Muskoka

  • *
  • 21
Tricopter issues
« on: May 31, 2016, 12:07:08 am »
Hmm, what's up with Libre and acro flying with a tricopter, proving pretty much impossible.

Have no issue at all with quads and Libre, flips and rolls are great. I have a home made wooden tricopter that had a KK2.1.5 on it and it flies great. Had a extra CC3D so I thought I'd try it, but it's been nothing but a nightmare. I've built 4 new arms already today, luckily nothing but arms and props have been broke. It flies fine in Rattitue and Rate mode, but is a mess when trying to flip, in either of those modes.

It's getting over "flipping" fine, but once coming out of the flip it starts to "tumble, rotate, spiral", and there's absolutely no way to recover. Just cut the throttle and watch it hit. I can't find any info out there about flipping tricopters with Libre, and there's not much if anything about prior / other versions. Starting to think it's a lost cause by the lack of info out there, but thought I'd ask first.

I'll add, the tricopter flips perfectly fine with the KK2, was just curious what the CC3D would do on the tricopter with Libre. So if anyone flips and rolls their tricopter with Libre I'm all ears as to how you set it up.

jrs

  • *
  • 3
Re: Tricopter issues
« Reply #1 on: May 31, 2016, 03:56:49 am »
I have a CC3D in a tricopter that I can flip and roll with no problems. I pretty much followed the setup in the following link:

http://fangin.com/blog/2014/04/14/multirotor-loops-and-rolls-just-got-easier-with-openpilot/

I set mine up with CCMA to 80 and deg/sec to 360. I have never tried it Rate mode, but in ratittude it work great.
« Last Edit: May 31, 2016, 04:01:50 am by jrs »

Muskoka

  • *
  • 21
Re: Tricopter issues
« Reply #2 on: May 31, 2016, 12:44:52 pm »
I have a CC3D in a tricopter that I can flip and roll with no problems. I pretty much followed the setup in the following link:

http://fangin.com/blog/2014/04/14/multirotor-loops-and-rolls-just-got-easier-with-openpilot/

I set mine up with CCMA to 80 and deg/sec to 360. I have never tried it Rate mode, but in ratittude it work great.

Still no luck here. It's like the yaw has no sense of direction when coming out of the flip, and starts spinning/correcting wildly, hence it puts you in a death spiral. The tri is a "T" configuration, not a "Y", but that should have absolutely no bearing, as they're geometrically the same as far as a controller is concerned. "H" quad vs "X" is geometrically different, "Y" to "T", the same.

It's setup with Rattitude and Rate modes. Tried Rattitude first with cruise control, as you can flip a quad with your eyes closed in that configuration. Thought maybe the cruise control, or the attitude part of Rattitude was causing the issue, so I then tried Rate mode with the same result. It flies much better in Rate mode, and was much easier to tune. In Rattitude I had to drop the Outer Loop gains to 0.7-0.8, from the default of 2.5, to get it to fly properly for instance.

I've tried Rate Mode Response of 450, 375, and I'm now at 300. Flips fine at all rates, but kept lowering it thinking it might have been too quick for the controller with a tricopter, unfortunately it hasn't made any difference. I run my quads at between 500 and 600 with no problems.

It's not a "setup" issue, as far as getting in the air. I've built enough quads with CC3D's loaded with OpenPilot, TauLabs, or Libre, and they all fly great. Was just curious if something "different/special" needed to be "tweaked/set", because we are currently scratching our heads on this one. It flies, flips just fine with a KK board, but is a spinning mess coming out of a flip with Libre.
« Last Edit: May 31, 2016, 12:51:54 pm by Muskoka »

jrs

  • *
  • 3
Re: Tricopter issues
« Reply #3 on: May 31, 2016, 01:19:04 pm »
The only thing that I have different from a quad setup is i had to lower the inner loop gain for the yaw from the default of about 600 down to the 300 ballpark.
The other two axis I left at the defaults.

Muskoka

  • *
  • 21
Re: Tricopter issues
« Reply #4 on: May 31, 2016, 03:02:01 pm »
The only thing that I have different from a quad setup is i had to lower the inner loop gain for the yaw from the default of about 600 down to the 300 ballpark.
The other two axis I left at the defaults.

Just tried again, another arm. >:(

I don't get it. It "flies" great. Just reset it to defaults all around, and it's solid as a rock in the air, until you try to flip. The tail/yaw starts to swing back and forth coming out of the loop, and that's it, it's done, death spiral. I'll play with the yaw gains and give it another go, but I'm really having doubts. Why it flips fine with other controllers and not CC3D with Libre makes no sense.

There's little to no wind here this morning so the elements are not an issue. This was built from stuff I just had laying around, the esc's are Hobbywing Flyfun, but are quite happy to run at 400hz, so I can't see that being the issue. If they were "the" problem, they'd be an issue in normal flight as well, and it's fine. Using a Turnigy coreless digital servo, and it's never been an issue, and there's no yaw issues in normal flight. I'm at a loss, and think this will just be a regular flier, no acrobatic stuff. Getting tired of building arms. :-\

jrs

  • *
  • 3
Re: Tricopter issues
« Reply #5 on: May 31, 2016, 09:06:24 pm »
I had a bad tail oscillation even during regular flying until I lowered the yaw gain, but once I lowered that it flies great. It would take off and fly with the higher setting, but then it would over correct and start oscillating to the point it was almost out of control. When I get home I will have a closer look at my settings and let you know if I notice anything else I have that's not default .

Muskoka

  • *
  • 21
Re: Tricopter issues
« Reply #6 on: June 01, 2016, 02:16:14 am »
I had a bad tail oscillation even during regular flying until I lowered the yaw gain, but once I lowered that it flies great. It would take off and fly with the higher setting, but then it would over correct and start oscillating to the point it was almost out of control. When I get home I will have a closer look at my settings and let you know if I notice anything else I have that's not default .

Thanks for the help. I tried lowering the yaw gain and it was alright, but the tail wandered too much for my liking. Went from default 620 down as low as 300 and felt the tail was too weak to hold.

I agree though, last flight this morning I tried some "aggressive pirouettes" and it was fine in one direction, but the other it got squirrelly and tumbled from about 10' up, of course broke another arm. Something is not right but I can't seem to get my finger on it. Some setting is off,  the defaults are alright, it flies fine, but doesn't like to piro, or flip.

Again, its much smoother in rate mode than any levelling mode but I'd like to get this flyable for someone who may not be comfy in rate mode. Pirouettes in Rattitude are ugly at best.

Muskoka

  • *
  • 21
Re: Tricopter issues
« Reply #7 on: June 10, 2016, 03:35:54 pm »
Dug a little further into this.

What I've found is the rear motor, yaw, does not have the same power output from the controller as the two front motors. The neutral point for that motor needs to be a lot higher (70 points on the output screen) than the front. It's not a hardware issue, definitely controller/firmware, as I've switched motors and esc's to various positions, as well as channels on the CC3D, and I see the same in all tests.

Why does the yaw motor, motor #3, have different power settings from the front two motors? I can now see why I was having issues trying to get this tri to flip. The tail "yaw" motor is that much slower, less power "signal" from the controller, compared to the front motors.

Or, are we supposed to keep cranking up the neutral point until it somewhat matches the 2 front motors. Never have this issue with a quad, all motors are always within 1 or 2 neutral points of each other on the output screen.

Just tried a qBrain to rule out my esc's and integrated board, although I've already tested all connection combinations, and I still see the same. The neutral point of the third (south) motor is at least 70 points behind the others.

Thoughts as to why this is the case?
« Last Edit: June 10, 2016, 04:17:35 pm by Muskoka »