Bit of a strange one this with my quad.
1. In loiter, altitude is gradually lost. Positioning is totally fine, it's just the alt. Raise it up to 8m and it will drift back to 3m in about 5 minutes. Sticks are exactly mid stick.
2. In auto, and alt hold, maneuvers can cause sudden alt drops. In Auto, I had to save it one time, because it went to WP1 ok, but turned to WP2, and it just seemed to let the alt slide down. I had Altitude adjust on, but it seemed to totally ignore that. In alt hold, I'm constantly having to pitch back - even on small movements - because it just drifts to ground. It will resume height when you do, but subjectively it's not resuming the *exact* height it did before.
3. I set the waypoint heights at 20m (and 30m on 2nd run). The quad appeared to make no effort to get to these heights. It only went to 50m (before I cancelled) on the last waypoint, because RTL was the last WP. (strange...I've got RTL set to 0, so it should have RTL'ed without gaining height.
Attached are my logs and PID's...does anything jump out?
Other data: Ambient temperature makes *a lot* of difference. Flying in 10'c - motors never exceeded 20'c. Flying in 26'c...they hit 50'c mid flight. Batteries and ESC's also mildly warm. Flight time: about 15mins. First 80% was loiter.
Replies
On my copter, I put black electrical tape on the top of the 2.5 case. This prevents light leaking on to the baro sensor. Without the tape, my copter will dip down several feet with certain angles of sunlight. Of course, since yours is doing this on the bench, your baro issue could be something different altogether.
I've been getting very similar problems, so far no luck with either the barometer or sonar.
Euan,
When you upgraded to 3.0 did you erase and reset your apm? We've been discussing some of the issues with upgrading that are causing some wacky problems.
Ok, I think I might have found my problem.
This is the Barometer output. It's completely wrong, and doesn't match what the Spektrum TM1000 collected, or my eyes saw. I made 3 height adjustments in the first 50%, and this was in loiter. These were made when the TM1000 (and my eye) said the quad was at 3m. The APM baro seems to think it held an absolutely flat height during this time.
It also thought it was at 20m when it started the Auto run...it was not. It was at 4m.
Possible electrical interference on the Baro from the PDB and ESC just below it maybe? And how to mitigate? Will do a bench test on altitude as well.