Dangerous divergence: FIXED and understood

The last Bastion to fine tune your FC matching your airframe & setup

Re: Dangerous divergence: FIXED and understood

Postby kinderkram » Wed Apr 24, 2013 4:21 pm

Good job, Angel!
btw: Max and Peter released a new version of the QGC for internal testing, focussing (not only) on better mixing functions.
From what I've seen, the mixer tables for Quads are now all based on +/-100 which should help to get rid of the divergence issue.

new-gui.jpg
new-gui.jpg (216.49 KiB) Viewed 5470 times
kinderkram
 
Posts: 2911
Joined: Fri Jun 22, 2012 7:47 am

Re: Dangerous divergence: FIXED and understood

Postby Kisssys » Wed Apr 24, 2013 6:39 pm

I've been playing with it today and it's very nice addition to the package.
Well Done!
Steve
Kisssys
Kisssys
 
Posts: 1340
Joined: Sat Jun 23, 2012 9:23 pm

Re: Dangerous divergence: FIXED and understood

Postby afernan » Wed Apr 24, 2013 6:56 pm

@kinderkram:

"Mix" tables should not be allways set a 100%!

It depends of the control supplied by the specific combinations motor/propeller and also the configuration used (X, +, quad, hexa, etc).

In some cases it´s enough with a 50%. In my case (for same quad, motors and default pids) in "+" mode it´s enough with a 50%. But when changed to "X" mode I need to increase mix to 80% and to increase ctrl_max to 400. If I use 100% in the "+" configuration I get oscillations of the control due to excesive control (like having a high P value)

The best way to check is to do a hard fly and check the LOG MOT_PITCH/ROLL as I´ve explained before in the post (to check we are still far of 300pwm so we still have tilt control margin in the pocket)

PD. Thanks for putting the technicar note about this in the download area

Angel
afernan
 
Posts: 1214
Joined: Sat Jul 07, 2012 5:24 pm
Location: Madrid, Spain

Re: Dangerous divergence: FIXED and understood

Postby LPR » Wed Apr 24, 2013 7:37 pm

Angel

I agree. If I use 100 on the quads that I built I get osculation and poor yaw control. For your quad with it's light weight and large props you need more than 50 pitch and roll and you could use less than 100 for yaw because large props and short motor to motor distance give more yaw authority than a more standard quad.

Larry
LPR
 
Posts: 1323
Joined: Tue Jun 26, 2012 1:29 pm
Location: MN, USA

Re: Dangerous divergence: FIXED and understood

Postby skew » Wed Apr 24, 2013 10:47 pm

Angel, I just got a quick flight in between rain storms,set everything to default,motor at 80%,cont set to 400. I banged the sticks full throw in every direction and flew it harder than I ever had before.She flew like a champ! Thanks for all of the time you put into this it was worth it!
skew
 
Posts: 181
Joined: Sun Oct 14, 2012 1:30 am
Location: Morehead, Ky

Re: Dangerous divergence: FIXED and understood

Postby Max » Wed Apr 24, 2013 11:10 pm

Hi Angel, nice troubleshooting. I looked over my quad X logs that's using +/-100 mixing on roll/pitch. The MOT_ROLL and PITCH figures are directly affected by Tilt Rate D setting. As you know. Somehow for me tuning the D value also brought the roll/pitch values to max. around +/- 100, vs. close to 300 with the higher D rate.

Quick tuning session, just adjusting Tilt Rate D:
quadX-mot-p-r-TRD-test.png
quadX-mot-p-r-TRD-test.png (11.89 KiB) Viewed 5443 times


Fairly aggressive flight afterwards:
quadX-mot-p-r-flight.png
quadX-mot-p-r-flight.png (11.33 KiB) Viewed 5443 times


-Max
Max
 
Posts: 2814
Joined: Mon Aug 13, 2012 9:45 pm
Location: Near Ithaca, NY, USA

Re: Dangerous divergence: FIXED and understood

Postby afernan » Thu Apr 25, 2013 9:32 am

@Max:
TILT_RATE_D was one of the first "suspect" (see first coments in this post) but I didn´t success really changing it, although we saw that decreasing it improved the situation a bit, but spoiling the tilt control also. TILT_RATE_D is the second derivative of the "error" so it´s nice to have as higher as possible for the best performance. Anyway it´s very interesant to see how interact different parameters in the performance.

The mix table %-tilt is very important also. You don´t want to have an "excesive" tilt control all time (ESC´s and motors suffer a lot, and battery consumtion increase also). I recomend you to make hand test with 20-50-100% on mix tilt-% to "feel" the reaction force in your hands (It´s a bit dangerous! but it worths)

Angel
afernan
 
Posts: 1214
Joined: Sat Jul 07, 2012 5:24 pm
Location: Madrid, Spain

Re: Dangerous divergence: FIXED and understood

Postby afernan » Sun Apr 28, 2013 10:56 am

More testing & tuning around the problem

Today tested a small (800gr) X-Quad with HP4215 (460kV) with carbon 12x4.5 prop in a really heavy wind (30knt+). Started with (80,400,7180) for the three main contibutors: mix tilt%, controller Max, Tilt_rate_D

I´ve monitored in real time with telemetry the MOT_PITCH/ROLL. With initial conf it was marginal: I reached the limit (400) quite offen. In order to reduce this control values I´ve increased the mix tilt up to 100% and also reduced the TILT_RATE_D. Then it was good (in the range 100-200 avrg)

Figures attached show MOT_PITCH before / after changes.
Attachments
New Picture (3).jpg
afernan
 
Posts: 1214
Joined: Sat Jul 07, 2012 5:24 pm
Location: Madrid, Spain

Re: Dangerous divergence: FIXED and understood

Postby Shogun » Sun Apr 28, 2013 1:18 pm

Thanks Angel, this was really a good catch!

With an 80%-mix I was now able to set my navigational pids to defaults with alt-speed im and om constraints reduced to 400. I have here a Quadrixette 30, A2212/10T 1400kv, 8045 Gemfan, 3S 2200mAh, AUW 800gr.

I did not try to raise Tilt-D yet, these are reduced to Tilt-Rate-D 3000 and Tilt-Angle-D 1200.
Shogun
 
Posts: 46
Joined: Tue Aug 14, 2012 10:53 am
Location: Aachen, Germany

Re: Dangerous divergence: FIXED and understood

Postby LPR » Sun Apr 28, 2013 1:54 pm

Angel

Does it fly more steady in the high wind?

Larry
LPR
 
Posts: 1323
Joined: Tue Jun 26, 2012 1:29 pm
Location: MN, USA

PreviousNext

Return to PID Tuning

Who is online

Users browsing this forum: No registered users and 4 guests

cron