[WatchDog Error]"CrashDump data detected" on ArduRover on a custom board

I’m using a custom carrier board, not from CUAV.

Setup

  • Board: custom carrier board (not CUAV), using original CUAV X7+ core

  • Firmware: ArduRover 4.6.3 (3fc7011a), ChibiOS 88b84600

  • Vehicle: ground rover

  • Power: USB only, no main battery

Symptom I don’t remember well when this symptom first started. The board had worked fine for over a week;I could arm it and drive the rover. At some point it stopped arming, and now every boot shows PreArm: CrashDump data detected and writes a crash_dump.bin to the SD card. If I delete crash_dump.bin, it comes back on the next boot.

Attached

Additional (you don’t need to read this) I can’t trace exactly when the problem started. It might have started after a log download in Mission Planner suddenly errored out and got cut off ; it seemed to stop working after that. Or it might have started after I changed all PWM output pins except one to Throttle. It might also be caused by insufficient USB power. For now I’m holding off on reflashing the firmware, so I can preserve the current state for an accurate root-cause analysis.

This log contains the watchdog(WDG) message :

Result of the decode_watchdog.py

Thanks for the report. There is a known issue on 4.6 with a watchdog while downloading logs.

There are some more details here:

1 Like

Before I changed the SD card I had a lot of trouble, sometimes the board wouldn’t even boot. After swapping it, everything works again and the errors are gone, so I assumed the SD card was the cause. I also got a GetLog() timeout error, like in the link. Could a failing SD card cause all of this, or could it be something else?

I also have some unexpected (WDT) resets on my ESP32 testplatform not always but very often at the moment I disarm (close logs). So I also assume the SD card handling.
Therefore I disabled logging (set LOG_BACKEND_TYPE to 0) and on tests today the problems are gone.
So to check you also can try to disable logging

1 Like