Align 800 Tuning Log and a couple of questions

ArduCopter Traditional Helicopter
1

Hey everyone,

I am working on getting an align 800 fully autonomous.
And given how little support there is available I figured I would post my logs and experiences up to help anyone else who is considering this. My log is below:

docs.google.com/document/d/1tg0 … sp=sharing

I will post the gains I am using once I connect to the pixhawk again this afternoon.

A couple of questions, the heli seems to bounce after a pitch movement in a hover, i.e. I do a hard forward pump on cyclic and then it seems to bounce a tiny bit while it is trying to converge, is this a P or an I gain issue? I’ve backed the I gain down and have started on the P but I can’t quite tell if its helping. I am unsure if it is the P because when I observed that on the roll axis, having too high a P didn’t make the oscillations converge, they seemed extremely underdampened and would worsen with time.
My I gain on the pitch is .15 currently and I believe the P is about .07 but I may be mistaken on that.

Another issue I noticed is that the collective behaves funny, the helicopter seems to ascend and descend by about a meter each direction during a hover, I am conscious of my stick inputs causing this but I can’t tell if it is because it is too touchy or if it is because the response is too slow and the gain inside of my brain is too high. Does anyone know which parameter I can adjust to edit these behaviors?

My last question that is much simpler is: what is the FF tuning for? I can tell that it helps, but I want to understand it better so I know what exactly look for in the heli’s behaviour.

I can keep posting updates if anyone cares to hear them, otherwise I’ll just post questions :slight_smile:

Thank you,

Sebastian

2

Hi Sebastian, when you say it bounces, is it like an oscillation in the pitch axis? I had a similar issue in tuning a 500, where after an aggressive cyclic input it would oscillate in the pitch axis. It would eventually stop after about 2 or 3. I managed to get rid of it by pulling back the pitch P parameter by about 1.0. In the default parameters both roll and pitch P are at 4.5. The wiki recommends a pitch P of 3.5 (although I think that was for an earlier version) and it makes sense as the oscillations only occurred after a user input, and not an error induced by an external factor. I also found that by increasing the head speed a little bit it also helped. I know of one other that solved a tail bob issue the same way. The P value is probably your biggest culprit though.
I hope this helps.

3

I actually discovered that it was a vibration issue. fixed that and it all got better!