Pixhawk Loiter issues

ArduCopter Copter 3.1
1

Please help.

Pixhawk with Arducopter 3.1.2 on Vulcan frame Hexa X
External Ublox compass on 10cm aluminum mast.

Stabilize & Alt hold works
PIDs done with auto-tune

Switch to loiter mode creates two problems:

  1. Toilet bowl
  2. Small oscilations - Loiter PIDS too high?

I have:
Done compass dance and made sure settings are correct - How do I check if the offsets are accurate though?

Vibration:
[attachment=2]Vibration.jpg[/attachment]

Mag Field:
[attachment=1]Mag Field.jpg[/attachment]

GPS:
[attachment=0]GPS.jpg[/attachment]

What do I do next?
Does Compassmot Work on Pixhawk yet or still the Nan values issue?
Thanks.

2

Please provide log files and next time, please post to the correct subforum.

3

I also experience same issue, though I’m not into toilet bowl thing.

4

Sorry Stefan.

Log attached. Starts with compass calibration then 20 mins flight and several attempts to engage loiter. If it is too long I’ll make a short flight with just the issues.
[attachment=0]Log.zip[/attachment]

5

So nobody else has chime’d in.

You’re logs look good to me. Vibs seem reasonable, DesRoll vs Roll and DesPitch vs Pitch all seem really reasonable, very good actually.

GPS seems fine.

Toiletbowl is almost always compass, so how’s that mounted, assuming you can’t compassmot with Pixhawk.

Put it on a stalk 3 inches above/away from stuff and I bet it stops… (Also is compass orientation correct?)

-MD

6

Thanks for taking the time to look mdisher!

Yup, she flies really well in Stabilize and Alt Hold. I auto-tuned. Was still a little fidgety so I brought down the rate P,I & D 10%

Just tested now and success!
Assuming the Pixhawk and Ublox was working, (Ublox was 10cm or so above everything on an aluminium stalk from the start) it had to be something causing greater than normal mag field. So I checked everything and re-run all the cables to the lowest position possible plus one of the ESC’s had two bullet connectors not in all the way. They connected well but the copper was little bit exposed. Fixed that, re-ran live calibration and Loiter works perfectly albeit a little fidgety.

Should I tune down the Loiter PIDs some?

Then I went through the compassmot motions - prop switch, secure chopper, throttle up etc - without actually running compassmot. There is still a slight variation in heading with the throttle at 75% and the aircraft held down in place.

More testing tomorrow.