AT25SF321

32-Mbit, 2.5V Minimum

SPI Serial Flash Memory with Dual-I/O and Quad-IO Support

Features
 Single 2.5V - 3.6V Supply
 Serial Peripheral Interface (SPI) Compatible
 Supports SPI Modes 0 and 3
 Supports Dual and Quad Output Read
 104MHz Maximum Operating Frequency
 Clock-to-Output (tV) of 6 ns
 Flexible, Optimized Erase Architecture for Code + Data Storage Applications
 Uniform 4-Kbyte Block Erase
 Uniform 32-Kbyte Block Erase
 Uniform 64-Kbyte Block Erase
 Full Chip Erase
 Hardware Controlled Locking of Protected Blocks via WP Pin
 3 Protected Programmable Security Register Pages
 Flexible Programming
 Byte/Page Program (1 to 256 Bytes)
 Fast Program and Erase Times
0.7ms Typical Page Program (256 Bytes) Time
70ms Typical 4-Kbyte Block Erase Time
300ms Typical 32-Kbyte Block Erase Time
600ms Typical 64-Kbyte Block Erase Time
 JEDEC Standard Manufacturer and Device ID Read Methodology
 Low Power Dissipation
2µA Deep Power-Down Current (Typical)
10µA Standby current (Typical)
4mA Active Read 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 (150-mil and 208-mil)
 8-pad Ultra Thin DFN (5 x 6 x 0.6 mm)
 Die in Wafer Form

DS-25SF321–047F–8/2017
 Description
The Adesto® AT25SF321 is a serial interface Flash memory device designed for use in 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 AT25SF321 is ideal for data storage as well, eliminating the need for
additional data storage devices.
The erase block sizes of the AT25SF321 have been optimized to meet the needs of today's code and data storage
applications. By optimizing the size of the 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 erase regions, the wasted and
unused memory space that occurs with 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 device also contains three pages of Security Register that can be used for purposes such as unique device
serialization, system-level Electronic Serial Number (ESN) storage, locked key storage, etc. These Security Register
pages can be individually locked.

1. Pin Descriptions and Pinouts

Table 1-1. Pin Descriptions

Asserted
Symbol Name and Function State Type

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.
CS Low Input
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.

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
SCK - Input
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.

SERIAL INPUT: 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 and Quad-Output Read commands, the SI Pin becomes an output pin
(I/O0) in conjunction with other pins to allow two or four bits of data on (I/O3-0) to be clocked
SI (I/O0) in on every falling edge of SCK - Input/Output
To maintain consistency with the SPI nomenclature, the SI (I/O0) pin will be referenced as
the SI pin unless specifically addressing the Dual-I/O and Quad-I/O modes in which case it
will be referenced as I/O0
Data present on the SI pin will be ignored whenever the device is deselected (CS is
deasserted).

AT25SF321 2
DS-25SF321–047F–8/2017
Table 1-1. Pin Descriptions (Continued)

Asserted
Symbol Name and Function State Type

SERIAL OUTPUT: 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-Output Read commands, the SO Pin remains an output pin (I/O0) in
conjunction with other pins to allow two bits of data on (I/O1-0) to be clocked in on every
falling edge of SCK
SO (I/O1) - Input/Output
To maintain consistency with the SPI nomenclature, the SO (I/O1) pin will be referenced as
the SO pin unless specifically addressing the Dual-I/O modes in which case it will be
referenced as I/O1
The SO pin will be in a high-impedance state whenever the device is deselected (CS is
deasserted).

WRITE PROTECT: The WP pin controls the hardware locking feature of the device.
With the Quad-Input Byte/Page Program command, the WP pin becomes an input pin (I/O2)
and, along with other pins, allows four bits (on I/O3-0) of data to be clocked in on every rising
edge of SCK. With the Quad-Output Read commands, the WP Pin becomes an output pin
(I/O2) in conjunction with other pins to allow four bits of data on (I/O33-0) to be clocked in on
WP
every falling edge of SCK.
(I/O2) - Input/Output
To maintain consistency with the SPI nomenclature, the WP (I/O2) pin will be referenced as
the WP pin unless specifically addressing the Quad-I/O modes in which case it will be
referenced as I/O2
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.

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. Please refer to
“Hold Function” on page 34 for additional details on the Hold operation.

HOLD With the Quad-Input Byte/Page Program command, the HOLD pin becomes an input pin
(I/O3) (I/O3) and, along with other pins, allows four bits (on I/O3-0) of data to be clocked in on every
- Input/Output
rising edge of SCK. With the Quad-Output Read commands, the HOLD Pin becomes an
output pin (I/O3) in conjunction with other pins to allow four bits of data on (I/O33-0) to be
clocked in on every falling edge of SCK.
To maintain consistency with the SPI nomenclature, the HOLD (I/O3) pin will be referenced
as the HOLD pin unless specifically addressing the Quad-I/O modes in which case it will be
referenced as I/O3
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.

DEVICE POWER SUPPLY: The VCC pin is used to supply the source voltage to the device.
VCC Operations at invalid VCC voltages may produce spurious results and should not be - Power
attempted.

GROUND: The ground reference for the power supply. GND should be connected to the
GND - Power
system ground.

AT25SF321 3
DS-25SF321–047F–8/2017
 Figure 1-1. 8-SOIC (Top View) Figure 1-2. 8-UDFN (Top View)

CS 1 8 VCC
CS 1 8 VCC
SO 2 7 HOLD
SO 2 7 HOLD
WP 3 6 SCK
WP 3 6 SCK
GND 4 5 SI
GND 4 5 SI

2. Block Diagram
Figure 2-1. Block Diagram

Control and I/O Buffers

Protection Logic and Latches
CS

SRAM
Data Buffer
SCK
Interface
Control
SI (I/O0) And
Logic Y-Decoder Y-Gating
SO (I/O1)
Address Latch

Flash
Memory
WP (I/O2) X-Decoder Array

HOLD (I/O3)

Note: I/O3-0 pin naming convention is used for Dual-I/O and Quad-I/O commands.

3. Memory Array
To provide the greatest flexibility, the memory array of the AT25SF321 can be erased in four levels of granularity
including a full chip erase. The size of the erase blocks is optimized for both code and data storage applications, allowing
both code and data segments to reside in their own erase regions. The Memory Architecture Diagram illustrates the
breakdown of each erase level.

AT25SF321 4
DS-25SF321–047F–8/2017
Figure 3-1. Memory Architecture Diagram
Block Erase Detail Page Program Detail

64KB 32KB 4KB 1-256 Byte

Block Address Page Address
Range Range

4KB 3FFFFFh – 3FF000h 256 Bytes 3FFFFFh – 3FFF00h

4KB 3FEFFFh – 3FE000h 256 Bytes 3FFEFFh – 3FFE00h
4KB 3FDFFFh – 3FD000h 256 Bytes 3FFDFFh – 3FFD00h
4KB 3FCFFFh – 3FC000h 256 Bytes 3FFCFFh – 3FFC00h
32KB
4KB 3FBFFFh – 3FB000h 256 Bytes 3FFBFFh – 3FFB00h
4KB 3FAFFFh – 3FA000h 256 Bytes 3FFAFFh – 3FFA00h
4KB 3F9FFFh – 3F9000h 256 Bytes 3FF9FFh – 3FF900h
4KB 3F8FFFh – 3F8000h 256 Bytes 3FF8FFh – 3FF800h
64KB
4KB 3F7FFFh – 3F7000h 256 Bytes 3FF7FFh – 3FF700h
Sector 63
4KB 3F6FFFh – 3F6000h 256 Bytes 3FF6FFh – 3FF600h
4KB 3F5FFFh – 3F5000h 256 Bytes 3FF5FFh – 3FF500h
4KB 3F4FFFh – 3F4000h 256 Bytes 3FF4FFh – 3FF400h
32KB
4KB 3F3FFFh – 3F3000h 256 Bytes 3FF3FFh – 3FF300h
4KB 3F2FFFh – 3F2000h 256 Bytes 3FF2FFh – 3FF200h
4KB 3F1FFFh – 3F1000h 256 Bytes 3FF1FFh – 3FF100h
4KB 3F0FFFh – 3F0000h 256 Bytes 3FF0FFh – 3FF000h
4KB 3EFFFFh – 3EF000h 256 Bytes 3FEFFFh – 3FEF00h
4KB 3EEFFFh – 3EE000h 256 Bytes 3FEEFFh – 3FEE00h
4KB 3EDFFFh – 3ED000h 256 Bytes 3FEDFFh – 3FED00h
4KB 3ECFFFh – 3EC000h 256 Bytes 3FECFFh – 3FEC00h
32KB
4KB 3EBFFFh – 3EB000h 256 Bytes 3FEBFFh – 3FEB00h
4KB 3EAFFFh – 3EA000h 256 Bytes 3FEAFFh – 3FEA00h
4KB 3E9FFFh – 3E9000h 256 Bytes 3FE9FFh – 3FE900h
4KB 3E8FFFh – 3E8000h 256 Bytes 3FE8FFh – 3FE800h
64KB
4KB 3E7FFFh – 3E7000h
Sector 62

•••
•••
4KB 3E6FFFh – 3E6000h
4KB 3E5FFFh – 3E5000h
4KB 3E4FFFh – 3E4000h 256 Bytes 0017FFh – 001700h
32KB
4KB 3E3FFFh – 3E3000h 256 Bytes 0016FFh – 001600h
4KB 3E2FFFh – 3E2000h 256 Bytes 0015FFh – 001500h
4KB 3E1FFFh – 3E1000h 256 Bytes 0014FFh – 001400h
4KB 3E0FFFh – 3E0000h 256 Bytes 0013FFh – 001300h
256 Bytes 0013FFh – 001300h
•••

•••

•••

256 Bytes 0013FFh – 001300h

256 Bytes 0012FFh – 001200h
4KB 00FFFFh–00F000h 256 Bytes 0011FFh – 001100h
4KB 00EFFFh–00E000h 256 Bytes 0010FFh – 001000h
4KB 00DFFFh–00D000h 256 Bytes 000FFFh – 000F00h
4KB 00CFFFh–00C000h 256 Bytes 000CFFh – 000C00h
32KB
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
64KB
4KB 007FFFh–007000h 256 Bytes 0007FFh – 000700h
Sector 0
4KB 006FFFh–006000h 256 Bytes 0006FFh – 000600h
4KB 005FFFh–005000h 256 Bytes 0005FFh – 000500h
4KB 004FFFh–004000h 256 Bytes 0004FFh – 000400h
32KB
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

AT25SF321 5
DS-25SF321–047F–8/2017
4. Device Operation
The AT25SF321 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 AT25SF321 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 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 AT25SF321
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 4-1. SPI Mode 0 and 3

CS

SCK

SI MSB LSB

SO MSB LSB

4.1 Dual Output Read

The AT25SF321 features a Dual-Output Read mode that allow two bits of data to be clocked out of the device every
clock cycle to improve throughput. To accomplish this, both the SI and SO pins are utilized as outputs for the transfer of
data bytes. With the Dual-Output Read Array command, the SI pin becomes an output along with the SO pin.

4.2 Quad Output Read

The AT25SF321 features a Quad-Output Read mode that allow four bits of data to be clocked out of the device every
clock cycle to improve throughput. To accomplish this, the SI, SO, WP, HOLD pins are utilized as outputs for the transfer
of data bytes. With the Quad-Output Read Array command, the SI, WP, HOLD pins become outputs along with the SO
pin.

5. 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 AT25SF321 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 AT25SF321 memory array is 3FFFFFh, address bits A23-A22 are always ignored by
the device.

AT25SF321 6
DS-25SF321–047F–8/2017
Table 5-1. Command Listing

Clock Address Dummy Data Section

Command Opcode Frequency Bytes Bytes Bytes Link

Read Commands

0Bh 0000 1011 Up to 85 MHz 3 1 1+

Read Array 6.1
03h 0000 0011 Up to 50 MHz 3 0 1+

Dual Output Read 3Bh 0011 1011 Up to 85 MHz 3 1 1+ 6.2

Dual I/O Read BBh 1011 1011 Up to 85 MHz 3 0 1+ 6.3

Quad Output Read 6Bh 0110 1011 Up to 85 MHz 3 1 1+ 6.4

Quad I/O Read EBh 1110 1011 Up to 85 MHz 3 1 1+ 6.5

1111 1111
Continuous Read Mode Reset - Dual FFFFh Up to 104 MHz 0 0 0 6.6
1111 1111

Continuous Read Mode Reset - Quad FFh 1111 1111 Up to 104 MHz 0 0 0 6.6

Program and Erase Commands

Block Erase (4 Kbytes) 20h 0010 0000 Up to 104 MHz 3 0 0

Block Erase (32 Kbytes) 52h 0101 0010 Up to 104 MHz 3 0 0 7.2

Block Erase (64 Kbytes) D8h 1101 1000 Up to 104MHz 3 0 0

60h 0110 0000 Up to 104 MHz 0 0 0

Chip Erase 7.3
C7h 1100 0111 Up to 104 MHz 0 0 0

Byte/Page Program (1 to 256 Bytes) 02h 0000 0010 Up to 104 MHz 3 0 1+ 7.1

Program/Erase Suspend 75h 0111 0101 Up to 104 MHz 0 0 0 7.4

Program/Erase Resume 7Ah 0111 1010 Up to 104 MHz 0 0 0 7.5

Protection Commands

Write Enable 06h 0000 0110 Up to 104 MHz 0 0 0 8.1

Write Disable 04h 0000 0100 Up to 104 MHz 0 0 0 8.2

Security Commands

Erase Security Register Page 44h 0100 0100 Up to 104 MHz 3 0 0 9.1

Program Security Register Page 42h 0100 0010 Up to 104 MHz 3 0 1+ 9.2

Read Security Register Page 48h 0100 1000 Up to 85MHz 3 1 1+ 9.3

Status Register Commands

Read Status Register Byte 1 05h 0000 0101 Up to 104 MHz 0 0 1

10.1
Read Status Register Byte 2 35h 0011 0101 Up to 104 MHz 0 0 1

Write Status Register 01h 0000 0001 Up to 104 MHz 0 0 1 or 2 10.2

Write Enable for Volatile Status

Register 50h 0101 0000 Up to 104MHz 0 0 0 10.3

AT25SF321 7
DS-25SF321–047F–8/2017
Table 5-1. Command Listing

Clock Address Dummy Data Section

Command Opcode Frequency Bytes Bytes Bytes Link

Miscellaneous Commands

Read Manufacturer and Device ID 9Fh 1001 1111 Up to 104MHz 0 0 3 11.1

Read ID 90h 1001 0000 Up to 104 MHz 0 3 2 11.2

Deep Power-Down B9h 1011 1001 Up to 104 MHz 0 0 0 11.3

Resume from Deep Power-Down ABh 1010 1011 Up to 104 MHz 0 0 0 11.4

Resume from Deep Power-Down and

ABh 1010 1011 Up to 104 MHz 0 3 1 11.4
Read ID

6. Read Commands

6.1 Read Array (0Bh and 03h)

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 is specified. The device incorporates an internal address
counter that automatically increments every clock cycle.
Two opcodes (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.
To perform the Read Array operation, the CS pin must first be asserted and the appropriate opcode (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
starting address location of the first byte to read within the memory array. Following the three address bytes, an
additional dummy byte needs to be clocked into the device if the 0Bh opcode is used for the Read Array operation.
After the three address bytes (and the dummy byte if using opcode 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
(3FFFFFh) 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 high-impedance state. The CS pin can
be deasserted at any time and does not require a full byte of data be read.

Figure 6-1. Read Array - 03h Opcode

&6

                        
6&.
23&2'( $''5(66%,76$$

6,         $ $ $ $ $ $ $ $ $
06% 06%

'$7$%<7(

+,*+,03('$1&(
62 ' ' ' ' ' ' ' ' ' '
06% 06%

AT25SF321 8
DS-25SF321–047F–8/2017
 Figure 6-2. Read Array - 0Bh Opcode

&6

                                
6&.
23&2'( $''5(66%,76$$ '21
7&$5(

6,         $ $ $ $ $ $ $ $ $ ; ; ; ; ; ; ; ;
06% 06% 06%

'$7$%<7(

+,*+,03('$1&(
62 ' ' ' ' ' ' ' ' ' '
06% 06%

6.2 Dual-Output Read Array (3Bh)

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 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 SI 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 seven of the first data byte is output on the SO pin, while bit six of the same
data byte is output on the SI pin. During the next clock cycle, bits five and four of the first data byte are output on the SO
and SI pins, respectively. The sequence continues with each byte of data being output after every four clock cycles.
When the last byte (3FFFFFh) 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 SI 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 6-3. Dual-Output Read Array

CS

0 1 2 3 4 5 6 7 8 9 10 11 12 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
SCK
OUTPUT OUTPUT
23&2'( $''5(66%,76$$ DON'
T &$5( DAT$%<7( DAT$%<7(

6,6,2 0 0 1 1 1 0 1 1 A A A A A A A A A X X X X X X X X D6 D4 D2 D0 D6 D4 D2 D0 D6 D4
MSB MSB MSB

HI
GH-
IMPEDANCE
SO D7 D5 D3 D1 D7 D5 D3 D1 D7 D5
MSB MSB MSB

AT25SF321 9
DS-25SF321–047F–8/2017
6.3 Dual-I/O Read Array (BBh)
The Dual-I/O Read Array command is similar to the Dual-Output 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
with two bits of address on each clock and two bits of data on every clock cycle.
The Dual-I/O Read Array command can be used at any clock frequency, up to the maximum specified by fRDDO. To
perform the Dual-I/O Read Array operation, the CS pin must first be asserted and then the opcode BBh 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 mode byte must also be
clocked into the device.
After the three address bytes and the mode byte have been clocked in, additional clock cycles will result in data being
output on both the SO and SI 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 seven of the first data byte is output on the SO pin, while bit six of the same
data byte is output on the SI pin. During the next clock cycle, bits five and four of the first data byte are output on the SO
and SI pins, respectively. The sequence continues with each byte of data being output after every four clock cycles.
When the last byte (3FFFFFh) 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 SI 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 6-4. Dual I/O Read Array (Initial command or previous M5, M4≠1,0)

CS

0 1 2 3 4 5 6 7 8 9 10 11 12 19 20 21 22 23 24 25 26 27
SCK
Address Bits Address Bits
Opcode A23-A16 A15-A8 A7-A0 M7-M0 Byte 1 Byte 2

I/O0 1 0 1 1 1 0 1 1 A22 A20 A18 A16 A14 A A0 M6 M4 M2 M0 D6 D4 D2 D0 D6

(SI) MSB MSB

I/O1 A23 A21 A19 A17 A15 A A 1 M 7 M 5 M 3 M1 D 7 D 5 D 3 D 1 D 7

(SO) MSB

6.3.1 Dual-I/O Read Array (BBh) with Continuous Read Mode

The Fast Read Dual I/O command can further reduce instruction overhead through setting the Continuous Read Mode
bits (M7-0) after the input Address bits (A23-0), as shown in Figure 6-5. The upper nibble of the (M7-4) controls the
length of the next Fast Read Dual I/O command through the inclusion or exclusion of the first byte instruction code. The
lower nibble bits of the (M3-0) are don't care ("x"). However, the IO pins should be high-impedance prior to the falling
edge of the first data out clock. If the "Continuous Read Mode" bits M5-4 = (1,0), then the next Fast Read Dual I/O
command (after CS is raised and then lowered) does not require the BBH instruction code, as shown in Figure 15. This
reduces the command sequence by eight clocks and allows the Read address to be immediately entered after CS is
asserted low. If the "Continuous Read Mode" bits M5-4 do not equal to (1,0), the next command (after CS is raised and
then lowered) requires the first byte instruction code, thus returning to normal operation. A Continuous Read Mode Reset
command can also be used to reset (M7-0) before issuing normal commands.

AT25SF321 10
DS-25SF321–047F–8/2017
 Figure 6-5. Dual-I/O Read Array (Previous command set M5, M4 = 1,0)

CS

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21
SCK
Address Bits Address Bits
A23-A16 A15-A8 A7-A0 M7-M0 Byte 1 Byte 2

I/O0 A22 A20 A18 A16 A14 A A 0 M 6 M4 M 2 M0 D 6 D 4 D 2 D 0 D 6

(SI) MSB

I/O1 A23 A21 A19 A17 A15 A A 1 M 7 M5 M3 M1 D 7 D 5 D 3 D 1 D 7

(SO) MSB

6.4 Quad-Output Read Array (6Bh)

The Quad-Output Read Array command is similar to the Dual-Output Read Array command. The Quad-Output Read
Array command allows four bits of data to be clocked out of the device on every clock cycle, rather than just one or two.
The Quad Enable bit (QE) of the Status Register must be set to enable for the Quad-Output Read Array instruction.
The Quad-Output Read Array command can be used at any clock frequency, up to the maximum specified by fRDQO. To
perform the Quad-Output Read Array operation, the CS pin must first be asserted and then the opcode 6Bh 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 the I/O3-0 pins. The data is always output with the MSB of a byte first and the MSB is always output on the I/O3
pin. During the first clock cycle, bit 7 of the first data byte will be output on the I/O3 pin while bits 6, 5, and 4 of the same
data byte will be output on the I/O2, I/O1, and I/O0 pins, respectively. During the next clock cycle, bits 3, 2, 1, and 0 of the
first data byte will be output on the I/O3, I/O2, I/O1 and I/O0 pins, respectively.
The sequence continues with each byte of data being output after every two clock cycles. When the last byte (3FFFFFh)
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 WP, HOLD, SO, SI 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.

AT25SF321 11
DS-25SF321–047F–8/2017
 Figure 6-6. Quad-Output Read Array

CS

0 1 2 3 4 5 6 7 8 9 10 11 12 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48
SCK
Byte 1 Byte 2 Byte 3 Byte 4 Byte 5
Opcode Address Bits A23-A0 Don't Care OUT OUT OUT OUT OUT

I/O0 0 1 1 0 1 0 1 1 A A A A A A A A A X X X X X X X X D4 D0 D4 D0 D4 D0 D4 D0 D4 D0
06% 06% 06%
(SI)
High-impedance
I/O1 D5 D1 D5 D1 D5 D1 D5 D1 D5 D1

(SO)
High-impedance
I/O2 D6 D2 D6 D2 D6 D2 D6 D2 D6 D2

(WP)
High-impedance
I/O3 D7 D3 D7 D3 D7 D3 D7 D3 D7 D3
MSB MSB MSB MSB MSB
(HOLD)

6.5 Quad-I/O Read Array(EBh)

The Quad-I/O Read Array command is similar to the Quad-Output Read Array command. The Quad-I/O Read Array
command allows four bits of address to be clocked into the device on every clock cycle, rather than just one.
The Quad-I/O Read Array command can be used at any clock frequency, up to the maximum specified by fRDQO. To
perform the Quad-I/O Read Array operation, the CS pin must first be asserted and then the opcode EBh 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 mode byte must also be
clocked into the device.
After the three address bytes, the mode byte and two dummy bytes have been clocked in, additional clock cycles will
result in data being output on the I/O3-0 pins. The data is always output with the MSB of a byte first and the MSB is always
output on the I/O3 pin. During the first clock cycle, bit 7 of the first data byte will be output on the I/O3 pin while bits 6, 5,
and 4 of the same data byte will be output on the I/O2, I/O1 and I/O0 pins, respectively. During the next clock cycle, bits 3,
2, 1, and 0 of the first data byte will be output on the I/O3, I/O2, I/O1 and I/O0 pins, respectively. The sequence continues
with each byte of data being output after every two clock cycles.
When the last byte (3FFFFFh) 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 I/O3, I/O2, I/O1 and I/O0 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.The
Quad Enable bit (QE) of the Status Register must be set to enable for the Quad-I/O Read Array instruction.

AT25SF321 12
DS-25SF321–047F–8/2017
 Figure 6-7. Quad-I/O Read Array (Initial command or previous M5, M4 ≠ 1,0)

CS

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
SCK

Opcode A23-A16 A15-A8 A7-A0 M7-M0 Dummy Byte 1 Byte 2

I/O0 1 1 1 0 1 0 1 1 A20 A16 A12 A8 A4 A0 M4 M0 D4 D0 D4 D0

(SI) MSB

I/O1 A21 A17 A13 A9 A5 A1 M5 M1 D5 D1 D5 D1

(SO)

I/O2 A22 A18 A14 A10 A6 A2 M6 M2 D6 D2 D6 D2

(WP)
I/O3 A23 A19 A15 A11 A7 A3 M7 M3 D7 D3 D7 D3
(HOLD)

6.5.1 Quad-I/O Read Array (EBh) with Continuous Read Mode

The Fast Read Quad I/O command can further reduce instruction overhead through setting the Continuous Read Mode
bits (M7-0) after the input Address bits (A23-0), as shown in Figure 6-6. The upper nibble (M7-4) of the Continuous Read
Mode bits controls the length of the next Fast Read Quad I/O command through the inclusion or exclusion of the first byte
instruction code. The lower nibble bits (M3-0) of the Continuous Read Mode bits are don't care. However, the IO pins
should be high-impedance prior to the falling edge of the first data out clock. If the Continuous Read Mode bits M5-4 =
(1,0), then the next Quad-I/O Read Array command (after CS is raised and then lowered) does not require the EBh
instruction code, as shown in Fig 6-8. This reduces the command sequence by eight clocks and allows the Read address
to be immediately entered after CS is asserted low. If the "Continuous Read Mode" bits M5-4 do not equal to (1,0), the
next command (after CS is raised and then lowered) requires the first byte instruction code, thus returning to normal
operation. A Continuous Read Mode Reset command can also be used to reset (M7-0) before issuing normal
commands.

Figure 6-8. Quad I/O Read Array with Continuous Read Mode (Previous Command Set M5-4 =1,0)

CS

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15
SCK

A23-A16 A15-A8 A7-A0 M7-M0 Dummy Byte 1 Byte 2

I/O0 A20 A16 A12 A8 A4 A0 M4 M0 D4 D0 D4 D0

(SI)

I/O1 A21 A17 A13 A9 A5 A1 M5 M1 D5 D1 D5 D1

(SO)

I/O2 A22 A18 A14 A10 A6 A2 M6 M2 D6 D2 D6 D2

(WP)
I/O3 A23 A19 A15 A11 A7 A3 M7 M3 D7 D3 D7 D3
(HOLD)

6.6 Continuous Read Mode Reset (FFh or FFFFh)

The Continuous Read Mode bits are used in conjunction with the Dual I/O Read Array and the Quad I/O Read Array
commands to provide the highest random Flash memory access rate with minimum SPI instruction overhead, thus
allowing more efficient XIP (execute in place) with this device family.

AT25SF321 13
DS-25SF321–047F–8/2017
 The "Continuous Read Mode" bits M7-0 are set by the Dual/Quad I/O Read Array commands. M5-4 are used to control
whether the 8-bit SPI instruction code (BBh or EBh) is needed or not for the next instruction. When M5-4 = (1,0), the next
instruction will be treated the same as the current Dual/Quad I/O Read Array command without needing the 8-bit
instruction code. When M5-4 do not equal (1,0), the device returns to normal SPI instruction mode, in which all
instructions can be accepted. M7-6 and M3-0 are reserved bits for future use; either 0 or 1 values can be used.
See Figure 6-9, the Continuous Read Mode Reset instruction (FFh or FFFFh) can be used to set M4 = 1, thus the device
will release the Continuous Read Mode and return to normal SPI operation.
To reset Continuous Read Mode during Quad I/O operation, only eight clocks are needed to shift in instruction FFh. To
reset Continuous Read Mode during Dual I/O operation, sixteen clocks are needed to shift in instruction FFFFh.

Figure 6-9. Continuous Read Mode Reset (Quad)

jz

W X Y Z [ \ ] ^
zjr
vwjvkl

zp 1 1 1 1 1 1 1 1
tzi

DON’T CARE
zv

Figure 6-10. Continuous Read Mode Reset (Dual)

jz

W X Y Z [ \ ] ^ 8 9 10 11 12 13 14 15
zjr
vwjvkl

zp 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
MSB

DON’T CARE
zv

7. Program and Erase Commands

7.1 Byte/Page Program (02h)

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 (06h)” on page 20) to set the Write Enable Latch (WEL) bit
of the Status Register to a logical “1” state.
To perform a Byte/Page Program command, an opcode of 02h must be clocked into the device followed by the three
address bytes denoting the first byte 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 to be programmed will apply. In this situation, any data

AT25SF321 14
DS-25SF321–047F–8/2017
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.
For 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). 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.
When the CS pin is deasserted, the device will take the data stored in the internal buffer and program it 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.
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 memory is in the
protected state (see “Non-Volatile Protection” on page 21), 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.
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.

Figure 7-1. Byte Program

&6

                       
6&.
23&2'( $''5(66%,76$$ '$7$,1

6,         $ $ $ $ $ $ $ $ $ ' ' ' ' ' ' ' '

06% 06% 06%

+,*+,03('$1&(
62

Figure 7-2. Page Program

&6

                    
6&.
23&2'( $''5(66%,76$$ '$7$,1%<7( '$7$,1%<7(Q

6,         $ $ $ $ $ $ ' ' ' ' ' ' ' ' ' ' ' ' ' ' ' '
06% 06% 06% 06%

+,*+,03('$1&(
62

AT25SF321 15
DS-25SF321–047F–8/2017
7.2 Block Erase (20h, 52h, or D8h)
A block of 4, 32, or 64 Kbytes 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. An opcode of 20h is used for a 4-Kbyte erase, an opcode of 52h is used
for a 32-Kbyte erase, or D8h is used for a 64-Kbyte 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
4- or 32- or 64-Kbyte 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 4-Kbyte erase, address bits A11-A0 will be ignored by the device and their values can be either a
logical “1” or “0”. For a 32-Kbyte erase, address bits A14-A0 will be ignored by the device. For a 64-Kbyte erase, address
bits A15-A0 will be ignored by the device. 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 byte boundary (multiples of eight bits); otherwise, the device will abort the operation and no erase
operation will be performed.
If the memory is in the protected 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.
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.

Figure 7-3. Block Erase

&6

                  
6&.
23&2'( $''5(66%,76$$

6, & & & & & & & & $ $ $ $ $ $ $ $ $ $ $ $

06% 06%

+,*+,03('$1&(
62

7.3 Chip Erase (60h or C7h)

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
must be clocked into the device. Since the entire memory array is to be erased, no address bytes need to be clocked into

AT25SF321 16
DS-25SF321–047F–8/2017
 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 byte boundary (multiples of eight bits); otherwise, no erase will be performed. In addition, if the memory
array is in the protected 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 the memory is in the protected 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.

Figure 7-4. Chip Erase

&6

       
6&.
23&2'(

6, & & & & & & & &

06%

+,*+,03('$1&(
62

7.4 Program/Erase Suspend (75h)

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 be suspended so that
other device operations can be performed. For example, by suspending an erase operation to a particular block, the
system can perform functions such as a program or read to a different block.
Chip Erase cannot be suspended. The Program/Erase Suspend command will be ignored if it is issued during a Chip
Erase.A program operation can be performed while an erase operation is suspended, but the program operation cannot
be suspended while an erase operation is currently suspended.
Other device operations, such as a Read Status Register, can also be performed while a program or erase operation is
suspended. Table 7-4 outlines the operations that are allowed and not allowed during a program or erase suspend.
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 the opcode of 75h 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
tSUSE. The Suspend (SUS) bit in the Write Status Register will 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 a byte boundary (multiples of eight bits). Otherwise, no suspend operation will be
performed.
If a read operation is attempted to a suspended area (page for programming or block for erasing), then the device will
output undefined data. Therefore, when performing a Read Array operation to an unsuspended area and the device's

AT25SF321 17
DS-25SF321–047F–8/2017
 internal address counter increments and crosses into the suspended area, the device will then start outputting undefined
data until the internal address counter crosses to an unsuspended area.
A program operation is not allowed to a block that has been erase suspended. If a program operation is attempted to an
erase suspended block, then the program operation will abort and the WEL bit in the Status Register will be reset back to
a logical "0" state. Likewise, an erase operation is not allowed to a block that included the page 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.
If an attempt is made to perform an operation that is not allowed during a program or erase suspend, such as a Write
Status Register operation, then the device will simply ignore the opcode and no operation will be performed. The state of
the WEL bit in the Status Register will not be affected.

Table 7-1. Operations Allowed and Not Allowed During a Program/Erase Suspend Command

Command During Program Suspend During Erase Suspend

Read Commands
Read Array (03h, 0Bh, 3Bh, BBh, 6Bh, EBh) Allowed (1) Allowed(1)
Continuous Read Reset (FFh) Allowed(1) Allowed(1)
Program and Erase Commands
Block Erase (20h, 52h, D8h) Not Allowed Not Allowed
Chip Erase (C7h, 60h) Not Allowed Not Allowed
Byte/Page Program (02h) Not Allowed Allowed(1)
Program/Erase Suspend (75h) Not Allowed Not Allowed
Program/Erase Resume (7Ah) Allowed Allowed
Protection Commands
Write Enable (06h) Allowed Allowed
Write Disable (04h) Allowed Allowed
Security Commands
Erase Security Register Page (44h) Not Allowed Not Allowed
Program Security Register Page (42h) Not Allowed Not Allowed
Read Security Register Page (48h) Allowed Allowed
Status Register Commands
Read Status Register (05h, 35h) Allowed Allowed
Write Status Register (01h) Not Allowed Not Allowed
Volatile Write Enable Status Register (50h) Not Allowed Not Allowed
Miscellaneous Commands
Read Manufacturer and Device ID (9Fh) Allowed Allowed
Read ID (90h) Allowed Allowed
Deep Power-Down (B9h) Not Allowed Not Allowed
(2)
Resume from Deep Power-Down (ABh) Allowed Allowed(2)

1. Allowed for all 64K-byte blocks other than the one currently suspended.
2. Allowed for reading Device ID.

AT25SF321 18
DS-25SF321–047F–8/2017
 Figure 7-5. Program/Erase Suspend
CS

0 1 2 3 4 5 6 7
SCK
OPCODE

SI 0 1 1 1 0 1 0 1
MSB

HI
GH-
IMPEDANCE
SO

7.5 Program/Erase Resume (7Ah)

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. The Program/Erase Resume instruction will
be accepted by the device only if the SUS bit in the Write Status Register equals 1 and the RDY/BSY bit equals 0. If the
SUS bit equals 0 or the RDY/BSY bit equals to 1, the Program/Erase Resume command will be ignored by the device. 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 Program/Erase Resume, the CS pin must first be asserted and opcode 7Ah 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 resume within a time of tRES. The SUS bit
in the Status Register will then be reset back to the logical "0" state to indicate 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 a byte boundary (multiples of eight bits). Otherwise, no resume operation will perform.
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 SUS bit in the Status Register to determine if the previously suspended
program or erase operation has resumed.

Figure 7-6. Program/Erase Resume.

0 1 2 3 4 5 6 7
SCK
OPCODE

SI 0 1 1 1 1 0 1 0
MSB

HI
GH-
IMPEDANCE
SO

AT25SF321 19
DS-25SF321–047F–8/2017
8. Protection Commands and Features

8.1 Write Enable (06h)

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, Program Security Register Pages, Erase Security Register
Pages 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 byte boundary
(multiples of eight bits); otherwise, the device will abort the operation and the WEL bit state will not change.

Figure 8-1. Write Enable

&6

       
6&.
23&2'(

6,        
06%

+,*+,03('$1&(
62

8.2 Write Disable (04h)

The Write Disable command is us0d 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, Program Security Register Page, and Write Status Register
commands will not be executed. Other conditions can also cause the WEL bit to be reset; for more details, refer to the
WEL bit section of the Status Register description.
To issue the Write Disable command, the CS pin must first be asserted and the opcode of 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 byte
boundary (multiples of eight bits); otherwise, the device will abort the operation and the WEL bit state will not change.

Figure 8-2. Write Disable

&6

       
6&.
23&2'(

6,        
06%

+,*+,03('$1&(
62

AT25SF321 20
DS-25SF321–047F–8/2017
8.3 Non-Volatile Protection
The device can be software protected against erroneous or malicious program or erase operations by utilizing the Non-
Volatile Protection feature of the device. Non-Volatile Protection can be enabled or disabled by using the Write Status
Register command to change the value of the Protection (CMP, SEC, TB, BP2, BP1, BP0) bits in the Status Register.
The following table outlines the states of the Protection bits and the associated protection area
.
Table 8-1. Memory Array with CMP=0

Protection Bits Memory Content

SEC TB BP2 BP1 BP0 Address Range Portion

X X 0 0 0 None None

0 0 0 0 1 3F0000h-3FFFFFh Upper 1/64

0 0 0 1 0 3E0000h-3FFFFFh Upper 1/32

0 0 0 1 1 3C0000h-3FFFFFh Upper 1/16

0 0 1 0 0 380000h-3FFFFFh Upper 1/8

0 0 1 0 1 300000h-3FFFFFh Upper 1/4

0 0 1 1 0 200000h-3FFFFFh Upper 1/2

0 1 0 0 1 000000h-00FFFFh Lower 1/64

0 1 0 1 0 000000h-01FFFFh Lower 1/32

0 1 0 1 1 000000h-03FFFFh Lower 1/16

0 1 1 0 0 000000h-07FFFFh Lower 1/8

0 1 1 0 1 000000h-0FFFFFh Lower 1/4

0 1 1 1 0 000000h-1FFFFFh Lower 1/2

X X 1 1 1 000000h-3FFFFFh ALL

1 0 0 0 1 3FF000h-3FFFFFh Upper 1/512

1 0 0 1 0 3FE000h-3FFFFFh Upper 1/256

1 0 0 1 1 3FC000h-3FFFFFh Upper 1/128

1 0 1 0 X 3F8000h-3FFFFFh Upper 1/64

1 0 1 1 0 3F8000h-3FFFFFh Upper 1/64

1 1 0 0 1 000000h-000FFFh Lower 1/512

1 1 0 1 0 000000h-001FFFh Lower 1/256

1 1 0 1 1 000000h-003FFFh Lower 1/128

1 1 1 0 X 000000h-007FFFh Lower 1/64

1 1 1 1 0 000000h-007FFFh Lower 1/64

AT25SF321 21
DS-25SF321–047F–8/2017
Table 8-2. Memory Array Protection with CMP=1

Protection Bits Memory Content

SEC TB BP2 BP1 BP0 Address Range Portion

X X 0 0 0 000000h-3FFFFFh All

0 0 0 0 1 000000h-3EFFFFh Lower 63/64

0 0 0 1 0 000000h-3DFFFFh Lower 31/32

0 0 0 1 1 000000h-3BFFFFh Lower 15/16

0 0 1 0 0 000000h-37FFFFh Lower 7/8

0 0 1 0 1 000000h-2FFFFFh Lower 3/4

0 0 1 1 0 000000h-1FFFFFh Lower 1/2

0 1 0 0 1 010000h-3FFFFFh Upper 63/64

0 1 0 1 0 020000h-3FFFFFh Upper 31/32

0 1 0 1 1 040000h-3FFFFFh Upper 15/16

0 1 1 0 0 080000h-3FFFFFh Upper 7/8

0 1 1 0 1 100000h-3FFFFFh Upper 3/4

0 1 1 1 0 200000h-3FFFFFh Upper 1/2

X X 1 1 1 NONE NONE

1 0 0 0 1 000000h-3FEFFFh Lower 1023/1024

1 0 0 1 0 000000h-3FDFFFh Lower 511/512

1 0 0 1 1 000000h-3FBFFFh Lower 255/256

1 0 1 0 X 000000h-3F7FFFh Lower 127/128

1 0 1 1 0 000000h-3F7FFFh Lower 127/128

1 1 0 0 1 001000h-3FFFFFh Upper 1023/1024

1 1 0 1 0 002000h-3FFFFFh Upper 511/512

1 1 0 1 1 004000h-3FFFFFh Upper 255/256

1 1 1 0 X 008000h-3FFFFFh Upper 127/128

1 1 1 1 0 008000h-3FFFFFh Upper 127/128

As a safeguard against accidental or erroneous protecting or unprotecting of the memory array, the Protection can be
locked from updates by using the WP pin (see “Protected States and the Write Protect Pin” on page 22 for more details).

8.4 Protected States and the Write Protect Pin

The WP pin is not linked to the memory array itself and has no direct effect on the protection status of the memory array.
Instead, the WP pin, is used to control the hardware locking mechanism of the device.
If the WP pin is permanently connected to GND, then the protection bits cannot be changed.

AT25SF321 22
DS-25SF321–047F–8/2017
9. Security Register Commands
The device contains three extra pages called Security Registers that can be used for purposes such as unique device
serialization, system-level Electronic Serial Number (ESN) storage, locked key storage, etc. The Security Registers are
independent of the main Flash memory.
Each page of the Security Register can be erased and programmed independently. Each page can also be
independently locked to prevent further changes.

9.1 Erase Security Registers (44h)

Before an erase Security Register Page 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 an Erase Security Register Page command, the CS pin must first be asserted and the opcode 44h must be
clocked into the device. After the opcode has been clocked in, the three address bytes specifying the Security Register
Page to be erased must be clocked in. 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 tBE.
Since the Erase Security Register Page command erases a region of bytes, the lower order address bits do not need to
be decoded by the device. Therefore address bits A7-A0 will be ignored by the device. 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 right after the last address bit (A0); otherwise, the device will
abort the operation and no erase operation will be performed.
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 tBE time to
determine if the device has finished erasing. At some point before the erase cycle completes, the RDY/BSY bit in the
Status Register will be reset back to the logical "0" state.
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.
The Security Registers Lock Bits (LB3-LB1) in the Status Register can be used to OTP protect the security registers.
Once a Lock Bit is set to 1, the corresponding Security Register will be permanently locked. The Erase Security Register
Page instruction will be ignored for Security Registers which have their Lock Bit set.

Table 9-1. Security Register Addresses for Erase Security Register Page Command

Address A23-A16 A15-A8 A7-A0

Security Register 1 00H 01H Don’t Care

Security Register 2 00H 02H Don’t Care
Security Register 3 00H 03H Don’t Care

Figure 9-1. Erase Security Register Page

&6

0 1 2 3 4 5 6 7     
6&.
OPCODE %,7$''5(66

6,         $ $ $ $ $
MSB

AT25SF321 23
DS-25SF321–047F–8/2017
9.2 Program Security Registers (42h)
The Program Security Registers command utilizes the internal 256-byte buffer for processing. Therefore, the contents of
the buffer will be altered from its previous state when this command is issued.
The Security Registers can be programmed in a similar fashion to the Program Array operation up to the maximum clock
frequency specified by fCLK. Before a Program Security Registers command can be started, the Write Enable command
must have been previously issued to the device (see “Write Enable (06h)” on page 20) to set the Write Enable Latch
(WEL) bit of the Status Register to a logical “1” state. To program the Security Registers, the CS pin must first be
asserted and the opcode of 42h 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 program within the Security Register.

Table 9-2. Security Register Addresses for Program Security Registers Command

Address A23-A16 A15-A8 A7-A0

Security Register 1 00H 01H Byte Address

Security Register 2 00H 02H Byte Address

Security Register 3 00H 03H Byte Address

Figure 9-2. Program Security Registers

&6

0 1 2 3 4 5 6 7 8 9 29 30 31 32 33 34 35 36 37 38 39
6&.
OPCODE ADDRESS BI
TS A23-
A0 DATAI
NBYTE1 DATAI
NBYTEn

6,   0   0 1  A A A A A A D D D D D D D D D D D D D D D D
MSB MSB MSB MSB

HI
GH-
IMPEDANCE
62

9.3 Read Security Registers (48h)

The Security Register can be sequentially read in a similar fashion to the Read Array operation up to the maximum clock
frequency specified by fCLK. To read the Security Register, the CS pin must first be asserted and the opcode of 48h 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 Security Register. Following the three address bytes, one
dummy byte must be clocked into the device before data can be output.
After the three address bytes and the dummy byte have been clocked in, additional clock cycles will result in Security
Register data being output on the SO pin. When the last byte (0003FFh) of the 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.

AT25SF321 24
DS-25SF321–047F–8/2017
 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.

Table 9-3. Security Register Addresses for Read Security Registers Command

Address A23-A16 A15-A8 A7-A0

Security Register 1 00H 01H Byte Address

Security Register 2 00H 02H Byte Address

Security Register 3 00H 03H Byte Address

Figure 9-3. Read Security Registers

&6

0 1 2 3 4 5 6 7 8 9 10 11 12 29 30 31 32 33 34 35 36
6&.
OPCODE ADDRESS BI
TS A23-
A0 DON'
TCARE

6, 0 1       A A A A A A A A A X X X X X X X X X
MSB MSB MSB

DATABYTE1

HI
GH-
IMPEDANCE
62 D D D D D D D D D D
MSB MSB

10. Status Register Commands

10.1 Read Status Register (05h and 35h)

The Status Register can be read to determine the device's ready/busy status, as well as the status of many other
functions such as Block Protection. The Status Register can be read at any time, including during an internally self-timed
program or erase operation.
To read Status Register Byte 1, the CS pin must first be asserted and the opcode of 05h must be clocked into the device.
After the opcode has been clocked in, the device will begin outputting Status Register Byte 1 data on the SO pin during
every subsequent clock cycle. After the last bit (bit 0) of Status Register Byte 1 has been clocked out, the sequence will
repeat itself starting again with bit 7 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. 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.
To read Status Register Byte 2, the CS pin must first be asserted and the opcode of 35h must be clocked into the device.
After the opcode has been clocked in, the device will begin outputting Status Register Byte 2 data on the SO pin during
every subsequent clock cycle. After the last bit (bit 0) of Status Register Byte 2 has been clocked out, the sequence will
repeat itself starting again with bit 7 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. 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.

AT25SF321 25
DS-25SF321–047F–8/2017
Table 10-1. Status Register Format - Byte 1

Bit(1) Name Type(2) Description

7 SRP0 Status Register Protection bit-0 R/W See Table 10-3 on Status Register Protection
6 SEC Block Protection R/W See Table 8-1 and 8-2 on Non-Volatile Protection
5 TB Top or Bottom Protection R/W See Table 8-1 and 8-2 on Non-Volatile Protection
4 BP2 Block Protection bit-2 R/W See Table 8-1 and 8-2 on Non-Volatile Protection
3 BP1 Block Protection bit-1 R/W See Table 8-1 and 8-2 on Non-Volatile Protection
2 BP0 Block Protection bit-0 R/W See Table 8-1 and 8-2 on Non-Volatile Protection
0 Device is not Write Enabled (default)
1 WEL Write Enable Latch Status R
1 Device is Write Enabled
0 Device is ready
0 RDY/BSY Ready/Busy Status R
1 Device is busy with an internal operation
Notes: 1. Only bits 7 through 2 of the Status Register can be modified when using the Write Status Register command.
2. R/W = Readable and writable
R = Readable only

Figure 10-1. Read Status Register Byte 1

CS

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
SCK
23&2'(

SI 0 0 0 0 0 1 0 1
MSB
67$7865(*,67(5 67$7865(*,67(5 67$7865(*,67(5
BYTE 1 BYTE 1 BYTE 1
HI
GH-
IMPEDANCE
SO D D D D D D D D D D D D D D D D D D D D D D D D
MSB MSB MSB

Table 10-2. Status Register Format – Byte 2

Bit(1) Name Type(2) Description

0 Device is not suspended (default)

7 SUS Suspend Status R
1 Device is suspended from erase

6 CMP Complement Block Protection R/W 0 See table on Block Protection

0 Security Register page-3 is not locked (default)

5 LB3 Lock Security Register 3 R/W
1 Security Register page-3 cannot be erased/programmed

0 Security Register page-2 is not locked (default)

4 LB2 Lock Security Register 2 R/W
1 Security Register page-2 cannot be erased/programmed

0 Security Register page-1 is not locked (default)

3 LB1 Lock Security Register 1 R/W
1 Security Register page-1 cannot be erased/programmed

2 RES Reserved for future use R 0 Reserved for future use

AT25SF321 26
DS-25SF321–047F–8/2017
Table 10-2. Status Register Format – Byte 2

Bit(1) Name Type(2) Description

0 HOLD and WP function normally (default)

1 QE Quad Enable R/W
1 HOLD and WP are I/O pins

0 SRP1 Status Register Protect bit-1 R/W See table on Status Register Protection
Notes: 1. Only bits 6 through 3, 1, and 0 of the Status Register can be modified when using the Write Status Register command
2. R/W = Readable and writable
R = Readable only.

Figure 10-2. Read Status Register Byte 2

CS

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
SCK
23&2'(

SI 0 0   0 1 0 1
MSB
67$7865(*,67(5 67$7865(*,67(5 67$7865(*,67(5
BYTE 2 BYTE 2 BYTE 2
HI
GH-
IMPEDANCE
SO D D D D D D D D D D D D D D D D D D D D D D D D
MSB MSB MSB

10.1.1 SRP1, SRP0 Bits

The SRP1 and SRP0 bits control whether the Status Register can be modified. The state of the WP pin along with the
values of the SRP1 and SRP0 determine if the device is software protected, hardware protected, or permanently
protected (see Table 10-3).

Table 10-3. Status Register Protection Table

SRP1 SRP0 WP Status Register Description

The Status Register can be written to after a Write Enable

0 0 X Software Protected
instruction, WEL=1.(Factory Default)

0 1 0 Hardware Protected WP=0, the Status Register is locked and cannot be written.

Hardware WP =1, the Status Register is unlocked and can be written to

0 1 1
Unprotected after a Write Enable instruction, WEL=1.

Power Supply Lock- Status Register is protected and cannot be written to again
1 0 X
Down (1) until the next Power-Down, Power-Up cycle.

Status Register is permanently protected and cannot be

1 1 X One Time Program
written to.

1. When SRP1, SRP0 = (1, 0), a Power-Down, Power-Up cycle will change SRP1, SRP0 to the (0, 0) state.

10.1.2 CMP, SEC, TB, BP2, BP1, BP0 Bits

The CMP, SEC, TB, BP2, BP1, and BP0 bits control which portions of the array are protected from erase and program
operations (see Tables 8-1 and 8-2).

AT25SF321 27
DS-25SF321–047F–8/2017
 The CMP bit complements the effect of the other bits.
The SEC bit selects between large and small block size protection.
The TB bit selects between top of the array or bottom of the array protection.
The BP2, BP1, and BP0 bits determine how much of the array is protected.

10.1.3 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, Program Security Register, Erase 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
 Program Security Register operation completes successfully or aborts
 Erase 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

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, Program Security Register,
Erase Security Register, or Write Status Register command must have been clocked into the device.

10.1.4 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”.

10.1.5 LB3, LB2, LB1 Bits

The LB3, LB2, and LB1 bits are used to determine if any of the three Security Register pages are locked.
The LB3 bit is in the logical “1” state if Security Register page-2 is locked and cannot be erased or programmed.
The LB2 bit is in the logical “1” state if Security Register page-1 is locked and cannot be erased or programmed.
The LB1 bit is in the logical “1” state if Security Register page-0 is locked and cannot be erased or programmed.

10.1.6 QE Bit
The QE bit is used to determine if the device is in the Quad Enabled mode. If the QE bit is in the logical “1” state, then the
HOLD and WP pins functions as input/output pins similar to the SI and SO. If the QE bit is in the logical “0” state, then the
HOLD pin functions as an input only and the WP pin functions as an input only.

10.2 Write Status Register (01h)

The Write Status Register command is used to modify the Block Protection, Security Register Lock-down, Quad Enable,
and Status Register Protection. Before the Write Status 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 issue the Write Status Register command, the CS pin must first be asserted and the opcode of 01h must be clocked
into the device followed by one or two bytes of data. The first byte of data consists of the SRP0, SEC, TB, BP2, BP1, BP0
bit values and 2 dummy bits. The second byte is optional and consists of 1 dummy bit, the CMP, LB3, LB2, LB1, 1
dummy bit, the QE, and 1 dummy bit. When the CS pin is deasserted, the bit values in the Status Register will be
modified, and the WEL bit in the Status Register will be reset back to a logical “0”.

AT25SF321 28
DS-25SF321–047F–8/2017
 The complete one byte or two bytes 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 Status Register bits will not change, memory protection status will not change, and the WEL bit in the
Status Register will be reset back to the logical “0” state

Table 10-4. Write Status Register Format.

Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Bit 0

SRPO SEC TB BP2 BP1 BP0 WEL RDY/BSY

Figure 10-3. Write Status Register

CS

0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
SCK

Opcode Status Register In - Byte 1 Status Register In - Byte 2

SI 0 0 0 0 0 0 0 1 D D D D D D D D D D D D D D D D
MSB MSB
High-Impedance
SO

Table 10-5. Write Status Register Byte 2

Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Bit 0

SUS CMP LB3 LB2 LB1 reserved QE SRP1

10.3 Write Enable for Volatile Status Register (50h)

The non-volatile Status Register bits described in Table 10-1 and Table 10-2 can also be written to as volatile bits. During
power up reset, the non-volatile Status Register bits are copied to a volatile version of the Status Register that is used
during device operation. This gives more flexibility to change the system configuration and memory protection schemes
quickly without waiting for the typical non-volatile bit write cycles or affecting the endurance of the Status Register non-
volatile bits.
To write the volatile version of the Status Register bits, the Write Enable for Volatile Status Register (50h) instruction
must be issued prior to each Write Status Registers (01h) instruction. Write Enable for Volatile Status Register instruction
will not set the Write Enable Latch bit. It is only valid for the next following Write Status Registers instruction, to change
the volatile Status Register bit values.

AT25SF321 29
DS-25SF321–047F–8/2017
 Figure 10-4. Write Enable for Volatile Status Register
&6

0 1 2 3 4 5 6 7
6&.
OPCODE

6,        
MSB

HI
GH-
IMPEDANCE
62

11. Other Commands and Functions

The AT25SF321 supports three different commands to access device identification that indicates the manufacturer,
device type, and memory density. The returned data bytes provide information as shown in Table 11-1.

Table 11-1. Manufacturer and Device ID Information

Dummy Manufacturer ID Device ID Device ID

Instruction Opcode Bytes (Byte #1) (Byte #2) (Byte #3)

Read Manufacturer and Device ID 9Fh 0 1Fh 87h 01h

Read ID (Legacy Command) 90h 3 1Fh 15h

Resume from Deep Power-Down and

ABh 3 15h
Read Device ID

11.1 Read Manufacturer and Device ID (9Fh)

Identification information can be read from the device to enable systems to electronically query and identify the device
while it is in system.
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 all devices used in the
application can be identified properly. Once the identification process is complete, the application can 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 the opcode of 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. 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 11-2. Manufacturer and Device ID Information

Byte No. Data Type Value

1 Manufacturer ID 1Fh

2 Device ID (Part 1) 87h

3 Device ID (Part 2) 01h

AT25SF321 30
DS-25SF321–047F–8/2017
Table 11-3. Manufacturer and Device ID Details

Hex
Data Type Bit 7 Bit 6 Bit 5 Bit 4 Bit 3 Bit 2 Bit 1 Bit 0 Value Details

JEDEC Assigned Code

Manufacturer ID 1Fh JEDEC Code: 0001 1111 (1Fh for Adesto)
0 0 0 1 1 1 1 1

Family Code Density Code Family Code:100 (AT25SFxxx series)

Device ID (Part 1) 87h
Density Code: 00111 (32-Mbit)
1 0 0 0 0 1 1 1

Sub Code Product Version Code Sub Code: 000 (Standard series)
Device ID (Part 2) 01h
Product Version: 00001
0 0 0 0 0 0 0 1

Figure 11-1. Read Manufacturer and Device ID

CS

0 6 7 8 14 15 16 22 23 24 30 31 32
SCK
OPCODE

SI 9Fh

HIGH-IMPEDANCE
SO 1Fh 87h 01h

MANUFACTURER ID DEVICE ID DEVICE ID

BYTE1 BYTE2

Note: Each transition shown for SI and SO represents one byte (8 bits)

11.2 Read ID (Legacy Command) (90h)

Identification information can be read from the device to enable systems to electronically query and identify the device
while it is in system. The preferred method for doing so is the JEDEC standard “Read Manufacturer and Device ID (9Fh)”
method described in Section 11.1 on page 30; however, the legacy Read ID command is supported on the AT25SF321
to enable backwards compatibility to previous generation devices.
To read the identification information, the CS pin must first be asserted and the opcode of 90h must be clocked into the
device, followed by three dummy bytes. After the opcode has been clocked in followed by three dummy bytes, 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 of 1Fh followed by a single byte of data representing a device code of 15h. After the
device code is output, the sequence of bytes will repeat.
Deasserting the CS pin will terminate the Read ID 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 read.

AT25SF321 31
DS-25SF321–047F–8/2017
 Figure 11-2. Read ID (Legacy Command)

&6

0 1 2 3 4 5 6 7 29 30 31 32 33 34 35 36 37 38 39
6&.
OPCODE '800<%<7(6

6, 1 0    0   X X X X X
MSB

'(9,&(,'
HI
GH-
IMPEDANCE
62 D D D D D D D D
MSB

11.3 Deep Power-Down (B9h)

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 of 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 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 11-3. Deep Power-Down

&6
W('3'

       
6&.
23&2'(

6,        
06%

+,*+,03('$1&(
62
$FWLYH&XUUHQW

,&&
6WDQGE\0RGH&XUUHQW
'HHS3RZHU'RZQ0RGH&XUUHQW

AT25SF321 32
DS-25SF321–047F–8/2017
11.4 Resume from Deep Power-Down (ABh)
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
recognize while in the Deep Power-Down mode.
To resume from the Deep Power-Down mode, the CS pin must first be asserted and the opcode of 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 byte
boundary (multiples of eight bits), then the device will abort the operation and return to the Deep Power-Down mode.

Figure 11-4. Resume from Deep Power-Down

&6
t
RDPD

0 1 2 3 4 5 6 7
6&.
OPCODE

6, 1 0 1 0 1 0 1 1
MSB

HI
GH-
IMPEDANCE
62
Act
iveCur
rent

,&&
St
andbyModeCur
rent
DeepPower
-DownModeCur
rent

11.4.1 Resume from Deep Power-Down and Read Device ID (ABh)

The Resume from Deep Power-Down command can also be used to read the Device ID.
When used to release the device from the Power-Down state and obtain the Device ID, the CS pin must first be asserted
and opcode of ABh must be clocked into the device, followed by 3 dummy bytes. The Device ID bits are then shifted out
on the falling edge of SCLK with most significant bit (MSB) first as shown in Figure 11-4. This command only outputs a
single byte Device ID. The Device ID value for the AT25SF321 is listed in Table 11-1.
After the last bit (bit 0) of the Device ID has been clocked out, the sequence will repeat itself starting again with bit 7 as
long as the CS pin remains asserted and the SCK pin is being pulsed. After CS is deasserted it must remain high for a
time duration of tRDPO before new commands can be received.
The same instruction may be used to read device ID when not in power down. In that case, /CS does not have to remain
high remain after it is deasserted.

AT25SF321 33
DS-25SF321–047F–8/2017
 Figure 11-5. Resume from Deep Power-Down and Read Device ID

&6

0 1 2 3 4 5 6 7 29 30 31 32 33 34 35 36 37 38 39
6&.
OPCODE '800<%<7(6
W5'32
6, 1 0 1 0 1 0 1 1 X X X X X
MSB

'(9,&(,'
HI
GH-
IMPEDANCE
62 D D D D D D D D
MSB

Act
i
veCur
rent

,&&

'HHS3RZHU'RZQ0RGH&XUUHQW 6WDQGE\0RGH&XUUHQW

11.5 Hold Function

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.
If the QE bit value in the Status Register has been set to logical “1”, then the HOLD pin does not function as a control pin.
The HOLD pin will function as an output for Quad-Output Read and input/output for Quad-I/O Read.
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.

AT25SF321 34
DS-25SF321–047F–8/2017
Figure 11-6. Hold Mode

CS

SCK

HOLD

Hold Hold Hold

AT25SF321 35
DS-25SF321–047F–8/2017
12. Electrical Specifications
12.1 Absolute Maximum Ratings*

Temperature under Bias. . . . . . . . . . . . . -55°C to +125°C *Notice: Stresses beyond those listed under “Absolute
Maximum Ratings” may cause permanent damage to
Storage Temperature . . . . . . . . . . . . . . . -65°C to +150°C the device. This is a stress rating only and functional
operation of the device at these or any other
conditions beyond those indicated in the operational
All Input Voltages (including NC Pins)
sections of this specification is not implied. Exposure
with Respect to Ground . . . . . . . . . . . . . . -0.6V to +4.1V
to absolute maximum rating conditions for extended
periods may affect device reliability.
All Output Voltages
with Respect to Ground . . . . . . . . . . -0.6V to VCC + 0.5V

12.2 DC and AC Operating Range

AT25SF321

Operating Temperature (Case) Industrial -40°C to 85°C

VCC Power Supply 2.5V to 3.6V

12.3 DC Characteristics

2.5V to 3.6V

Symbol Parameter Condition Min Typ Max Units

CS, HOLD, WP = VIH

IDPD Deep Power-Down Current 2 5 µA
All inputs at CMOS levels

CS, HOLD, WP = VIH

ISB Standby Current 13 25 µA
All inputs at CMOS levels

f = 20MHz; IOUT = 0mA 3 6 mA

ICC1 (1)(3) Active Current, Read (03h,

f = 50MHz; IOUT = 0mA 4 7 mA
0Bh) Operation

f = 85MHz; IOUT = 0mA 5 8 mA

f = 50MHz; IOUT = 0mA 5 8 mA

Active Current,(3Bh, BBh
ICC2(1)
Read Operation (Dual)
f = 85MHz; IOUT = 0mA 6 10 mA

f = 50MHz; IOUT = 0mA 6 10 mA

Active Current,(6Bh, EBh
ICC3(1)
Read Operation (Quad)
f = 85MHz; IOUT = 0mA 8 14 mA

Active Current,
ICC4(1) CS = VCC 10 16 mA
Program Operation

Active Current,
ICC5(1) CS = VCC 10 16 mA
Erase Operation

AT25SF321 36
DS-25SF321–047F–8/2017
 2.5V to 3.6V

Symbol Parameter Condition Min Typ Max Units

ILI Input Load Current All inputs at CMOS levels 1 1 µA

ILO Output Leakage Current All inputs at CMOS levels 1 1 µA

VIL Input Low Voltage VCC x 0.3 V

VIH Input High Voltage VCC x 0.7 V

VOL Output Low Voltage IOL = 1.6mA; VCC = 2.5V 0.4 V

VOH Output High Voltage IOH = -100µA VCC - 0.2V V

1. Typical values measured at 3.0V @ 25°C for the 2.5V to 3.6V range

12.4 AC Characteristics - Maximum Clock Frequencies

2.5V to 3.6V 2.7V to 3.6V

Symbol Parameter Min Typ Max Min Typ Max Units

fCLK Maximum Clock Frequency for All Operations 104 104 MHz
(excluding opcodes below)
fRDLF Maximum Clock Frequency for 03h Opcode 50 50 MHz

fRDHF Maximum Clock Frequency for 0Bh Opcode 70 85 MHz

fRDDO Maximum Clock Frequency for 3B, BBh 70 85 MHz

Opcode
fRDQO Maximum Clock Frequency for 6B, EBh 70 85 MHz
Opcode

12.5 AC Characteristics - All Other Parameters

2.5V to 3.6V 2.7V to 3.6V
Symbol Parameter Min Typ Max Min Typ Max Units

tCLKH Clock High Time 5 5 ns

tCLKL Clock Low Time 5 5 ns

tCLKR(1) Clock Rise Time, Peak-to-Peak (Slew Rate) 0.1 0.1 V/ns

tCLKF(1) Clock Fall Time, Peak-to-Peak (Slew Rate) 0.1 0.1 V/ns

tCSH Chip Select High Time 10 10 ns

tCSLS Chip Select Low Setup Time (relative to Clock) 5 5 ns

tCSLH Chip Select Low Hold Time (relative to Clock) 5 5 ns

tCSHS Chip Select High Setup Time (relative to Clock) 5 5 ns

AT25SF321 37
DS-25SF321–047F–8/2017
12.5 AC Characteristics - All Other Parameters
2.5V to 3.6V 2.7V to 3.6V
Symbol Parameter Min Typ Max Min Typ Max Units

tCSHH Chip Select High Hold Time (relative to Clock) 5 5 ns

tDS Data In Setup Time 2 2 ns

tDH Data In Hold Time 2 2 ns

tDIS(1) Output Disable Time 7 6 ns

Output Valid Time (03h, 0Bh) 7 6 ns

tV Output Valid Time (3Bh, BBh - Dual) 7 7

Output Valid Time (6Bh, EBh - Quad) 8 8

tOH Output Hold Time 0 0 ns

tHLS HOLD Low Setup Time (relative to Clock) 5 5 ns

tHLH HOLD Low Hold Time (relative to Clock) 5 5 ns

tHHS HOLD High Setup Time (relative to Clock) 5 5 ns

tHHH HOLD High Hold Time (relative to Clock) 5 5 ns

tHLQZ (1) HOLD Low to Output High-Z 6 6 ns

tHHQZ(1) HOLD High to Output High-Z 6 6 ns

tWPS(1) (2) Write Protect Setup Time 20 20 ns

(1)(2)
tWPH Write Protect Hold Time 100 100 ns

tEDPD(1) Chip Select High to Deep Power-Down 1 1 µs

tRDPD(1) Chip Select High to Standby Mode 5 5 µs

tRDPO(1) Resume Deep Power-Down, CS High to ID 5 5 µs

1. Not 100% tested (value guaranteed by design and characterization).

2. Only applicable as a constraint for the Write Status Register command when BPL = 1.

12.6 Program and Erase Characteristics

2.5 to 3.6V 2.7V to 3.6V

Symbol Parameter Min Typ Max Min Typ Max Units

tPP (1) Page Program Time (256 Bytes) 0.7 5 0.7 3.0 ms

tBP Byte Program Time 5 5 µs

4 Kbytes 60 300 60 300

tBLKE(1) Block Erase Time 32 Kbytes 300 1300 300 1300 ms

64 K bytes 500 3000 500 3000

tCHPE(1) (2) Chip Erase Time 25 60 25 60 sec

AT25SF321 38
DS-25SF321–047F–8/2017
12.6 Program and Erase Characteristics
2.5 to 3.6V 2.7V to 3.6V

Symbol Parameter Min Typ Max Min Typ Max Units

tSRP(1) Security Register Program Time 2.5 2.5 ms

tSRP(1) Security Register Erase Time 15 15 ms

tWRSR(2) Write Status Register Time 15 15 ms

1. Maximum values indicate worst-case performance after 100,000 erase/program cycles.

2. Not 100% tested (value guaranteed by design and characterization).

12.7 Power-Up Conditions

Symbol Parameter Min Max Units

tVCSL Minimum VCC to Chip Select Low Time 20 µs

tPUW Power-up Device Delay Before Program or Erase Allowed 10 ms

VPOR Power-on Reset Voltage 2.3 V

AT25SF321 39
DS-25SF321–047F–8/2017
12.8 Input Test Waveforms and Measurement Levels
AC 0.9VCC AC
DRIVING VCC/2 MEASUREMENT
LEVELS 0.1VCC LEVEL

tR, tF < 2 ns (10% to 90%)

12.9 Output Test Load

Device
Under
Test

30pF

13. AC Waveforms
Figure 13-1. Serial Input Timing
W&6+

&6
W&6/6 W&6/+ W&6++
W&/.+ W&/./ W&6+6
6&.
W'6 W'+

6, 06% /6% 06%

+,*+,03('$1&(
62

Figure 13-2. Serial Output Timing

&6
W&/.+ W&/./ W',6

6&.

6,
W2+
W9 W9
62

AT25SF321 40
DS-25SF321–047F–8/2017
Figure 13-3. WP Timing for Write Status Register Command When BPL = 1

CS
t WPS t WPH

WP

SCK

SI 0 0 0 X MSB
MSB OF LSB OF MSB OF
WRITE STATUS REGISTER WRITE STATUS REGISTER NEXT OPCODE
OPCODE DATA BYTE

HIGH-IMPEDANCE
SO

Figure 13-4. HOLD Timing – Serial Input

CS

SCK
tHHH tHLS
tHLH tHHS
HOLD

SI

HIGH-IMPEDANCE
SO

Figure 13-5. HOLD Timing – Serial Output

CS

SCK
tHHH tHLS
tHLH tHHS
HOLD

SI
tHLQZ tHHQX

SO

AT25SF321 41
DS-25SF321–047F–8/2017
14. Ordering Information

14.1 Ordering Code Detail

AT 2 5 S F 3 2 1 – SSHD – B

Designator Shipping Carrier Option

B = Bulk (tubes)
T = Tape and reel
Y = Tray

Product Family Operating Voltage

D = 2.5V to 3.6V

Device Grade
H = Green, NiPdAu lead finish,
Device Density Industrial temperature range
(–40°C to +85°C)
32 = 32-megabit
Package Option
Interface M = 8-pad, 5 x 6 x 0.6 mm UDFN
1 = Serial SS = 8-lead, 0.150" wide SOIC
S = 8-lead, 0.208" wide SOIC
DWF = Die in Wafer Form

Max. Freq.
Ordering Code (1) Package Operating Voltage (MHz) Operation Range

AT25SF321-SSHD-B
8S1
AT25SF321-SSHD-T

AT25SF321-SHD-B Industrial
8S2 2.5V to 3.6V 85MHz
AT25SF321-SHD-T (-40°C to +85°C)

AT25SF321-MHD-T 8MA1
(2)
AT25SF321-DWF DWF

1. The shipping carrier option code is not marked on the devices.

2. Contact Adesto for mechanical drawing or Die Sales information.

Package Type

8S1 8-lead, 0.150" Wide, Plastic Gull Wing Small Outline Package (JEDEC SOIC)

8S2 8-lead, 0.208" Wide, Plastic Gull Wing Small Outline Package (EIAJ SOIC)

8MA1 8-pad (5 x 6 x 0.6mm body), Thermally Enhanced Plastic Ultra Thin Dual Flat No-lead (UDFN)

DWF Die in Wafer Form

AT25SF321 42
DS-25SF321–047F–8/2017
15. Packaging Information

15.1 8S1 – 8-lead, .150” JEDEC SOIC

E E1

N L

Ø
TOP VIEW
END VIEW
e b
A COMMON DIMENSIONS
(Unit of Measure = mm)

SYMBOL MIN NOM MAX NOTE

A1
A 1.35 – 1.75
A1 0.10 – 0.25
b 0.31 – 0.51
C 0.17 – 0.25
D 4.80 – 5.05
D
E1 3.81 – 3.99
E 5.79 – 6.20
SIDE VIEW e 1.27 BSC

Notes: This drawing is for general information only. L 0.40 – 1.27

Refer to JEDEC Drawing MS-012, Variation AA Ø 0° – 8°
for proper dimensions, tolerances, datums, etc.

5/19/10

® TITLE GPC DRAWING NO. REV.

8S1, 8-lead (0.150” Wide Body), Plastic Gull
Package Drawing Contact: SWB 8S1 F
Wing Small Outline (JEDEC SOIC)
contact@adestotech.com

AT25SF321 43
DS-25SF321–047F–8/2017
15.2 8S2 – 8-lead, .208” EIAJ SOIC

E E1

L
N

TOP VIEW θ

END VIEW
e b COMMON DIMENSIONS
A (Unit of Measure = mm)

SYMBOL MIN NOM MAX NOTE

A1 A 1.70 2.16
A1 0.05 0.25
b 0.35 0.48 4
C 0.15 0.35 4
D 5.13 5.35
D
E1 5.18 5.40 2
E 7.70 8.26
L 0.51 0.85
SIDE VIEW θ 0° 8°
e 1.27 BSC 3
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 aren't 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 .021 mm.

4/15/08
® TITLE GPC DRAWING NO. REV.
8S2, 8-lead, 0.208” Body, Plastic Small
Package Drawing Contact: Outline Package (EIAJ) STN 8S2 F
contact@adestotech.com

AT25SF321 44
DS-25SF321–047F–8/2017
15.3 8MA1 – UDFN

E
C

Pin 1 ID

D SIDE VIEW

TOP VIEW
A1
A

K E2
0.45 Option A
8 1 Pin #1
Pin #1 Notch Chamfer COMMON DIMENSIONS
(0.20 R) (C 0.35)
(Unit of Measure = mm)
(Option B)
7 2 SYMBOL MIN NOM MAX NOT E
A 0.45 0.55 0.60
D2 e
A1 0.00 0.02 0.05
6 3
b 0.35 0.40 0.48
C 0.152 REF
5 4 D 4.90 5.00 5.10
D2 3.80 4.00 4.20
b
E 5.90 6.00 6.10
BOTTOM VIEW
L E2 3.20 3.40 3.60
e 1.27
L 0.50 0.60 0.75
y 0.00 – 0.08
K 0.20 – –

4/15/08
TITLE GPC DRAWING NO. REV.
® Package Drawing Contact: 8MA1, 8-pad (5 x 6 x 0.6 mm Body), Thermally
contact@adestotech.com Enhanced Plastic Ultra Thin Dual Flat No Lead YFG 8MA1 D
Package (UDFN)

AT25SF321 45
DS-25SF321–047F–8/2017
16. Revision History
Revision Level – Release Date History

A – June 2015 Initial release

Updated operation description in Section 7.4 (Program/Erase

B – July 2015 Suspend) and references in Table 7-1. Updated Table 8-2. Updated
specifications for Icc3 at 85 MHz and Chip Erase Time (max).

C – August 2015 Added Die in Wafer Form ordering option.

Updated specification for TPP. Changed document status from

D – January 2016
Preliminary to Complete.

Updated 90h opcode description (added 3 dummy bytes). Added

E – April 2016
Suspend bit description.

F – August 2017 Updated corporate address.

AT25SF321 46
DS-25SF321–047F–8/2017
Corporate Office
California | USA
Adesto Headquarters
3600 Peterson Way
Santa Clara, CA 95054
Phone: (+1) 408.400.0578
Email: contact@adestotech.com

© 2017 Adesto Technologies. All rights reserved. / Rev.: DS-25SF321–047F–8/2017

Adesto®, the Adesto logo, CBRAM®, and DataFlash® are registered trademarks or trademarks of Adesto Technologies. All other marks are the property of their respective
owners.

Disclaimer: Adesto Technologies Corporation makes no warranty for the use of its products, other than those expressly contained in the Company's standard warranty which is detailed in Adesto's Terms
and Conditions located on the Company's web site. The Company assumes no responsibility for any errors which may appear in this document, reserves the right to change devices or specifications
detailed herein at any time without notice, and does not make any commitment to update the information contained herein. No licenses to patents or other intellectual property of Adesto are granted by the
Company in connection with the sale of Adesto products, expressly or by implication. Adesto's products are not authorized for use as critical components in life support devices or systems.