RE: Intermitting problem with the nRF5340

JONATHAN LL — Wed, 03 May 2023 14:56:21 GMT

Want to add one additional comment:

Errata 160 is less likely at room temps, are the devices used somewhere cold?

I would say that the number one suspect from the info you have shared so gar is the errata 161.

Are you using approtect? If so then you cant watch what's happening when the device gets stuck.
If you set CONFIG_NRF_APPROTECT_LOCK=n then you should be able to connect a debugger and see what is happening when it gets stuck.

And are you using watchdog timers to reset the device ? could maybe be a workaround.

Regards,
Jonathan

RE: Intermitting problem with the nRF5340

JONATHAN LL — Tue, 02 May 2023 20:35:25 GMT

Hi Jeff and thanks for posting!

Will gladly do my best to help here so thanks for all the details.

[quote user=""]Our FW has also been tested and verified to work on the nRF5340-DK board but not over extended periods of time.[/quote]

Would be good to know if the issue is reproducible on the nRF5340DK then I would be very confident to exclude firmware if it does not happen.

[quote user=""]It seems all units will eventually stop working if allowed to operate long enough.[/quote]

Do you have any data on witch devices has experienced the issue and how often it. Could be good to know if some units are more prone to the issue.

Have you been in contact with Fanstel regarding the issue to see if they have any info on this.

If this is a due to the chip having a bug or faulty from production in some way then we would like to have the SoC build code. Could you provide this to us. (We can make a private ticket and do any info exchange there that you don't want public)

I can do a HW review of device and see if there is anything to comment on.

You could also share you application and we could have a look but its often more difficult to find any issues this way if we don't have the device to test on. But if the issue is reproducible on the nRF5340DK then this is one way forward.

Edit:
Are you using DCDC? if so then qwertynoon comment could be something to test on a couple of units. Currently not aware of a anything DCDC related that could fit the issue you describe.
Do you have CONFIG_RESET_ON_FATAL_ERROR=y ?

And are you following the workaround for Errata 161? If not then if you have an assert that is triggered or after a system off it can put the device in a stuck state. The workaround was added to main a month ago, https://github.com/NordicSemiconductor/nrfx/blame/master/hal/nrf_reset.h#L177 , make sure to incorporate this.

Regards,
Jonathan

RE: Intermitting problem with the nRF5340

qwertynoon — Tue, 02 May 2023 06:42:23 GMT

I would advise you to run your test with DCDC turned off (CONFIG_BOARD_ENABLE_DCDC=n) and whether this bug will happen. I have a bitter experience with this functionality.