nrf5340 resets but does not halt on "mon reset" command in debugger

[randomdude]

Hi,

I have two nearly identical revisions of a board with an nrf5340 which behave differently in a curious way:

On one revision I can issue a "mon reset" command in gdb, and it will reset the chip and then wait until a "continue" command is issued.

However, doing the same on the other revision will still reset the chip, but start running immediately, without waiting for a "continue" command.

I have flashed both boards with the exact same image (and with "--recover"), and attached the debugger in the same way via "west attach".

Also, this has been observed with multiple boards across these two revisions, and they all show the same behavior respectively.

What could be the reason for this?

[Priyanka]

H,

Are you working with the nRF5340 DK? Or is it some other board?

If it's the DK, could you tell us the revisions of both boards?

Regards,

Priyanka

[randomdude]

Hi Priyanka, Thanks for answering!

It's our own design, but I largely followed the nrf5340 QKAA reference layout 1_2 suggestion on both revisions.

Regards,

Al

[Priyanka]

Hi,

Apologies for the delay in response. We are a bit understaffed at the moment due to the summer vacations here in Norway.

Thank you for understanding.

Since you are flashing the same firmware, same flashing method as well as debugging procedure, then it could only be a hardware difference in the two revisions that is causing this issue.

Please make sure that the REST pin is not floating in either revisions. Could you tell me what are the differences in both design?

I guess it would be a bit helpful if you could upload the schematics of both revisions so that I can take a look at it.

Regards,

Priyanka

[randomdude]

Hi Priyanka,

as this is contract work for a client, I can't share a complete schematic of the project. The differences between both revisions are marginal and only apply to a few GPIO pins. The power supply and clock configuration hasn't changed between revisions, and matches the PDF I uploaded.

I actually did have the RESET (I assume this is what you meant by " REST"?) pin floating in both revisions, but I added a 10k pull-up to the board that was misbehaving, and it didn't make a difference.

But I've got a few more details about the behavior:

Apparently, my initial observation turned out to be wrong, and the behavior is not the same across all boards. Some seem to behave correctly, while others don't.

I can't make heads or tails of this, it all seems very odd. I still hope this helps a bit anyway.

Regards,

Al

nrf53-fanout.pdf