Need help to analyze a crash reason


it is a small model i have that still seems to be a source of non-stop issues.
I replaced all ESCs on it, new motors, got all ESC signal wires replaced with coax shielded ones - and it was flying fine. It sat on a shelf for a month or so, and i wanted to run autotune on it today, so, in alt hold mode it flipped and crashed less than a min in the air, then after next takeoff, it went quite a bit into the process, finished roll part, went into pitch part, i moved it slowly back to the spot, soind was fine, no visual issues, and upon moving closer to the spot it suddenly flipped in the air same way as first time and crashed.

Here are log, if anybody could assist to look at them to see what ESC did that or if it was something else it would be very helpful. I am not sure anymore what that is. it had those motor twitches on it from the beginning - with different motors and different ESCs, so i don`t know what to think - is it a damaged DMA that interrupts DSHOT stream? is it even possible?
Or if it is some loose wire again, but, which ESC then? I looked at it all and see nothing wrong at all, all looks fine. Yet, it crashed twice in a row, so it has to be some reason…

Only part left not replaced on it now is the FC chip itself.

and i see this crazy picture - current spike was at the time of the crash. where is ALL ESC telem is? how come all current from ESCs is gone while primary FC current sensor registered this spike?

i took it to dev thread - all issues got fixed after adding NODMA to all UARTs in the hwdef and rebuilding the code.
in my opinion it looked like a DMA failure to deliver DSHOT stream to ESCs - be aware.

Not related to your issues but why use old school althold auto tune instead of improved poshold AutoTune

because it is how i prefer it done.
also due to the fact that i autotune on the big field and do not care for craft to burn extra battery to constantly re-position itself. in alt hold it finishes faster.

1 Like