S25FS512S QSPI FLASH - Why is the 8-contact-WSON package slower than the 16-pin-SOIC package

Tip / Sign in to post questions, reply, level up, and achieve exciting badges. Know more

cross mob
Anonymous
Not applicable

We are designing in  a S25FS512S QSPI Flash memory chip into a mass-produced product.

  • On version 1 of our PCB we have the 16-pin SOIC
  • On version 2 of the same PCB we have replaced the 16-pin SOIC   with    8-contact WSON

The problem is the write and erase is much slower on the WSON version (test results table below).

  Temp is the same around 27 deg C.

  Programmed/erased cycles low, all chips are new.

  Vcc should be with 10% of 1V8

We need support to understand why the WSON package is slower.

Looking forward to getting some support.

Test results:

Version 1 with 16-pin SOIC

Version 2 with 8-contact WSON

Test 1

: writing 1Mb file into the start of the flash

VER1 256k average block write takes - ~11 s

VER2 256k average block write takes - ~18 s

VER1 average erase block takes - ~700ms

VER2 average erase block takes - ~1.5s

Test2:

Erasing three consecutive blocks (256k) , writing 8K of data into beginning of each block, reading the data to verify it is correct.

Times below are in ms.

Writing to QSPI-Flash... block=79

  erase time =740 write time =530 bytes=8191

Writing to QSPI-Flash... block=80

  erase time =620 write time =530 bytes=8191

Writing to QSPI-Flash... block=81

  erase time =620 write time =530 bytes=8191

Reading from QSPI-Flash... block=79

read time =50 bytes=8191

Reading from QSPI-Flash... block=80

read time =50 bytes=8191

Reading from QSPI-Flash... block=81

read time =40 bytes=8191

Writing to QSPI-Flash... block=79

erase time =1150 write time =650 bytes=8191

Writing to QSPI-Flash... block=80

erase time =1100 write time =630 bytes=8191

Writing to QSPI-Flash... block=81

erase time =1150 write time =640 bytes=8191

Reading from QSPI-Flash... block=79

read time =40 bytes=8191

Reading from QSPI-Flash... block=80

read time =40 bytes=8191

Reading from QSPI-Flash... block=81

read time =50 bytes=8191

0 Likes
1 Solution
BushraH_91
Moderator
Moderator
Moderator
750 replies posted 50 likes received 250 solutions authored

Hello Prakash,

The packaging type does not have any impact on the flash program/erase performance.

Since test conditions are identical to both the packaging type devices, most likely the performance difference comes from the variety of individual devices. Different devices and even different sectors within the same device may have different program/erase performance. Please refer to attachment for  FS512S program/erase performance table. The performance of most devices are around typical (Typ) value , but some devices close to the Max value is also expectable.

Thank you

Regards,

Bushra

View solution in original post

0 Likes
4 Replies
BushraH_91
Moderator
Moderator
Moderator
750 replies posted 50 likes received 250 solutions authored

Hello Prakash,

Thank you for contacting Cypress Community Forum.

What is the speed grade for SOIC package and WSON package?

Is it WSON version came from fresh part with no previous program/erase?

Thank you

Regards,

Bushra

0 Likes

Hello Prakash,

  1. Replacing 16-pin SOIC with 8-contract WSON is only package change, should not have any performance differences. Are you working at the same condition?
  2. Same Vcc on both boards?
  3. Both work at QSPI mode?
  4. Please check waveform if there is any extra delays in between clock groups?
  5. How do you measure the program/erase time?

You mentioned: “Vcc should be with 10% of 1V8” – What does this mean? If it means Vcc could be 10% lower than 1.8V, which could be reach 1.62V, it will be a problem. FS512S requires 1.7V < Vcc < 2.0V..

Thank you

Regards,

Bushra

0 Likes
Anonymous
Not applicable

Greetings Bushra,

Thank you for your help on this support request.

  • Both the 16-pin SOIC and the 8-contract WSON are speed grade AG which is 133 MHz.
  • Vcc is stable at 1.8 V during the tests, I have checked this on the board during a test.
  • Both package types are programmed and working in the same QSPI mode i.e. the same software is running on both boards.
  • The ambient temperature around the chip is 27 deg C during the tests
  • The program/erase time is measured by the software (using an ARM debugger),  using the same code on both boards,

Probing all the pins and connecting a scope to analyze the protocol timing would be a last resort.

I do have some more test results for the Version 2 board:

Capture.JPG

I need to understand why the WSON package is slower, but note I only have one test results for board version1 (SOIC ).

Can you explain what is the expected variation in erase time i.e. by how many milliseconds can the sector erase time vary on the same chip and on different chips  (with Vcc stable at 1.8 V and temp 27 deg C).

Kind regards

Prakash

0 Likes
BushraH_91
Moderator
Moderator
Moderator
750 replies posted 50 likes received 250 solutions authored

Hello Prakash,

The packaging type does not have any impact on the flash program/erase performance.

Since test conditions are identical to both the packaging type devices, most likely the performance difference comes from the variety of individual devices. Different devices and even different sectors within the same device may have different program/erase performance. Please refer to attachment for  FS512S program/erase performance table. The performance of most devices are around typical (Typ) value , but some devices close to the Max value is also expectable.

Thank you

Regards,

Bushra

0 Likes