Bad Motor Output with APM!

Hello everybody,
After struggling with taking of for about 2 weeks, I think that I finally found out that my APM card makes BAD outputs for the motors.
In fact, when I look at the logs I can see that :

  • With auto-analysis I have a WARNING message (highlighted in blue in the first screenshot, link below) : Test: Motor Balance = WARN - Motor channel averages = [1248, 1268, 1149, 1209] Average motor output = 1218
    Difference between min and max motor averages = 119

  • By analysing manually, I can see that in fact, the outputs are bad (on the second screenshot, you can see the legend with the arrows I drawed in green : arrow 1 shows the legend , arrows 2&3 the big differences between the motor outputs in response to the " RCin " and arrow 5 shows what I looked at.

  • I could confirm that something was wrong when I looked at the the “failsafe” section in MP (under “mandatory hardware” ), the radio 1-2-3-4 outputs in the right culumn (called “2nd flight mode”) reacts a bit weirdly ( 3rd and 4th screenshots show what I have when I put throttle at maximum) even without props, and even without the connection to the ESCs, so I don’t think it’s because of them…

I’m quite frustrated because I can’t take off, the quad reacts as if it was drunk, and eventually always crashes as soon as I stop to compensate with the joysticks)… even if everything seems really fine … (I have been through the troubleshooting guide of ardupilot many times, recalibrated and rebuilt the quad 3 times, checked everything was alright). Rmq: I have a 2-3% difference in the minimum throttle between 1 and the other 3 motors, but I calibrated the ESCs 10 times and can’t make it any better…and I read that this should be an acceptable difference.

–> So does andbody have an idea about HOW to correct the motor outputs of the APM ?

Thanks a lot for your time


Have you checked if the quad is balanced, so the weight is distributed equally for all the motor?

The quad is balanced, the CofG being underneath the APM and more on the back than on the front.
But anyhow, do you think this is really a hardware issue, as I have the same Warning in the logs when not connecting the ESCs into the output ? (maybe the card kind of saves what it “felt” before and compensate what doesn’t exist anymore, meaning that you are right and I should balance better the weight to the motors in front ! Does the CofG have to be exactly in the middle?)