This is a discussion re the ArduCopter-3.1 AutoTune feature. A place for people to report successes and failures, ask questions and get answers.
The AutoTune wiki page is here.
You need to be a member of diydrones to add comments!
This reply was deleted.
Replies
OK. Several more tests:
Removing gimbal and testing gave similar poor results but with roll and pitch values closer.
restoring the default orrientations of 0 and moving esc connections as appropriate to default front/back and recalibrating acc and compass and testing autotune gave similar poor results.
I measured my frame and since it is only about a cm off from square I changed the frame type to X, recalibrated acc and compas (back to orrientation of 4. Similar poor results.
Manually entered default pid values. Autotune tests return to level more cleanly, but still poor results.
Restested with 2200mAh instead of 8000mAH 3C pack. All up weight now down to 1.2kG. Success! This time the resulting parameters look much more reasonable (see image below). Of course it is a little sluggish (but stable) when I put the heavier pack back on for a second test flight with the autotuned pids.
My conclusion is that the extra weight put the quadcopter's physical characteristics out of the range of the ability of autotune to work correctly even though it hovers at about 50-60% throttle with all that weight.. Is this a known issue? I'm happy to help test any code changes which might help.
Hi Brad,
I just completed an auto-tune with my quadcopter which uses 100KV motors, 28Inch Props and had an all up weight of over 8kg. Hover thrust was approximately 65%. The Autotune was done in calm wind conditions and was a complete success. I am absolutely amazed at the difference. Based on my experience, weight was not an issue but that may be because of the individual components in my unit.
Regards,
Oliver
There is one more non-ideal aspect of my setup I noticed this morning. The big battery was so wide that it was able rock a little on the roll axis which I can see might confuse autotune especially given the large percentage of mass which is accounted for by the battery. I reworked the battery strap to have a wider attachment and the big pack feels much more firmly attached to the frame now.
Unfortunately, the winds are forcast to be high for the next several days at least so I don't know when I will be able to test it.
The wind was finally down so I retested autotune with the much more secure battery strap configuration. This time the Pitch look better and the D values did not seem to go to minimum, but the Roll response in flying is still a big disappointment with overshoot pretty obvious and nasty oscillation when descending at 1.5m/s. The camera and gimbal were not attached so the mass is not as spread out along the roll axis. I would have expected the roll values to be much closer to the pitch.
I'm pretty convinced there is a problem with the code or algorithm that causes trouble for a Quad with my configuration. I'm giving up for now and going back to something closer to what I got with the smaller battery.
I'm happy to make further tests if that will help.
Thanks,
--Brad
69.BIN.trunc
I have two 5000 mah 3c batteries, my pitch and roll values are different perhaps batt weigh distribution I have horrible weather conditinons too, but when I can, I try again Thank's Brad
Thank's for post your experience, I'm beguin with this great piece of the code now, I learn a lot of things from my cuads doing autotune but they can give me more, I'm going to try with one battery
I habe similar issues with my Y6B and autotune. Its seems to be to heavy (Hover at 63%)to get good Rate D values and regarding to them the right P and I values. I reduced the D values 70% manually and the P,I values about 40%. Looks much better now, but still not perfect.
Hi all, long time lurker but first time poster. Its time for me to get connected and involved with this WONDERFUL community.
I just wrote out a really nice reply - but then my...trusty... macbook thought it would be a good time to freeze up. So im starting all over so ill be (trying) to keep it short and sweet.
I have completed many autotunes with my bird. Majority successfully completed. The reason for today's post is for poor flight performance in windy conditions. More specifically, I was on a mountain side where there were turbulent updrafts. I had just recently done an AT which had considerably increased my Stab_P values. (which I think are to blame for the poor windy flight performance), and slight increased the PIDs (still acceptable values for this frame).
Symptoms in WIND:
-Difficult to combat occasional yaw drifts up to 30deg within a second or two.
-Horribly unstable and bouncy/shaky/rough flight in stabilize / loiter / alt hold. (much worse in loiter)
-I did not feel in control of the aircraft.
-rock solid when climbing in RTL.....
-Flys like a champion without wind at home.
I can post resulting video of flight if requested. (Currently uploading)
I have been unable to find much literature on the Stab_P values. My understanding is that the higher the number the faster the rate of change toward the desired angle. I have spent plenty of time tuning P I and Ds but never fiddled much with the Stab_P. In recent AT the values keeps rising and rising.
These are acceptable PID values for this frame, but I am unsure of the Stab_P. Could these high values be the cause of poor stabilization in windy conditions? Comparing Roll to RollIN and Pitch to PitchIN, they do not track themselves as well as id imagine they could/have been told they should. Is this the stab_P causing the system to overreact to the wind?
APC 10*4.7 props on 2212 Sunnyskys via iPeaka 40A on 3s. APM2.6 on 3.1.5
Also, while digging through logs i've noticed several times on several occasions that RTL and DRIFT are momentarily engaged unintentionally. Drift is one of my 6 flight modes - but NEVER gets used on my behalf. RTL is on ch7. In most cases I am unaware during flight that it has happened. the tlogs do not supply PPM values so im not sure how to diagnose this problem either.
Attached are logs from the AT flight and the poor performance flight afterwards.
Please let me know if I have missed any information for diagnostics.
I would just go out and fiddle with the Stab_P settings inflight and see what I can see, but it's generally not windy here and it's a good drive back to the mountains...
I thank you all in advanced for everything you have already done for the drone world and any help you may be able to offer me!
2014-10-15 Autotune_157.log
2014-10-18 Bumpy lemon flight 2_162.kmz
You are not alone with this problem, there is a post about it, I have good results tappering air entrances, I have good performance with stabilize mode in windy days but not so good in modes that requires the barometer, first of all check your vibrations. In my modest opinion (I'm not an expert only tryng to help) stab P looks a little hight? I have the idea to read that non equidistant frames like yours needs different roll and pitch values (not mine)
RTL and Drift are near in radio values? look that PWM are not very near http://copter.ardupilot.com/wiki/flight-modes/
Sorry my bad english perhaps someone else can help more; this something to start.
Here's a test I did very recently. From what i've read on the wiki, this is well within the limits (red bars).
Ill put stab_P on a pot and play with it this weekend and see what happens.
-
25
-
26
-
27
-
28
-
29
of 55 Next