EVSPIN32G4 - Frequent Overcurrents

[grisharevzin]

Hi.

I'm driving a PMSM with FOC (Workbench file attached for reference) with MCSDK 6.3.0 in torque mode.

Main parameters:

• 1400 Hz electrical frequency max -- therefore 16 kHz PWM
• 100 Hz mechanical frequency max
• ~0.03 mH line-to-line inductance
• ~9 V/Hz Kv 
• 0.1 Ohm winding resistance
• up to 20 A torque sustained
• around 48 V input

With the default settings of OCP -- 45 Apk and 47.06 ns filter -- OCP is reliably triggered when Iq reference goes above about 7.5 A. To get to 20 A I have to seriously increase both the filter duration and the threshold, which feels a bit unsafe...

Measured Iq and Id plots have very noticeable first-order oscillations, but everything else seems to be reasonably normal, and the torque response (and the speed control as well) is good:

The magnitude of Iq and Id's oscillations can be somewhat reduced by a more agressive PID tune (essentially increasing the cut-off frequency of the current loop) but this eventually causes to loop to become unstable (as expected).

I have two questions:

1. Is the OCP triggering at this level something expected (short current peaks of this type are normal?) or is it indicative of a fault somewhere? 
2. Is this type of Iq/Id oscillations normal as well? 

[GMA]

Hello @grisharevzin,

For the short current peaks, you can check with a scope the comparator input signals that are available on the board at TP4, TP5 and TP6 test points.

On your display, Iq/Id harmonic seems to be 2 times the motor frequency. Motor cogging torque could be a root cause this behavior.

[grisharevzin · Accepted solution]

Following up on this -- it was a faulty motor with internal short circuits.

Some noise on Iq/Id is expected, but the amplitude of it shouldn't be strongly related to the value of reference Iq/Id as was the case here. E. g. what I learned in relation to this motor is that you shouldn't get ±5A noise on 10A Iqref and ±10A noise on 20A Iqref.

Here's the same plot with a non-fauty motor (7A torque setpoint):

...and with 50A torque setpoint:

[GMA]

Hello @grisharevzin,

Thank you for the feedback.