8693B–DFLASH–5/12
Features
Single 2.7V - 3.6V supply
Serial Peripheral Interface (SPI) compatible
Supports SPI Modes 0 and 3
Supports Atmel® RapidS™ operation
Supports Dual-Input Program and Dual-Output Read
Very high operating frequencies
100MHz for RapidS
85MHz for SPI
Clock-to-output time (tV) of 5ns maximum
Flexible, optimized erase architecture for code + data storage applications
Uniform 4KB, 32KB, and 64KB Block Erase
Full Chip Erase
Individual sector protection with Global Protect/Unprotect feature
128 Sectors of 64KB each
Hardware controlled locking of protected sectors via WP pin
Sector Lockdown
Make any combination of 64KB sectors permanently read-only
128-byte One-Time Programmable (OTP) Security Register
64 bytes factory preprogrammed
64 bytes user programmable
Flexible programming
Byte/Page Program (1 to 256 bytes)
Fast program and erase times
2.5ms typical Page Program (256 bytes) time
75ms typical 4KB Block Erase time
300ms typical 32KB Block Erase time
600ms typical 64KB Block Erase time
Program and Erase Suspend/Resume
Automatic checking and reporting of erase/program failures
Software controlled reset
JEDEC Standard Manufacturer and Device ID Read Methodology
Low power dissipation
25mA Active Read current (typical at 20MHz)
5μA Deep Power-Down current (typical)
Endurance: 100,000 program/erase cycles
Data retention: 20 years
Complies with full industrial temperature range
Industry standard green (Pb/Halide-free/RoHS compliant) package options
8-lead SOIC (0.208” wide)
8-pad Ultra-thin DFN (5 x 6 x 0.6mm)
Atmel AT25DF641A
64-Mbit, 2.7V Minimum SPI Serial Flash Memory
with Dual–I/O Support
DATASHEET
2
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
1. Description
The Atmel AT25DF641A is a 2.7V minimum, serial-interface Flash memory ideally suited for a wide variety of
high-volume consumer-based applications in which program code is shadowed from Flash memory into embedded or
external RAM for execution. The flexible erase architecture of the AT25DF641A, with its erase granularity as small as
4KB, makes it ideal for data storage as well, eliminating the need for additional data storage EEPROM devices.
The physical sectoring and the erase block sizes of the AT25DF641A have been optimized to meet the needs of today's
code and data storage applications. By optimizing the size of the physical sectors and erase blocks, the memory space
can be used much more efficiently. Because certain code modules and data storage segments must reside by
themselves in their own protected sectors, the wasted and unused memory space that occurs with large sectored and
large block erase Flash memory devices can be greatly reduced. This increased memory space efficiency allows
additional code routines and data storage segments to be added while still maintaining the same overall device density.
The AT25DF641A also offers a sophisticated method for protecting individual sectors against erroneous or malicious
program and erase operations. By providing the ability to individually protect and unprotect sectors, a system can
unprotect a specific sector to modify its contents while keeping the remaining sectors of the memory array securely
protected. This is useful in applications where program code is patched or updated on a subroutine or module basis, or in
applications where data storage segments need to be modified without running the risk of errant modifications to the
program code segments. In addition to individual sector protection capabilities, the AT25DF641A incorporates Global
Protect and Global Unprotect features that allow the entire memory array to be either protected or unprotected all at
once. This reduces overhead during the manufacturing process since sectors do not have to be unprotected one by one
prior to initial programming.
To take code and data protection to the next level, the AT25DF641A incorporates a sector lockdown mechanism that
allows any combination of individual 64KB sectors to be locked down and become permanently read-only. This
addresses the need of certain secure applications that require portions of the Flash memory array to be permanently
protected against malicious attempts at altering program code, data modules, security information or
encryption/decryption algorithms, keys, and routines. The device also contains a specialized OTP (One-Time
Programmable) Security Register that can be used for purposes such as unique device serialization, system-level
Electronic Serial Number (ESN) storage, locked key storage, or other purposes.
Specifically designed for use in 3V systems, the AT25DF641A supports read, program, and erase operations with a
supply voltage range of 2.7V to 3.6V. No separate voltage is required for programming and erasing.
3
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
2. Pin Descriptions and Pinouts
Table 2-1. Pin Descriptions
Symbol Name and Function
Asserted
State Type
CS
Chip Select: Asserting the CS pin selects the device. When the CS pin is
deasserted, the device will be deselected and normally be placed in standby mode
(not Deep Power-Down mode), and the SO pin will be in a high-impedance state.
When the device is deselected, data will not be accepted on the SI pin.
A high-to-low transition on the CS pin is required to start an operation, and a
low-to-high transition is required to end an operation. When ending an internally
self-timed operation such as a program or erase cycle, the device will not enter the
standby mode until the completion of the operation.
Low Input
SCK
Serial Clock: This pin is used to provide a clock to the device and is used to control
the flow of data to and from the device. Command, address, and input data present
on the SI pin is always latched in on the rising edge of SCK, while output data on the
SO pin is always clocked out on the falling edge of SCK.
— Input
SI (SIO)
Serial Input (Serial Input/Output): The SI pin is used to shift data into the device.
The SI pin is used for all data input including command and address sequences.
Data on the SI pin is always latched in on the rising edge of SCK.
With the Dual-Output Read Array command, the SI pin becomes an output pin (SIO)
to allow two bits of data (on the SO and SIO pins) to be clocked out on every falling
edge of SCK. To maintain consistency with SPI nomenclature, the SIO pin will be
referenced as SI throughout the document with exception to sections dealing with
the Dual-Output Read Array command in which it will be referenced as SIO.
Data present on the SI pin will be ignored whenever the device is deselected
(CS is deasserted).
— Input/Output
SO
(SOI)
Serial Output (Serial Output/Input): The SO pin is used to shift data out from the
device. Data on the SO pin is always clocked out on the falling edge of SCK.
With the Dual-Input Byte/Page Program command, the SO pin becomes an input pin
(SOI) to allow two bits of data (on the SOI and SI pins) to be clocked in on every
rising edge of SCK. To maintain consistency with SPI nomenclature, the SOI pin will
be referenced as SO throughout the document with exception to sections dealing
with the Dual-Input Byte/Page Program command in which it will be referenced as
SOI.
The SO pin will be in a high-impedance state whenever the device is deselected
(CS is deasserted).
— Output/Input
WP
Write Protect: The WP pin controls the hardware locking feature of the device.
Please refer to “Protection Commands and Features” on page 21 for more details on
protection features and the WP pin.
The WP pin is internally pulled-high and may be left floating if hardware controlled
protection will not be used. However, it is recommended that the WP pin also be
externally connected to VCC whenever possible.
Low Input
4
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
Figure 2-1. Pinouts
HOLD
Hold: The HOLD pin is used to temporarily pause serial communication without
deselecting or resetting the device. While the HOLD pin is asserted, transitions on
the SCK pin and data on the SI pin will be ignored, and the SO pin will be in a
high-impedance state.
The CS pin must be asserted, and the SCK pin must be in the low state in order for
a Hold condition to start. A Hold condition pauses serial communication only and
does not have an effect on internally self-timed operations such as a program or
erase cycle. See “Hold” on page 46 for additional details on the Hold operation.
The HOLD pin is internally pulled-high and may be left floating if the Hold function
will not be used. However, it is recommended that the HOLD pin also be externally
connected to VCC whenever possible.
Low Input
VCC
Device Power Supply: The VCC pin is used to supply the source voltage to the
device.
Operations at invalid VCC voltages may produce spurious results and should not be
attempted.
— Power
GND Ground: The ground reference for the power supply. GND should be connected to
the system ground. — Power
Table 2-1. Pin Descriptions (Continued)
Symbol Name and Function
Asserted
State Type
CS
SO (SOI)
WP
GND
VCC
HOLD
SCK
SI (SIO)
8
7
6
5
1
2
3
4
8-pad UDFN
(Top View)
1
2
3
4
8
7
6
5
CS
SO (SOI)
WP
GND
VCC
HOLD
SCK
SI (SIO)
8-lead SOIC
(Top View)
5
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
3. Block Diagram
Figure 3-1. Block Diagram
Note: SIO and SOI pin naming convention is used for Dual-I/O commands.
CS
SCK
SO (SOI)
SI (SIO)
WP
HOLD
Interface
Control
and
Logic
Control and
Protection Logic
I/O Buffers
and Latches
SRAM
Data Buffer
Y-Decoder Y-Gating
Flash
Memory
Array
X-Decoder
Address Latch
6
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
4. Memory Array
To provide the greatest flexibility, the AT25DF641A memory array can be erased in four levels of granularity including a
full Chip Erase. In addition, the array has been divided into physical sectors of uniform size, which can be individually
protected from program and erase operations. The size of the physical sectors is optimized for both code and data
storage applications, allowing both code and data segments to reside in their own isolated regions. The Memory
Architecture Diagram illustrates the breakdown of each erase level as well as the breakdown of each physical sector.
Figure 4-1. Memory Architecture Diagram
Internal Sectoring for 64KB 32KB 4KB 1-256 byte
Protection, Lockdown, Block Erase Block Erase Block Erase Page Program
and Suspend Functions (D8h Command) (52h Command) (20h Command) (02h Command)
4KB 7FFFFFh–7FF000h 256 bytes 7FFFFFh–7FFF00h
4KB 7FEFFFh–7FE000h 256 bytes 7FFEFFh–7FFE00h
4KB 7FDFFFh – 7FD000h 256 bytes 7FFDFFh – 7FFD00h
4KB 7FCFFFh–7FC000h 256 bytes 7FFCFFh– 7FFC00h
4KB 7FBFFFh–7FB000h 256 bytes 7FFBFFh–7FFB00h
4KB 7FAFFFh – 7FA000h 256 bytes 7FFAFFh – 7FFA00h
4KB 7F9FFFh – 7F9000h 256 bytes 7FF9FFh –7FF900h
4KB 7F8FFFh – 7F8000h 256 bytes 7FF8FFh –7FF800h
4KB 7F7FFFh – 7F7000h 256 bytes 7FF7FFh –7FF700h
4KB 7F6FFFh – 7F6000h 256 bytes 7FF6FFh –7FF600h
4KB 7F5FFFh – 7F5000h 256 bytes 7FF5FFh –7FF500h
4KB 7F4FFFh – 7F4000h 256 bytes 7FF4FFh –7FF400h
4KB 7F3FFFh – 7F3000h 256 bytes 7FF3FFh –7FF300h
4KB 7F2FFFh – 7F2000h 256 bytes 7FF2FFh –7FF200h
4KB 7F1FFFh – 7F1000h 256 bytes 7FF1FFh –7FF100h
4KB 7F0FFFh – 7F0000h 256 bytes 7FF0FFh –7FF000h
4KB 7EFFFFh– 7EF000h 256 bytes 7FEFFFh– 7FEF00h
4KB 7EEFFFh– 7EE000h 256 bytes 7FEEFFh– 7FEE00h
4KB 7EDFFFh – 7ED000h 256 bytes 7FEDFFh – 7FED00h
4KB 7ECFFFh– 7EC000h 256 bytes 7FECFFh– 7FEC00h
4KB 7EBFFFh– 7EB000h 256 bytes 7FEBFFh–7FEB00h
4KB 7EAFFFh – 7EA000h 256 bytes 7FEAFFh–7FEA00h
4KB 7E9FFFh – 7E9000h 256 bytes 7FE9FFh – 7FE900h
4KB 7E8FFFh – 7E8000h 256 bytes 7FE8FFh – 7FE800h
4KB 7E7FFFh – 7E7000h
4KB 7E6FFFh – 7E6000h
4KB 7E5FFFh – 7E5000h
4KB 7E4FFFh – 7E4000h 256 bytes 0017FFh – 001700h
4KB 7E3FFFh – 7E3000h 256 bytes 0016FFh – 001600h
4KB 7E2FFFh – 7E2000h 256 bytes 0015FFh – 001500h
4KB 7E1FFFh – 7E1000h 256 bytes 0014FFh – 001400h
4KB 7E0FFFh – 7E0000h 256 bytes 0013FFh – 001300h
256 bytes 0012FFh – 001200h
256 bytes 0011FFh – 001100h
256 bytes 0010FFh – 001000h
4KB 00FFFFh – 00F000h 256 bytes 000FFFh – 000F00h
4KB 00EFFFh – 00E000h 256 bytes 000EFFh – 000E00h
4KB 00DFFFh – 00D000h 256 bytes 000DFFh – 000D00h
4KB 00CFFFh – 00C000h 256 bytes 000CFFh – 000C00h
4KB 00BFFFh – 00B000h 256 bytes 000BFFh – 000B00h
4KB 00AFFFh – 00A000h 256 bytes 000AFFh – 000A00h
4KB 009FFFh – 009000h 256 bytes 0009FFh – 000900h
4KB 008FFFh – 008000h 256 bytes 0008FFh – 000800h
4KB 007FFFh – 007000h 256 bytes 0007FFh – 000700h
4KB 006FFFh – 006000h 256 bytes 0006FFh – 000600h
4KB 005FFFh – 005000h 256 bytes 0005FFh – 000500h
4KB 004FFFh – 004000h 256 bytes 0004FFh – 000400h
4KB 003FFFh – 003000h 256 bytes 0003FFh – 000300h
4KB 002FFFh – 002000h 256 bytes 0002FFh – 000200h
4KB 001FFFh – 001000h 256 bytes 0001FFh – 000100h
4KB 000FFFh – 000000h 256 bytes 0000FFh – 000000h
64KB
• • •
64KB
32KB
32KB
• • •
Range
• • •
• • •
Range
Page Program Detail
Page AddressBlock Address
64KB
64KB
(Sector 127)
Block Erase Detail
32KB
32KB
64KB
(Sector 0)
32KB
32KB
• • •
64KB
(Sector 126)
7
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
5. Device Operation
The AT25DF641A is controlled by a set of instructions that are sent from a host controller, commonly referred to as the
SPI master. The SPI master communicates with the AT25DF641A via the SPI bus which is comprised of four signal lines:
Chip Select (CS), Serial Clock (SCK), Serial Input (SI), and Serial Output (SO).
The AT25DF641A features a Dual-Input Program mode in which the SO pin becomes an input. Similarly, the device also
features a Dual-Output Read mode in which the SI pin becomes an output. In the Dual-Input Byte/Page Program
command description, the SO pin will be referred to as the SOI (Serial Output/Input) pin, and in the Dual-Output Read
Array command, the SI pin will be referenced as the SIO (Serial Input/Output) pin.
The SPI protocol defines a total of four modes of operation (Mode 0, 1, 2, or 3) with each mode differing in respect to the
SCK polarity and phase and how the polarity and phase control the flow of data on the SPI bus. The AT25DF641A
supports the two most common modes, SPI Modes 0 and 3. The only difference between SPI Modes 0 and 3 is the
polarity of the SCK signal when in the inactive state (when the SPI Master is in standby mode and not transferring any
data). With SPI Modes 0 and 3, data is always latched in on the rising edge of SCK and always output on the falling edge
of SCK.
Figure 5-1. SPI Mode 0 and 3
6. Commands and Addressing
A valid instruction or operation must always be started by first asserting the CS pin. After the CS pin has been asserted,
the host controller must then clock out a valid 8-bit opcode on the SPI bus. Following the opcode, instruction-dependent
information, such as address and data bytes, would then be clocked out by the host controller. All opcode, address, and
data bytes are transferred with the Most-Significant Bit (MSB) first. An operation is ended by deasserting the CS pin.
Opcodes not supported by the AT25DF641A will be ignored by the device and no operation will be started. The device
will continue to ignore any data presented on the SI pin until the start of the next operation (CS pin being deasserted and
then reasserted). In addition, if the CS pin is deasserted before complete opcode and address information is sent to the
device, then no operation will be performed, and the device will simply return to the idle state and wait for the next
operation.
Addressing of the device requires a total of three bytes of information to be sent, representing address bits A23-A0.
Since the upper address limit of the AT25DF641A memory array is 7FFFFFh, address bit A23 is always ignored by the
device.
SCK
CS
SI
SO
MSB LSB
MSB LSB
8
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
Table 6-1. Command Listing
Command Opcode
Clock
Frequency
Address
Bytes
Dummy
Bytes
Data
Bytes
Read Commands
Read Array
1Bh 0001 1011 Up to 100MHz 3 2 1+
0Bh 0000 1011 Up to 85MHz 3 1 1+
03h 0000 0011 Up to 40MHz 3 0 1+
Dual-Output Read Array 3Bh 0011 1011 Up to 65MHz 3 1 1+
Program and Erase Commands
Block Erase (4KB) 20h 0010 0000 Up to 100MHz 3 0 0
Block Erase (32KB) 52h 0101 0010 Up to 100MHz 3 0 0
Block Erase (64KB) D8h 1101 1000 Up to 100MHz 3 0 0
Chip Erase 60h 0110 0000 Up to 100MHz 0 0 0
C7h 1100 0111 Up to 100MHz 0 0 0
Byte/Page Program (1 to 256 bytes) 02h 0000 0010 Up to 100MHz 3 0 1+
Dual-Input Byte/Page Program (1 to 256 bytes) A2h 1010 0010 Up to 100MHz 3 0 1+
Program/Erase Suspend B0h 1011 0000 Up to 100MHz 0 0 0
Program/Erase Resume D0h 1101 0000 Up to 100MHz 0 0 0
Protection Commands
Write Enable 06h 0000 0110 Up to 100MHz 0 0 0
Write Disable 04h 0000 0100 Up to 100MHz 0 0 0
Protect Sector 36h 0011 0110 Up to 100MHz 3 0 0
Unprotect Sector 39h 0011 1001 Up to 100MHz 3 0 0
Global Protect/Unprotect Use Write Status Register Byte 1 Command
Read Sector Protection Registers 3Ch 0011 1100 Up to 100MHz 3 0 1+
Security Commands
Sector Lockdown 33h 0011 0011 Up to 100MHz 3 0 1
Freeze Sector Lockdown State 34h 0011 0100 Up to 100MHz 3 0 1
Read Sector Lockdown Registers 35h 0011 0101 Up to 100MHz 3 0 1+
Program OTP Security Register 9Bh 1001 1011 Up to 100MHz 3 0 1+
Read OTP Security Register 77h 0111 0111 Up to 100MHz 3 2 1+
Status Register Commands
Read Status Register 05h 0000 0101 Up to 100MHz 0 0 1+
Write Status Register Byte 1 01h 0000 0001 Up to 100MHz 0 0 1
Write Status Register Byte 2 31h 0011 0001 Up to 100MHz 0 0 1
Miscellaneous Commands
Reset F0h 1111 0000 Up to 100MHz 0 0 1
Read Manufacturer and Device ID 9Fh 1001 1111 Up to 85 MHz 0 0 1 to 4
Deep Power-Down B9h 1011 1001 Up to 100 MHz 0 0 0
Resume from Deep Power-Down ABh 1010 1011 Up to 100 MHz 0 0 0
9
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
7. Read Commands
7.1 Read Array
The Read Array command can be used to sequentially read a continuous stream of data from the device by simply
providing the clock signal once the initial starting address has been specified. The device incorporates an internal
address counter that automatically increments on every clock cycle.
Three opcodes (1Bh, 0Bh, and 03h) can be used for the Read Array command. The use of each opcode depends on the
maximum clock frequency that will be used to read data from the device. The 0Bh opcode can be used at any clock
frequency up to the maximum specified by fCLK, and the 03h opcode can be used for lower frequency read operations, up
to the maximum specified by fRDLF. The 1Bh opcode allows the highest read performance possible and can be used at
any clock frequency up to the maximum specified by fMAX; however, use of the 1Bh opcode at clock frequencies above
fCLK should be reserved to systems employing the RapidS protocol.
To perform the Read Array operation, the CS pin must first be asserted and then the appropriate opcode (1Bh, 0Bh, or
03h) must be clocked into the device. After the opcode has been clocked in, the three address bytes must be clocked in
to specify the location of the first byte to read within the memory array. Following the three address bytes, additional
dummy bytes may need to be clocked into the device, depending on which opcode is used for the Read Array operation.
If the 1Bh opcode is used, then two dummy bytes must be clocked into the device after the three address bytes. If the
0Bh opcode is used, then a single dummy byte must be clocked in after the address bytes.
After the three address bytes (and the dummy bytes or byte if using opcodes 1Bh or 0Bh) have been clocked in,
additional clock cycles will result in data being output on the SO pin. The data is always output with the MSB of a byte
first. When the last byte (7FFFFFh) of the memory array has been read, the device will continue reading back at the
beginning of the array (000000h). No delays will be incurred when wrapping around from the end of the array to the
beginning of the array.
Deasserting the CS pin will terminate the read operation and put the SO pin into a high-impedance state. The CS pin can
be deasserted at any time and does not require that a full byte of data be read.
10
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
Figure 7-1. Read Array – 1Bh Opcode
Figure 7-2. Read Array – 0Bh Opcode
Figure 7-3. Read Array – 03h Opcode
SCK
CS
SI
SO
MSB MSB
2310
00011011
675410119812 394243414037 3833 36353431 3229 30 44 47 484645 50 5149 52 55 565453
Opcode
AAAA AAAAA
MSB
XXXXXXXX
MSB MSB
DDDDDDDDDD
Address Bits A23-A0 Dummy Byte
MSB
XXXXXXXX
Dummy Byte
Data Byte 1
High-impedance
SCK
CS
SI
SO
MSB MSB
2310
00001011
675410119812 394243414037 3833 36353431 3229 30 44 47 484645
Opcode
AAAA AAAAA
MSB
XXXXXXXX
MSB MSB
DDDDDDDDDD
Address Bits A23-A0 Dummy Byte
Data Byte 1
High-impedance
SCK
CS
SI
SO
MSB MSB
2310
00000011
675410119812 373833 36353431 3229 30 39 40
Opcode
AAAA AAAAA
MSB MSB
DDDDDDDDDD
Address Bits A23-A0
Data Byte 1
High-impedance
11
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
7.2 Dual-Output Read Array
The Dual-Output Read Array command is similar to the standard Read Array command and can be used to sequentially
read a continuous stream of data from the device by simply providing the clock signal once the initial starting address has
been specified. Unlike the standard Read Array command, however, the Dual-Output Read Array command allows two
bits of data to be clocked out of the device on every clock cycle, rather than just one.
The Dual-Output Read Array command can be used at any clock frequency up to the maximum specified by fRDDO. To
perform the Dual-Output Read Array operation, the CS pin must first be asserted and then the opcode of 3Bh must be
clocked into the device. After the opcode has been clocked in, the three address bytes must be clocked in to specify the
location of the first byte to read within the memory array. Following the three address bytes, a single dummy byte must
also be clocked into the device.
After the three address bytes and the dummy byte have been clocked in, additional clock cycles will result in data being
output on both the SO and SIO pins. The data is always output with the MSB of a byte first, and the MSB is always output
on the SO pin. During the first clock cycle, bit 7 of the first data byte is output on the SO pin, while bit 6 of the same data
byte is output on the SIO pin. During the next clock cycle, bits 5 and 4 of the first data byte are output on the SO and SIO
pins, respectively. The sequence continues with each byte of data being output after every four clock cycles. When the
last byte (7FFFFFh) of the memory array has been read, the device will continue reading from the beginning of the array
(000000h). No delays will be incurred when wrapping around from the end of the array to the beginning of the array.
Deasserting the CS pin will terminate the read operation and put the SO and SIO pins into a high-impedance state. The
CS pin can be deasserted at any time and does not require that a full byte of data be read.
Figure 7-4. Dual-Output Read Array
2310
00111011
675410119812 394243414037 3833 36353431 3229 30 44 47 484645
AAAA AAAAA XXXXXXXX
D7
D6
D5
D4
D3
D2
D1
D0
D7
D6
D5
D4
D7
D6
D5
D4
D3
D2
D1
D0
MSB MSB MSB
Opcode Address Bits A23-A0 Dummy Byte
High-impedance
SCK
CS
SIO
SO
Output
Data Byte 1
Output
Data Byte 2
MSB MSB MSB
12
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
8. Program and Erase Commands
8.1 Byte/Page Program
The Byte/Page Program command allows anywhere from a single byte of data to 256 bytes of data to be programmed
into previously erased memory locations. An erased memory location is one that has all eight bits set to the Logical 1
state (a byte value of FFh). Before a Byte/page Program command can be started, the Write Enable command must
have been previously issued to the device (see “Write Enable” on page 21) to set the Write Enable Latch (WEL) bit of the
Status Register to a Logical 1 state.
Note: The internal programming operation of the AT25DF641A occurs on a nibble-wide basis, and all four bits of
the nibble must be in an erased state (Logic 1 state) prior to the programming of any of the four bits. If any
one of the four bits is in the programmed state (Logic 0 state) and an attempt is made to program one of the
other four bits in the nibble from a Logic 1 to a Logic 0, then the contents of the nibble cannot be guaranteed
and may contain erroneous data.
Example 1: A 4KB Block Erase is performed to erase the first 4KB block of memory (all bytes set to FFh). The
application then programs the first byte location with a value of 7Fh (0111 1111). Without erasing the 4KB
block again, the application then attempts to program the same byte location with BFh (1011 1111)
expecting that the resulting byte value stored in memory will be 3Fh (0011 1111). However, because of the
way the AT25DF641A programs bytes internally and because all four bits of the most-significant nibble were
not in the erased state prior to the programming of the BFh value, the most-significant nibble will not contain
the value of 0011b but will instead contain a different value.
Example 2: A 4KB Block Erase is performed to erase the first 4KB block of memory (all bytes set to FFh). The
application then programs the first byte location with a value of 7Fh (0111 1111). Without erasing the 4KB
block again, the application then attempts to program the same byte location with FCh (1111 1100),
expecting that the resulting byte value stored in memory will be 7Ch (0111 1100). The resulting byte value
stored in the memory will indeed be 7Ch because no additional bits in the most-significant nibble were being
programmed from a Logic 1 to a Logic 0, and all four bits in the least-significant nibble were in the erased
state prior to the programming of the 7Ch byte value.
To perform a Byte/Page Program command, a 02h opcode must be clocked into the device followed by the three address
bytes denoting the first location of the memory array to begin programming at. After the address bytes have been clocked
in, data can then be clocked into the device and will be stored in an internal buffer.
If the starting memory address denoted by A23-A0 does not fall on an even 256-byte page boundary (A7-A0 are not
all 0), then special circumstances regarding which memory locations are to be programmed will apply. In this situation,
any data that is sent to the device that goes beyond the end of the page will wrap around to the beginning of the same
page. In addition, if more than 256 bytes of data are sent to the device, then only the last 256 bytes sent will be latched
into the internal buffer.
Example: If the starting address denoted by A23-A0 is 0000FEh, and three bytes of data are sent to the device, then
the first two bytes of data will be programmed at addresses 0000FEh and 0000FFh, while the last byte of
data will be programmed at address 000000h. The remaining bytes in the page (addresses 000001h
through 0000FDh) will not be programmed and will remain in the erased state (FFh).
When the CS pin is deasserted, the device will take the data stored in the internal buffer into the appropriate memory
array locations based on the starting address specified by A23-A0 and the number of data bytes sent to the device. If
less than 256 bytes of data were sent to the device, then the remaining bytes within the page will not be programmed and
will remain in the erased state (FFh). The programming of the data bytes is internally self-timed and should take place in
a time of tPP or tBP if only programming a single byte.
13
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
The three address bytes and at least one complete byte of data must be clocked into the device before the CS pin is
deasserted, and the CS pin must be deasserted on even byte boundaries (multiples of eight bits); otherwise, the device
will abort the operation and no data will be programmed into the memory array. In addition, if the address specified by
A23-A0 points to a memory location within a sector that is in the protected state (see “Protect Sector” on page 23) or
locked down (see “Sector Lockdown” on page 29), then the Byte/Page Program command will not be executed, and the
device will return to the idle state once the CS pin has been deasserted. The WEL bit in the Status Register will be reset
back to the Logical 0 state if the program cycle aborts due to an incomplete address being sent, an incomplete byte of
data being sent, the CS pin being deasserted on uneven byte boundaries, or because the memory location to be
programmed is protected or locked down.
While the device is programming, the Status Register can be read and will indicate that the device is busy. For faster
throughput, it is recommended that the Status Register be polled rather than waiting the tBP or tPP time to determine if the
data bytes have finished programming. At some point before the program cycle completes, the WEL bit in the Status
Register will be reset back to the Logical 0 state.
The device also incorporates an intelligent programming algorithm that can detect when a byte location fails to program
properly. If a programming error arises, it will be indicated by the EPE bit in the Status Register.
Figure 8-1. Byte Program
Figure 8-2. Page Program
SCK
CS
SI
SO
MSB MSB
2310
00000010
675410119812 3937 3833 36353431 3229 30
Opcode
High-impedance
AAAA AAAAA
MSB
DDDDDDDD
Address Bits A23-A0 Data In
SCK
CS
SI
SO
MSB MSB
2310
00000010
6754983937 3833 36353431 3229 30
Opcode
High-impedance
AA AAAA
MSB
DDDDDDDD
Address Bits A23-A0 Data In Byte 1
MSB
DDDDDDDD
Data In Byte n
14
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
8.2 Dual-Input Byte/Page Program
The Dual-Input Byte/Page Program command is similar to the standard Byte/Page Program command and can be used
to program anywhere from a single byte of data up to 256 bytes of data into previously erased memory locations. Unlike
the standard Byte/Page Program command, however, the Dual-Input Byte/Page Program command allows two bits of
data to be clocked into the device on every clock cycle rather than just one.
Note: The internal programming operation of the AT25DF641A occurs on a nibble-wide basis, and all four bits of
the nibble must be in an erased state (Logic 1 state) prior to the programming of any of the four bits. If any
one of the four bits is in the programmed state (Logic 0 state) and an attempt is made to program one of the
other four bits in the nibble from a Logic 1 to a Logic 0, then the contents of the nibble cannot be guaranteed
and may contain erroneous data.
Example 1: A 4KB Block Erase is performed to erase the first 4KB block of memory (all bytes set to FFh). The
application then programs the first byte location with a value of 7Fh (0111 1111). Without erasing the 4KB
block again, the application then attempts to program the same byte location with BFh (1011 1111)
expecting that the resulting byte value stored in memory will be 3Fh (0011 1111). However, because of the
way the AT25DF641A programs bytes internally and because all four bits of the most-significant nibble were
not in the erased state prior to the programming of the BFh value, the most-significant nibble will not contain
the value of 0011b but will instead contain a different value.
Example 2: A 4KB Block Erase is performed to erase the first 4KB block of memory (all bytes set to FFh). The
application then programs the first byte location with a value of 7Fh (0111 1111). Without erasing the 4KB
block again, the application then attempts to program the same byte location with FCh (1111 1100),
expecting that the resulting byte value stored in memory will be 7Ch (0111 1100). The resulting byte value
stored in the memory will indeed be 7Ch because no additional bits in the most-significant nibble were being
programmed from a Logic 1 to a Logic 0, and all four bits in the least-significant nibble were in the erased
state prior to the programming of the 7Ch byte value.
Before the Dual-Input Byte/Page Program command can be started, the Write Enable command must have been
previously issued to the device (see “Write Enable” on page 21) to set the Write Enable Latch (WEL) bit of the Status
Register to a Logical 1 state. To perform a Dual-Input Byte/Page Program command, an A2h opcode must be clocked
into the device followed by the three address bytes denoting the first location of the memory array to begin programming
at. After the address bytes have been clocked in, data can then be clocked into the device two bits at a time on both the
SOI and SI pins.
The data is always input with the MSB of a byte first, and the MSB is always input on the SOI pin. During the first clock
cycle, bit 7 of the first data byte is input on the SOI pin while bit 6 of the same data byte is input on the SI pin. During the
next clock cycle, bits 5 and 4 of the first data byte are input on the SOI and SI pins, respectively. The sequence continues
with each byte of data being input after every four clock cycles. Like the standard Byte/Page Program command, all data
clocked into the device is stored in an internal buffer.
If the starting memory address denoted by A23-A0 does not fall on an even 256-byte page boundary (A7-A0 are
not all 0), then special circumstances regarding which memory locations are to be programmed will apply. In this
situation, any data that is sent to the device that goes beyond the end of the page will wrap around back to the beginning
of the same page. In addition, if more than 256 bytes of data are sent to the device, then only the last 256 bytes sent will
be latched into the internal buffer.
Example: If the starting address denoted by A23-A0 is 0000FEh, and three bytes of data are sent to the device, then
the first two bytes of data will be programmed at addresses 0000FEh and 0000FFh while the last byte of
data will be programmed at address 000000h. The remaining bytes in the page (addresses 000001h
through 0000FDh) will not be programmed and will remain in the erased state (FFh).
When the CS pin is deasserted, the device will program the data stored in the internal buffer into the appropriate memory
array locations based on the starting address specified by A23-A0 and the number of data bytes sent to the device. If
less than 256 bytes of data are sent to the device, then the remaining bytes within the page will not be programmed and
will remain in the erased state (FFh). The programming of the data bytes is internally self-timed and should take place in
a time of tPP or tBP if only programming a single byte.
15
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
The three address bytes and at least one complete byte of data must be clocked into the device before the CS pin is
deasserted, and the CS pin must be deasserted on even byte boundaries (multiples of eight bits); otherwise, the device
will abort the operation and no data will be programmed into the memory array. In addition, if the address specified by
A23-A0 points to a memory location within a sector that is in the protected state (see “Protect Sector” on page 23) or
locked down (see “Sector Lockdown” on page 29), then the Byte/Page Program command will not be executed, and the
device will return to the idle state once the CS pin has been deasserted. The WEL bit in the Status Register will be reset
back to the Logical 0 state if the program cycle aborts due to an incomplete address being sent, an incomplete byte of
data being sent, the CS pin being deasserted on uneven byte boundaries, or because the memory location to be
programmed is protected or locked down.
While the device is programming, the Status Register can be read and will indicate that the device is busy. For faster
throughput, it is recommended that the Status Register be polled rather than waiting the tBP or tPP time to determine if the
data bytes have finished programming. At some point before the program cycle completes, the WEL bit in the Status
Register will be reset back to the Logical 0 state.
The device also incorporates an intelligent programming algorithm that can detect when a byte location fails to program
properly. If a programming error arises, it will be indicated by the EPE bit in the Status Register.
Figure 8-3. Dual-Input Byte Program
Figure 8-4. Dual-Input Page Program
SCK
CS
SI
SOI
MSB MSB
2310
10100010
675410119812 33353431 3229 30
Opcode
AAAA AAAAA
Address Bits A23-A0
MSB
D7
D6
D5
D4
D3
D2
D1
D0
Input
Data Byte
High-impedance
SCK
CS
SI
SOI
MSB MSB
2310
10100010
675410119812 3937 3833 36353431 3229 30
Opcode
AAAA AAAAA
MSB MSB
D7
D6
D5
D4
D3
D2
D1
D0
D7
D6
D5
D4
D3
D2
D1
D0
Address Bits A23-A0
Input
Data Byte 1
MSB
D7
D6
D5
D4
D3
D2
D1
D0
Input
Data Byte n
Input
Data Byte 2
High-impedance
16
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
8.3 Block Erase
A block of 4KB, 32KB, or 64KB can be erased (all bits set to the Logical 1 state) in a single operation by using one of
three different opcodes for the Block Erase command. A 20h opcode is used for a 4KB erase, a 52h opcode is used for a
32KB erase, and a D8h opcode is used for a 64KB erase. Before a Block Erase command can be started, the Write
Enable command must have been previously issued to the device to set the WEL bit of the Status Register to a Logical 1
state.
To perform a Block Erase, the CS pin must first be asserted and the appropriate opcode (20h, 52h, or D8h) must be
clocked into the device. After the opcode has been clocked in, the three address bytes specifying an address within the
4KB, 32KB, or 64KB block to be erased must be clocked in. Any additional data clocked into the device will be ignored.
When the CS pin is deasserted, the device will erase the appropriate block. The erasing of the block is internally
self-timed and should take place in a time of tBLKE.
Since the Block Erase command erases a region of bytes, the lower order address bits do not need to be decoded by the
device. Therefore, for a 4KB erase, address bits A11-A0 will be ignored by the device and their values can be either a
Logical 1 or 0. For a 32KB erase, address bits A14-A0 will be ignored, and for a 64KB erase, address bits A15-A0 will be
ignored. Despite the lower order address bits not being decoded by the device, the complete three address bytes must
still be clocked into the device before the CS pin is deasserted, and the CS pin must be deasserted on an even byte
boundary (multiples of eight bits); otherwise, the device will abort the operation and no erase operation will be performed.
If the address specified by A23-A0 points to a memory location within a sector that is in the protected or locked down
state, then the Block Erase command will not be executed, and the device will return to the idle state once the CS pin has
been deasserted.
The WEL bit in the Status Register will be reset back to the Logical 0 state if the erase cycle aborts due to an incomplete
address being sent, the CS pin being deasserted on uneven byte boundaries, or because a memory location within the
region to be erased is protected or locked down.
While the device is executing a successful erase cycle, the Status Register can be read and will indicate that the device
is busy. For faster throughput, it is recommended that the Status Register be polled rather than waiting the tBLKE time to
determine if the device has finished erasing. At some point before the erase cycle completes, the WEL bit in the Status
Register will be reset back to the Logical 0 state.
The device also incorporates an intelligent erase algorithm that can detect when a byte location fails to erase properly. If
an erase error occurs, it will be indicated by the EPE bit in the Status Register.
Figure 8-5. Block Erase
SCK
CS
SI
SO
MSB MSB
2310
CCCCCCCC
675410119812 3129 3027 2826
Opcode
AAAA AAAAA AAA
Address Bits A23-A0
High-impedance
17
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
8.4 Chip Erase
The entire memory array can be erased in a single operation by using the Chip Erase command. Before a Chip Erase
command can be started, the Write Enable command must have been previously issued to the device to set the WEL bit
of the Status Register to a Logical 1 state.
Two opcodes, 60h and C7h, can be used for the Chip Erase command. There is no difference in device functionality
when utilizing the two opcodes, so they can be used interchangeably. To perform a Chip Erase, one of the two opcodes
(60h or C7h) must be clocked into the device. Since the entire memory array is to be erased, no address bytes need to
be clocked into the device, and any data clocked in after the opcode will be ignored. When the CS pin is deasserted, the
device will erase the entire memory array. The erasing of the device is internally self-timed and should take place in a
time of tCHPE.
The complete opcode must be clocked into the device before the CS pin is deasserted, and the CS pin must be
deasserted on an even byte boundary (multiples of eight bits); otherwise, no erase will be performed. In addition, if any
sector of the memory array is in the protected or locked down state, then the Chip Erase command will not be executed,
and the device will return to the idle state once the CS pin has been deasserted. The WEL bit in the Status Register will
be reset back to the Logical 0 state if the CS pin is deasserted on uneven byte boundaries or if a sector is in the protected
or locked down state.
While the device is executing a successful erase cycle, the Status Register can be read and will indicate that the device
is busy. For faster throughput, it is recommended that the Status Register be polled rather than waiting the tCHPE time to
determine if the device has finished erasing. At some point before the erase cycle completes, the WEL bit in the Status
Register will be reset back to the Logical 0 state.
The device also incorporates an intelligent erase algorithm that can detect when a byte location fails to erase properly. If
an erase error occurs, it will be indicated by the EPE bit in the Status Register.
Figure 8-6. Chip Erase
SCK
CS
SI
SO
MSB
2310
CCCCCCCC
6754
Opcode
High-impedance
18
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
8.5 Program/Erase Suspend
In some code-plus-data storage applications, it is often necessary to process certain high-level system interrupts that
require relatively immediate reading of code or data from the Flash memory. In such an instance, it may not be possible
for the system to wait the microseconds or milliseconds required for the Flash memory to complete a program or erase
cycle. The Program/Erase Suspend command allows a program or erase operation in progress to a particular 64KB
sector of the Flash memory array to be suspended so that other device operations can be performed.
By suspending an erase operation to a particular sector, the system can perform a program or read operation within
another 64KB sector in the device. Other device operations, such as a Read Status Register, can also be performed
while a program or erase operation is suspended. Table 8-1 outlines the operations that are allowed and not allowed
while a program or erase operation is suspended.
Since the need to suspend a program or erase operation is immediate, the Write Enable command does not need to be
issued prior to the Program/Erase Suspend command being issued. Therefore, the Program/Erase Suspend command
operates independently of the state of the WEL bit in the Status Register.
To perform a Program/Erase Suspend, the CS pin must first be asserted and then the opcode B0h must be clocked into
the device. No address bytes need to be clocked into the device, and any data clocked in after the opcode will be
ignored. When the CS pin is deasserted, the program or erase operation currently in progress will be suspended within a
time of tSUSP. The Program Suspend (PS) bit or the Erase Suspend (ES) bit in the Status Register will then be set to the
Logical 1 state to indicate that the program or erase operation has been suspended. In addition, the RDY/BSY bit in the
Status Register will indicate that the device is ready for another operation. The complete opcode must be clocked into
the device before the CS pin is deasserted, and the CS pin must be deasserted on an even byte boundary (multiples of
eight bits); otherwise, no suspend operation will be performed.
Read operations are not allowed to a 64KB sector that has had its program or erase operation suspended. If a read is
attempted to a suspended sector, then the device will output undefined data. Therefore, when performing a Read Array
operation to an unsuspended sector, and the device’s internal address counter increments and crosses the sector
boundary to a suspended sector, the device will then start outputting undefined data continuously until the address
counter increments and crosses a sector boundary to an unsuspended sector.
A program operation is not allowed on a sector that has been erase suspended. If a program operation is attempted on
an erase suspended sector, then the program operation will abort and the WEL bit in the Status Register will be reset
back to the Logical 0 state. Likewise, an erase operation is not allowed to a sector that has been program suspended. If
attempted, the erase operation will abort and the WEL bit in the Status Register will be reset to a Logical 0 state.
During an Erase Suspend, a program operation to a different 64KB sector can be started and subsequently suspended.
This results in a simultaneous Erase Suspend/Program Suspend condition, which will be indicated by the ES and PS bits
in the Status Register being set to the Logical 1 state.
If a Reset operation (see “Reset” on page 41) is performed while a sector is erase suspended, the suspend operation will
abort and the contents of the block in the suspended sector will be left in an undefined state. However, if a Reset is
performed while a sector is program suspended, the suspend operation will abort but only the contents of the page that
was being programmed and subsequently suspended will be undefined. The remaining pages in the 64KB sector will
retain their previous contents.
If an attempt is made to perform an operation that is not allowed while a program or erase operation is suspended, such
as a Protect Sector command, then the device will simply ignore the opcode and no operation will be performed. The
state of the WEL bit in the Status Register, as well as the SPRL (Sector Protection Registers Locked) and SLE (Sector
Lockdown Enabled) bits, will not be affected.
19
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
Table 8-1. Operations Allowed and Not Allowed During a Program or Erase Suspend
Figure 8-7. Program/Erase Suspend
Command
Operation During
Program Suspend
Operation During
Erase Suspend
Read Commands
Read Array (All Opcodes) Allowed Allowed
Program and Erase Commands
Block Erase Not Allowed Not Allowed
Chip Erase Not Allowed Not Allowed
Byte/Page Program (All Opcodes) Not Allowed Allowed
Program/Erase Suspend Not Allowed Allowed
Program/Erase Resume Allowed Allowed
Protection Commands
Write Enable Not Allowed Allowed
Write Disable Not Allowed Allowed
Protect Sector Not Allowed Not Allowed
Unprotect Sector Not Allowed Not Allowed
Global Protect/Unprotect Not Allowed Not Allowed
Read Sector Protection Registers Allowed Allowed
Security Commands
Sector Lockdown Not Allowed Not Allowed
Freeze Sector Lockdown State Not Allowed Not Allowed
Read Sector Lockdown Registers Allowed Allowed
Program OTP Security Register Not Allowed Not Allowed
Read OTP Security Register Allowed Allowed
Status Register Commands
Read Status Register Allowed Allowed
Write Status Register (All Opcodes) Not Allowed Not Allowed
Miscellaneous Commands
Reset Allowed Allowed
Read Manufacturer and Device ID Allowed Allowed
Deep Power-Down Not Allowed Not Allowed
Resume from Deep Power-Down Not Allowed Not Allowed
SCK
CS
SI
SO
MSB
2310
10110000
6754
Opcode
High-impedance
20
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
8.6 Program/Erase Resume
The Program/Erase Resume command allows a suspended program or erase operation to be resumed and continue
programming a Flash page or erasing a Flash memory block where it left off. As with the Program/Erase Suspend
command, the Write Enable command does not need to be issued prior to the Program/Erase Resume command being
issued. Therefore, the Program/Erase Resume command operates independently of the state of the WEL bit in the
Status Register.
To perform a Program/Erase Resume, the CS pin must first be asserted and then the opcode D0h must be clocked into
the device. No address bytes need to be clocked into the device, and any data clocked in after the opcode will be
ignored. When the CS pin is deasserted, the program or erase operation currently suspended will be resumed within a
time of tRES. The PS bit or the ES bit in the Status Register will then be reset back to the Logical 0 state to indicate that
the program or erase operation is no longer suspended. In addition, the RDY/BSY bit in the Status Register will indicate
that the device is busy performing a program or erase operation. The complete opcode must be clocked into the device
before the CS pin is deasserted, and the CS pin must be deasserted on an even byte boundary (multiples of eight bits);
otherwise, no resume operation will be performed.
During a simultaneous Erase Suspend/Program Suspend condition, issuing the Program/Erase Resume command will
result in the program operation resuming first. After the program operation has been completed, the Program/Erase
Resume command must be issued again in order for the erase operation to be resumed.
While the device is busy resuming a program or erase operation, any attempts at issuing the Program/Erase Suspend
command will be ignored. Therefore, if a resumed program or erase operation needs to be subsequently suspended
again, the system must either wait the entire tRES time before issuing the Program/Erase Suspend command, or it must
check the status of the RDY/BSY bit or the appropriate PS or ES bit in the Status Register to determine if the previously
suspended program or erase operation has resumed.
Figure 8-8. Program/Erase Resume
SCK
CS
SI
SO
MSB
2310
11010000
6754
Opcode
High-impedance
21
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
9. Protection Commands and Features
9.1 Write Enable
The Write Enable command is used to set the Write Enable Latch (WEL) bit in the Status Register to a Logical 1 state.
The WEL bit must be set before a Byte/Page Program, Erase, Protect Sector, Unprotect Sector, Sector Lockdown,
Freeze Sector Lockdown State, Program OTP Security Register, or Write Status Register command can be executed.
This makes the issuance of these commands a two-step process, thereby reducing the chances of a command being
accidentally or erroneously executed. If the WEL bit in the Status Register is not set prior to the issuance of one of these
commands, then the command will not be executed.
To issue the Write Enable command, the CS pin must first be asserted and the opcode of 06h must be clocked into the
device. No address bytes need to be clocked into the device, and any data clocked in after the opcode will be ignored.
When the CS pin is deasserted, the WEL bit in the Status Register will be set to a Logical 1. The complete opcode must
be clocked into the device before the CS pin is deasserted, and the CS pin must be deasserted on an even byte
boundary (multiples of eight bits); otherwise, the device will abort the operation and the state of the WEL bit will not
change.
Figure 9-1. Write Enable
SCK
CS
SI
SO
MSB
2310
00000110
6754
Opcode
High-impedance
22
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
9.2 Write Disable
The Write Disable command is used to reset the Write Enable Latch (WEL) bit in the Status Register to the Logical 0
state. With the WEL bit reset, all Byte/Page Program, erase, Protect Sector, Unprotect Sector, Sector Lockdown, Freeze
Sector Lockdown State, Program OTP Security Register, and Write Status Register commands will not be executed.
Other conditions can also cause the WEL bit to be reset; for more details, see Section 11.1.5, “WEL Bit” on page 37 in
“Status Register Commands”.
To issue the Write Disable command, the CS pin must first be asserted and then the opcode 04h must be clocked into
the device. No address bytes need to be clocked into the device, and any data clocked in after the opcode will be
ignored. When the CS pin is deasserted, the WEL bit in the Status Register will be reset to a Logical 0. The complete
opcode must be clocked into the device before the CS pin is deasserted, and the CS pin must be deasserted on an even
byte boundary (multiples of eight bits); otherwise, the device will abort the operation and the state of the WEL bit will not
change.
Figure 9-2. Write Disable
SCK
CS
SI
SO
MSB
2310
00000100
6754
Opcode
High-impedance
23
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
9.3 Protect Sector
Every physical 64KB sector of the device has a corresponding single-bit Sector Protection Register that is used to control
the software protection of a sector. Upon device power-up, each Sector Protection Register will default to the Logical 1
state indicating that all sectors are protected and cannot be programmed or erased.
Issuing the Protect Sector command to a particular sector address will set the corresponding Sector Protection Register
to the Logical 1 state. The following table outlines the two states of the Sector Protection Registers.
Table 9-1. Sector Protection Register Values
Before the Protect Sector command can be issued, the Write Enable command must have been previously issued to set
the WEL bit in the Status Register to a Logical 1. To issue the Protect Sector command, the CS pin must first be
asserted, and then the opcode 36h must be clocked into the device followed by three address bytes designating any
address within the sector to be protected. Any additional data clocked into the device will be ignored. When the CS pin is
deasserted, the Sector Protection Register corresponding to the physical sector addressed by A23-A0 will be set to the
Logical 1 state, and the sector itself will then be protected from program and erase operations. In addition, the WEL bit in
the Status Register will be reset back to the Logical 0 state.
The three complete address bytes must be clocked into the device before the CS pin is deasserted, and the CS pin must
be deasserted on an even byte boundary (multiples of eight bits); otherwise, the device will abort the operation. When the
device aborts the Protect Sector operation, the state of the Sector Protection Register will be unchanged, and the WEL
bit in the Status Register will be reset to a Logical 0.
As a safeguard against accidental or erroneous protecting or unprotecting of sectors, the Sector Protection Registers can
themselves be locked from updates by using the SPRL (Sector Protection Registers Locked) bit of the Status Register
(please refer to the Status Register description for more details). If the Sector Protection Registers are locked, then any
attempts to issue the Protect Sector command will be ignored, and the device will reset the WEL bit in the Status
Register back to a Logical 0 and return to the idle state once the CS pin has been deasserted.
Figure 9-3. Protect Sector
Value Sector Protection Status
0 Sector is unprotected and can be programmed and erased.
1 Sector is protected and cannot be programmed or erased. This is the default state.
SCK
CS
SI
SO
MSB MSB
2310
00110110
675410119812 3129 3027 2826
Opcode
AAAA AAAAA AAA
Address Bits A23-A0
High-impedance
24
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
9.4 Unprotect Sector
Issuing the Unprotect Sector command to a particular sector address will reset the corresponding Sector Protection
Register to the Logical 0 state (see Table 9-4 for Sector Protection Register values). Every physical sector of the device
has a corresponding single-bit Sector Protection Register that is used to control the software protection of a sector.
Before the Unprotect Sector command can be issued, the Write Enable command must have been previously issued to
set the WEL bit in the Status Register to a Logical 1. To issue the Unprotect Sector command, the CS pin must first be
asserted and then the opcode of 39h must be clocked into the device. After the opcode has been clocked in, the three
address bytes designating any address within the sector to be unprotected must be clocked in. Any additional data
clocked into the device after the address bytes will be ignored. When the CS pin is deasserted, the Sector Protection
Register corresponding to the sector addressed by A23-A0 will be reset to the Logical 0 state, and the sector itself will be
unprotected. In addition, the WEL bit in the Status Register will be reset back to the Logical 0 state.
The three complete address bytes must be clocked into the device before the CS pin is deasserted, and the CS pin must
be deasserted on an even byte boundary (multiples of eight bits); otherwise, the device will abort the operation, the state
of the Sector Protection Register will be unchanged, and the WEL bit in the Status Register will be reset to a Logical 0.
As a safeguard against accidental or erroneous locking or unlocking of sectors, the Sector Protection Registers can
themselves be locked from updates by using the SPRL (Sector Protection Registers Locked) bit of the Status Register
(see Section 11., “Status Register Commands” on page 35 for more details). If the Sector Protection Registers are
locked, then any attempts to issue the Unprotect Sector command will be ignored, and the device will reset the WEL bit
in the Status Register back to a Logical 0 and return to the idle state once the CS pin has been deasserted.
Figure 9-4. Unprotect Sector
SCK
CS
SI
SO
MSB MSB
2310
00111001
675410119812 3129 3027 2826
Opcode
AAAA AAAAA AAA
Address Bits A23-A0
High-impedance
25
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
9.5 Global Protect/Unprotect
The Global Protect and Global Unprotect features can work in conjunction with the Protect Sector and Unprotect Sector
functions.
Example: A system can globally protect the entire memory array and then use the Unprotect Sector command to
individually unprotect certain sectors and individually reprotect them later by using the Protect Sector
command. Likewise, a system can globally unprotect the entire memory array and then individually protect
certain sectors as needed.
Performing a Global Protect or Global Unprotect is accomplished by writing a certain combination of data to the Status
Register using the Write Status Register Byte 1 command (see “Write Status Register Byte 1” on page 39 for command
execution details). The Write Status Register command is also used to modify the SPRL (Sector Protection Registers
Locked) bit to control hardware and software locking.
To perform a Global Protect, the appropriate WP pin and SPRL conditions must be met, and the system must write a
Logical 1 to bits 5, 4, 3, and 2 of the first byte of the Status Register. Conversely, to perform a Global Unprotect, the same
WP and SPRL conditions must be met but the system must write a Logical 0 to bits 5, 4, 3, and 2 of the first byte of the
Status Register. Table 9-2 details the conditions necessary for a Global Protect or Global Unprotect to be performed.
Sectors that have been erase or program suspended must remain in the unprotected state. If a Global Protect operation
is attempted while a sector is erase or program suspended, the protection operation will abort, the protection states of all
sectors in the Flash memory array will not change, and WEL bit in the Status Register will be reset back to a Logical 0.
Essentially, if the SPRL bit of the Status Register is in the Logical 0 state (Sector Protection Registers are not locked),
then writing a 00h to the first byte of the Status Register will perform a Global Unprotect without changing the state of the
SPRL bit. Similarly, writing a 7Fh to the first byte of the Status Register will perform a Global Protect and keep the SPRL
bit in the Logical 0 state. The SPRL bit can, of course, be changed to a Logical 1 by writing an FFh if software-locking or
hardware-locking is desired along with the Global Protect.
If the desire is to only change the SPRL bit without performing a Global Protect or Global Unprotect, then the system can
simply write a 0Fh to the first byte of the Status Register to change the SPRL bit from a Logical 1 to a Logical 0 provided
the WP pin is deasserted. Likewise, the system can write an F0h to change the SPRL bit from a Logical 0 to a Logical 1
without affecting the current sector protection status (no changes will be made to the Sector Protection Registers).
When writing to the first byte of the Status Register, bits 5, 4, 3, and 2 will not actually be modified, but will be decoded by
the device for the purposes of the Global Protect and Global Unprotect functions. Only bit 7, the SPRL bit, will actually be
modified. Therefore, when reading the first byte of the Status Register, bits 5, 4, 3, and 2 will not reflect the values written
to them but will instead indicate the status of the WP pin and the sector protection status. Please see “Read Status
Register” on page 35 and Table 11-1 on page 35 for details on the Status Register format and what values can be read
for bits 5, 4, 3, and 2.
26
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
Table 9-2. Valid SPRL and Global Protect/Unprotect Conditions
WP
State
Current
SPRL
Value
New Write Status
Register Byte 1
Data
Protection Operation
New
SPRL
Value
Bit
7 6 5 4 3 2 1 0
0 0
0 x 0 0 0 0 x x
0 x 0 0 0 1 x x
0 x 1 1 1 0 x x
0 x 1 1 1 1 x x
1 x 0 0 0 0 x x
1 x 0 0 0 1 x x
1 x 1 1 1 0 x x
1 x 1 1 1 1 x x
Global Unprotect – all Sector Protection Registers reset to 0.
No change to current protection.
No change to current protection.
No change to current protection.
Global Protect – all Sector Protection Registers set to 1.
Global Unprotect – all Sector Protection Registers reset to 0.
No change to current protection.
No change to current protection.
No change to current protection.
Global Protect – all Sector Protection Registers set to 1.
0
0
0
0
0
1
1
1
1
1
0 1 x x x x x x x x
No change to the current protection level. All sectors currently protected will remain protected
and all sectors currently unprotected will remain unprotected.
The Sector Protection Registers are hard-locked and cannot be changed when the WP pin is
LOW and the current state of SPRL is 1. Therefore, a Global Protect/Unprotect will not occur.
In addition, the SPRL bit cannot be changed (the WP pin must be High in order to change
SPRL back to a 0).
1 0
0 x 0 0 0 0 x x
0 x 0 0 0 1 x x
0 x 1 1 1 0 x x
0 x 1 1 1 1 x x
1 x 0 0 0 0 x x
1 x 0 0 0 1 x x
1 x 1 1 1 0 x x
1 x 1 1 1 1 x x
Global Unprotect – all Sector Protection Registers reset to 0.
No change to current protection.
No change to current protection.
No change to current protection.
Global Protect – all Sector Protection Registers set to 1.
Global Unprotect – all Sector Protection Registers reset to 0.
No change to current protection.
No change to current protection.
No change to current protection.
Global Protect – all Sector Protection Registers set to 1.
0
0
0
0
0
1
1
1
1
1
1 1
0 x 0 0 0 0 x x
0 x 0 0 0 1 x x
0 x 1 1 1 0 x x
0 x 1 1 1 1 x x
1 x 0 0 0 0 x x
1 x 0 0 0 1 x x
1 x 1 1 1 0 x x
1 x 1 1 1 1 x x
No change to the current protection level. All sectors currently protected will
remain protected, and all sectors currently unprotected will remain unprotected.
The Sector Protection Registers are soft-locked and cannot be changed when
the current state of SPRL is 1. Therefore, a Global Protect/Unprotect will not
occur. However, the SPRL bit can be changed back to a 0 from a 1 since the WP
pin is High. To perform a Global Protect/Unprotect, the Write Status Register
command must be issued again after the SPRL bit has been changed from
1 to 0.
0
0
0
0
0
1
1
1
1
1
27
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
9.6 Read Sector Protection Registers
The Sector Protection Registers can be read to determine the current software protection status of each sector. Reading
the Sector Protection Registers, however, will not determine the status of the WP pin.
To read the Sector Protection Register for a particular sector, the CS pin must first be asserted and then the opcode 3Ch
must be clocked in. Once the opcode has been clocked in, three address bytes designating any address within the sector
must be clocked in. After the last address byte has been clocked in, the device will begin outputting data on the SO pin
during every subsequent clock cycle. The data being output will be a repeating byte of either FFh or 00h to denote the
value of the appropriate Sector Protection Register.
At clock frequencies above fCLK, the first byte of data output will not be valid. Therefore, if operating at clock frequencies
above fCLK, at least two bytes of data must be clocked out from the device in order to determine the correct status of the
appropriate Sector Protection Register.
Table 9-3. Read Sector Protection Register - Output Data
Deasserting the CS pin will terminate the read operation and put the SO pin into a high-impedance state. The CS pin can
be deasserted at any time and does not require that a full byte of data be read.
In addition to reading the individual Sector Protection Registers, the Software Protection Status (SWP) bits in the Status
Register can be read to determine if all, some, or none of the sectors are software protected (See “Read Status Register”
on page 35 for more details).
Figure 9-5. Read Sector Protection Register
Output Data Sector Protection Register Value
00h Sector Protection Register value is 0 (sector is unprotected).
FFh Sector Protection Register value is 1 (sector is protected).
SCK
CS
SI
SO
MSB MSB
2310
00111100
675410119812 373833 36353431 3229 30 39 40
Opcode
AAAA AAAAA
MSB MSB
DDDDDDDDDD
Address Bits A23-A0
Data Byte
High-impedance
28
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
9.7 Protected States and the Write Protect (WP) Pin
The WP pin is not linked to the memory array itself and has no direct effect on the protection status or lockdown status of
the memory array. Instead, the WP pin, in conjunction with the SPRL (Sector Protection Registers Locked) bit in the
Status Register, is used to control the hardware locking mechanism of the device. For hardware locking to be active, two
conditions must be met-the WP pin must be asserted and the SPRL bit must be in the Logical 1 state.
When hardware locking is active, the Sector Protection Registers are locked and the SPRL bit itself is also locked.
Therefore, sectors that are protected will be locked in the protected state, and sectors that are unprotected will be locked
in the unprotected state. These states cannot be changed as long as hardware locking is active, so the Protect Sector,
Unprotect Sector, and Write Status Register commands will be ignored. In order to modify the protection status of a
sector, the WP pin must first be deasserted, and the SPRL bit in the Status Register must be reset back to the Logical 0
state using the Write Status Register command. When resetting the SPRL bit back to a Logical 0, it is not possible to
perform a Global Protect or Global Unprotect at the same time because the Sector Protection Registers remain
soft-locked until after the Write Status Register command has been executed.
If the WP pin is permanently connected to GND, then once the SPRL bit is set to a Logical 1, the only way to reset the bit
back to the Logical 0 state is to power-cycle the device. This allows a system to power-up with all sectors software
protected but not hardware locked. Therefore, sectors can be unprotected and protected as needed and then hardware
locked at a later time by simply setting the SPRL bit in the Status Register.
When the WP pin is deasserted, or if the WP pin is permanently connected to VCC, the SPRL bit in the Status Register
can still be set to a Logical 1 to lock the Sector Protection Registers. This provides a software locking ability to prevent
erroneous Protect Sector or Unprotect Sector commands from being processed. When changing the SPRL bit to a
Logical 1 from a Logical 0, it is also possible to perform a Global Protect or Global Unprotect at the same time by writing
the appropriate values into bits 5, 4, 3, and 2 of the first byte of the Status Register.
Tables 9-4 and 9-5 detail the various protection and locking states of the device.
Table 9-4. Sector Protection Register States
Note: 1. “n” represents a sector number.
Table 9-5. Hardware and Software Locking
WP
Sector Protection Register
n(1)
Sector
n(1)
X
(Don’t Care)
0 Unprotected
1 Protected
WP SPRL Locking SPRL Change Allowed Sector Protection Registers
0 0 Can be modified from 0 to 1
Unlocked and modifiable using the Protect and Unprotect
Sector commands. Global Protect and Unprotect can also be
performed.
0 1 Hardware
Locked Locked
Locked in current state. Protect and Unprotect Sector
commands will be ignored. Global Protect and Unprotect
cannot be performed.
1 0 Can be modified from 0 to 1
Unlocked and modifiable using the Protect and Unprotect
Sector commands. Global Protect and Unprotect can also be
performed.
1 1 Software
Locked Can be modified from 1 to 0
Locked in current state. Protect and Unprotect Sector
commands will be ignored. Global Protect and Unprotect
cannot be performed.
29
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
10. Security Commands
10.1 Sector Lockdown
Certain applications require that portions of the Flash memory array be permanently protected against malicious
attempts at altering program code, data modules, security information or encryption/decryption algorithms, keys, and
routines. To address these applications, the device incorporates a sector lockdown mechanism that allows any
combination of individual 64KB sectors to be permanently locked so that they become read-only. Once a sector is locked
down, it can never be erased or programmed again, and it can never be unlocked from the locked down state.
Each 64KB physical sector has a corresponding single-bit Sector Lockdown Register that is used to control the lockdown
status of that sector. These registers are nonvolatile and will retain their state even after a device power-cycle or reset
operation. The following table outlines the two states of the Sector Lockdown Registers.
Table 10-1. Sector Lockdown Register Values
Issuing the Sector Lockdown command to a particular sector address will set the corresponding Sector Lockdown
Register to the Logical 1 state. Each Sector Lockdown Register can only be set once; therefore, once set to the Logical 1
state, a Sector Lockdown Register cannot be reset back to the Logical 0 state.
Before the Sector Lockdown command can be issued, the Write Enable command must have been previously issued to
set the WEL bit in the Status Register to a Logical 1. In addition, the Sector Lockdown Enabled (SLE) bit in the Status
Register must have also been previously set to the Logical 1 state by using the Write Status Register Byte 2 command
(see “Write Status Register Byte 2” on page 40). To issue the Sector Lockdown command, the CS pin must first be
asserted and then the opcode 33h must be clocked into the device followed by three address bytes designating any
address within the 64KB sector to be locked down. After the three address bytes have been clocked in, a confirmation
byte of D0h must also be clocked in immediately following the three address bytes. Any additional data clocked into the
device after the first byte of data will be ignored. When the CS pin is deasserted, the Sector Lockdown Register
corresponding to the sector addressed by A23-A0 will be set to the Logical 1 state, and the sector itself will then be
permanently locked down from program and erase operations within a time of tLOCK. In addition, the WEL bit in the
Status Register will be reset back to the Logical 0 state.
The three complete address bytes and the correct confirmation byte value of D0h must be clocked into the device before
the CS pin is deasserted, and the CS pin must be deasserted on an even byte boundary (multiples of eight bits);
otherwise, the device will abort the operation. When the device aborts the Sector Lockdown operation, the state of the
corresponding Sector Lockdown Register as well as the SLE bit in the Status Register will be unchanged; however, the
WEL bit in the Status Register will be reset to a Logical 0.
As a safeguard against accidental or erroneous locking down of sectors, the Sector Lockdown command can be enabled
and disabled as needed by using the SLE bit in the Status Register. In addition, the current sector lockdown state can be
frozen so that no further modifications to the Sector Lockdown Registers can be made (see “Freeze Sector Lockdown
State”). If the Sector Lockdown command is disabled or if the sector lockdown state is frozen, then any attempts to issue
the Sector Lockdown command will be ignored, and the device will reset the WEL bit in the Status Register back to a
Logical 0 and return to the idle state once the CS pin has been deasserted.
Value Sector Lockdown Status
0Sector is not locked down and can be programmed and erased. This is the default state.
1 Sector is permanently locked down and can never be programmed or erased again.
30
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
Figure 10-1. Sector Lockdown
10.2 Freeze Sector Lockdown State
The current sector lockdown state can be permanently frozen so that no further modifications to the Sector Lockdown
Registers can be made; therefore, the Sector Lockdown command will be permanently disabled, and no additional
sectors can be locked down aside from those already locked down. Any attempts to issue the Sector Lockdown
command after the sector lockdown state has been frozen will be ignored.
Before the Freeze Sector Lockdown State command can be issued, the Write Enable command must have been
previously issued to set the WEL bit in the Status Register to a Logical 1. In addition, the Sector Lockdown Enabled
(SLE) bit in the Status Register must have also been previously set to the Logical 1 state. To issue the Freeze Sector
Lockdown State command, the CS pin must first be asserted and the opcode of 34h must be clocked into the device,
followed by three command-specific address bytes of 55AA40h. After the three address bytes have been clocked in, a
confirmation byte of D0h must be clocked in immediately following the three address bytes. Any additional data clocked
into the device will be ignored. When the CS pin is deasserted, the current sector lockdown state will be permanently
frozen within a time of tLOCK. In addition, the WEL bit in the Status Register will be reset back to the Logical 0 state, and
the SLE bit will be permanently reset to a Logical 0 to indicate that the Sector Lockdown command is permanently
disabled.
The three complete and correct address bytes and the confirmation byte must be clocked into the device before the CS
pin is deasserted, and the CS pin must be deasserted on an even byte boundary (multiples of eight bits); otherwise, the
device will abort the operation. When the device aborts the Freeze Sector Lockdown State operation, the WEL bit in the
Status Register will be reset to a Logical 0; however, the state of the SLE bit will be unchanged.
Figure 10-2. Freeze Sector Lockdown State
SCK
CS
SI
SO
MSB MSB
2310
00110011
6754983937 3833 36353431 3229 30
Opcode
High-impedance
AA AAAA
MSB
11010000
Address Bits A23-A0 Confirmation Byte In
SCK
CS
SI
SO
MSB MSB
2310
00110100
6754983937 3833 36353431 3229 30
Opcode
High-impedance
01 0000
MSB
11010000
Address Bits A23-A0 Confirmation Byte In
31
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
10.3 Read Sector Lockdown Registers
The Sector Lockdown Registers can be read to determine the current lockdown status of each physical 64KB sector. To
read the Sector Lockdown Register for a particular 64KB sector, the CS pin must first be asserted and then the opcode
35h must be clocked in. Once the opcode has been clocked in, three address bytes designating any address within the
64KB sector must be clocked in. After the address bytes have been clocked in, data will be output on the SO pin during
every subsequent clock cycle. The data being output will be a repeating byte of either FFh or 00h to denote the value of
the appropriate Sector Lockdown Register.
At clock frequencies above fCLK, the first byte of data output will not be valid. Therefore, if operating at clock frequencies
above fCLK, at least two bytes of data must be clocked out from the device in order to determine the correct status of the
appropriate Sector Lockdown Register.
Table 10-2. Read Sector Lockdown Register – Output Data
Deasserting the CS pin will terminate the read operation and put the SO pin into a high-impedance state. The CS pin can
be deasserted at any time and does not require that a full byte of data be read.
Figure 10-3. Read Sector Lockdown Register
Output Data Sector Lockdown Register Value
00h Sector Lockdown Register value is 0 (Sector is not locked down).
FFh Sector Lockdown Register value is 1 (Sector is permanently locked down).
SCK
CS
SI
SO
MSB MSB
2310
00110101
675410119812 394243414037 3833 36353431 3229 30 44 47 484645
Opcode
AAAA AAAAA
MSB
XXXXXXXX
MSB MSB
DDDDDDDDDD
Address Bits A23-A0 Dummy Byte
Data Byte
High-impedance
32
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
10.4 Program OTP Security Register
The device contains a specialized OTP (One-Time Programmable) Security Register that can be used for purposes such
as unique device serialization, system-level Electronic Serial Number (ESN) storage, locked key storage, etc. The OTP
Security Register is independent of the main Flash memory array and is comprised of a total of 128 bytes of memory
divided into two portions. The first 64 bytes (byte locations 0 through 63) of the OTP Security Register are allocated as a
one-time user-programmable space. Once these 64 bytes have been programmed, they cannot be erased or
reprogrammed. The remaining 64 bytes of the OTP Security Register (byte locations 64 through 127) are factory
programmed by Atmel and will contain a unique value for each device. The factory programmed data is fixed and cannot
be changed.
Table 10-3. OTP Security Register
The user-programmable portion of the OTP Security Register does not need to be erased before it is programmed. In
addition, the Program OTP Security Register command operates on the entire 64-byte user-programmable portion of the
OTP Security Register at one time. Once the user-programmable space has been programmed with any number of
bytes, the user-programmable space cannot be programmed again; therefore, it is not possible to only program the first
two bytes of the register and then program the remaining 62 bytes at a later time.
Before the Program OTP Security Register command can be issued, the Write Enable command must have been
previously issued to set the WEL bit in the Status Register to a Logical 1. To program the OTP Security Register, the CS
pin must first be asserted and then the opcode 9Bh must be clocked into the device followed by the three address bytes
denoting the first byte location of the OTP Security Register to begin programming at. Since the size of the
user-programmable portion of the OTP Security Register is 64 bytes, the upper order address bits do not need to be
decoded by the device. Therefore, address bits A23-A6 will be ignored by the device and their values can be either a
Logical 1 or 0. After the address bytes have been clocked in, data can then be clocked into the device and will be stored
in the internal buffer.
If the starting memory address denoted by A23-A0 does not start at the beginning of the OTP Security Register memory
space (A5-A0 are not all 0), then special circumstances regarding which OTP Security Register locations to be
programmed will apply. In this situation, any data that is sent to the device that goes beyond the end of the 64-byte
user-programmable space will wrap around back to the beginning of the OTP Security Register. In addition, if more than
64 bytes of data are sent to the device, then only the last 64 bytes sent will be latched into the internal buffer.
Example: If the starting address denoted by A23-A0 is 00003Eh, and three bytes of data are sent to the device, then
the first two bytes of data will be programmed at OTP Security Register addresses 00003Eh and 00003Fh
while the last byte of data will be programmed at address 000000h. The remaining bytes in the OTP
Security Register (addresses 000001h through 00003Dh) will not be programmed and will remain in the
erased state (FFh).
When the CS pin is deasserted, the device will take the data stored in the internal buffer and program it into the
appropriate OTP Security Register locations based on the starting address specified by A23-A0 and the number of data
bytes sent to the device. If less than 64 bytes of data were sent to the device, then the remaining bytes within the OTP
Security Register will not be programmed and will remain in the erased state (FFh). The programming of the data bytes
is internally self-timed and should take place in a time of tOTPP. It is not possible to suspend the programming of the OTP
Security Register.
Security Register Byte Number
0 1 . . . 62 63 64 65 . . . 126 127
One-Time User Programmable Factory Programmed by Atmel
33
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
The three address bytes and at least one complete byte of data must be clocked into the device before the CS pin is
deasserted, and the CS pin must be deasserted on even byte boundaries (multiples of eight bits); otherwise, the device
will abort the operation and the user-programmable portion of the OTP Security Register will not be programmed. The
WEL bit in the Status Register will be reset back to the Logical 0 state if the OTP Security Register program cycle aborts
due to an incomplete address being sent, an incomplete byte of data being sent, the CS pin being deasserted on uneven
byte boundaries, or because the user-programmable portion of the OTP Security Register was previously programmed.
While the device is programming the OTP Security Register, the Status Register can be read and will indicate that the
device is busy. For faster throughput, it is recommended that the Status Register be polled rather than waiting the tOTPP
time to determine if the data bytes have finished programming. At some point before the OTP Security Register
programming completes, the WEL bit in the Status Register will be reset back to the Logical 0 state.
If the device is powered-down during the OTP Security Register program cycle, then the contents of the 64-byte
user-programmable portion of the OTP Security Register cannot be guaranteed and cannot be programmed again.
The Program OTP Security Register command utilizes the 256-byte internal buffer for processing. Therefore, the
contents of the buffer will be altered from its previous state when this command is issued.
Figure 10-4. Program OTP Security Register
SCK
CS
SI
SO
MSB MSB
2310
10011011
6754983937 3833 36353431 3229 30
Opcode
High-impedance
AA AAAA
MSB
DDDDDDDD
Address Bits A23-A0 Data In Byte 1
MSB
DDDDDDDD
Data In Byte n
34
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
10.5 Read OTP Security Register
The OTP Security Register can be sequentially read in a similar fashion to the Read Array operation up to the maximum
clock frequency specified by fMAX. To read the OTP Security Register, the CS pin must first be asserted and then the
opcode 77h must be clocked into the device. After the opcode has been clocked in, the three address bytes must be
clocked in to specify the starting address location of the first byte to read within the OTP Security Register. Following the
three address bytes, two dummy bytes must be clocked into the device before data can be output.
After the three address bytes and the dummy bytes have been clocked in, additional clock cycles will result in OTP
Security Register data being output on the SO pin. When the last byte (00007Fh) of the OTP Security Register has been
read, the device will continue reading back at the beginning of the register (000000h). No delays will be incurred when
wrapping around from the end of the register to the beginning of the register.
Deasserting the CS pin will terminate the read operation and put the SO pin into a high-impedance state. The CS pin can
be deasserted at any time and does not require that a full byte of data be read.
Figure 10-5. Read OTP Security Register
SCK
CS
SI
SO
MSB MSB
2310
01110111
675410119812 3336353431 3229 30
Opcode
AAAA AAAAA XXX
MSB MSB
DDDDDDDDDD
Address Bits A23-A0
MSB
XXXXXX
Dummy Byte
Data Byte 1
High-impedance
35
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
11. Status Register Commands
11.1 Read Status Register
The 2-byte Status Register can be read to determine the device’s ready/busy status, as well as the status of many other
functions such as Hardware Locking and Software Protection. The Status Register can be read at any time, including
during an internally self-timed program or erase operation.
To read the Status Register, the CS pin must first be asserted and then the opcode 05h must be clocked into the device.
After the opcode has been clocked in, the device will begin outputting Status Register data on the SO pin during every
subsequent clock cycle. After the second byte of the Status Register has been clocked out, the sequence will repeat
itself, starting again with the first byte of the Status Register as long as the CS pin remains asserted and the clock pin is
being pulsed. The data in the Status Register is constantly being updated, so each repeating sequence will output new
data. The RDY/BSY status is available for both bytes of the Status Register and is updated for each byte.
At clock frequencies above fCLK, the first two bytes of data output from the Status Register will not be valid. Therefore, if
operating at clock frequencies above fCLK, at least four bytes of data must be clocked out from the device in order to read
the correct values of both bytes of the Status Register.
Deasserting the CS pin will terminate the Read Status Register operation and put the SO pin into a high-impedance
state. The CS pin can be deasserted at any time and does not require that a full byte of data be read.
Table 11-1. Status Register Format – Byte 1
Notes: 1. Only bit 7 of Status Register Byte 1 will be modified when using the Write Status Register Byte 1 command.
2. R/W = Readable and Writeable
R = Readable only
Bit(1) Name Type(2) Description
7 SPRL Sector Protection Registers Locked R/W
0 Sector Protection Registers are unlocked (default).
1 Sector Protection Registers are locked.
6 RES Reserved for future use R 0 Reserved for future use.
5 EPE Erase/Program Error R
0 Erase or program operation was successful.
1 Erase or program error detected.
4 WPP Write Protect (WP) Pin Status R
0 WP is asserted.
1 WP is deasserted.
3:2 SWP Software Protection Status R
00 All sectors are software unprotected (all Sector
Protection Registers are 0).
01
Some sectors are software protected. Read
individual Sector Protection Registers to determine
which sectors are protected.
10 Reserved for future use.
11 All sectors are software protected (all Sector
Protection Registers are 1 – default).
1 WEL Write Enable Latch Status R
0 Device is not write enabled (default).
1 Device is write enabled.
0 RDY/BSY Ready/Busy Status R
0Device is ready.
1 Device is busy with an internal operation.
36
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
Table 11-2. Status Register Format – Byte 2
Notes: 1. Only bits 4 and 3 of Status Register Byte 2 will be modified when using the Write Status Register Byte 2 command.
2. R/W = Readable and Writeable
R = Readable only
11.1.1 SPRL Bit
The SPRL bit is used to control whether the Sector Protection Registers can be modified or not. When the SPRL bit is in
the Logical 1 state, all Sector Protection Registers are locked and cannot be modified with the Protect Sector and
Unprotect Sector commands (the device will ignore these commands). In addition, the Global Protect and Global
Unprotect features cannot be performed. Any sectors that are presently protected will remain protected, and any sectors
that are presently unprotected will remain unprotected.
When the SPRL bit is in the Logical 0 state, all Sector Protection Registers are unlocked and can be modified (the
Protect Sector and Unprotect Sector commands, as well as the Global Protect and Global Unprotect features, will be
processed as normal). The SPRL bit defaults to the Logical 0 state after device power-up. The Reset command has no
effect on the SPRL bit.
The SPRL bit can be modified freely whenever the WP pin is deasserted. However, if the WP pin is asserted, then the
SPRL bit may only be changed from a Logical 0 (Sector Protection Registers are unlocked) to a Logical 1 (Sector
Protection Registers are locked). In order to reset the SPRL bit back to a Logical 0 using the Write Status Register
Byte 1 command, the WP pin will have to first be deasserted.
The SPRL bit is the only bit of Status Register Byte 1 that can be user modified via the Write Status Register Byte 1
command.
Bit(1) Name Type(2) Description
7 RES Reserved for future use R 0 Reserved for future use.
6 RES Reserved for future use R 0 Reserved for future use.
5 RES Reserved for future use R 0 Reserved for future use.
4 RSTE Reset Enabled R/W
0 Reset command is disabled (default).
1 Reset command is enabled.
3 SLE Sector Lockdown Enabled R/W
0Sector Lockdown and Freeze Sector Lockdown State
commands are disabled (default).
1Sector Lockdown and Freeze Sector Lockdown State
commands are enabled.
2 PS Program Suspend Status R
0 No sectors are program suspended (default).
1 A sector is program suspended.
1 ES Erase Suspend Status R
0 No sectors are erase suspended (default).
1 A sector is erase suspended.
0 RDY/BSY Ready/Busy Status R
0Device is ready.
1 Device is busy with an internal operation.
37
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
11.1.2 EPE Bit
The EPE bit indicates whether the last erase or program operation completed successfully or not. If at least one byte
during the erase or program operation did not erase or program properly, then the EPE bit will be set to the Logical 1
state. The EPE bit will not be set if an erase or program operation aborts for any reason such as an attempt to erase or
program a protected region or a locked down sector, an attempt to erase or program a suspended sector, or if the WEL
bit is not set prior to an erase or program operation. The EPE bit will be updated after every erase and program
operation.
11.1.3 WPP Bit
The WPP bit can be read to determine if the WP pin has been asserted or not.
11.1.4 SWP Bits
The SWP bits provide feedback on the software protection status for the device. There are three possible combinations
of the SWP bits that indicate whether none, some, or all of the sectors have been protected using the Protect Sector
command or the Global Protect feature. If the SWP bits indicate that some of the sectors have been protected, then the
individual Sector Protection Registers can be read with the Read Sector Protection Registers command to determine
which sectors are in fact protected.
11.1.5 WEL Bit
The WEL bit indicates the current status of the internal Write Enable Latch. When the WEL bit is in the Logical 0 state,
the device will not accept any Byte/Page Program, erase, Protect Sector, Unprotect Sector, Sector Lockdown, Freeze
Sector Lockdown State, Program OTP Security Register, or Write Status Register commands. The WEL bit defaults to
the Logical 0 state after a device power-up or reset operation. In addition, the WEL bit will be reset to the Logical 0 state
automatically under the following conditions:
Write Disable operation completes successfully
Write Status Register operation completes successfully or aborts
Protect Sector operation completes successfully or aborts
Unprotect Sector operation completes successfully or aborts
Sector Lockdown operation completes successfully or aborts
Freeze Sector Lockdown State operation completes successfully or aborts
Program OTP Security Register operation completes successfully or aborts
Byte/Page Program operation completes successfully or aborts
Block Erase operation completes successfully or aborts
Chip Erase operation completes successfully or aborts
Hold condition aborts
If the WEL bit is in the Logical 1 state, it will not be reset to a Logical 0 if an operation aborts due to an incomplete or
unrecognized opcode being clocked into the device before the CS pin is deasserted. In order for the WEL bit to be reset
when an operation aborts prematurely, the entire opcode for a Byte/Page Program, erase, Protect Sector, Unprotect
Sector, Sector Lockdown, Freeze Sector Lockdown State, Program OTP Security Register, or Write Status Register
command must have been clocked into the device.
38
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
11.1.6 RSTE Bit
The RSTE bit is used to enable or disable the Reset command. When the RSTE bit is in the Logical 0 state (the default
state after power-up), the Reset command is disabled and any attempts to reset the device using the Reset command
will be ignored. When the RSTE bit is in the Logical 1 state, the Reset command is enabled.
The RSTE bit will retain its state as long as power is applied to the device. Once set to the Logical 1 state, the RSTE bit
will remain in that state until it is modified using the Write Status Register Byte 2 command or until the device has been
power cycled. The Reset command itself will not change the state of the RSTE bit.
11.1.7 SLE Bit
The SLE bit is used to enable and disable the Sector Lockdown and Freeze Sector Lockdown State commands. When
the SLE bit is in the Logical 0 state (the default state after power-up), the Sector Lockdown and Freeze Sector Lockdown
commands are disabled. If the Sector Lockdown and Freeze Sector Lockdown commands are disabled, then any
attempts to issue the commands will be ignored. This provides a safeguard for these commands against accidental or
erroneous execution. When the SLE bit is in the Logical 1 state, the Sector Lockdown and Freeze Sector Lockdown
State commands are enabled.
Unlike the WEL bit, the SLE bit does not automatically reset after certain device operations. Therefore, once set, the SLE
bit will remain in the Logical 1 state until it is modified using the Write Status Register Byte 2 command or until the device
has been power cycled. The Reset command has no effect on the SLE bit.
If the Freeze Sector Lockdown State command has been issued, then the SLE bit will be permanently reset in the
Logical 0 state to indicate that the Sector Lockdown command has been disabled.
11.1.8 PS Bit
The PS bit indicates whether or not a sector is in the Program Suspend state.
11.1.9 ES Bit
The ES bit indicates whether or not a sector is in the Erase Suspend state.
11.1.10 RDY/BSY Bit
The RDY/BSY bit is used to determine whether or not an internal operation, such as a program or erase, is in progress.
To poll the RDY/BSY bit to detect the completion of a program or erase cycle, new Status Register data must be
continually clocked out of the device until the state of the RDY/BSY bit changes from a Logical 1 to a Logical 0.
Figure 11-1. Read Status Register
SCK
CS
SI
SO
MSB
2310
00000101
675410119812 212217 20191815 1613 14 23 24 28 29272625 30
Opcode
MSB MSB
DDDDDD DDDD
MSB
DDDDDDDD D DDD D D
Status Register
Byte 1
Status Register
Byte 1
Status Register
Byte 2
High-impedance
39
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
11.2 Write Status Register Byte 1
The Write Status Register Byte 1 command is used to modify the SPRL bit of the Status Register and/or to perform a
Global Protect or Global Unprotect operation. Before the Write Status Register Byte 1 command can be issued, the Write
Enable command must have been previously issued to set the WEL bit in the Status Register to a Logical 1.
To issue the Write Status Register Byte 1 command, the CS pin must first be asserted and the opcode of 01h must be
clocked into the device followed by one byte of data. The one byte of data consists of the SPRL bit value, a dummy bit,
four data bits to denote whether a Global Protect or Unprotect should be performed, and two additional dummy bits (see
Table 11-3). Any additional data bytes that are sent to the device will be ignored. When the CS pin is deasserted, the
SPRL bit in the Status Register will be modified, and the WEL bit in the Status Register will be reset back to a
Logical 0. The values of bits 5, 4, 3, and 2 and the state of the SPRL bit before the Write Status Register Byte 1
command was executed (the prior state of the SPRL bit) will determine whether or not a Global Protect or Global
Unprotect will be performed. See “Global Protect/Unprotect” on page 25 for more details.
The complete one byte of data must be clocked into the device before the CS pin is deasserted, and the CS pin must be
deasserted on even byte boundaries (multiples of eight bits); otherwise, the device will abort the operation, the state of
the SPRL bit will not change, no potential Global Protect or Unprotect will be performed, and the WEL bit in the Status
Register will be reset back to the Logical 0 state.
If the WP pin is asserted, then the SPRL bit can only be set to a Logical 1. If an attempt is made to reset the SPRL bit to
a Logical 0 while the WP pin is asserted, then the Write Status Register Byte 1 command will be ignored, and the WEL bit
in the Status Register will be reset back to the Logical 0 state. In order to reset the SPRL bit to a Logical 0, the WP pin
must be deasserted.
Table 11-3. Write Status Register Byte 1 Format
Figure 11-2. Write Status Register Byte 1
Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Bit 0
SPRL X Global Protect/Unprotect X X
SCK
CS
SI
SO
MSB
2310
0000000
6754
Opcode
10 119814151312
1
MSB
DXDDDDXX
Status Register In
Byte 1
High-impedance
40
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
11.3 Write Status Register Byte 2
The Write Status Register Byte 2 command is used to modify the RSTE and SLE bits of the Status Register. Using the
Write Status Register Byte 2 command is the only way to modify the RSTE and SLE bits in the Status Register during
normal device operation, and the SLE bit can only be modified if the sector lockdown state has not been frozen. Before
the Write Status Register Byte 2 command can be issued, the Write Enable command must have been previously issued
to set the WEL bit in the Status Register to a Logical 1.
To issue the Write Status Register Byte 2 command, the CS pin must first be asserted and then the opcode 31h must be
clocked into the device followed by one byte of data. The one byte of data consists of three dummy bits, the RSTE bit
value, the SLE bit value, and three additional dummy bits (see Table 11-4). Any additional data bytes that are sent to the
device will be ignored. When the CS pin is deasserted, the RSTE and SLE bits in the Status Register will be modified,
and the WEL bit in the Status Register will be reset back to a Logical 0. The SLE bit will only be modified if the Freeze
Sector Lockdown State command has not been previously issued.
The complete one byte of data must be clocked into the device before the CS pin is deasserted, and the CS pin must be
deasserted on even byte boundaries (multiples of eight bits); otherwise, the device will abort the operation, the state of
the RSTE and SLE bits will not change, and the WEL bit in the Status Register will be reset back to the Logical 0 state.
Table 11-4. Write Status Register Byte 2 Format
Figure 11-3. Write Status Register Byte 2
Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Bit 0
X X X RSTE SLE X X X
SCK
CS
SI
SO
MSB
2310
0011000
6754
Opcode
10 119814151312
1
MSB
XXXDDXXX
Status Register In
Byte 2
High-impedance
41
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
12. Other Commands and Functions
12.1 Reset
In some applications, it may be necessary to prematurely terminate a program or erase cycle early rather than wait the
hundreds of microseconds or milliseconds necessary for the program or erase operation to complete normally. The
Reset command allows a program or erase operation in progress to be ended abruptly and returns the device to an idle
state. Since the need to reset the device is immediate, the Write Enable command does not need to be issued prior to the
Reset command being issued. Therefore, the Reset command operates independently of the state of the WEL bit in the
Status Register.
The Reset command can only be executed if the command has been enabled by setting the Reset Enabled (RSTE) bit in
the Status Register to a Logical 1. If the Reset command has not been enabled (the RSTE bit is in the Logical 0 state),
then any attempts at executing the Reset command will be ignored.
To perform a Reset, the CS pin must first be asserted and then the opcode F0h must be clocked into the device. No
address bytes need to be clocked in, but a confirmation byte of D0h must be clocked into the device immediately after the
opcode. Any additional data clocked into the device after the confirmation byte will be ignored. When the CS pin is
deasserted, the program or erase operation currently in progress will be terminated within a time of tRST. Since the
program or erase operation may not complete before the device is reset, the contents of the page being programmed or
the block being erased cannot be guaranteed to be valid.
The Reset command has no effect on the states of the Sector Protection Registers, the Sector Lockdown Registers, or
the SPRL, RSTE, and SLE bits in the Status Register. The WEL, PS, and ES bits, however, will be reset back to their
default states. If a Reset operation is performed while a sector is erase suspended, the suspend operation will abort, and
the contents of the block being erased in the suspended sector will be left in an undefined state. If a Reset is performed
while a sector is program suspended, the suspend operation will abort, and the contents of the page that was being
programmed and subsequently suspended will be undefined. The remaining pages in the 64KB sector will retain their
previous contents.
The complete opcode and confirmation byte must be clocked into the device before the CS pin is deasserted, and the CS
pin must be deasserted on an even byte boundary (multiples of eight bits); otherwise, no Reset operation will be
performed.
Figure 12-1. Reset
SCK
CS
SI
SO
MSB
2310
1111000
6754
Opcode Confirmation Byte In
10 119814151312
0
MSB
11010000
High-impedance
42
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
12.2 Read Manufacturer and Device ID
Identification information can be read from the device to enable systems to electronically query and identify the device
while it is in system. The identification method and the command opcode comply with the JEDEC standard for
“Manufacturer and Device ID Read Methodology for SPI Compatible Serial Interface Memory Devices”. The type of
information that can be read from the device includes the JEDEC-defined Manufacturer ID, the vendor-specific Device
ID, and the vendor-specific Extended Device Information.
The Read Manufacturer and Device ID command is limited to a maximum clock frequency of fCLK. Since not all Flash
devices are capable of operating at very high clock frequencies, applications should be designed to read the
identification information from the devices at a reasonably low clock frequency to ensure that all devices to be used in the
application can be identified properly. Once the identification process is complete, the application can then increase the
clock frequency to accommodate specific Flash devices that are capable of operating at the higher clock frequencies.
To read the identification information, the CS pin must first be asserted and then the opcode 9Fh must be clocked into
the device. After the opcode has been clocked in, the device will begin outputting the identification data on the SO pin
during the subsequent clock cycles. The first byte that will be output will be the Manufacturer ID, followed by two bytes of
Device ID information. The fourth byte output will be the Extended Device Information (EDI) String Length, which will be
01h indicating that one byte of EDI data follows. After the one byte of EDI data is output, the SO pin will go into a
high-impedance state; therefore, additional clock cycles will have no affect on the SO pin and no data will be output. As
indicated in the JEDEC standard, reading the EDI String Length and any subsequent data is optional.
Deasserting the CS pin will terminate the Manufacturer and Device ID read operation and put the SO pin into a
high-impedance state. The CS pin can be deasserted at any time and does not require that a full byte of data be read.
Table 12-1. Manufacturer and Device ID Information
Table 12-2. Manufacturer and Device ID Details
Table 12-3. EDI Data
Byte No. Data Type Value
1 Manufacturer ID 1Fh
2 Device ID (Part 1) 48h
3 Device ID (Part 2) 00h
4 [Optional to Read] Extended Device Information (EDI) String Length 01h
5 [Optional to Read] EDI Byte 1 00h
Data Type Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Bit 0
Hex
Value Details
Manufacturer ID
JEDEC Assigned Code
1Fh JEDEC Code: 0001 1111 (1Fh for Atmel)
0 0 0 1 1 1 1 1
Device ID (Byte 1)
Family Code Density Code
48h Family Code: 010 (SPI or Dual-I/O)
Density Code: 01000 (64-Mbit)
0 1 0 0 1 0 0 0
Device ID (Byte 2)
Sub Code Product Variant
00h Sub Code: 000 (Standard Series)
Product Variant: 00000
0 0 0 0 0 0 0 0
Byte Number Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Bit 0
Hex
Value Details
1
RFU Device Revision
00h RFU: Reserved for Future Use
Device Revision: 00000 (Initial Version)
00000000
43
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
Figure 12-2. Read Manufacturer and Device ID
SCK
CS
SI
SO
60
9Fh
87 46
Opcode
1Fh 48h 00h 01h 00h
Manufacturer ID Device ID
Byte 1
Device ID
Byte 2
EDI
String Length
EDI
Data Byte 1
High-impedance
14 1615 22 2423 38 403930 3231
Note: Each transition shown for SI and SO represents one byte (8 bits)
44
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
12.3 Deep Power-Down
During normal operation, the device will be placed in the standby mode to consume less power as long as the CS pin
remains deasserted and no internal operation is in progress. The Deep Power-Down command offers the ability to place
the device into an even lower power consumption state called the Deep Power-Down mode.
When the device is in the Deep Power-Down mode, all commands including the Read Status Register command will be
ignored with the exception of the Resume from Deep Power-Down command. Since all commands will be ignored, the
mode can be used as an extra protection mechanism against program and erase operations.
Entering the Deep Power-Down mode is accomplished by simply asserting the CS pin, clocking in the opcode B9h, and
then deasserting the CS pin. Any additional data clocked into the device after the opcode will be ignored. When the CS
pin is deasserted, the device will enter the Deep Power-Down mode within the maximum time of tEDPD.
The complete opcode must be clocked in before the CS pin is deasserted, and the CS pin must be deasserted on an
even byte boundary (multiples of eight bits); otherwise, the device will abort the operation and return to the standby mode
once the CS pin is deasserted. In addition, the device will default to the standby mode after a power-cycle.
The Deep Power-Down command will be ignored if an internally self-timed operation such as a program or erase cycle is
in progress. The Deep Power-Down command must be reissued after the internally self-timed operation has been
completed in order for the device to enter the Deep Power-Down mode.
Figure 12-3. Deep Power-Down
SCK
CS
SI
SO
MSB
ICC
2310
10111001
6754
Opcode
High-impedance
Standby Mode Current
Active Current
Deep Power-Down Mode Current
tEDPD
45
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
12.4 Resume from Deep Power-Down
In order to exit the Deep Power-Down mode and resume normal device operation, the Resume from Deep Power-Down
command must be issued. The Resume from Deep Power-Down command is the only command that the device will
recognized while in the Deep Power-Down mode.
To resume from the Deep Power-Down mode, the CS pin must first be asserted and then the opcode ABh must be
clocked into the device. Any additional data clocked into the device after the opcode will be ignored. When the CS pin is
deasserted, the device will exit the Deep Power-Down mode within the maximum time of tRDPD and return to the standby
mode. After the device has returned to the standby mode, normal command operations such as Read Array can be
resumed.
If the complete opcode is not clocked in before the CS pin is deasserted, or if the CS pin is not deasserted on an even
byte boundary (multiples of eight bits), then the device will abort the operation and return to the Deep Power-Down
mode.
Figure 12-4. Resume from Deep Power-Down
SCK
CS
SI
SO
MSB
ICC
2310
10101011
6754
Opcode
High-impedance
Deep Power-Down Mode Current
Active Current
Standby Mode Current
tRDPD
46
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
12.5 Hold
The HOLD pin is used to pause the serial communication with the device without having to stop or reset the clock
sequence. The Hold mode, however, does not have an affect on any internally self-timed operations such as a program
or erase cycle. Therefore, if an erase cycle is in progress, asserting the HOLD pin will not pause the operation, and the
erase cycle will continue until it is finished.
The Hold mode can only be entered while the CS pin is asserted. The Hold mode is activated simply by asserting the
HOLD pin during the SCK low pulse. If the HOLD pin is asserted during the SCK high pulse, then the Hold mode won’t be
started until the beginning of the next SCK low pulse. The device will remain in the Hold mode as long as the HOLD pin
and CS pin are asserted.
While in the Hold mode, the SO pin will be in a high-impedance state. In addition, both the SI pin and the SCK pin will be
ignored. The WP pin, however, can still be asserted or deasserted while in the Hold mode.
To end the Hold mode and resume serial communication, the HOLD pin must be deasserted during the SCK low pulse. If
the HOLD pin is deasserted during the SCK high pulse, then the Hold mode won’t end until the beginning of the next SCK
low pulse.
If the CS pin is deasserted while the HOLD pin is still asserted, then any operation that may have been started will be
aborted, and the device will reset the WEL bit in the Status Register back to the Logical 0 state.
Figure 12-5. Hold Mode
SCK
CS
HOLD
Hold HoldHold
47
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
13. Atmel RapidS Implementation
To implement RapidS and operate at clock frequencies higher than what can be achieved in a viable SPI implementation,
a full clock cycle can be used to transmit data back and forth across the serial bus. The AT25DF641A is designed to
always clock its data out on the falling edge of the SCK signal and clock data in on the rising edge of SCK.
For full clock cycle operation to be achieved, when the AT25DF641A is clocking data out on the falling edge of SCK, the
host controller should wait until the next falling edge of SCK to latch the data in. Similarly, the host controller should clock
its data out on the rising edge of SCK in order to give the AT25DF641A a full clock cycle to latch the incoming data in on
the next rising edge of SCK.
Implementing RapidS allows a system to run at higher clock frequencies since a full clock cycle is used to accommodate
a device’s clock-to-output time, input setup time, and associated rise/fall times.
Example: If the system clock frequency is 100MHz (10ns cycle time) with a 50% duty cycle, and the host controller
has an input setup time of 2ns, then a standard SPI implementation would require that the slave device be
capable of outputting its data in less than 3ns to meet the 2ns host controller setup time [(10ns x 50%) - 2ns]
not accounting for rise/fall times. In an SPI Mode 0 or 3 implementation, the SPI master is designed to clock
in data on the next immediate rising edge of SCK after the SPI slave has clocked its data out on the
preceding falling edge. This essentially makes SPI a half-clock cycle protocol and requires extremely fast
clock-to-output times and input setup times in order to run at high clock frequencies. With a RapidS
implementation of this example, however, the full 10ns cycle time is available which gives the slave device
up to 8ns, not accounting for rise/fall times, to clock its data out. Likewise, with RapidS, the host controller
has more time available to output its data to the slave since the slave device would be clocking that data in
a full clock cycle later.
Figure 13-1. Atmel RapidS Operation
SCK
MOSI
MISO
tV
1
234567
81
234567
8
MOSI = Master Out, Slave In MISO = Master In, Slave Out
The Master is the ASIC/MCU and the Slave is the memory device.
The Master always clocks data out on the rising edge of SCK and always clocks data in on the falling edge of SCK.
The Slave always clocks data out on the falling edge of SCK and always clocks data in on the rising edge of SCK.
A. Master clocks out first bit of BYTE A on the rising edge of SCK
B. Slave clocks in first bit of BYTE A on the next rising edge of SCK
C. Master clocks out second bit of BYTE A on the same rising edge of SCK
D. Last bit of BYTE A is clocked out from the Master
E. Last bit of BYTE A is clocked into the slave
F. Slave clocks out first bit of BYTE B
G. Master clocks in first bit of BYTE B
H. Slave clocks out second bit of BYTE B
I. Master clocks in last bit of BYTE B
A
B
C D
E
F
G
1
H
BYTE A
MSB LSB
BYTE B
MSB LSB
Slave CS
I
48
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
14. Electrical Specifications
14.1 Absolute Maximum Ratings*
14.2 DC and AC Operating Range
14.3 DC Characteristics
Temperature under Bias . . . . . . . . . . -55C to +125C
Storage Temperature . . . . . . . . . . . . -65C to +150C
All input voltages(including NC pins)
with respect to ground . . . . . . . . . . . . . .-0.6V to +4.1V
All output voltages
with respect to ground . . . . . . . . . .-0.6V to VCC + 0.5V
*Notice: Stresses beyond those listed under “Absolute Maximum
Ratings” may cause permanent damage to the device.
Functional operation of the device at these or any other
conditions beyond those indicated in the operational sections
of this specification is not implied. Exposure to absolute
maximum rating conditions for extended periods may affect
device reliability. Voltage extremes referenced in the “Absolute
Maximum Ratings” are intended to accomodate short duration
undershoot/overshoot conditions and does not imply or
guarantee functional device operation at these levels for any
extended period of time.
Atmel AT25DF641A
Operating Temperature (Case) Industrial -40C to 85C
VCC Power Supply 2.7V to 3.6V
Symbol Parameter Condition Min Typ Max Units
ISB Standby Current CS, WP, HOLD = VCC,
all inputs at CMOS levels 25 50 μA
IDPD Deep Power-down Current CS, WP, HOLD = VCC,
all inputs at CMOS levels 5 20 μA
ICC1 Active Current, Read Operation
f = 100MHz; IOUT = 0mA;
CS = VIL, VCC = Max 30 40
mA
f = 85MHz; IOUT = 0mA;
CS = VIL, VCC = Max 29 38
f = 66MHz; IOUT = 0mA;
CS = VIL, VCC = Max 28 37
f = 50MHz; IOUT = 0mA;
CS = VIL, VCC = Max 27 36
f = 33MHz; IOUT = 0mA;
CS = VIL, VCC = Max 26 35
f = 20MHz; IOUT = 0mA;
CS = VIL, VCC = Max 25 34
ICC2 Active Current, Program Operation CS = VCC, VCC = Max 12 18 mA
ICC3 Active Current, Erase Operation CS = VCC, VCC = Max 12 18 mA
ILI Input Leakage Current VIN = CMOS levels 1 μA
ILO Output Leakage Current VOUT = CMOS levels 1 μA
VIL Input Low Voltage 0.3 x VCC V
VIH Input High Voltage 0.7 x VCC V
VOL Output Low Voltage IOL = 1.6mA; VCC = Min 0.4 V
VOH Output High Voltage IOH = -100μA; VCC = Min VCC - 0.2V V
49
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
14.4 AC Characteristics - Maximum Clock Frequencies
14.5 AC Characteristics – All Other Parameter
Notes: 1. Not 100% tested (value guaranteed by design and characterization).
2. 15pF load at frequencies above 70MHz, 30pF otherwise.
3. Only applicable as a constraint for the Write Status Register Byte 1 command when SPRL = 1.
Symbol Parameter Min Max Units
Atmel RapidS and SPI Operation
fMAX
Maximum Clock Frequency for All Operations – Atmel RapidS Operation Only
(excluding 03h, 0Bh, 3Bh, and 9F opcodes) 100 MHz
fCLK Maximum Clock Frequency for All Operations (excluding 03h opcode) 85 MHz
fRDLF Maximum Clock Frequency for 03h Opcode (Read Array – Low Frequency) 40 MHz
fRDDO Maximum Clock Frequency for 3Bh Opcode (Dual-Output Read) 65 MHz
Symbol Parameter Min Max Units
tCLKH Clock High Time 4.3 ns
tCLKL Clock Low Time 4.3 ns
tCLKR(1) Clock Rise Time, Peak-to-peak (Slew Rate) 0.1 V/ns
tCLKF(1) Clock Fall Time, Peak-to-peak (Slew Rate) 0.1 V/ns
tCSH Chip Select High Time 50 ns
tCSLS Chip Select Low Setup Time (Relative to Clock) 5 ns
tCSLH Chip Select Low Hold Time (Relative to Clock) 5 ns
tCSHS Chip Select High Setup Time (Relative to Clock) 5 ns
tCSHH Chip Select High Hold Time (Relative to Clock) 5 ns
tDS Data In Setup Time 2 ns
tDH Data In Hold Time 1 ns
tDIS(1) Output Disable Time 5 ns
tV(2) Output Valid Time 5 ns
tOH Output Hold Time 2 ns
tHLS HOLD Low Setup Time (Relative to Clock) 5 ns
tHLH HOLD Low Hold Time (Relative to Clock) 5 ns
tHHS HOLD High Setup Time (Relative to Clock) 5 ns
tHHH HOLD High Hold Time (Relative to Clock) 5 ns
tHLQZ(1) HOLD Low to Output High-Z 5 ns
tHHQX(1) HOLD High to Output Low-Z 5 ns
tWPS(1)(3) Write Protect Setup Time 20 ns
tWPH(1)(3) Write Protect Hold Time 100 ns
tSECP(1) Sector Protect Time (From Chip Select High) 20 ns
tSECUP(1) Sector Unprotect Time (From Chip Select High) 20 ns
tLOCK(1) Sector Lockdown and Freeze Sector Lockdown State Time (From Chip Select High) 200 μs
tEDPD(1) Chip Select High to Deep Power-Down 1 μs
tRDPD(1) Chip Select High to Standby Mode 50 μs
tRST Reset Time 30 μs
50
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
14.6 Program and Erase Characteristics
Notes: 1. Maximum values indicate worst-case performance after 100,000 erase/program cycles.
2. Not 100% tested (value guaranteed by design and characterization).
14.7 Power-Up Conditions
14.8 Input Test Waveforms and Measurement Levels
14.9 Output Test Load
Symbol Parameter Min Typ Max Units
tPP(1) Page Program Time (256 bytes) 2.5 6.0 ms
tBP Byte Program Time 30 μs
tBLKE(1) Block Erase Time
4KB 75 200
ms32KB 300 600
64KB 600 1100
tCHPE(1)(2) Chip Erase Time 70 150 sec
tSUSP Program/Erase Suspend Time Program 10 20
μs
Erase 25 40
tRES Program/Erase Resume Time Program 10 20
μs
Erase 12 20
tOTPP(1) OTP Security Register Program Time 200 500 μs
tWRSR(2) Write Status Register Time 200 ns
Symbol Parameter Min Max Units
tVCSL Minimum VCC to Chip Select Low Time 70 μs
tPUW Power-Up Device Delay Before Program or Erase Allowed 10 ms
VPOR Power-On Reset Voltage 1.5 2.5 V
AC
Driving
Levels
AC
Measurement
Level
0.1VCC
VCC/2
0.9VCC
tR, tF < 2 ns (10% to 90%)
Device
Under
Test 15pF (frequencies above 70MHz)
or
30pF
51
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
15. AC Waveforms
Figure 15-1. Serial Input Timing
Figure 15-2. Serial Output Timing
Figure 15-3. WP Timing for Write Status Register Byte 1 Command When SPRL = 1
CS
SI
SCK
SO
MSB
High-impedance
MSBLSB
tCSLS
tCLKH tCLKL tCSHS
tCSHH
tDS tDH
tCSLH
tCSH
CS
SI
SCK
SO
t
V
t
CLKH
t
CLKL
t
DIS
t
V
t
OH
WP
SI
SCK
SO
000
High-impedance
MSBX
t
WPS
t
WPH
CS
LSB of
Write Status Register
Data Byte
MSB of
Write Status Register
Byte 1 Opcode
MSB of
Next Opcode
52
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
Figure 15-4. HOLD Timing – Serial Input
Figure 15-5. HOLD Timing – Serial Output
CS
SI
SCK
SO
tHHH tHLS
tHLH tHHS
HOLD
High-impedance
CS
SI
SCK
SO
tHHH tHLS
tHLQZ
tHLH tHHS
HOLD
tHHQX
53
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
16. Ordering Information
16.1 Ordering Detail
16.2 Green Package Options (Pb/Halide-free/RoHS Compliant)
Notes: 1. B = Bulk
2. T = Tape and reel
SOIC = 4K per reel
UDFN = 5K per reel
3. The shipping carrier option code is not marked on the device.
Atmel Designator
Product Family
Interface
Device Density
Device Revision
Shipping Carrier Option
Device Grade
Package Option
1 = Serial
64 = 64-megabit
B = Bulk (tubes)
Y = Bulk (trays)
T = Tape and reel
H = Green, NiPdAu lead finish,
Industrial temperature range
(-40°C to +85°C)
S = 8-lead, 0.208” wide SOIC
M = 8-pad, 5 x 6 x 0.6mm UDFN
AT25DF641A-SH-B
Ordering Code Package Lead Finish Operating Voltage fSCK Device Grade
AT25DF641A-SH-B(1)(3)
8S2
NiPdAu 2.7V to 3.6V 100MHz Industrial
(-40°C to +85°C)
AT25DF641A-SH-T(2)(3)
AT25DF641A-MH-Y(3)
8MA1
AT25DF641A-MH-T(2)(3)
Package Type
8S2 8-lead, 0.208” wide, Plastic Gull Wing Small Outline (EIAJ SOIC)
8MA1 8-pad (5 x 6 x 0.6mm body), Thermally Enhanced Plastic Ultra Thin Dual Flat No Lead (UDFN)
54
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
17. Packaging Information
17.1 8S2 — 8-lead EIAJ SOIC
Package Drawing Contact:
packagedrawings@atmel.com
DRAWING NO. REV. TITLE GPC
1
1
N
N
E
E
q
q
C
C
E1
E1
L
L
A
A
b
b
A1
A1
e
e
D
D
Notes: 1. This drawing is for general information only; refer to
EIAJ Drawing EDR-7320 for additional information.
2. Mismatch of the upper and lower dies and resin burrs
are not included.
3. Determines the true geometric position.
4. Values b, C apply to plated terminal. The standard
thickness of the plating layer shall measure between
0.007 to .021mm.
8S2 F
4/15/08
8S2, 8-lead, 0.208” Body, Plastic Small
Outline Package (EIAJ) STN
End View
Top View
Side View
COMMON DIMENSIONS
(Unit of Measure = mm)
SYMBOL MIN NOM MAX NOTE
1.70
0.05
0.35
0.15
5.13
5.18
7.70
0.51
0˚
A
A1
b
C
D
E1
E
L
q
e 1.27 BSC
2.16
0.25
0.48
0.35
5.35
5.40
8.26
0.85
8˚
4
4
2
3
55
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
17.2 8MA1 — 8-pad UDFN
Package Drawing Contact:
packagedrawings@atmel.com
DRAWING NO. REV. TITLE GPC
Notes: 1. This package conforms to JEDEC reference MO-229,
Saw Singulation.
2. The terminal #1 ID is a Laser-marked Feature.
8MA1 D
4/15/08
8MA1, 8-pad (5 x 6 x 0.6mm Body), Thermally
Enhanced Plastic Ultra Thin Dual Flat No Lead
Package (UDFN)
YFG
Top View
Side View
COMMON DIMENSIONS
(Unit of Measure = mm)
SYMBOL MIN NOM MAX NOTE
0.45
0.00
0.35
4.90
3.80
5.90
3.20
0.50
0.00
0.20
A
A1
b
C
D
D2
E
E2
e
L
y
K
0.55
0.02
0.40
0.152 REF
5.00
4.00
6.00
3.40
1.27
0.60
–
–
0.60
0.05
0.48
5.10
4.20
6.10
3.60
0.75
0.08
–
Pin 1 ID
E
D
A1
A
y
C
Bottom View
E2
D2
L
b
e
1
2
3
4
8
7
6
5
Pin #1 Notch
(0.20 R)
K
Pin #1
Chamfer
(C 0.35)
Option A
(Option B)
56
Atmel AT25DF641A [DATASHEET]
8693B–DFLASH–5/12
18. Revision History
Doc. Rev. Date Comments
8693B 05/2012
Increase Page Program typical time from 1.0ms to 2.5ms and maximum time from 3ms to 6ms.
Increase Block Erase 64KB typical time from 400ms to 600ms and maximum time from
950ms to 1100ms.
Decrease Read Array 03h maximum clock frequency from 50MHz to 40MHz.
Decrease Dual-Output Read Array 3Bh maximum clock frequency from 85MHz to 65MHz.
Increase Deep Power-Down current maximum from 10μA to 20μA.
Increase active curent, read operation:
100MHz from typical 12mA to 30mA and maximum 19mA to 40mA.
85MHz from typical 10mA to 29mA and maximum 17mA to 38mA.
66MHz from typical 8mA to 28mA and maximum 14mA to 37mA.
50MHz from typical 7mA to 27mA and maximum 12mA to 36mA.
33MHz from typical 6mA to 26mA and maximum 10mA to 35mA.
20MHz from typical 5mA to 25mA and maximum 8mA to 34mA.
Increase maximum tRDPD time from 30μs to 50μs.
Increase Byte Program typical time from 7μs to 30μs.
Increase Chip Erase typical time from 55 sec to 70 sec and maximum from 100 sec to 150 sec.
Add note and two examples to Byte/Page Program section.
Various typopgraphical edits throughout document.
Change from preliminary to complete/released status.
Update template.
8693A 08/2010 Initial document release.
Atmel Corporation
2325 Orchard Parkway
San Jose, CA 95131
USA
Tel: (+1) (408) 441-0311
Fax: (+1) (408) 487-2600
www.atmel.com
Atmel Asia Limited
Unit 01-5 & 16, 19F
BEA Tower, Millennium City 5
418 Kwun Tong Roa
Kwun Tong, Kowloon
HONG KONG
Tel: (+852) 2245-6100
Fax: (+852) 2722-1369
Atmel Munich GmbH
Business Campus
Parkring 4
D-85748 Garching b. Munich
GERMANY
Tel: (+49) 89-31970-0
Fax: (+49) 89-3194621
Atmel Japan G.K.
16F Shin-Osaki Kangyo Bldg
1-6-4 Osaki, Shinagawa-ku
Tokyo 141-0032
JAPAN
Tel: (+81) (3) 6417-0300
Fax: (+81) (3) 6417-0370
© 2012 Atmel Corporation. All rights reserved. / Rev.: 8693B–DFLASH–5/12
Disclaimer: The information in this document is provided in connection with Atmel products. No license, express or implied, by estoppel or otherwise, to any intellectual property right is granted by this
document or in connection with the sale of Atmel products. EXCEPT AS SET FORTH IN THE ATMEL TERMS AND CONDITIONS OF SALES LOCATED ON THE ATMEL WEBSITE, ATMEL ASSUMES
NO LIABILITY WHATSOEVER AND DISCLAIMS ANY EXPRESS, IMPLIED OR STATUTORY WARRANTY RELATING TO ITS PRODUCTS INCLUDING, BUT NOT LIMITED TO, THE IMPLIED
WARRANTY OF MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE, OR NON-INFRINGEMENT. IN NO EVENT SHALL ATMEL BE LIABLE FOR ANY DIRECT, INDIRECT,
CONSEQUENTIAL, PUNITIVE, SPECIAL OR INCIDENTAL DAMAGES (INCLUDING, WITHOUT LIMITATION, DAMAGES FOR LOSS AND PROFITS, BUSINESS INTERRUPTION, OR LOSS OF
INFORMATION) ARISING OUT OF THE USE OR INABILITY TO USE THIS DOCUMENT, EVEN IF ATMEL HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH DAMAGES. Atmel makes no
representations or warranties with respect to the accuracy or completeness of the contents of this document and reserves the right to make changes to specifications and products 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 products are not intended, authorized, or warranted for use as components in applications intended to support or sustain life.
Atmel®, Atmel logo and combinations thereof, Enabling Unlimited Possibilities®, and others are registered trademarks, RapidS™, and others are trademarks of
Atmel Corporation or its subsidiaries. Other terms and product names may be trademarks of others.
