I have the rate gains tuned, tuned in stabilize mode (default stabilize gains). Started to move on to AltHold but when I switched into AltHold it aggressively jerked up and down several inches without a noticeable overall change in altitude. I started backing down the throttle gains until I hit the minimums and basically was unable to stop the oscillation/jerking although it seemed lower the frequency of the oscillations when the gains were bottomed out. I'm at a total loss here, the only thing left to try is increasing the gains from default which logically thinking makes no sense.
Please send some advise.
Setup:
Trex 500 TT DFC
Pixhawk mounted 90* roll
Arducopter 3.2.1
head speed 2350rpm
symmetrical blades
6s 4000mah battery
Replies
Hi,
maybe this helps:
RSC = headspeed
'Throttle' = collective pitch
'Extended Tuning' Throttle Accel P = 'Full Parameter List/Tree' THR_ACCEL_P or ACCEL_Z_P
After setting RATE tuning parameters you should switch to AltHold and do the ACCEL_Z_xx tuning, on all my helis I had to lower the ACCEL_Z_P from 0.5 to 0.3 to 0.4 (collective pitch +/- 12°). Too high P and D makes fast oscillations, I makes slow oscillations.
Check for smooth operation of the servos, i.e. do they follow collective pitch input continuously or do the jump. If they follow in steps lower the RATE_xx_D.
If this doesn't help check for vertical play. I have Pixhawk mounted 90° / 270° without problems using the provided foam.
Be VERY careful using MP 'Extended Tuning' as it wants to set limits which are for copters and not for helis - I prefer 'Full Parameter List/Tree'.
http://www.nextfpv.com.au/products/apm-pixhawk-anti-vibration-mount
Hi,
I think the APM-Pixhawk-anti-vibration-mount is good when mounted horizontal on a gasser heli, according to the vibration logs I never needed one with an electric motor. In case you mounted the Pixhawk using the provided 3DR-foam on the anti-vibration-mount you could try to fix the top plate of the anti-vibration-mount to the bottom plate of the anti-vibration-mount (=disable the anti-vibration-mount) and compare the results in AltHold and vibration logs.
I tested with EKF enable and got the same results. I then as a last ditch effort reduced the two throttle P gain settings to almost zero, was warned the values were below the minimum limit. The oscilations went away but the heli sank in altitude. I then incremented them both up by 0.5 and althold worked as expected. Still need to continue tuning but atleast it is now working without jerking up and down. I want to give a BIG thanks for all the help. The documentation is in serious need of updating, HAHA. Also the views in MP are very misleading for a traditional heli.
Thanks again!
Thanks for the info, I'll give that a go and let you know how it turns out.
Well, I went out and ran some test flights. I could not find an Accel_Z_P but did find Pilot_Accel_Z and WayNav_Accel_Z. Setting limits were 50-300 or close there about. I ended up setting those as well as the throttle accels all to minimum values by the time I was done and still did not get rid of the oscillation in the Z when switching from stabilize to althold. The behavior was that the heli would ascend and descend as well as hold altitude when commanded but the whole time it would oscillate almost violently. I did notice that in stabilize the collective had to be one mark beyond mid stick to maintain altitude in hover. At mid stick it would fall at what I am estimating to be 1 MPS. Could the mid stick setting cause this oscillation? Should I increase the PWM setting at mid to have it hold altitude?
Thanks for the previous response and further help will be GREATLY appreciated.
Trey
Hi,
On the 'Extended Tuning' screen in the box marked 'Throttle Accel' there is 'P' (default 0.5). This is the same as 'THR_ACCEL_P' (3.2.1 version) in 'Full Parameter List/Tree'. In 3.3 this is renamed to 'ACCEL_Z_P' that's why you cannot find it in 3.2.1 (therefore I wrote THR_ACCEL_P or ACCEL_Z_P).
On my helis I did the rate tuning and experienced the vertical oscillations using the default setting. Changing THR_ACCEL_P from 0.5 to 0.3 - 0.4 was enough to cure the Z-oscillations in AltHold. I mounted the Pixhawk(s) with the provided foam on a sheet of PVC or aluminum glued/bolted to the frame. I think your mounting location is good, but I am not sure about your vibration damping mount. Please check/post your vibration logs. Too soft is as bad as too stiff, as the Pixhawk could 'overshoot'.
Trey