Dangerous divergence: FIXED and understood

[LPR]

Angel

I'm creating a parameter file now for a small quad with 9" props and weighs 750 grams without a battery. It flies without the parameter file but not well.

Larry

[teramax]

Angel, sure it is not only your problem.
This kind of reaction i had the first time a week ago, on a quad that flys firmware tests for nearly a year.

The only things i changed were a 350g brushless gimbal underneath and latest firmware (r76).
I first thougt it was my fault cause i hung this 350g on the quad without looking at the pid´s, removed the weight, tested again and here it was again.
(always the same procedure: set home-switch RTH- shoot away the rig 10 meters than center sticks - and let the quad bounce back like on a rubber string to it´s home point)

Befor i stumbled over this tread i replaced motors (RCT 5010) and ESC (SmartDrones 4in1), thought they are to slow (weak) for a stable and fast reaktion -with the MT Navigators and SimonK ESC.

So, rebuilding is done and i am on testing yours and Bills theory.
First thing i can say, divergence issue is also present but more powerfull, faster and more scary but not causing a crash like befor.
Testrig: 1800g AUW, MT MN4010-9, RCT CF 15x5 Props, Flyduino NFET 20A ESC, AQ6 r63, 3s8000

[LPR]

Angel

The small quad flies very well, maybe a little better then the heavier larger quad I built that has 15" props. The small props make the difference I think. The props have to turn fast to hover so they can slow down quickly when needed.

Larry

[Jdmagoo]

Has this issue been resolved? Is this a common problem with all AQ boards or just a few? Interested to here from the effected users

Cheers

James

[afernan]

@James:
I guess I´ve got long fligh experience with this divergence since I was forcing it (in flight) many times trying to get a good PID combination (without success till now).

I´ve already seen this problem in three different AQ boards and quite a lot of frames, combination of motors, ESC and propellers:

Motors: : T-Motor MT2212 750kv, 5010-14 360KV , HP4215 460KV 24N18P
ESC´s: Flyduino Simon-K, ESC32
Propellers::10x4.5,12x4.5 carbon, 13x4 carbon, 15x5.5 carbon
Frames: quad (0.8, 1, 1.2 Kg rtf), Hexa (2.6Kg rft)

I´ve played a lot also with PID´s without success. So, I can say the problem is still there.

Basically: if you flight smooth >> no problem. If you fly hard >> you can get the devil and crash (I´ve got A LOT already).

I´m trying to understand the main attitude control loop and compare with other codes much more simple that don´t show this divergence (Arducopter, for instance). They use nested PID (pos/rate) (AQ not, although is prepared for that). Then, I´ll try to flight again, but the problem is that every flight test normally ends with a crash.

Angel

[Jdmagoo]

Thanks for the detailed explanation

[LPR]

Angel

I have not had a problem with the kind of divergence you are describing ever since I learned to set the Tilt Rate D lower till the AQ multicopter would behave.
To my surprise when I add a second set of batteries to my octo I suddenly had to be careful when flying in Alt/Posi Hold. I did not see any problems when flying a mission.
Luckly the octo attitude is slower changing then small multies and I was able to switch to manual to stop oscillations.
Changing PIDs did not resolve the problem. Removed the extra set of batteries.

Larry

[chschmid]

Hi Larry

It would be very interesting to have some more info bout that system change.
With one set of batt, what is your factor throttle. and what about with your 2nd set of batt. I ask because IMHO I believe that the tilt rate D and tilt angele P ar very dependent of how much power the system has in relation to weight. It seems that the more overpower (factorthrottle < 0.700) the system has, the less probles occur although tilt rate D and tilt angele P have to be adjusted.

Best
Christof

[afernan]

@Larry.
your comments are in line with my experiences. The problem is still there even for big machines, although things happens slowly due to the big size. Also, normally we use to fly a big copter smoother than a small one so this reduce the problem.

TILT_RTE_D must be changed with care. Too high can increase the divergence problem (also creates additional oscillations) . Too low will not stop the oscillations generated by the TILT_ANGLE controls. I´ve flight tested a lot this changes.

Angel

[LPR]

Angle

I totally agree with your opinion and Christof might have some of the answer to the problem. I don't think I readjusted the Throttle Factor when I tried using two sets of batteries. Not having the right amount of power to be able to tilt the octo quick enough could cause the problem with control. I'll have to try the two sets of batteries again and set the throttle factor.

I did have a bad experience a few days ago with having the throttle factor set a little too high. I converted the Tx so the throttle joy stick will spring to center just like the pitch and roll joy stick. That conversion works great for Alt/Pos Hold mode and for flying missions but in manual you have to be very careful to not let go of the throttle joy stick before you disarm.
I set the throttle factor to where the octo would not fly over a couple of feet off the ground in manual mode and that worked well.
My problem came when I had just charged the batteries and as I walked into the house I thought I disarmed the octo. I set down the Tx and changed some navi PIDs and then the ground station said "LOST LINK". I looked outside and no octo was there. After twenty minuets of panic I found it over 1/2 km away. Only a motor and prop and three arms were broken.
I now I set the throttle factor so the octo will not go above two feet, in manual mode, with a battery that has been just charged.

Larry