Plane: quadplane tilt rate issues

[tridge]

There are some issues with the quadplane tilt down rate for tilt-enabled quadplanes.

• the quadplane.assisted_flight flag oscillates during the TRANSITION_TIMER phase of the fwd transition
• if Q_TRANSITION_MS is smaller than the time it takes to go from Q_TILT_MAX to full tilt with the configured Q_TILT_RATE parameters then the last part will happen at 90 deg/s
• when we complete the transition there can be a sudden throttle change on the tilted motors, much faster than the THR_SLEWRATE should allow

assisted_flight oscillation

The assisted flight flag oscillates as SLT_Transition::update() will set it to false once airspeed is reached due to should_assist() returning false. It then gets set back to true in the TRANSITION_TIMER further down in the same function. The impact is not significant, but it is confusing.

too fast slew

If Q_TRANSITION_MS is smaller than the time it takes to go from Q_TILT_MAX to full tilt then the last part of the tilt will happen at 90 deg/s due to the fast_tilt logic and TRANSITION_TIMER no longer being true. Possibly we should stay in TRANSITION_TIMER until the full tilt is reached?

instant throttle change

when the transition has completed and the VTOL motors are turned off the tilted motors move from VTOL tilt control to fwd throttle control. When this happens the throttle changes instantly and the sudden throttle change can disturb the aircraft pitch as the pitch PID does not have time to compensate.

[timtuxworth]

Is this what makes transitioning on a tilt rotor so hmmm - frustrating? It feels like it's trying hard to transition but won't quite get there? I haven't flown my VTBird for a while this is the last flight but I always had this feeling.
BTW it was always with manually transitioning to FBWA - the auto mode transition always seems to be much smoother and quicker and I always wished I could duplicate it. Here's the log:
https://www.dropbox.com/scl/fi/tv8sl0va1xhdti691zrv2/log_154_2024-8-31-18-24-52.bin?rlkey=b7ujssije9iulhdgho9ky603m&dl=0

[priseborough]

I think there is also a potential issue with the way that TECS can dip the throttle demand at the time that the tilting rotor throttle control hand over to TECS from the motors library. For vehicles where the propwash over the front wing contributes significantly to lift, we should be limiting throttle to no less than the trim value initially.

[priseborough]

I've created a branch which contains a modification to the forward transition completion criteria to make it wait until rotor tilt forward has completed.

https://github.com/priseborough/ardupilot/tree/pr-quadplane-tilt-rate

I'll have a look at what we can do to prevent a step in throttle for those setups with a front wing that benefits from prop-wash energisation at lower speeds.

[priseborough]

See #29696 for fix to throttle glitch.