Steady state PWM of 2 motor phases and current measurments

[tprohas]

Hi All,

I am driving a 3 phase PMSM as follows:

I set my voltage to 12 volts. I PWM a single High Side (phase A) FET, at low 10% duty cycle (20KHz freq). I Turn ON a single LOW side FET PHase B. This is done for a steady state test. I measure the current via a Phase B low side shunt. This is effectively 2 inductors in series, with series resistance.

I hand calculate the expected currents, taking into account inductor resistance, shunt resistance, etc. If I Assume 12 volts, and resistance of 0.140 ohms, max possible current is 86Amps. At 10% I would expect about 8.5Amps. At low duty I'm close to my expected value.

The problem is if I increase the duty cycle, to 20 or 30 %, I have a wider margin of error, over 2-4 Amps. I'm trying to understand where the error is. I don't know if reactance of the inductor is causing the issue, and if it is, is there any info on determining this for a PWM duty cycle value (not a Sinusoidal Reactance!).

Any ideas or suggestions?


Tom

 

[moffy]

Let me guess that the current is on the high side? If it is then the inductors could be saturating at peak current.

 

[tprohas]

Yes, currents are higher than expected. My goal is to predict the current based on a given voltage and PWM duty cycle. Then when I apply the voltage and pwm duty for a short time (10ms), verify the value is as expected. The indutance is ~1000uH(micro H).

I plot the current ramp up and it reaches steady state in ~2ms. As Duty cycle increases, the current goes up accordingly.

I'm not clear on how saturation comes into play. If I saturate, then do I have pure resistance? But then why is resistance increasing with duty cycle. I don't think heat is a factor, but I need to run a few more tests.

 

[moffy]

When you run into saturation it looks like the inductance decreases. Normal magnetic cored inductors are non linear, unless they have large air gaps. But you can also get very large current spikes which can be destructive.