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.

 

Views: 74288

Reply to This

Replies to This Discussion

Has anyone tried AutoTune on heavy octos? Any reason or special considerations to try or not try it ? (besides "this is beta software", testing, etc ...)

I've got a 5kg (without gimbal) flat 8, 1100mm,  flys great on rc5, sub .1g vibrations, solid arms and frame, CG well centered, throttle_mid under 500. Thinking of trying AT with gimbal replaced by equivalent mass of lead. Already did AT on a smaller copter, worked great. But am a bit apprehensive :-)   ...  Also concerned about board ressources consumed by Autotune. Would that be an issue on an octo?  Maybe turn-off all logging for starters?

Hi Wessie,

1. It isn't actually wind that causes the problem it is that you can't tell when your copter has built up airspeed. On a calm day you can see if the copter is starting to move much in one direction and stop it. Wind gusts on the other hand will mess autotune up. Exactly how all this will affect the gains is anybodies guess. I can say that they should still be safe no matter what conditions you do them in.

2. I don't know, I would need to see the autotune logs and then know how the copter was misbehaving.

3. Yes, doing it indoors is fine and will give good results if you keep what Rob said in mind. I have done it in my lounge room but it is a little cramped :) (It was raining and windy and I needed to test some changes before letting Randy release it)

Hi Raph,

Sorry but Yaw tuning won't be in there until 3.2 at least.

It is very hard to make any comment about your roll/pitch performance during descent from the video. Can you post a flashlog?

As for your poor yaw on decent, this may be coupling between your roll/pitch corrections and yaw. Again the logs will help but to test this reduce Stab_P as low as you are happy going and if the problem doesn't get better then it is just in the yaw controller.

Video looks great though!!

Hi Leonard,

I just was out on the field to try autotune with my Y6 without cam and with the batteries hard mounted n the frame.

The result was not good.

It did not as bad during the autotune as last time. No scary moves any more (or I am used to it now), but after the autotune I switched of Ch7 and on again, to fly with the new PIDs and it was shaky as hell. Was the same like last time. Then I said to me "o.k., back to the original values" and switched Ch7 off, but it did not change anything. Then I landed and disarmed with autotune off. So normally it must have restored the old values. After changing batteries, I gave it a try, but the PIDs from autotune were stored. So it flew with very bad shakes.

So I have to issues.

1st autotune does not work on my frame

2nd it does not restore the old values when disarming with autotune off

I attached the values after the autotune.

If you wish I could also upload the log file.

Attachments:

Hi Criro,

Glad you got it working!!

Let us know how your test flights go.

You should find that your autotune will run faster if you run it again from the tuned values. Be careful on your first flight because your pitch values may be affected by a lower battery.

Hi Jay,

If you send me your flash log I can tell you what the gains from your Autotune is.

We have had heavy copters do autotune and so far they have done very well with very few problems. There was one problem but that is fixed in rc5 and 7.

I would suggest leaving the lead off during Autotune to begin with (assuming it is safe to fly like that). The reason I say that is the heavier the copter the more aggressive the gains can be. So by doing Autotune in the lightest condition you should have a tune that is very reliable once the gimbal is added.

We are currently working through a issue where some frames have some vibration when the gimbal is reattached. This doesn't happen with all copters though. I have suggested an experiment to two pilots but I have not got any feedback yet.

If you decide to give it a go, make sure you let us know how you go!!

Thanks!

I used to fly one 2220 lipo, but sometimes I put 2 x 2200 in parallel for longer fly.

If adding the second lipo as supplemental payload, do I need a new autotune?

Thanx!

I have plugged Raph's settings into mine, they were VASTLY different to what I had. (STAB_X was 5.12 vs 8.25 now)  Ive flown those settings, still deciding if I like that feel or not, it feels VERY sensitive, may reduse a pid here and there still, will try it in the wind later today.  Loiter seemed to work fine , althoug drifted around more than i am used to, but generally stayed within a 5m radius, where it was 1-2 metres before.

Looks like i am still trying to Wrap my head around pid tuning, can you tell me if I am way off by saying : (i am trying to put this in laymans terms here.

Think I may still try AT on RC7, since you guys are saying that its just about foolproof now. (wink Wink)

Hi Devs,

Why can't I find any info about the rc7 and were is the 3.0.1 giant forum were I could find firmware release info

And maybe its only me but I find the forum system the worst forum system I ever seen, are there any plans for introducing a forum system like fpvlab/forum were you can make topics and with a good overview of the topics with last reactions.

Sorry to hear that mate!!!

Log file would be very helpful.

You should be fine with extra weight, but going the other way is more risky.

Autotune should give very safe numbers though so you should be a long way from losing control anyway (when compared to the way we manually tune).

Reply to Discussion

RSS

© 2019   Created by Chris Anderson.   Powered by

Badges  |  Report an Issue  |  Terms of Service