nRF52832 PAN16 is still present

[florentb]

We use nRF52832-QFAA-B0 and we found that PAN16 issue is still present, at least in some corner cases.

It happens when we put the chip in low-power mode with WFI ; at wakeup, we find random bitflips in the lower part of the RAM.

This usually result in hardfault later.

This bug happens only if all peripherals are disabled and the chip is not under debug mode. For example if the UARTE is left enabled, the bug will not happen.

Our system_nrf52.c is up-to-date. But if we remove the test : if(errata_16()) and force the 0x4007C074 register change, the issue does not happen any more.

[haakonsh]

I was erroneous and I have deleted my previous response.

This has been redefined as PAN[108] RAM: RAM content cannot be trusted upon waking up from System ON Idle or System OFF mode

[florentb]

Yes, we confirmed that the errata for PAN108 works, the issue is not reproduced.

[haakonsh]

Glad to hear! :)