Yaw spin during VTOL transition

ArduPlane Plane 4.6
1

I have an QuadPlane running ArduPlane 4.6.3 here (4 motors, with the front two tiltable). During the transition from QLOITER to FBWA or CRUISE, it starts spinning fast around the yaw axis. Fortunately, each time I was able to land it safely by switching back to QLOITER.
The log shows one motor diagonal ramping up while the other ramps down. However, I don’t think it’s a desync or a motor spinning out of control—I likely would have heard that. A mechanical failure is also unlikely, given that it is absolutely stable and controllable in QLOITER.
I am also certain that I have assigned the motors correctly, as the motor test in Mission Planner activates the right motors.
I have already found other posts describing very similar problems:
Massive yaw spin during transition of VTOL
QuadPlane（ 2 front motor tilt）transition yaw swing
Quadplane Transition Problems
At least one user reports that switching to vectored tilt solved the issue. But I only want to do that if nothing else works.

Here are my parameters:

VTOL.param (27.7 KB)

Bin log here:

Thanks for any help in advance!

2

Change motor rotation direction to H mixing. (also parameter)

In X mixing, the two front motors act as if they form a strong false CANT angle during the transition process.

Q_FRAME_TYPE = 3 (H)

ㅁ
ㅁ327×359 66.8 KB

1 Like
3

Any explanation, why do you think so?
What I see in the code (AP_MotorsMatrix.cpp in setup_quad_matrix()) is this comment:

case MOTOR_FRAME_TYPE_H: {
    // H frame set-up - same as X but motors spin in opposite directions

I can try it anyways, but I would prefer to understand what I’m doing. :slight_smile:

4

In Quad X Mixing

if you want to yaw the plane to a Left, the output of motors (3) and (4) will go up.

1
11920×1398 521 KB

The motor is slightly Tilt during Transition, but it is still VTOL control.

The plane wants to yaw to the left, so it raises the motor(4) output , but Tilt causes thrust deflection, so start yawing to the right.

2
21920×1263 329 KB

The plane turns up the motor(4) output for the left yaw, and the plane does the right yawing faster.

Therefore, it is resolved with QUAD H Mixing.

I’m not sure if the translator is speaking properly.:melting_face:

2 Likes
5

It was tried and proven. H worked. You just need to take some time and think about what he explained with pictures. You will realize suddenly.

6

Ok, thanks I will try it out later, as soon as the wetter will get better here. :slight_smile:

7

@socke7 Hi Andreas. A comment, a question:

Is there a specific reason you chose Q_TILT_TYPE 0 (continuous tiltrotor) even though you obviously have two tilt servos? Why not Q_TILT_TYPE 2 (Vectored Yaw)?

During hover, motors 3 and 4 require significantly less power than motors 1 and 2. Is it possible that the motors aren’t aligned properly, causing the propellers to rotate at an angle during hover and thus generating torque around the vertical axis?

Screenshot 2026-07-06 180738
Screenshot 2026-07-06 1807382776×1350 115 KB

I would definitely look into this, regardless of whether it’s an H or X configuration.

Rolf

1 Like
8

You nailed it! I just had an amazing first flight with transition.
This point really should have been mentioned in the documentation – or did I overlook it?

However, thank you very much - you saved my day! :fireworks:

9

Hi Rolf,
yes, there is a reason I chose Q_TILT_TYPE 0 (continuous tiltrotor). Before I build this specific aircraft, I had a number of tricopter VTOL frames. These had continously all kind of trouble with the tilt servos. So I decided to minimize the servo movements and hopefully all the trouble here with the current frame. Do you think this is disadvantageous?

And you are right regarding motor 3 and 4. This is something I need to have an eye on it. But currently I have no idea what is going on. At least front and back motor arms measure the same length from CG. Arms are solid CFK tube and shouldn’t have significant flex.