Thanks for the tips @sanderux. Turned out that setting the transition speed to 13 m/s seemed to solve the problem of climb during the transition. Thanks!
However, unfortunately, I also got a repeat of the drop of 10m after the transition.
Log is at http://review.px4.io/plot_app?log=72490c96-96c8-4869-92bc-25c40dc994e8 – if I read it right, after the transition, even though the plane is flying quite level, it decides to cut the pusher throttle to 0 briefly while applying some down elevator. The plane plunges 10m then recovers by turning the pusher throttle back on and applying up elevator.
I guess it’s a combination of a couple configuration problems?
-
Even though it is flying level in the transition, it seems to think it’s pitched up a couple degrees. I guess I have to be more careful about level horizon during calibration.
-
Although the control action after the transition (setting the target pitch to -6 degrees, cutting the throttle to 0, and apply down elevator) seems pretty extreme to me, I guess it’s just an issue of being too sensitive to the control inputs. I guess I can make it less sensitive by adjusting the center of mass?
So I will try again with a more accurate level horizon calibration and center of mass adjustment but would appreciate any other advice on whether that 0 throttle is the expected behavior given the sensor inputs.