CC2540: CC2540 program loss issue

user4206199

Part Number: CC2540

Dear

Customer use our cc2540 in their product,about 1/200 product appear program loss issue lately，and it become nomal work after reload program，pls help provide some suggestions

over 7 years ago

0 Katie Pier over 7 years ago

TI__Guru* 79861 points

Hi,

Can you please provide some more details to try to help diagnose the issue: if you read out the flash of an affected board using a programming tool instead of just reloading it, do you see flash as being totally blank, or is it simply different than the expected memory? If it's different, can you provide info about what differences there are (e.g. blank sections, individual bad bytes, and are bits always going 1->0 on the failing units)?

Do you have OAD functionality in your application?

Do you use any locking/protection mechanism on the flash?

Regards,
Katie

0 juntao liu over 7 years ago in reply to Katie Pier

Prodigy 100 points

Hi,

Thanks for your answer first.And the differences between normal and bad image file as follows:

1.from 00001000 to 00015000 total 80KB image codes lost ,instead by FF.

2.The right crc shadow 0x2012 ocupied by 0xFFFF,and the file name space has the same problem.

Please help us solve the issue.

Best regards.

0 Katie Pier over 7 years ago in reply to juntao liu

TI__Guru* 79861 points

Hi,

Thanks for providing the information. It looks like some sort of erasure has occurred (FF is the erased state for Flash). Since it is an erase in such a large block, can you comment on if there is any sort of field firmware upgrade functionality implemented in your application, or if your application does any in-system flash writes?

Regards,
Katie

0 juntao liu over 7 years ago in reply to Katie Pier

Prodigy 100 points

Hi,

We use Universal Boot Loader for SOC-8051 by USB-MSD updating image run code. The UBL receives image data and erases the flash.I doubt that the problem occured in writing data to the flash,but i do not know why this would happen.

0 TER over 7 years ago in reply to juntao liu

TI__Guru**** 317180 points

One thing that is not clear for me: When does this occur? Is it in production, is it after n days in the field?

0 juntao liu over 7 years ago in reply to TER

Prodigy 100 points

Situation one is our customer using it good for the beginning ,but after several days later during using it on work,it suddenly did not work.I read out the contents from the flash and find out that the right crc shadow 0x2012 ocupied by 0xFFFF,and the file name space has the same problem.
Situation two is the first time using the dongle it works fine,but serval days later inserting it in the usb interface for the second time,it does not work.I read out the contents from the flash and find out that from 00001000 to 00015000 total 80KB image codes lost ,instead by FF.

0 TER over 7 years ago in reply to juntao liu

TI__Guru**** 317180 points

For me it looks like one or more flash pages have been erased. Does your program contain code that do a flash erase? If so, the question is why that piece of code runs.

0 juntao liu over 7 years ago in reply to TER

Prodigy 100 points

No,I just use NV page by calling the api,but never do a flash erase.

0 TER over 7 years ago in reply to juntao liu

TI__Guru**** 317180 points

The NVS driver could erase a page, do you have an example of the API calls you use?

0 juntao liu over 7 years ago in reply to TER

Prodigy 100 points

The first time I use the read flash api as follows:

The firsttime I use the write flash api as follows:

Api function osal_snv_read() and osal_snv_write（）you can find it in the stack BLE-CC254x-1.3.2.

0 user4206199 over 7 years ago in reply to juntao liu

Intellectual 470 points

TER

Any suggestion for this issue?thanks

0 Aslak N. over 7 years ago in reply to user4206199

TI__Mastermind 23440 points

Hi user4206199,

Does this happen on any computer, or a specific one?

The USB-MSD bootloader can be configured to appear as a mass storage device for some seconds after boot, or if a GPIO is held high/low. Is this happening here?

Is it possible that some program, or the operating system itself, tries to write to the drive when it appears as a mass storage device, and that this causes the erase?

Best regards,
Aslak

0 juntao liu over 7 years ago in reply to Aslak N.

Prodigy 100 points

Hi Aslak N，

This happens both on labtop and desktop.

I configured 0 second delay as a mass storage device after boot.

Which GPIO you mean is held high/low,and what affact if so?

0 J Lindh over 7 years ago in reply to juntao liu

TI__Guru 57865 points

Hi,

I am curious to know what happens if you increase the delay. If you change it to 300ms, for example, do you still observe the issue?

0 juntao liu over 7 years ago in reply to J Lindh

Prodigy 100 points

I used to configure the delay to 1 second,and do a test pull in and out the USB interface with a dongle for 1000 times.The dongle worked fun.But It happened program loss when our costumers using it.

0 J Lindh over 7 years ago in reply to juntao liu

TI__Guru 57865 points

If it only happens on a few devices, I'm interested in knowing the details around the hardware design. Have you followed out HW reference design for USB dongle? (CC2540EMK-USB)

0 juntao liu over 7 years ago in reply to J Lindh

Prodigy 100 points

Our dongle Circuit schematic diagram as follows.Are there any problems?

0 TER over 7 years ago in reply to juntao liu

TI__Guru**** 317180 points

From what I can see the CFG-DISCS section (referring to Universal Boot Loader for SoC-8051 by USB-MSD Developer's Guide) can set a GPIO. What is this struct defined as in your case? If you have defined a GPIO that is floating this could lead to issues.

I read through some of the thread again but have to verify a detail: Does the "flash pages get overwritten be 0xFF" only happen when the dongle is inserted or could it happen after it's been in use some time? Not familiar with this bootloader but it should only do a flash write the first time the dongle is in use?