nRF52832 module dead once being mounted on custom PCB

[Avadacadabara]

Dear all,

We're developing a custom HW with several nRF52832 chips on one board. For convenience we applied third party module instead of building minimum system by ourselves.

But one of the module at certain position cannot be detected by JLINK. Peripheral circuits are similar for all six modules but only the one in the middle comes up with such problem. We've tried the module independently before mount it to PCB and it works. But the module is dead once we soldered it to PCB.

What might be the cause of that?

Thanks ahead,

Ava

FYR, following pics are scopes our SCH and PCB file of two nearby modules(above description indicates the module on right hand side):

[ketiljo]

I'd suspect a short somewhere. Have you measured all supply voltage pins? If your DEC1 pin is same as the DEC4 pin on the nRF52832, the caps to ground, C23 and C24, mist be 1 µF each for the DCDC converter to work. 

[Avadacadabara]

Vcc of all modules are connected to the same bus, but only this module fails to work. We did measure supply pins(those "TXDB" pads) but no short detected. Do we have to test other Vcc that haven't been lead out from module since they might have different behavior?

C23,C24 are 100nF based on reference circuit but we left the L5,6,11,12 empty. DEC4 cap seems already integrated in module but not quite sure the value. Do I need to adjust cap on DEC1 and DEC4 if we don't intend to turn on DCDC converter?

We tried to left the whole PCB empty and solder only this module and its peripherals, it's still dead...

Thanks

[Avadacadabara]

Solved, we mounted inductors to enable the DCDC and it can be detected by JLINK.

But only this module has such restriction.

[ketiljo]

If you tested with the DCDC enabled before you mounted the module and the DCDC inductor was missing or not correct, this will happen as the DCDC can't supply the device if it's not working. If the DCDC wasn't enabled, it should work with the internal LDO.