Have any of you seen this happen
after loading new code and reseting APM . setting declination my APM board magneto meter seems to be drifting .
my board is on the table pointing north and in the APM planner you can see it does a full 360 drift while still pointing north . does doesnt change even if make the declination from - to + any body else getting this ? this is reconfirmed with my code always yawing badley .
What magnetic declination are you using?
tried from -18. to -18.99
I had a similiar issue recently: http://www.diydrones.com/forum/topics/forever-yawing.
For me if you click on the Status tab, the yaw value was constantly counting down while the compass counted down as well. In your video we cannot tell if the yaw value is changing. I did fix my issue by going into CLI, 'setup, 'erase', 'reset' and then reloading the firmware. Now when sitting on the table the yaw value and compass direction no longer count down... and best of all it flies normal again.
thanx for the tip checked it now the yaw value IS counting down . but erase and reset doesnt work
Stupid question, but is your magnetometer activated in the setup and working?
Apparently, just because its working in the CLI, doesnt mean that DCM is using it to correct yaw drift.
I remember back in the day (when I got my first apm from you)
I got this exact issue, but it was because I didnt have a magnetometer. Chris said its caused by gyro drift...
yes it is enabled checked it in the CLI , but strange doesnt matter what i do with the declination it doesnt change
perhaps its a bug in the new code. Suggest try loading an older version and see if the problem goes away. If it does, you know what to do, if it doesnt, you know what to do :)
Also are you using the relaxpatch version of arduino?
think got the same problem...
does it take off ?
it only takes of becuase i fight it when taking of already but if i just give power , no it yaws soon after take of and crashes badley
It might be a bad solder joint, ground, or an i2c bus short circuit. Check or redo the solderin to the compass, and clean both boards very well. Specially greasy residues from anti oxydant grease.
i'm having the same problem as Johann with my APM1 + mag.
Since i updated to version 2.1.* the copter started to yaw for quite a while after takeoff, but recovered after ~30 seconds. In loiter mode is also doing circles. Moving to version 2.2 didn't help and also going back to 2.0.45 (which i used before) didn't.
My declination setting is correct (+1,26)
I already resoldered my magnetomter to the board.
When i read the heading of the magnetometer in flight planner i get wrong readings compared to my phones compass. When i turn my apm by 90 degress it only reports a change in heading of about 90 degrees but then drifts back by -20 degrees, while the apm is at rest. Then if i rotate it 90 deg further it reports a change of 170 and then also drifts by +20 (almost back to normal, which would be 180). Interesting thing is, that it seems to get north always right.
I already tried moving the copter around in circles after startup, but that didn't change anything.
So i checked the magnetic offsets in the CLI (setup->compass) and they were all 0. Then after some waving around of the copter they changed to some values. Then i did a reset and waved the apm around and after quite a while i got new mag offset values... but these were different than before (heading still being wrong).
So i tried the AP_compass test in Arduino.
Still wrong offsets like in the FlightPlanner.
While running the test i waved the apm around until the offset values of the output didn't change any more. Then i put these
offset values into the code of the AP_compass test where the mag is beeing initialized and suddenly the compass readings are correct.
For me it seems, that the Arducopter Code forgot how the calibrate the mag offsets.
But strange thing is, that going back to the previous version doesn't change anything.
Is there a possibility to enter the mag offsets into arducopter code manually?