Dangerous divergence: FIXED and understood

[bn999]

[afernan]

@Bill:
I´ve flown using TILT_ANGLE_I = 0 (equivalent than limiting IM) without any significant result, when forcing the divergence.

It should be nice to log the evolution of the P,I,D terms for both rate and angle tilt, in order to analyce them later. I think there is the key to solve the problem. We´ve done that in Arducopter loging and even printing and plotting P-I-D terms in rear time to tune the Loiter (we had problems with PID´s there) and was very useful to solve the problem.

Angel

[bn999]

Setting the ITERM to 0 will certainly cause trouble. You need the ITERM, you just don't want it to wind up too much. You could also play with limiting the PTERM and/or the total output. Small, incremental changes are best when tuning a system which is already close to being stable.

In the past I have logged a PID. There is a lot of data as it is updated at 400Hz. With some hacking, it would be possible to setup a log file.

[afernan]

I´ve tryed also small I-terms (0.0004, 0.0003) without success.

It makes sense to play with PM, IM, or even DM or the OM. I´ll try that next time.

In your opinion, what´s the main term is inducing this overshoot? Why Pitch can overpass the angle limited by Factor_Pitch? (Probably I don´t understand well that parameters meant)


Angel

[LPR]

Angel

My quad weighs a little more then yours 1.6 kg with a 4s 5 a/h battery. Without a battery 1 kg. I have seen the same problem as your seeing when the string between the motors gets loose. The aluminum arms cause more vibrations then the carbon arms but there cheap and good for testing. With the string tight the control problems go away for me. The string I'm using is for sail boats and is made of Vectran which is known to absorb vibration.

Larry

[afernan]

@Larry:

thanks for your comments. I agree that joining the tips of the arms by a pre-tensioned string can cancel some of the vibrations, let say in the high freq range (the range of the motors rotation). If a lack of arm stiffness were the problem, this should be solved by CFRP arms of large Alu sectior arms, which is not the case.
Now I´m flying a "heavy dutty" quad with arms 20 x 20 x 2 mm section (280mm long) which is a kind of "beast". So, no doubt about lack of stiffness...and the divergence problem is still exactly the same. I don´t really think this´s an "aero-elasticity" problem (I use to deal with that in my dayly work). The solution is inside the code and/or PID´s.



Angel

[LPR]

Angel

Your most likely right that the very light weight and slow turning props are the problem with your quad. You could look at using braking to slow the motor because the slow turning 5010 have little drag. The large diameter of the 5010 gives high angular inertia also.

Larry

[afernan]

[LPR]

Angel

Thrust control could give a solution for you.

viewtopic.php?f=26&t=1731&p=7634&hilit=ESC32+igor#p7634

Larry

[afernan]

@Larry: thanks for the link. It´s very interesant; I´ll try that.

Anyway, this divergence is not "my" problem. I think it´s an AQ problem at this moment affecting to more people also. Flying smooth or big copter shouldn´t be an issue, but that´s not always the case.

Angel