SMALL - espritmodel.com SMALL - Telemetry SMALL - Radio
Reply
Thread Tools
Old Feb 26, 2014, 09:27 PM
Registered User
United States, CA, Ukiah
Joined Jan 2011
266 Posts
Discussion
vtol tailsitter starter PID values?

Hi all,
I'm working on a tailsitter vtol uav and getting some oscillation from the elevons, I think it's feedback from the multiwii board but not sure yet. I'm hoping I can minimize with adjustments to the PID values but I'm just making a guess until I know more about how PID control works. Any suggestions for starting out?

Thanks,
Devin
devinvagt is offline Find More Posts by devinvagt
Reply With Quote
Sign up now
to remove ads between posts
Old Feb 26, 2014, 11:49 PM
Registered User
Australia, NSW, Sydney
Joined Sep 2012
350 Posts
Is it fast or slow oscillation? If fast, lower p gain. If slow, lower I gain
jolyboy is online now Find More Posts by jolyboy
Reply With Quote
Old Feb 27, 2014, 09:09 AM
Registered User
Ran D. St. Clair's Avatar
United States, CA, Sunnyvale
Joined Dec 2006
864 Posts
Your pitch and roll gains are going to be very different. Pitch gains will be much lower. When I say roll, I mean ailerons in airplane mode. You might call it yaw in hover mode. I assume you are using differential thrust for airplane yaw, hover roll. That will be different as well.

All of the control surface based stability and control will be weak in hover mode for that design. You lack sufficient lever arm to do much, especially in pitch.

Assuming you can get it to hover, it won't handle hardly any wind. It should be fun to play with though.
Ran D. St. Clair is offline Find More Posts by Ran D. St. Clair
Reply With Quote
Old Feb 28, 2014, 12:03 PM
Registered User
United States, CA, Ukiah
Joined Jan 2011
266 Posts
Thanks for the help.
I can get it to hover pretty well in horizon mode with the multiwii, but the pitch and yaw are indeed a little weak. I may switch to a delta wing setup to increase the pitch control more. putting the motors and control surfaces nearer to the wing tips should help with the yaw.
The oscillations are fast, typically i can get it to oscillate by putting some gentle pressure on a control surface. I think the flexibility of the frame is increasing it.
devinvagt is offline Find More Posts by devinvagt
Reply With Quote
Old Feb 28, 2014, 08:58 PM
Registered User
Australia, NSW, Sydney
Joined Sep 2012
350 Posts
Maybe use larger elevons for better control, the ones in the photo look tiny although it is hard to see them.

Are the elevons in the airstream of the props too?
jolyboy is online now Find More Posts by jolyboy
Reply With Quote
Old Mar 01, 2014, 03:52 PM
Registered User
United States, CA, Ukiah
Joined Jan 2011
266 Posts
Yes, they are in the wash of the props. That's a good idea, I may make them a little bigger. I am also going to cut down the wing tips all the way to the elevons to lower the moment.
Quote:
Originally Posted by jolyboy View Post
Maybe use larger elevons for better control, the ones in the photo look tiny although it is hard to see them.

Are the elevons in the airstream of the props too?
devinvagt is offline Find More Posts by devinvagt
Reply With Quote
Old Mar 07, 2014, 12:40 AM
Registered User
United States, CA, Ukiah
Joined Jan 2011
266 Posts
update: it flies! multiwii PID values still wonky

This is how it looks now! apparently it is called an << H Wing >> by some.
Here's what's happening: No matter what I do to the PID for pitch, it's still way too aggressive in bringing it back to level. And no matter how much I turn up the PID on roll, it still takes a few seconds to bring it back to level. Essentially, if I could switch these two behaviors, the H wing would fly perfectly. Any ideas?? Just a limitation of the code?
Also, lowering the P for pitch below 3 causes oscillations, as does increasing the P for the roll past 9. Changing I and D for roll seems to have no effect.
Currently:
P I D
Roll 9 0.25 20
Pitch 4 0.03 0
Yaw 7 0.03 0
devinvagt is offline Find More Posts by devinvagt
Reply With Quote
Old Mar 07, 2014, 02:53 PM
Registered User
RCvertt's Avatar
Joined Aug 2007
4,107 Posts
Can you show us where the CG is? I would also cut those vertical stabs so they look like triangles. Too much vertical surface in front of the CG currently for yaw stability.

For your situation you may want the pitch gyro to be sensitive but constrain the servos movement range to a smaller value so it doesn't over correct. Lengthening your servo horn is one way to shorten its movement range.

If you need more roll response, scaling your roll gyro command in code will do that.
RCvertt is offline Find More Posts by RCvertt
RCG Plus Member
Latest blog entry: Chupacabra sighting !!!
Reply With Quote
Old Mar 07, 2014, 03:44 PM
Registered User
United States, CA, Ukiah
Joined Jan 2011
266 Posts
The CG is roughly in the middle. It hovers with minimal adjustment.
How do I scale the gyro in the code?
I've been reluctant to change the control horns on the elevons as these also control yaw, and yaw response is aleady kind of slow.
Also, the multiwii gui lets you increase rc rate for pitch and roll, but not yaw. Looks like I need to look into the code again.
devinvagt is offline Find More Posts by devinvagt
Reply With Quote
Old Mar 07, 2014, 03:58 PM
Registered User
RCvertt's Avatar
Joined Aug 2007
4,107 Posts
CG sounds a little aft for a straight wing. Has it flown forward yet?
RCvertt is offline Find More Posts by RCvertt
RCG Plus Member
Last edited by RCvertt; Mar 07, 2014 at 07:45 PM.
Reply With Quote
Old Mar 07, 2014, 10:23 PM
Registered User
United States, CA, Ukiah
Joined Jan 2011
266 Posts
It may indeed be a little aft, I'm still trying to iron out the aggressive pitch correction before I try putting it horizontal again.
I set the control horns down some ( less throw) and it just slows down the rate of oscillation and makes it a little worse. So far no combination of P I D has managed to eliminate the overshoot.
devinvagt is offline Find More Posts by devinvagt
Reply With Quote
Old Mar 07, 2014, 11:59 PM
Registered User
RCvertt's Avatar
Joined Aug 2007
4,107 Posts
Got some hover video we can look at?
RCvertt is offline Find More Posts by RCvertt
RCG Plus Member
Latest blog entry: Chupacabra sighting !!!
Reply With Quote
Old Mar 08, 2014, 01:09 AM
Registered User
United States, CA, Ukiah
Joined Jan 2011
266 Posts
I'll upload some video the next time I test in the day time! I usually test at night for safety.
I minimized the overshoot by *increasing* the P value for pitch. About 7 gives a nice return to horizontal in about 1 second with minimal overshoot., I and D are zero. It seems like P value is inversely proportional to response time. I can't understand this but it is handling nicely now.
devinvagt is offline Find More Posts by devinvagt
Reply With Quote
Reply


Thread Tools

Similar Threads
Category Thread Thread Starter Forum Replies Last Post
Idea G.N.A.T.S., EPP VTOL Tailsitter ** FREE PLANS** leadfeather Foamies (Scratchbuilt) 88 Feb 04, 2014 03:52 PM
Build Log F-22 based VTOL tailsitter with pseudo thrust vectoring Phaedra Scratchbuilt Indoor and Micro Models 14 Jan 28, 2014 01:00 AM
Discussion Fan fold VTOL twin tailsitter Helno VTOLs 1 Jan 12, 2014 11:28 PM
Discussion Blade CX tailsitter Foamy VTOL test flight in smoke/air tiltrotorcoaxial Foamies (Scratchbuilt) 4 Jan 15, 2011 12:12 PM
Help! tiny VTOL/Tailsitter using RC plane parts nanobot9000 VTOLs 5 Mar 04, 2010 02:08 PM