Lately I've been having a problem with random yaw in my hex where it wasn't before.
Two problems. First, it always wants to yaw 10-15 degrees at take-off, but I don't think it's a crooked arm/motor as it doesn't always yaw the same direction. In fact, the direction it yaws appears to be a function of the direction it's facing; if I'm facing north or east it wants to yaw clockwise; if I'm facing south or west it wants to yaw counterclockwise.
Second, there's mild yaw drift in flight. Again, the direction is not always the same, sometimes it yaws counter-clockwise and sometimes it yaws clockwise. In this case, the direction doesn't seem to be a function of direction the copter is facing, but the yaw is more likely to occur when I'm rolling or pitching the copter rather than just hovering.
I've tried updating the declination of the magnetometer, but that doesn't seem to solve the problem.
I've been flying the copter with this problem for 20-30 minutes a day for 3-4 days now, so I don't think it's a calibration issue.
Any suggestions for what else might be causing the problem or what I might try to fix it?
Are you sure the Mag is working?
A good method to test this is to power up the board facing different directions and connect to the planner to see what is the direction is pointing.
If it is working, try to disable and re-enable the mag. You could have bad offsets. It happened to me yesterday, the quad was constantly yawing left or right based on my first input. After resetting everything was fine again.
Same here yestarday, i solved when disable and re-enable the mag.
Sometime i've cyclic and costant slow rotation of the compass (i noticed this behavior in the planner) and the drone was still on the table.
Now i thinks there's another little bug somewhere... :P
Thanks for the suggestions. I re-installed 2.055, cleared eeprom, and re-setup the copter. I also switched power for APM from ESC to a dedicated UBEC, just in case. That seemed to do the trick - yaw drift is gone now.
My guess is that the compass offsets were wacky; I've seen some mention of that problem in other threads.
However, I do want to switch back to the APM power via ESC to see if that might have been responsible. If I find that the yaw drift comes back, I'll be sure to report it as another reason to power APM through a dedicated source, rather than ESC.