Narrow THROTTLE_MIN/MAX range makes tilt-rotor transition impossible due to down scaling of Q_TILT_MAX

ArduPlane Plane 4.5
, ,
1

Hello,

I have a large (15kg) fixed wing tilt-motor VTOL aircraft in a wing/canard configuration where all 4 motors tilt. I have done initial flight testing transitioning to FBWA to characterize performance for TECS tuning. I have established a good profile and these transitions have gone fairly well. However, I am running into a major problem with the setting of the tilt angle logic during transition to FBWA/B after configuring THROTTLE_MIN/MAX.

After performing some initial maneuvers in FBWA for TECS tuning with a wide THROTTLE_MIN/MAX range, I have determined that the aircraft requires the following values for safe flight:

THROTTLE_MIN = 37
THROTTLE_MAX = 70

My understanding of the transition logic suggests the following should happen:

Switch to FBWA/B:

  • If THROTTLE > 50%, tilt all motors to Q_TILT_MAX
  • If THROTTLE < 50%, scale tilt linearly to an intermediate angle

However, this is what appears to be actually happening

Switch to FBWA/B:

  • If THROTTLE > 50 percentage points above THROTTLE_MIN, tilt all motors to Q_TILT_MAX
  • If THROTTLE < 50 percentage points above THROTTLE_MIN, scale tilt linearly to an intermediate angle

My problem is that with a narrow THROTTLE_MIN/MAX range, it creates a situation where it is impossible to tilt the motors all the way to Q_TILT_MAX. As I only have 33 percentage points in my throttle range, it will only tilt the motors to 66% of the way to Q_TILT_MAX. In this situation, transition will fail or take an excessively long time to reach AIRSPEED_MIN (even with Q_TILT_MAX=80).

I don’t have many workarounds for this problem, besides increasing THROTTLE_MAX to at or above 87%, and evaluating climb rates, or setting Q_TILT_MAX above 80. However, with a high power to weight ratio VTOL aircraft, I am concerned about overspeed in high throttle situations.

The logic appears to be tied to the “set tilt” calculation in the tiltrotor code: ardupilot/ArduPlane/tiltrotor.cpp at ab68066e9c7a671b6a032289d5eda382b3d3245b · ArduPilot/ardupilot · GitHub

And also described in this forum post: FireFLY6 Failed Transitions

I can provide some log files and more information about the aircraft if that would be beneficial. Any help is appreciated.

Thank you,
Neil

2

This is on my list of things to fix. I have some other stuff in the queue first, so if someone else out there wants to take this please do.

I’ll let you know when I make a patch that you can test out.

3

Hi Robert, I have a tilt made in Realflight if you need one for testing.

4

I was gonna do it on the Convergence RF SITL model, but can’t hurt to test it out on two different models. Shoot me a link

5

Will do. I will get it linked asap and let you know.

6

Sent to your Discord PM if needed.

7

Hey Bob,

A teammate of mine has pushed a recent PR with a revised calculation where it scales the tilt angle based on the difference (thr_max - thr_min), rather than by simply 50 points. Seems to work in SITL: