Curious as I am I loaded the code with the latest MP. All went well but when I armed the motors and fly , only two motors started. And the quad kept turning to the right without liftoff.
Then I loaded 2.0.19 with arduino, and did CLI setup. I couldn't arm the motors. When I used the setup of the planner it worked. Also used Xbee with 115200 and works great.
First flight was very wobbly so lowered P to 0.2 I to 0,02 and D to 0.09 in stable.
I was able to fly more stable. But the problem I keep having is when I bank left or right the yaw changes without stick input.
This is something I had since using 2.0.2.
Could this be a PID setting?
Replies
Just switched to version 20 and yaw problem seems gone!!!!!!
Did a reset and loaded everything with arduino and also did setup with CLI.
Then lowered PID in stable and the quad flies very stable. Some little more tuning needed. Its allmost like the ardupirateNG code.
Mavlink works also great. Just one issue with altitude. This doesn't show very correct. When I'm 2 meters of the ground it says 10m in tge GCS. At 4 m the GCS says about 20m. Or is this an pid setting.
Didn't try alt hold yet ;-)
I haven't tried with mag enabled. But in the previous versions it did'nt work neither. Nevertheless I will give it a shot this evening.
When I fly slow the yaw stays fix without any problem. and doesn't change. Only when I bank hard left or right and stop the quad by compensating a bit in the other direction it starts turning a bit.
Its like it wants to compensate. Its very difficult to orientate immediately after the turn.
Furthermore I would like to add the the pid tuning goes rather well. Its going in the right direction. The development that is ;-)
I am afraid to fly this like that... but maybe this could help finding the YAW issue.
BTW, is there an easy way to undo the offsets calibration?
We're flew 2.0.19 and the next version 2.0.20 today, in all modes, and it was absolutely flawless. GPS hold, alt hold, waypoints, simple mode, etc. Everything was rock solid. Jason even did flips!
He'll probably commit the latest code to the SVN tonight, but I think we're very close to coming out of beta. Jason is tweaking a few setup processes to make it easier to use, and the Mission Planner is being updated accordingly. But based on a couple quads being put through the paces today, I think we've got the most solid ArduCopter code ever.
Just look how relaxed Jason was watching it! It's staying within a 5m box in 15-20mph winds!
I would try 2.0.16 code (SVN version 2421). I had it all fairly stable until recent code updates which seem to affect my YAW.
Ollie.