Just loaded my standard '+' Quad with version 2.0.13 using the APM Planner.
I didn't try 2.0.12 but since 2.0.11 I can tell there is a definite improvement in the yaw movement although I probably need to tweak the PIDs to get it really crisp. I also notice that yawing right it seems to take longer to stop rotating than yawing left. This was true with all the previous versions too although I don't know why it happens.
When I tried Loiter is raced off into the air for some reason - not sure why. Because I was in a fairly limited flying area I decided to switch back to Stabilise mode rather than find out where it was going!
Nigel
Replies
Lots of problems with yaw was solved when I uncommented compass orientation (components down, pins forward).
There is some problem with declination: If I write declination on first setup (lets say 7.9) than in PID tuning window (right column) it is not 7.9, but something like 1 / 7.9. I think there is some mistake. I will travel with my hexa for AP, so I need adjust declination often and I will use PID tuning window (right column) ... so I do not know what to write there - declination or 1/declination?
p.s. I know that my declination is wrong because in Loiter mode copter is not correctly aiming loiter point, but some 45 degrees right. I have downloaded android application to determine declination in every gps location.
p.s.2 For Altitude hold I have p = 0.25 I = 0 D = 0
2.0.13 fixed an issue where Yad Dterm was computed twice. This was causing most of the issue folks were having.
Also, I noted that my copter was a bit banged up and that it was holding yaw with a 16° steady state error. That mean it was fighting to maintain Hold against the copter's natural tendancy to Yaw right. This is due to motors that aren't at a 90° angle to the floor. My system was sending +90pwm and -90pwm to the motors for Yaw control. This simply isn't in the range of acceptability as it eats all the headroom for decent yaw control. Please check your logs to see how big your steady state error is. Ultimately you want the quad to be perfectly balanced and not using the motors to keep it on target.
For Alt hold: Please try and put some felt over the BMP sensor to prevent airflow around it. Then adjust your PID settings till you get something that works well for you. The equation I use and Pirate's use are functionally equivalent. I got mine from Berkeley's Quad and I have been able to get it to perform very well in the past. I just think the tuning needs some attention.
The I-term may be ramping too high and too much. Pirates code ramps very slowly and has little effect.
The D-term .03 is probably fine
The P term - I have not idea what is the best choice. I think Baro P is too high. Try .1
Fly my hexa with 2.0.6 beta... my yaw work stable. Also test the Simple mode... what can I tell? I'm in love with simple mode :D
I will load 2.0.13 tonight
I will decrease loiter radius from 10 meters to 2. I think this might help.
Until some days ago I flashed every new update there was for ACM.
Since 2.08 I stopped flashing any more, because since ArduPiratesNG I wasn't able to fly any more with my selfmade frame and Roxxy 2827-34 motors and 10x4.7 APC props.
And after reading about that not-working LOITER (position hold) function I am confident I made the right choice.
I know this doesn't help improving the software and I know every version released is still beta. I just hope you guys will be able to finish the first "final" version for me to use safely before this summer ends.