• Howdy! we're looking for donations to finish custom knowledgebase software for this forum. Please see our Funding drive thread

Problem with BLDC and belt (Power transmission vibration)

ok, I was wrong about the Settings not working in "BLDC" mode.
I changed the transition to FOC to 20000 ERPM so it definitely stays in Block commuting "BLDC" mode,
and the setup runs great! The belt causes no problem at all. Very good.
But in FOC it struggles under load.
I don't think detection gives me reasonable results. I ran it several times.
Also it fails to detect one parameter (stays red)
I wish I knew what to enter for this motor manually...
I am posting the results here:


  • IMG_3043.JPG
    8.6 MB · Views: 2
  • IMG_3044.jpg
    7.3 MB · Views: 4
I think Lq-Ld was only introduced in vesc 6 so that may be why, I don't know that it matters though as I think it's only used for drive modes that have D axis current and that motor probably has so little Lq-Ld you don't want to use any of those, or can't with that hardware.

Poor detection from the other variables though could very easily be the issue. I've heard on some hardware lowering the frequency to 5khz while running detection can improve results, you can turn it back to whatever it was before for actually running the motor. I think it's called zero vector frequency now but used to be called switching frequency. Also can help if you increase the duty cycle for the test but only at reasonable levels for the motor, as in if those are set too small like you pick a motor size too much smaller than your motor detection will be poor but of course setting then way to high can damage things. You should write down the detected variables to see if they change with detection changes. Obviously they shouldn't change much so if they are they probably aren't detecting well I think it's often the flux linkage that is the culprit.

And yes knowing what the variables should be would be helpful but unfortunately I have no configured any similar enough motor to even make a guess and the only other way is to measure them with other tools.

I've also heard that lowering the time constant a bit and recalculating the gains will often improve things running stably.