Is there systematic approach to tune a tricopter?

Log Analysis -- deprecated Copter Logs -- deprecated
1

I have a fresh build of quanum trifecta with micro apm.

Initially the tail servo was moving left and right for no reason and causing the whole thing to shake violently as soon as it takes off the ground.

After some experimenting I found that setting the YAW P value to anything in the recommended range (I believe that was 0.15-0.25) completely removes the tail shake just like that.

BUT now it just rotates around it’s center constantly.

I experimented with few D values and all P values from 0.15 to 0.25 (0.01 increments) that didn’t help and it feels like shooting in the dark hoping I’ll hit something.

I suppose there is better more systematic approach to find the right combination? Can’t find any guides on tuning tricopters though? I don’t believe they are THAT exotic?

p.s.

how important is the compas for that in stablize mode?
I attempted to cut the SJ10 thing that is supposed to make it use external one but it still uses the internal one with the benefit that I cant’ calibrate it now. Says not enough data points.

But in APM planner it moves correctly when I move the trifecta.

So at this point I’m not sure if I have 2 issues

2

Try
Yaw stabilize: 1.2
Yaw p rate: 0.280
Yaw i rate: 0.038
Yaw d rate: 0.010

After much experimentation these worked for me, and have worked for a few other tricopters.

Compass is vital any GPS mode like Loiter, RTL, PosHold, Auto, etc. If you have no GPS or will never use any of those mode then compass isn’t very important. (BUT then failsafe can’t be used either). The FC uses inertial navigation (gyro and acceleromter) for heading hold and compass for direction finding.

3

Thanks
I started with those and experimented for half hour.
I decreased the tail wobble a lot
now i can raise it a bit but it instantly goes to the back and rotates CW so much that still can’t be flown
and eventually the tail wobble issue appears again

i eventually got to P=0.065, I=0.010, D=0.00
P stab=1.2

better like not the violent tail wobble as before but still unflyable
is there a systematic approach to get to the perfect ones?

or it might be from something else?