S132 NRF_FAULT_ID_SD_ASSERT during GAP scan

[mkuyper]

I am using S132 v7.0.1 on an NRF52832 to do a basic scan for BLE advertisements for 30 seconds (10 seconds on each primary advertising channel):

#define SCANTIME        10 // sec
#define sec2slots(s)    ((s) * 1600)
#define sec2tenms(s)    ((s) *  100)

static const ble_gap_scan_params_t scanparams = {
    .interval = sec2slots(SCANTIME),
    .window   = sec2slots(SCANTIME),
    .timeout  = sec2tenms(3 * SCANTIME),
};

This works fine and as expected on my current development board.

I am now in the middle of bringing up a new revision of the production board, and I am now on this new board getting an SD fault handler call during the GAP scan operation:

sd_fault_handler id=1, pc=0x00014CBC, info=0

I realized that this error happens exactly 10 seconds after the start of the scan operation, i.e. right when the scan interval/window expires and the SD should switch to the next channel. If I shorten the scan interval/window to, e.g., 5 seconds, the problem goes away. Coupled with the fact that this only occurs on the new hardware, this leads me to believe it could be a timing-related issue, such as the 32 kHz oscillator possibly being out-of-spec causing a timing check to throw an assertion.

Any information as to what causes the assertion at the above-mentioned PC would be appreciated in finding the root cause!

Thanks,
Mike

[tesc]

Hi,

I can confirm this is a timing check throwing the assertion. The SoftDevice thinks that an internal event overstayed its planned duration. Clock source issue may be the reason, or if you are e.g. disabling interrupts for too long, overstaying timeslots, etc.

Regards,
Terje