"Failed to power up DAP" successfully resolved, possible causes?

[MAlen]

Hi all,

we recently had a problem with couple of our boards based on Raytac MDBT50Q-1MV2 (nrf52840 + osc. + antenna). We base our app on Zephyr. After programming some boards (cca 5% of the batch) using Segger EDU mini they became unresponsive, and trying to connect RTT viewer resulted in "Failed to power up DAP". Based on other tickets we immediately suspected nRST line. On one board there was indeed a short to GND and we scrapped it. However, other boards had no short, only not HIGH level on the pin (most had 0.01 - 0.3 V), but one had 1.5 V. This one behaved peculiar since it worked fine in 1 out of 20-30 power-ONs. After shorting nRST to VCC we successfully erased the flash and reprogrammed all the boards, all work fine now.

My question is how could have this happened, or rather, how to avoid it?

My hunch tells me that somehow the pull down got enabled on pin P0.18 (default nRST and we use it as such) and caused permanent reset state. On peculiar one perhaps both pull down and pull up? The troubling thing is we do not use P0.18 anywhere in our app, it is not even defined.

Thank you for any hint/idea/suggestion.

[Kenneth]

You should look into whether you may have CONFIG_GPIO_AS_PINRESET in your defines, if it is defined, then in system_nrf52.c it will be used to enable pin reset:
https://infocenter.nordicsemi.com/topic/ps_nrf52840/uicr.html#register.PSELRESET-0-1 

If it's enabled you can check by:

Unfortunately if your code set this pin output low and/or configure pull-down, then it may be that the the chip is stuck in a infinite pin reset. The only way is to force it high externally in such case. I can't think of other reasons why it was low (other than electrical problems externally though, e.g. soldering residue).

[Matevz B]

Thank you very much for this hint. It has saved many hours of debugging. The board is no longer brciked!