Page 1 of 1

Tricopter Yaw bouncing back

Posted: Sat Oct 11, 2014 1:21 pm
by Arakon
I'm running a mini tricopter. I finally got it to fly decently, but I have some issues with yaw still.
The front props are counter rotating. When I turn left, it turns slower than right.. that's normal.
However, if I turn only slightly left (i.e. just adjusting the "nose" a little to the left, or turning very slowly) and release the stick, the copter turns back to the right a bit. Worst case, if I turn it just very slightly left, it may turn right further than even the initial position from which I turned.
Which of the PID values is responsible for this?
I actually remember seeing a tutorial that described this bounce back effect specifically, but I haven't been able to find it anymore.
Previously it actually turned a good quarter turn to the right just lifting off, I got rid of that by increasing I gain quite a bit.
Current PID yaw values are 3.6/0.058/5 .

Re: Tricopter Yaw bouncing back

Posted: Sun Nov 16, 2014 10:35 pm
by pelotron
I'm having this same issue on a mid sized tricopter. The copter turns right much more easily than it turns left, and tends to slowly turn right on its own during hover.

Re: Tricopter Yaw bouncing back

Posted: Mon Nov 17, 2014 7:03 am
by nebbian
Couple of things:
1) Note the position of your tail rotor in a stable hover. Now land, and set up your mechanism so that when the servo is centred, the tail rotor is in this position. This will stop the liftoff twist and slow yaw issue.
2) Add a whole heap more D to your tune. I've got mine up at 50 or so. Then bump up P until it's sweet.
3) A faster tail servo makes a lot of difference.

Re: Tricopter Yaw bouncing back

Posted: Mon Nov 24, 2014 12:18 am
by pelotron
Thanks for the tips. I will try these this week.

Re: Tricopter Yaw bouncing back

Posted: Sun Nov 30, 2014 6:28 am
by pelotron
The yaw performance of my copter is greatly improved now. I think the biggest key was to adjust the angle at which the rear motor sits when the servo is in the neutral position. Before I did that I had weird behavior where if I gave a little bit of left rudder the craft would turn right, but if I gave close to full left then it would start turning left. I think this might be because at that time the motor was straight up at 90 degrees when the servo was in its neutral position, and in hover MultiWii would tilt it to about 70 degrees. But when I gave radio input, the servo might have reset and then processed the rudder input as an offset from its neutral position rather than from the angle set by MultiWii, so with a little bit of left rudder it might shift to 80 degrees which would still produce a right yaw motion.

If that is what is happening then that means that rudder input ignores or overrides the influence from the gyro, which seems incorrect to me. I'm going to dig into the source to see if that is really the case; I think combining the inputs of rudder and gyro would be a better approach than letting one override the other. But if it's been this way for this long there are probably difficulties involved that I'm not aware of yet.