Two motors always spin faster. Autotune got horrible tunning (LOG)

Hi all!

I’ve trying to solve some problems with my F450 Chinese clone with no success. It fly correctly but suddenly comes into instability and seems it’s going to crash. I use Pixhawk from aliexpress and cheap motors and ESCs.

I tried autotune but results are worse. It is now very unstable.

RCOUT for motors 1 and 2 are always lower than for motors 3 and 4. Motors 1 and 2 are sometimes even almost shutdown.

I’ve noticed that arms 3 and 4 are slightly more flexible than 1 and 2.

Results after autotune are the following:

Here are the logs:
Autotune roll: https://drive.google.com/file/d/1sIndIY-pT_pzXDgy44AAxv2CpOJvVEXg/view?usp=sharing
Autotune pitch: https://drive.google.com/file/d/1P-Fk6scU5DWUm3fOWMKZmm-LztDmNkDk/view?usp=sharing
Autotune yaw: https://drive.google.com/file/d/1II3QtZB5jhLokGGX4plV6A6IrZ2n6-NQ/view?usp=sharing
Unstable flight after autotune all axes: https://drive.google.com/file/d/1EnGFhUChra1T68caY1ya4flAN-CL-vLE/view?usp=sharing

How can I tune my drone if autotune fails? Why autotune fails?

Does it necessarily mean that something is wrong the fact that 2 motors spin always faster? Any suggestions to solve it?

Thanks!!

Perhaps you have unsync esc

re-calibrate your ESCs

Thanks for the answers!

amilcarlucas I’ve recalibrated the escs many times with the “all at the same time” method and I do not think this is the problem.

About unsync esc, to be honest I do not know what it means and how to detect that. I’ve done a quick search on google and still not clear to me. I will investigate it in deep, but if anybody can explain to me what are the symptoms and consequences of unsync esc and how to fix, I would be gratefully!

Thanks again!

I think the 4 esc are in sync because when I calibrate them and move a little up the throttle, all 4 motors start spinning at the same time.

I doubt motors 1&2 are not perpendicular.

This happens when your centre of gravity is wrong. To keep the aircraft horizontal two motors have to work harder… I am not saying that is your problem though.

I did an autotune on a 600mm motor to motor aircraft and the resultat is brillant… Just saying this to say that if your aircraft is able to do do position hold, then autotune does a great job :slight_smile:

If CG is the problem, Motors 1 & 3 (front heavy), motors 1 & 4 (Right side heavy), etc…, will work harder than the opposite pair. If diagonally opposite motors work hard, they are fighting the yaw induced by the not so straight motors, in this case 1&2.
clone 450 frame will sometimes have twisted arms. Can be corrected by placing Shim under the motor Screws.
By trial & error , and graphing RC Out 1-4 , till the lines overlap (in a perfect hover).
OR
Buy a new S500 frame.

1 Like

That has a lot of sense! Thank you!

Do you think this explains the bad autotune and the unstability that sometimes the drone has?

May be, But an autotune expert will be the right person to comment.!

you are flying normally and, for moment, the quad become to move crazy near out of control, some esc aren´t adecuate to some motors so you need to change them but first be shure that isn´t other the issue, the log indicates that is something relate with your harware, frame not enought rigid can give similar issues,

1 Like

Hi cala2! Thanks for your answer!

By “hardware”, do you mean electronics hardware? maybe some sensor is wrong in pixhawk? or you mean something with the frame, the motors or the esc?

Something related your frame, motors , esc. Your hardware couldn"track the desired attitude, as soon I can I share graphic. First verify that all is well stiff on your ship, try to move arms if they are enought rigid

Thanks cala2! I wait impatiently the graphs!

The arms and the whole frame are a little bit flexible, but with my little experience I think it is normal. I also want to do some testing for checking esc and motors unsyc issues, but I’m not sure how to do so I’m searching and learning on youtube.

I also think a good tuning will help. Do you think that the hardware can not track the desired attitude can be explained solely by bad PID tuning?