What caused my quadcopter flyaway?

ArduCopter
, ,
1

Hi, everyone! This has happened once before (with a different drone), but it happened again today, and clearly my previous attempts to fix it did not work.

Essentially, I was flying an auto mission with a target velocity of 5 m/s, when it just shot upwards, and nothing I did could stop it. Eventually it stopped, and although I briefly was in control again, it ended up free falling into the ground from 115 meters.

I can’t find anything in the logs that could cause that. I’ve checked for GPS glitches (they only occur after it begins its rapid ascent), abnormal voltage/current readings (I was wondering if the Pixhawk lost power?), and everything else I can think of, but nothing seems to be at fault.

The only correlation I could find was that the pitch was deeper than at any other point in the flight, right when the drone took off, but it was only -29 degrees. With my other two (nearly identical drones), I’ve reached angles of up to -40 degrees without issues. On the previous crash, the drone dipped to -41 degrees right when it took off upwards. On the other hand, it reached +47 degrees without problems.

Does anyone have any ideas as to why this happened? Here’s a link to the log:

Zipped logfile

Thanks in advance!

2

The usual cause of vertical flyaways is vibration, and yours is no exception.

40%20am
Screen Shot 2019-08-14 at 7.11.40 am.png2278×888 458 KB

You can see here that once the clipping starts the altitude goes up.
I would then correlate this with direct IMU readings but you didn’t log any IMU values.

3

I would guess the copter fell apart in flight. I am sure its hard to check now perhaps you could find something lose.

4

So what’s the physical mechanism behind vibration causing a flyaway? Are power connections being shaken loose (temporarily)? Why does that cause it to shoot upwards of all directions?

Thanks for the quick responses!