Help on Analysis of log after crash

#1

Hi Everyone

I am now working on a log of one of our drone that crashed while doing a test mission waypoint
https://logs.px4.io/plot_app?log=50718125-e642-4d41-b0da-5aa69abb9e64
The drone start behaving badly (falling down) around 11:05, we switched to position mode just before 11:09 to try bringing it up, but not visible effects
I also see hat around that time, in the Power figure, Battery Current drops lower than previously in the flight
Furthermore, the difference between Pitch/Yam/Roll setpoint vs estimated grows bigger and bigger, and setpoint value is mostly 0
And, in Actuator Control 0, ar 11:05 Thrust goes max (it clips), and Pitch/Yaw/Roll are going down

My question is : I have doubts about what is a cause or a consequence. Is the current drop the cause of motors going slower and drone dropping ? Or is battery current going down because motors are controlled to slow down wrongly ? Is it indeterminable ?

I should add that this machine has propellers 15" and that PID had not been configured yet (default value for that flight)

Thanks in advance
Bastien

#2

First of all, it looks like your FC is under a lot of vibration especially on z axis. Raw acceleration plot shows about +/- 10 m/s^s and at some point it becomes even +/- 20 m/s^s which is bad for FC to estimate its postion and attitude.

Per the magnetometer, it’s data get distorted at around 11:10 into flying. Not sure whether this is due from user input or from external, but the drone needs to be properly calibrated its compass before flying.

These are all I can guess for now. Cheers!