[newtoon]

Hi !

I've small issue with my quadcopter which is not symetrical.

Its nose dives when I throttle up for instance, of have sudde instability with some wind. I can't find where to lower rear
motors sensitivity compared to the front.

I try different combinations (PIDs) and would like to experiment with the mixer too, to compensate the fact that the weight
is distributed "in line" (battery at the rear, FPV system at the front).

Well, I just would like to know and can't find simple information anywhere. If I lower the mixer cursor for pitch, what does it do on the FC ?

Thanks in advance. 

[NicholasDavid]

Nose dive under throttle sounds like either a PID problem, weight distribution, or hardware problem.

If it's hardware then it's like one or two of your motor/esc are outputting more than the rest. Could be a calibration issue but it's kind of rare unless they got damages from a crash or something

Could be a weight distribution problem. As in not balanced front to back and air resistance and drag are causing it to pitch when throttle applied. Try moving the battery back and forth and see if it helps

And lastly it's most likely a PID issue. Specifically your pitch I gain.  The I gain will make the copter hold its angle and stay there when you punch it. Try raising your pitch I gain some and fly again.

[newtoon]

Thanks !

I have already raised I quite a lot (and I'm in attitude mode). Weight ? Well, the CG is not so badly placed (the battery in the back is compensated by the FPV
system and when I hold my quad on a finger, it's not bad compared to the board. The CC3D board is not excactly in the center of thrust (one cm max difference so
not bad).

Anyway, I'm interested in the mixer first. I just want to know what (theoretically) happens if I change the mixer for, let's say, Pitch.

I already tried it at home with Pitch at 40 instead of 50, but it still flies OK. Some hesitations when I put 60.

There 's quite a lot of info on mixer for cleanflight, but very scarce (a lot of matrixes) for Openpilot/ Librepilot. Anyway, even if I used matrixes,
I saw that the motor layout is not the only thing to consider. Mass is another one. Anyone knows ?

[f5soh]

No need to play with mixers, if you want more authority on Pitch axis you just need to increase the Pitch mix slider.
Save and look at Custom tab to display changes.



You can also simply increase the P,I,D for Pitch axis in Stabilization tab.

[newtoon]

Thanks !

I still do not understand what does a setting at 40 % in pitch in the mixer sliders. 

Besides, my pitch issue (that I described) is still there even if I put my CC3D one cm back.

This is no drag issue. It does that in my own small living room or even when starting from the ground.

It may be my odd frame configuration, but still, very strange that it is quick to nose down but slow to nose up.

[liftbag]

Please, do this simplte task: set pitch Integral inner loop to zero and observe how it behaves in hovering.
If you have a forward/backward drift, then adjusts the battery position until you have a neutral behavior.
Once you have corrected the center of gravity with this method, set roll/pitch I-Term to three times the Proportional.
If you continue to have anomalous behavior, then you need to investigate for hardware problems.

[NicholasDavid]

Quick question.  Is I term set at 3x P gains pretty standard for Libre Pilot.

As I understand, I term at double the P is kind of the standard starting point for tuning b

Is 3x what a lot of I terms end up at?

Obivously depending on components...but in general.

[liftbag]

Quick question.  Is I term set at 3x P gains pretty standard for Libre Pilot.

As I understand, I term at double the P is kind of the standard starting point for tuning b

Is 3x what a lot of I terms end up at?

Obivously depending on components...but in general.

High I-Term give a better locked-in feeling. I set it 4x P-term on most of my multi rotors.

[NicholasDavid]

Interesting. I have a 3s quad I'm tuning at the moment and with 300ish P gains and I in the mid 700's it still pitches back when I cut throttle. All other behavior seems fine. I'll up the #'s and see what happens. Thanks

[TheOtherCliff]

Unsymmetrical or out of balance multicopters either need a strong compensation mechanism (I term in PID must work well) or if too badly unsymmetric, need to have adjustments made to the mixers (thrust) (not recommended because you must use "custom" and then it doesn't know it is a quadcopter which confuses other issues).

If I had one really out of balance I would reduce the thrust mixing on the lightly loaded motors (the ones that lift too much when you climb hard).  E.g. if it tilts forward when climbing, I would reduce the thrust on the back motors.  Makes sense, right? 

In the example, for it to be perfect, you would reduce the pitch mixing on the back motors as well, but this is way way less important than the thrust mixing.