I’m getting some very odd intermittent twitches with my Tricopter which I can’t seem to solve. The image below shows Chan4 out (output to rear motor ESC) and appears to show the output spike BEFORE the pitch change. Does that mean it’s software related as if it was hardware the spike to the motor would surely occur after the pitch change?
Also it yaws slowly at times without input, e.g. between line 20x and 21x
Any ideas appreciated 
p.s. The flight controller is a AUAV-X1 from Arsovtech
Hi Graham,
It looks like the magnetometer might be the source of the problem (is it enabled?):
[attachment=0]mag_pitch_gyro.png[/attachment]
It appears that the spikes in magZ precede the gyro response (when there’s a high correlation). The pitch measurement doesn’t make a lot of sense to me though. This is my first look at an Arducopter log, and I don’t know what the units are. No labels… Very nice GUI though.
Is there something onboard causing DC current spikes at that frequency?
Slightly off-topic question: Is the log file raw mavlink binary? My python mavlink parser rejects every record saying “bad prefix”.
73, Mark
Mag is enabled, don’t think Copter can fly without it. Surely the mag wouldn’t affect the pitch or output to ESC in such a drastic way?
Don’t think there are any DC spikes, the Vcc value is pretty smooth at 5.01V - 5.16V
Graham,
I suspect the issue is somehow related to the set-up or tuning of the tail servo. It probably doesn’t have enough range to keep the yaw under control.
Could you downgrade the software to AC3.1.5? If this is done and the problem persists then we immediately know its’ a set-up issue and not an issue with AC3.2.
Tested earlier and both problems exist with v3.1.5, the slow yaw I can live with or hunt down but the sudden 20°-30° rearward pitch is scary at times.
So does that mean the rearward pitching is some kind of setup or hardware issue?
Here’s the log from this morning.
In this plot from the last log, the CH4 output to the rear motor ESC appears to occur before the pitch change so the pitch change seems to be commanded.
If it was a fauly ESC I would expect the pitch change first and the ESC output after that as presumably the controller would try offset the sudden pitch change? I’m happy to be wrong though.
I don’t have another of the same ESC but could, with some surgery swop it with one from another arm, I just don’t want to waste the time if it’s not the cause.
Changed the rear ESC and that seems to have solved the issue.
