Strange oscillation ONLY in Auto !Track dependent!

Hi I woud like to have someone look in to this.
My plane (wing 2 elevons) is very small. All packed in such a small plane. It is difficult for the compass to get proper offsets but the plane after autotune flies very well in cruise, WBWA, loiter, RTL, stab, no problem there.
When in auto it is a different case. After a while it starts oscillate so badly. Why should this happen if in other modes it is ok? What more it starts only when inbound to the north halve of the headings. If it flies south halves of the headings it is goone.
The log is to be found here:

I’m not a tuning expert but I’ll share what I’ve learned. I’m fighting with an ARwing that has a similar problem. I found it has something to do with the I and D terms of the PID tuning. And right now, I’m leaning more to D term being the issue. The wiki does state that having the D term too high will cause rapid servo movement. Autotune for me produced a tune that created very similar results to what you’re seeing. Violent oscillations, but only in certain conditions.

After autotune RLL2SRV_D = 0.11701048910617828. The wiki mentions not going above 0.1. This may be a good place to start, and turn that down.

Otherwise, except for the oscillations, I’d say your tune looks really good. Pitch and roll seem to be tracking well with desired pitch and roll.

Hi JeyPi,

I found three different compasses in the log file. All of them are calibrated very differently (not or bad?) to each other. All three compasses are enabled. I would disable at least two of them.

Hope this helps.