BLDC CW/CCW rotation issue

Question

I wrote the post with BLDC delta winding commutation sequence, and to this day a believe the findings are true.

But i have an issue that is occurring only in CCW rotation, at about 97% up to 100% duty cycle the motor starts to run rough but at a slight higher RPM (2600 to 3000) in short bursts (like it would have timing advance for short periods).

This is not happening on CW rotation, the RPM is the same in CW vs CCW up to 97% duty cycle, but above that in CCW there is the simptome described.

The strange thing is that this is NOT related to load nor to supply voltage (it is doing the same at 10, 20 even at 100 Amps and at 24, 37, 74 Volts supply), the only relation to the simptome i found is the duty cycle (above 97%).

Note the motor is run in sensor mode with no timing advance (not in SW not in HW), there is no cross conduction, all the drive signals on the gates check out fine the only thing that is a bit strange is that the phase wave forms look different (some pulses in the wrong places) in CW vs CCW, and i can not figure out why. I added a few scope images, they are measured with ground on U phase and channel A and B on V and W phase of the motor.

The question is, what causes the described simptom?

Photo nr 1 CW rotation duty cycle 100% NO simptome.

Photo nr 2 CCW rotation duty cycle 100% WITH simptome.

Photo nr 3 CW rotation duty cycle about 95% NO simptome.

Photo nr 4 CCW rotation duty cycle about 85% NO simptome.

Photo nr 5 CCW rotation duty cycle about 97% WITH simptome.

Answer 1

Well, after 3 months of head bashing, I concluded that there is nothing wrong with the controller.

If I swap the motor phases, I can make the motor run ok in CCW mode on controller and motor spinning CW, and make it run bad on CW mode on controller and motor spinning CCW.

So when the motor spins CCW it seems the sensor output is not the same as spinning CW, according to BEMF and hall output, for a 5.358 ms pulse when motor spins CW there is 2.5 degres or 4.1% of timing advance, and in CCW the same values are timing retard, witch apparently is enough to cause the symptom.

And one thing that I did not pay attention, even that the simptom was present at low supply voltages as well as high voltages, when increasing voltage the simptom was worsening because the motor was increasing speed and the timing retard is fixed (halls physical position) so the faster is spins the greater the timing retard.