Hello everyone,
I’m having a problem with a VTOL (2 tilt motors, 1 tail motor). The aircraft flies perfectly in hover and fixed-wing modes. However, during the transition, the motor control oscillates, and I have no idea what’s causing it:
In the screenshot, the current line is shown in orange to represent the motor control.
Once the switch to fixed-wing mode is complete, the oscillation in motor power disappears and the motors are controlled correctly.
I would be grateful if someone could give me a tip on which parameters influence this behavior.
Attached is a link to the log for the experts.
The log file shows only hover. But it has very strong oscillations in roll and yaw. It seems that tilt servoes are working very hard and oscillate. That is probably due to incorrectly tuned yaw. The oscillations in yaw also cause the roll oscillations which are corrected by the motors, but most probably you have to sort the yaw issue only. During transition, yaw instability generally increases and causes more roll instability, because the motors tilt forward…
So instability will be much more apparent during transition, but as is, it is also present in hover.
So, not sure if you have tuned the vtol using qwick tune. If not, you should tune or retune the yaw. Note the values yaw has before and after tuning. Most probably you have to reduce the yaw parameters (PIDs) at least by half. No matter swhat qwick tune will suggest. It is not dangerous (not like roll/pitch), it could make the copter not very responsive in yaw, but not crash it.
Tricopters are notorious for tuning problems and usually there is no solution, i.e. either you get mushy yaw, or you get oscillations, botgh of which increase during the transition. So the tuning always will be a compromise. Personally, I feel yaw is much less important than stable hover, so I would sacrifice yaw.
Thank you very much for your analysis, your suggestions, and the effort you put into responding to me! Now I’m thinking about how I can improve the flight performance.
I configured the aircraft using Qwick Tune. I also have backups of the data from before and after Qwick Tune. I’ll take another look at those.
I notice that only DCM.ErrYaw was recorded throughout the entire flight (takeoff, flight, and landing); ANG.DesYaw and ANG.Yaw drop off after the transition at 2:28.
ATT.Yaw is recorded throughout the entire flight, while ATT.DesYaw also drops off after the transition.
Based on which values did you conclude that the aircraft is yaw oscillating?
Furthermore, I still can’t find a clear answer to the question of why the left engine (RCOU.C3) and the right engine (RCOU.C4) are “throttling up” more or less evenly and then throttling back down during the transition or in the phase where the aircraft is accelerating to reach minimum airspeed. If they had to compensate for yaw, they would have to “throttle up” in opposite directions to counteract the rotational movement. But my initial question was about this “throttling up.”
Do you have any idea?
You have RCOU for tilt motors oscillating constantly. That indicates that the yaw is most probably incorrectly tuned. It is very likely that after you tune yaw, reducing PIDs, not only yaw oscillations will improve, but also the roll. Your log shows only hover, not transition, bythe way.
Also, if you look at ATT, desired Yaw and Yaw, Yaw also shows oscillations.
If you mean throttling up continuosly, that happens because initially, the motors tilt forward, so they have to divide power between lifting and forward thrust, increasing overall power. After a short while, when the plane gains speed, the wing begins to generate lift and the hover power decreases.
1 Like
You’re absolutely right. I have to apologize - I linked the wrong file, which is why we were talking past each other. I was already wondering about that. The file I uploaded and that you analyzed is much older and was created before Qwick-Tune. Here is the correct file, which is also the one the image above refers to.
Interestingly, you can clearly see how Qwick-Tune improves PID control.
My original question and the screenshot above refer to this file.
