Elevator input reversed after crash

Hello all,

After several succesfull long flights with my medium range delta airframe I had a rather strange crash last tuesday. Some facts:

  • The aircraft was not carrying a camera on its nose, putting the CG slightly further backwards then it was tuned for using autotune. But still the CG was within the envelope of the airframe.

  • I checked all control surface deflections and directions before take-off, I always do this every flight as it became habit.

  • The Plane has flown for approximately 5 hours already without any issues or strange behaviour.

  • Directly after the lauch of this flight i noticed some Pitch oscilations, this is maybe due to the CG and thus the servo trimming not being where it was with the camera on the nose? (my assumption)

What happened:
I threw the aircraft in stabilzed mode and it flew away fine, made a circle and put the plane into FWBA and then loiter mode which it also performed as expected for some circles. I then put on my FPV goggles and switched into Cruise mode, this is where I first saw the pitch oscilations.
Then switched into Stabilized to check if the oscilations would be the same or dampen out, (should have probably switched into acro at this point though).

At this point the aircraft suddenly nose dived very steeply and crashed into a tree before I could switch to Acro to save it. It took some effort to retrieve the craft and after checking it and plugging it in to retrieve the logfiles I found that my Elevator stick command was suddenly Reversed!?!
Looking a the logfiles the input still seems to be fine during the crash though, I firstly pulled up, then the dive started and then I pushed down since the craft was almost upsidedown and that would have been the best way out at that moment. it however already hit the tree at that point. But that behaviour is clearly visible in the recorded logfile.

To be clear here, after the flight i did not at any moment touch my transmitter so there is no chance I accidentally reversed it. I checked the Parameter list from before the flight with the one currently in the FC and there is nothing reversed as well? If I put the aircraft in stabilized the elevons still compensate in the right direction.

Does anybody have an explanation to this behaviour or had the same issue?
Iam really puzzled about how the pitch input can be reversed suddenly without any changed parameters.

Logfile form the flight as well as param lists before and after the flight can be found in my google drive: Ardupilot logfiles crash austria - Google Drive

Additionally: The dive may be the direct result of bad cg icw the tuning, but I don’t understand the pitch input reverse.

Same here close to similar situation. had crash and logs has no explanation only experienced hunch.( servo issue or FC, after crash still working ok). My case still under investigation by Ardu developers.

Looking forward to see the cause.

Hi Sunny, thanks for your reaction!

Did you see the same pitch oscilations as well at the start of the flight or before the crash?

Can you link to the forum discussion about your crash?

Another update:

I reversed the elevator output on my rc and flew again today to just check if everything else worked as expected. And everything did work!

One thing I found was that terrain following while flying in a hilly environment should not be turned on since it is following the slopes way to aggressively. I suspect this to be the cause of the first crash together with the more aft cg. Where the plane just wanted to climb while loosing airspeed causing it to stall eventually.

Today I flew with the camera attached so the cg was o.k. the strange pitch wobble was now gone as well.

But the one thing that still puzzles me is why and how can my elevator input suddenly be reversed from what it was before the crash?!? I really don’t understand.

Can this be caused for instance by the dropping voltage while the plane was stuck in the tree for a whole night? At some point the STM32 is probably shutting down and rebooting allot while the voltage is right on the edge of not working and working? Can this cause some anomalies in the firmware or some registers? Iam just guessing here.

That was very unexpected crash. Elevator full up.