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
I just wanted to give an update about my problems with autotune I had 4 months ago (which unfortunately crashed my octo). The advice was I had a problem with my (SimonK) ESCs and you were correct!
During the winter-time I ordered some new ESCs and flashed them on my own with the WII-ESC-Firmware.
Today I gave another try to autotune (second flight after the complete rebuild of my octo) and all the problems were gone! The copter finished autotune without any critical situation and I have the feeling some more problems (regarding uncontrollable drops of 30 m nearly to the ground) are gone now. My octo feels safe as it never did before :-)
My feedback for autotune: Stabilize pitch / roll was way too high (autotune set it to 8,6, I reduced it to 5), but that may be my preference for flying. Rate_P pitch / roll was also a bit too high (0,2, I reduced it to 1,9), but the result is really impressive now!
Ok, I found your post.
I suspect Autotune isn't coming up with gains that are too high for your copter, I think the problem is you are flying the copter in a different configuration to the one you used to autotune.
If you are flying with a large gimbal that is free to move on the frame you will probably find that auto tune doesn't work. If you do autotune with a load firmly attached to your copter then replace it with a gimbal your parameters will probably be too high and verging on unstable.
The best approach to using autotune in this situation is to remove the gimbal and do autotune with the lightest setup you can (that doesn't cause the battery voltage to sag). This should give more conservative pids that will be safer with your gimbal.
However, the final performance depends on the resonate frequency of the gimbal and suspension system. Autotune is not designed to tune a copter with this extra degree of freedom. In the end you must ensure the copter is stable yourself and make manual adjustments if needed.
Darrell Burkey,
Can you post a link or tell me what name you go by on that forum?
Has anyone tried this with the 3DR quads? I tried it on mine yesterday and the performance seems to be pretty good after auto tuning, but the values seem to be a fair way off compared to the already optimized versions on the APM?
Before: https://dl.dropboxusercontent.com/u/203420/Quadcopter/PIDs.png
After: https://dl.dropboxusercontent.com/u/203420/Quadcopter/ATPIDs.png
Seems that it also has different values for roll/pitch, even though they should be locked? The wind did pick up a little during the final pitch testing, whether that might have had an effect. It started raining as soon as the tuning had finished, so I had to bail, but bringing it back in to land it seemed pretty solid. Will go out and try it again soon.
Your tune looks good and reflects what we see from copters normally.
Hi Randy, Leonard and all the developers who made ac3.1 possible, it is solid, thank you!
However, It seems that I keep getting in and out of problems while attempting autotune, btw stock setting are fine, just wanted to see what can be improved by autotune. As you might recall from my posts here couple of weeks ago my previous problem was that autotune was going for far too long than what I read on this forum it is supposed to take 8min+.
Ok, after flying on stock PIDs for a while now after the last crush (2 weeks ago) I decided to give another shot at autotune.
I had good 3d GPS lock with 10+ satellites, almost 0 wind, checked baro vs gps alt looked almost the same, flew in stab a little and copter was rock stable, switched to alt hold, perfect, no ups/downs, tried loiter and the copter got "glued" to the spot where I left it (thank you developers for the awesome work!).
Switched back to alt hold and turned on autotune. 4min and it was done with rolls, perfect thought I, the last time tried autotune copter was rolling straight for 8min+, now came the pitches, after about 30 sec of pitching I was able to adjust the position of the copter just fine, however, it started to make them more aggressive and moving too much so I tried to adjust the position at this point I had 0 response on the sticks and copter moving towards a tree, so I dropped throttle and due to alt hold "dead zone" copter landed rather gracefully on the wet grass (dew). looking back at the log I see that after autotune was done with rolls apm got through all the modes I enabled before getting back to alt hold. I know that I did not move the mode switch is this an intended behavior? also, even though I said I had "dead sticks" it seemed that pitch and roll got mixed together, so when I was hard pushing back the copter was moving to the right and a bit forward (could be wind?) again it was a very "dampened" response this is why "dead sticks" I suspect it could be sudden wind though but I do not recall any.
p.s. this the first time I am using my x8r with s.bus to ppm converter? could this be an issue? if anyone knows on the frsky s.bus to cppm what is the correct jumper position for apm?
2014-02-11 09-06.log
Hi Leonard,
I tried AT on my new Pixhawk powered Tarot960 Hexa for the first time yesterday, before AT the Hexa was very jittery as standard settings too high,,,
after autotune, (very successful) Hexa flew great! so disarm with AT still on,, to save settings, refreshed MP to check new settings (surprised) but happy days, flies great, take off again, still good....
BBUUUUUUTTTTT on next boot up,, new battery,, Hexa flew all jittery like before AT??? refreshed parameters.. and the new autotuned settings are still there.. but flies like on default??
I repeated the process over and over,,, did 5 autotunes, tried refreshing full parameter list, saving it, and reloading it on new battery boot up... but everytime it seems the pixhawk boots into its default settings,,,,, even though it shows autotuned pids...
weird weird weird
any ideas??
Cheers
BOB
Thanks for the reply,
I can see that I never got out of RTL now from the logs, my mistake for sure, whoops! I've never used it before but apparently it's not working properly as demonstrated by my crash. I broke my no name ebay GPS / compass (the mobius is OK) so I bought a 3DR one, hopefully that solves any RTL issues once I get it.
I do have a couple of questions now though. The crash altitude would've been the same as the take off altitude. It's very cold here, -15c so I wonder if I didn't give the quad time to cool off before power up, could that cause it to have a bad baro reading?
Also, if the autotune finished why did it not attempt to do the pitch tuning and just stopped after the roll tuning? You can see that I waited 10 seconds before attempting (and failing..) to gain control of the quad again and it never tries to do the pitch wobbles.
So I have done more investigation into my issue and wanted to refine my questions.
The autotune stopped with a code of DATA_AUTOTUNE_ABANDONED. Google pointed me to the source code and I believe this is the reason it stopped:
In my log the RPGain starts at 0.055 and gets down to 0.01, then the next line the autotune is that it abandoned. Why would this happen? I have done autotune in the past, maybe it's not good to do it again without setting the PID's to defaults first?
The other issue is one that you raised with the baro alt. You can see that it tracks well against RelAlt for most of the flight until the end when it goes way out of whack. Why would this happen? My baro sensor is covered with cotton padding, then with a piece of black electrical tape, then with a chunk of dark foam.
If it helps I've attached a pic of my quad, it's an ADS 400Q, 1.5kg AUW. Thanks for any help.
Hi Muhammad Abdul Hadi ,
This is what I would suggest the problem is too. If this doesn't work, post a data flash log here and we will have a look.
-
42
-
43
-
44
-
45
-
46
of 55 Next