TMS320F28022-Q1: What are root causes for pre-condition failed?

Will Gao

Hi BU,

Customer side met 2 chips can't be erased (returned from customer's field, i.e. car OEM). They tried to use On-Chip Flash tool to erase the chips, and 1 chip returns erasing error -21, that is pre-condition failure.

Since pre-condition is after pre-compaction, so that means this chip is not in over-erased or "depletion" state. Am I right?

And I let them give me the Flash contents for this chip, you can check from: https://e2e.ti.com/cfs-file/__key/communityserver-discussions-components-files/171/C1_5F00_sections_5F00_not_5F00_erasable.bin

As you can see from the above bin file, the section B for this chip is almost 0's, that is the clear stage for erasing process. However, there is some bits can't be cleared as following fig shows:

I want to check with you what are the possible root causes for this pre-condition error? Could it be Flash memory cell HW failure? If it is, we need to do failure analysis and at the mean time need to check the application root cause for this failure. Please give your guidance, thanks.

Regards,

Will

over 2 years ago

0 Will Gao over 2 years ago

TI__Intellectual 2335 points

And there is a pattern, the bit can't be cleared in each 0x??E0 raw of sector B:

Maybe it is the pattern of Flash memory organization.

0 Will Gao over 2 years ago in reply to Will Gao

TI__Intellectual 2335 points

Hi BU,

Any comments?

0 Will Gao over 2 years ago in reply to Will Gao

TI__Intellectual 2335 points

BU experts,

Customer is pushing me, please put this issue in high priority, thanks.

0 Charles Roberson over 2 years ago in reply to Will Gao

TI__Genius 15080 points

Will,

>Since pre-condition is after pre-compaction, so that means this chip is not in over-erased or "depletion" state. Am I right?

Correct, as programming zero values occurs first before the erase step.

>What are the possible root causes for this pre-condition error?

Programming with the flash pump, however I'm not certain this is what's going on.

>Could it be Flash memory cell HW failure?

Will help to resolve this and refer hardware experts as well.

Thanks and regards,

Charles

0 Will Gao over 2 years ago in reply to Charles Roberson

TI__Intellectual 2335 points

Hi Charles,

May I know the progress for this issue now? Thanks.

Will

0 Charles Roberson over 2 years ago in reply to Will Gao

TI__Genius 15080 points

Hi Will,

This issue has been moved to the previous flash technology expert for this device. Can expect an update for this week.

Thanks and regards,

Charles

0 Will Gao over 2 years ago in reply to Charles Roberson

TI__Intellectual 2335 points

Hi Charles,

any updates?

0 Nima Eskandari over 2 years ago in reply to Will Gao

TI__Guru 52040 points

Charles Roberson, Matt, can you look into this?

0 MatthewPate over 2 years ago in reply to Nima Eskandari

TI__Guru* 83050 points

Will,

Looking at the data, even though the fail bits are 0x100 apart, I think that these are really 0x80(128) 16-bit words apart since the C28x is a word based device. I need to check, but I believe this may be occurring in a common location in each row.

Before we send the devices back to TI for analysis, let me ask one of the customer return engineers if there is any other information we need from the customer.

Best,

Matthew

0 Will Gao over 2 years ago in reply to MatthewPate

TI__Intellectual 2335 points

Hi Matthew,

Yes, for C28x, it should be 0x80 apart.

Could you tell me the possible root causes for this kind of issue, i.e. pre-condition failed? If this is an error code for erasing operation, that means we have this error's pattern, please share to me.

Regards,

Will

+1 MatthewPate over 2 years ago in reply to Will Gao

TI__Guru* 83050 points

Will,

We will need to ask the customer to submit these units back to TI for Failure Analysis. Through re-test we should be able to isolate why these bits are having an issue with pre-condition. Given the pattern to the data, this may not be in the physical flash bit(s), but rather in the selection logic of the bits in the flash array. FA should be able to tell us more.

Let me know if you need help to submit these back, or if you have done this before.

Best,

Matthew

0 Will Gao over 2 years ago in reply to MatthewPate

TI__Intellectual 2335 points

Hi Matthew,

Got it. I will let customer to submit CPR. Will let you know the progress via email. Thanks.

Regards,

Will

C2000™︎ microcontrollers

C2000 microcontrollers forum

TMS320F28022-Q1: What are root causes for pre-condition failed?