Plane: fixed ramp of motors on back transition in tilt quadplanes

[tridge]

this removes the special case for SHUT_DOWN state of the motors in tilt quadplanes. The old code seemed to assume that the output_motor_mask() doesn't work in SHUT_DOWN state, but in fact it does, so we can get proper spoolup behavious on back transition just by removing the special case

this PR builds on #26248 to give us a quadplane-tilt frame for testing

[IamPete1]

I suspect we could remove the tilt-rotor motors_active concept all together, the only other user is here:

ardupilot/ArduPlane/quadplane.cpp

Line 2090 in 1d08662

if (motors->get_throttle() > 0.01f || tiltrotor.motors_active()) {

I would like to do the same for tailsitter, then it remove the special case completely. But that can wait for a new PR.

get_desired_spool_state is used quite a lot, so that might give a different result now, but since SLT's don't have a problem I don't think it will be a issue.

[tridge]

I suspect we could remove the tilt-rotor motors_active concept all together, the only other user is here:

I think we need to keep that user though, or throttle suppression could break for tilt rotors

[IamPete1]

I had a bit more of a in-depth look. I don't think this will work if slew limits are set (Q_M_SLEW_UP_TIME and Q_M_SLEW_DN_TIME).

The problem is that the output mask method slews the outputs here:

ardupilot/libraries/AP_Motors/AP_MotorsMulticopter.cpp

Line 740 in d5e4d19

set_actuator_with_slew(_actuator[i], thrust + diff_thrust);

The output to motors method zeros that _actuator value if in SpoolState::SHUT_DOWN:

ardupilot/libraries/AP_Motors/AP_MotorsMatrix.cpp

Line 152 in d5e4d19

_actuator[i] = 0.0f;

So we would only ever get a throttle of 1 loops worth of spool up.

I'm fairly sure we actually have the same issue now. Except it will be limited to one loops change from the output values from the mixer that are set in the SpoolState::THROTTLE_UNLIMITED.

The esay fix is just to say that we don't care about spool limits and remove them in the output mask. However, that would result in a very big jump in individual outputs as they swap. It would also be nice for the spool up to come up to meet the slewed actuators rather than them swapping from one method to the other instantly.

Alternately we could add a "has mask override" method that would skip setting the actuator in the shut down case. I suspect that will allow us to get the best result, but it does mean were leaking QP logic into motors a bit. But it might be a nice path have the slew limits applied in motor test for example.

[tridge]

Alternately we could add a "has mask override" method that would skip setting the actuator in the shut down case. I suspect that will allow us to get the best result, but it does mean were leaking QP logic into motors a bit. But it might be a nice path have the slew limits applied in motor test for example.

I've implemented that in the PR

[rmackay9]

This looks OK to me, at least it's better than duplicating the ramp up code in the quadplane code.

[tridge]

I flew this on a GriffinPro today, no issues

[IamPete1]

Also "copter" tailsitters.

[IamPete1]

Might want to do the same in ground idle too?

[IamPete1]

There is a off by one bug in tricopter than means the slewed actuators don't match between mask and normal output.

ardupilot/libraries/AP_Motors/AP_MotorsTri.cpp

Lines 100 to 102 in 728d9a9

	set_actuator_with_slew(_actuator[1], actuator_spin_up_to_ground_idle());
	set_actuator_with_slew(_actuator[2], actuator_spin_up_to_ground_idle());
	set_actuator_with_slew(_actuator[4], actuator_spin_up_to_ground_idle());

[IamPete1]

This is a fix for the tricopter motor indexing thing. #26313

[IamPete1]

The is a small conflict with the tricopter indexing fixup.

fixed