, once a 5V level has been detected to the VBUS pad, a resid-
ual level (about 3V) can be measured on the VBUS pin.
None.
3. Spike on TWI pins when TWI is enabled

100 ns negative spike occurs on SDA and SCL pins when TWI is enabled.
No known workaround, enable AT90USB64/128 TWI first versus the others nodes of the
TWI network.
2. High current consumption in sleep mode

If a pending interrupt cannot wake the part up from the selected mode, the current consump-
tion will increase during sleep when executing the SLEEP instruction directly after a SEI
instruction.
Before entering sleep, interrupts not used to wake up the part from the sleep mode should
be disabled.
1. Asyncrhonous timer interrupt wake up from sleep generates multiple interrupts

If the CPU core is in sleep and wakes-up from an asynchronous timer interrupt and then go
back in sleep again it may wake up multiple times.
A software workaround is to wait with performing the sleep instruction until
TCNT2>OCR2+1.
425
7593D–AVR–07/06
35. Datasheet Revision History for AT90USB64/128
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.
35.1 Changes from 7593A to 7593B

1. Changed default configuration for fuse bytes and security byte.
2. Suppression of timer 4,5 registers which does not exist.
Updated typical application schematics in USB section
35.2 Changes from 7593B to 7593C

1. Update to package drawings, MQFP64 and TQFP64.
35.3 Changes from 7593C to 7593D

1. For further product compatibility, changed USB PLL possible prescaler configurations.
Only 8MHz and 16MHz crystal frequencies allows USB operation (See Table 6-13 on
page 51).
426 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
1 Pin Configurations ................................................................................... 3
1.1Disclaimer ..................................................................................................................4
2 Overview ................................................................................................... 4
2.1Block Diagram ...........................................................................................................5
2.2Pin Descriptions .........................................................................................................6
3 About Code Examples ............................................................................. 8
4 AVR CPU Core ........................................................................................ 10

4.1Introduction ..............................................................................................................10
4.2Architectural Overview .............................................................................................10
4.3ALU – Arithmetic Logic Unit .....................................................................................11
4.4Status Register ........................................................................................................12
4.5General Purpose Register File ................................................................................13
4.6Stack Pointer ...........................................................................................................14
4.7Instruction Execution Timing ...................................................................................15
4.8Reset and Interrupt Handling ...................................................................................16
5 AVR AT90USB64/128 Memories ........................................................... 19

5.1In-System Reprogrammable Flash Program Memory .............................................19
5.2SRAM Data Memory ................................................................................................20
5.3EEPROM Data Memory ..........................................................................................23
5.4I/O Memory ..............................................................................................................29
5.5External Memory Interface ......................................................................................30
6 System Clock and Clock Options ......................................................... 39

6.1Clock Systems and their Distribution .......................................................................39
6.2Clock Sources .........................................................................................................40
6.3Low Power Crystal Oscillator ...................................................................................41
6.4Low Frequency Crystal Oscillator ............................................................................44
6.5Calibrated Internal RC Oscillator .............................................................................44
6.6128 kHz Internal Oscillator ......................................................................................46
6.7External Clock .........................................................................................................47
6.8Clock Output Buffer .................................................................................................48
6.9Timer/Counter Oscillator ..........................................................................................48
6.10System Clock Prescaler ........................................................................................48
6.11PLL ........................................................................................................................50
7 Power Management and Sleep Modes ................................................. 53
1
7593D–AVR–07/06
7.1Idle Mode .................................................................................................................54
7.2ADC Noise Reduction Mode ....................................................................................54
7.3Power-down Mode ...................................................................................................54
7.4Power-save Mode ....................................................................................................54
7.5Standby Mode .........................................................................................................55
7.6Extended Standby Mode .........................................................................................55
7.7Power Reduction Register .......................................................................................56
7.8Minimizing Power Consumption ..............................................................................57
8 System Control and Reset .................................................................... 59

8.1Internal Voltage Reference ......................................................................................64
8.2Watchdog Timer ......................................................................................................64
9 Interrupts ................................................................................................ 70
9.1Interrupt Vectors in AT90USB64/128 ......................................................................70
10 I/O-Ports .................................................................................................. 74
10.1Introduction ............................................................................................................74
10.2Ports as General Digital I/O ...................................................................................75
10.3Alternate Port Functions ........................................................................................79
10.4Register Description for I/O-Ports ..........................................................................92
11 External Interrupts ................................................................................. 96
12 Timer/Counter0, Timer/Counter1, and Timer/Counter3 Prescalers . 100

12.1Internal Clock Source ..........................................................................................100
12.2Prescaler Reset ...................................................................................................100
12.3External Clock Source .........................................................................................100
12.4General Timer/Counter Control Register – GTCCR ............................................101
13 8-bit Timer/Counter0 with PWM .......................................................... 102

13.1Overview ..............................................................................................................102
13.2Timer/Counter Clock Sources .............................................................................103
13.3Counter Unit ........................................................................................................103
13.4Output Compare Unit ...........................................................................................104
13.5Compare Match Output Unit ................................................................................106
13.6Modes of Operation .............................................................................................107
13.7Timer/Counter Timing Diagrams .........................................................................111
13.88-bit Timer/Counter Register Description ............................................................112
14 16-bit Timer/Counter (Timer/Counter1 and Timer/Counter3) ........... 119
2 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
14.1Overview ..............................................................................................................119
14.2Accessing 16-bit Registers ..................................................................................121
14.4Counter Unit ........................................................................................................124
14.5Input Capture Unit ...............................................................................................126
14.6Output Compare Units .........................................................................................128
14.1016-bit Timer/Counter Register Description ........................................................140
15 8-bit Timer/Counter2 with PWM and Asynchronous Operation ...... 151

15.1Overview ..............................................................................................................151
15.3Counter Unit ........................................................................................................152
15.4Output Compare Unit ...........................................................................................153
15.88-bit Timer/Counter Register Description ............................................................162
15.9Asynchronous operation of the Timer/Counter ....................................................167
15.10Timer/Counter Prescaler ...................................................................................171
16 Output Compare Modulator (OCM1C0A) ........................................... 172

16.1Overview ..............................................................................................................172
16.2Description ...........................................................................................................172
17 Serial Peripheral Interface – SPI ......................................................... 174

17.1SS Pin Functionality ............................................................................................178
17.2Data Modes .........................................................................................................181
18 USART ................................................................................................... 183

18.1Overview ..............................................................................................................183
18.2Clock Generation .................................................................................................184
18.3Frame Formats ....................................................................................................187
18.4USART Initialization .............................................................................................189
18.5Data Transmission – The USART Transmitter ....................................................190
18.6Data Reception – The USART Receiver .............................................................192
18.7Asynchronous Data Reception ............................................................................196
3
7593D–AVR–07/06
18.8Multi-processor Communication Mode ................................................................199
18.9USART Register Description ...............................................................................200
18.10Examples of Baud Rate Setting .........................................................................205
19 USART in SPI Mode ............................................................................. 208

19.1Overview ..............................................................................................................208
19.2Clock Generation .................................................................................................209
19.3SPI Data Modes and Timing ................................................................................209
19.4Frame Formats ....................................................................................................210
19.5Data Transfer .......................................................................................................212
19.6USART MSPIM Register Description ..................................................................214
19.7AVR USART MSPIM vs. AVR SPI ......................................................................216
20 2-wire Serial Interface .......................................................................... 218

20.1Features ..............................................................................................................218
20.22-wire Serial Interface Bus Definition ..................................................................218
20.3Data Transfer and Frame Format ........................................................................219
20.4Multi-master Bus Systems, Arbitration and Synchronization ...............................222
20.5Overview of the TWI Module ...............................................................................223
20.6TWI Register Description .....................................................................................226
20.7Using the TWI ......................................................................................................229
20.8Transmission Modes ...........................................................................................233
20.9Multi-master Systems and Arbitration ..................................................................246
21 USB controller ...................................................................................... 248

21.1Features ..............................................................................................................248
21.2Block Diagram .....................................................................................................248
21.3Typical Application Implementation .....................................................................249
21.4General Operating Modes ...................................................................................253
21.5Power modes .......................................................................................................257
21.6Speed Control ......................................................................................................258
21.7Memory management ..........................................................................................259
21.8PAD suspend .......................................................................................................260
21.9OTG timers customizing ......................................................................................261
21.10Plug-in detection ................................................................................................261
21.11ID detection .......................................................................................................262
21.12Registers description .........................................................................................263
21.13USB Software Operating modes .......................................................................267
4 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
22 USB Device Operating modes ............................................................ 269
22.1Introduction ..........................................................................................................269
22.2Power-on and reset .............................................................................................269
22.3Speed identification on startup ............................................................................269
22.4Endpoint reset .....................................................................................................270
22.5USB reset ............................................................................................................270
22.6Endpoint selection ...............................................................................................270
22.7Endpoint activation ..............................................................................................270
22.8Address Setup .....................................................................................................271
22.9Suspend, Wake-up and Resume .........................................................................272
22.10Detach ...............................................................................................................272
22.11Remote Wake-up ...............................................................................................273
22.12STALL request ...................................................................................................273
22.13CONTROL endpoint management ....................................................................274
22.14OUT endpoint management ..............................................................................275
22.15IN endpoint management ..................................................................................277
22.16Isochronous mode .............................................................................................278
22.17Overflow ............................................................................................................279
22.18Interrupts ...........................................................................................................279
22.19Registers ...........................................................................................................281
23 USB Host Operating Modes ................................................................ 293

23.1Pipe description ...................................................................................................293
23.2Detach .................................................................................................................293
23.3Power-on and Reset ............................................................................................293
23.4Device Detection .................................................................................................294
23.5Pipe Selection ......................................................................................................294
23.6Pipe Configuration ...............................................................................................294
23.7USB Reset ...........................................................................................................296
23.8Address Setup .....................................................................................................296
23.9Remote Wake-Up detection ................................................................................296
23.10USB Pipe Reset .................................................................................................296
23.11Pipe Data Access ..............................................................................................296
23.12Control Pipe management .................................................................................297
23.13OUT Pipe management .....................................................................................297
23.14IN Pipe management .........................................................................................298
23.15Interrupt system .................................................................................................299
5
7593D–AVR–07/06
23.16Registers ...........................................................................................................300
24 Analog Comparator .............................................................................. 313

24.1Analog Comparator Multiplexed Input .................................................................315
25 Analog to Digital Converter - ADC ...................................................... 316

25.1Features ..............................................................................................................316
25.2Operation .............................................................................................................317
25.3Starting a Conversion ..........................................................................................318
25.4Prescaling and Conversion Timing ......................................................................319
25.5Changing Channel or Reference Selection .........................................................322
25.6ADC Noise Canceler ...........................................................................................323
25.7ADC Conversion Result .......................................................................................327
25.8ADC Register Description ....................................................................................329
26 JTAG Interface and On-chip Debug System ...................................... 335

26.1Overview ..............................................................................................................335
26.2Test Access Port – TAP ......................................................................................335
26.3TAP Controller .....................................................................................................337
26.4Using the Boundary-scan Chain ..........................................................................338
26.5Using the On-chip Debug System .......................................................................338
26.6On-chip Debug Specific JTAG Instructions .........................................................339
26.7On-chip Debug Related Register in I/O Memory .................................................340
26.8Using the JTAG Programming Capabilities .........................................................340
26.9Bibliography .........................................................................................................340
27 IEEE 1149.1 (JTAG) Boundary-scan ................................................... 341

27.1Features ..............................................................................................................341
27.2System Overview .................................................................................................341
27.3Data Registers .....................................................................................................341
27.4Boundary-scan Specific JTAG Instructions .........................................................343
27.5Boundary-scan Related Register in I/O Memory .................................................344
27.6Boundary-scan Chain ..........................................................................................345
27.7AT90USB64/128 Boundary-scan Order ..............................................................348
27.8Boundary-scan Description Language Files ........................................................351
28 Boot Loader Support – Read-While-Write Self-Programming ......... 352

28.1Boot Loader Features ..........................................................................................352
28.2Application and Boot Loader Flash Sections .......................................................352
6 AT90USB64/128
7593D–AVR–07/06
AT90USB64/128
28.3Read-While-Write and No Read-While-Write Flash Sections ..............................352
28.4Boot Loader Lock Bits .........................................................................................355
28.5Entering the Boot Loader Program ......................................................................356
28.6Addressing the Flash During Self-Programming .................................................360
28.7Self-Programming the Flash ................................................................................361
29 Memory Programming ......................................................................... 368

29.1Program And Data Memory Lock Bits .................................................................368
29.2Fuse Bits ..............................................................................................................369
29.3Signature Bytes ...................................................................................................371
29.4Calibration Byte ...................................................................................................371
29.5Parallel Programming Parameters, Pin Mapping, and Commands .....................371
29.6Parallel Programming ..........................................................................................374
29.7Serial Downloading ..............................................................................................382
29.8Serial Programming Pin Mapping ........................................................................383
29.9Programming via the JTAG Interface ..................................................................387
30 Electrical Characteristics .................................................................... 400

30.1Absolute Maximum Ratings* ...............................................................................400
30.2DC Characteristics ...............................................................................................400
30.3External Clock Drive Waveforms .........................................................................402
30.4External Clock Drive ............................................................................................402
30.5Maximum speed vs. VCC ..............................................................................................................................402
30.62-wire Serial Interface Characteristics .................................................................403
30.7SPI Timing Characteristics ..................................................................................405
30.8Hardware Boot EntranceTiming Characteristics ..................................................406
30.9ADC Characteristics – Preliminary Data ..............................................................407
30.10External Data Memory Timing ...........................................................................408
31 Register Summary ............................................................................... 414
32 Instruction Set Summary ..................................................................... 418
33 Ordering Information ........................................................................... 421

33.1TQFP64 ...............................................................................................................422
33.2QFN64 .................................................................................................................423
34 Errata ..................................................................................................... 424

34.1Rev B ...................................................................................................................424
34.2Rev A ...................................................................................................................424
7
7593D–AVR–07/06
35 Datasheet Revision History for AT90USB64/128 .............................. 426
35.1Changes from 7593A to 7593B ...........................................................................426
35.2Changes from 7593B to 7593C ...........................................................................426
35.3Changes from 7593C to 7593D ...........................................................................426
8 AT90USB64/128
7593D–AVR–07/06
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 anyintellectu-
alproperty 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 STATUTORYWAR-
RANTY RELATING TO ITS PRODUCTS INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICU-
LARPURPOSE, 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 INFORMA-
TION) ARISING OUTOF THE USE OR INABILITY TO USE THIS DOCUMENT, EVEN IF ATMEL HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAM-
AGES. Atmel makes norepresentationsor warranties with respect to the accuracy or completeness of the contents of this document and reserves the right to make
changes to specificationsand 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 useas components in applications intended to support or sustainlife.
© Atmel Corporation 2006. All rights reserved. Atmel ®, logo and combinations thereof, are registered trademarks, and Everywhere You Are ®
are the trademarks of Atmel Corporation or its subsidiaries. Other terms and product names may be trademarks of others.
Printed on recycled paper.
7593D–AVR–07/06

90usb646 647 1286 1287

Uploaded by

Copyright:

Available Formats

90usb646 647 1286 1287

Uploaded by

Document Information

Original Title

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

90usb646 647 1286 1287

Uploaded by

Copyright:

Available Formats

Features

• High Performance, Low Power AVR® 8-Bit Microcontroller

(INT.7/AIN.1/UVcon) PE7 2 47 PA4 (AD4)

UGnd 6 43 PE2 (ALE/HWB)

UCap 7 42 PC7 (A15/IC.3/CLK0)

VBus 8 AVR USB 41 PC6 (A14/OC.3A)

(IUID) PE3 9 TQFP64 40 PC5 (A13/OC.3B)

(SS/PCINT0) PB0 10 39 PC4 (A12/OC.3C)

(PCINT1/SCLK) PB1 11 38 PC3 (A11/T.3)

(PDI/PCINT2/MOSI) PB2 12 37 PC2 (A10)

(PDO/PCINT3/MISO) PB3 13 36 PC1 (A9)

(PCINT4/OC.2A) PB4 14 35 PC0 (A8)

(PCINT5/OC.1A) PB5 15 34 PE1 (RD)

(PCINT6/OC.1B) PB6 16 33 PE0 (WR)

2.1 Block Diagram

Figure 2-1. Block Diagram

VCC PORTF DRIVERS PORTA DRIVERS PORTC DRIVERS

8-BIT DATA BUS

PROGRAM SRAM MCU CONTROL TIMING AND

PORTE DRIVERS PORTB DRIVERS PORTD DRIVERS PORTG DRIVERS

PE7 - PE0 PB7 - PB0 PD7 - PD0 PG4 - PG0

2.2 Pin Descriptions

2.2.3 Port A (PA7..PA0)

2.2.5 Port C (PC7..PC0)

2.2.6 Port D (PD7..PD0)

2.2.7 Port E (PE7..PE0)

2.2.8 Port F (PF7..PF0)

3. About Code Examples

4.2 Architectural Overview

Figure 4-1. Block Diagram of the AVR Architecture

Data Bus 8-bit

Data I/O Module 2

4.3 ALU – Arithmetic Logic Unit

• 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

4.5 General Purpose Register File

Figure 4-2. AVR CPU General Purpose Working Registers

4.5.1 The X-register, Y-register, and Z-register

4.6 Stack Pointer

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

4.7 Instruction Execution Timing

Figure 4-5. The Parallel Instruction Fetches and Instruction Executions

Register Operands Fetch

ALU Operation Execute

Result Write Back

4.8 Reset and Interrupt Handling

Assembly Code Example

4.8.1 Interrupt Response Time

Table 5-1. Memory Mapping.

Notes: 1. Byte address.

5.1 In-System Reprogrammable Flash Program Memory

Figure 5-1. Program Memory Map

Application Flash Section

Boot Flash Section