VTOL Skywalker 1900 - errors and strange behaviour following crash repair


Hope someone can offer advice.
Recently repaired my Tiltrotor VTOL (Pixhawk 2.1) after a bad crash where I switched from Qstabilise to manual accidentally and I was too close to the ground.
Anyway it used to fly fine apart from a slight yaw oscillation in Qstab which I was ok with.
I’ve tidied the internal wiring and repaired but I’m getting some strange behaviour.
I successfully took off in QSTAB the transitioned to FBWA but I was fighting to keep the plane in the air. The throttle seemed restricted in some way and I couldn’t maintain altitude. Quickly switched back to QSTAB and landed.
Then I decided to hover and test QRTL - it rolled to one side and looked like it was going into a toilet bowl situation but i didn’t wait to find out and switched back to QSTAB in time to recover.
I’ve got the Dataflash log and telemetry log and there appears to be a number of errors namely “EKF yaw reset”, “error pos horiz variance” and “error compass variance”.
I would like to interrogate these logs to diagnose the issues but I have no idea where to start.
Could anyone give me some pointers about what to look for in the logs?

Hi Michael, why not share your dataflash log file here because the wonderful community members could possibly help you?


I was going to upload my files but thought I should have a go at figuring out things myself. However I suppose I’ve tried that and don’t really know where to start.
New things I’ve noticed or read are that some have said it could be compass issues and in the auto analysis the compass failed- in re-routing the wiring and tidying it up is it possible I’ve routed the GPS wiring too close to something else which is causing noise in the GPS signal?
I’ll upload the telemetry log and the DF log. Note you should ignore the telemetry until 58% because nothing much happens. The plane was on the ground whilst I figured out another problem. At 78.7% in the telemetry log I’ve noticed something strange - when switching to QRTL the log says that it switches to RTL first. This was not commanded by me so not sure why this would happen?
As for the DF log I’ve looked at it but not really sure where to start. There appear to be a few issues where for example the demand doesn’t match the output with some of the controls like throttle or roll. However this might be normal behaviour in a fly by wire mode?
Ok, can’t upload so I’ll paste link:
Telemetry log:

DF log: