Using the SDK11.0.0 development product, it was found that the 52832 restarted randomly, with a reboot time of 7 days or longer.

[Lion]

Hi everyone,

Using the SDK11.0.0 development product, it was found that the 52832 restarted randomly, with a reboot time of 7 days or longer.

App code condition:

1. SDK11.0.0

2. Enable watchdog,5 second timeout

3. Apptimer 500ms and 1s timer reset watchdog registers

4. undef DEBUG

Reset register value is 0x0a

1. Reset from watchdog detected

2. Reset from CPU lock-up detected

Please help me analyze the reasons for the 52 restart, can you locate the issue of the code?

In order to solve 52 random restart issue.

Thanks.

[Einar Thorsrud]

Hi,

As you write, RESETREAS value 0x0A indicate a watchdog reset (DOG field set) and CPU lockup (LOCKUP field set). Do you clear the RESETREAS register after each reset, so that you know that both bits were set right before the previous reset? Or could one of the bits be older? The reason I ask is that the RESETREAS register is special in the way that it is only reset to the reset value (all zeros) after a power-on reset or brownout reset.

The CPU lockup does not trigger a reset if the device is in debug interface mode. So, if CPU lockup occurs in this case, the system will stay in the fault handler until the watchdog reset occurs. Do you have a debugger connected to the device?

 

[Lion]

Hi Einor.

Thank you for your reply.

I reset the register every time.

err_code = sd_power_reset_reason_get(&reset_reasons);
APP_ERROR_CHECK(err_code);
err_code = sd_power_reset_reason_clr(0x000F000F);
APP_ERROR_CHECK(err_code);

I understand what you mean by debugging to find out where the CPU lockup, but CPU lockup time is too long (7 days or more). Therefore, it cannot be in debug mode consistently.

So, is there any other way?

Thanks.

[Einar Thorsrud]

I see. Can you confirm that there is no debugger connected to the nRF (active or not), as that would help rule out the possibility of it somehow accidentally entered debug mode?