Noisy ANGLE spot value - SOLVED [SOLVED]

[ChazFisher]

I've gotten my adapted V2.0 controller to spin the GM e-Assist motor using the power stage of a Honda IMA inverter. BUT.... I'm seeing one thing that puzzles me. The rotor angle read back over the web interface seems very noisy, even when the motor is not spinning. Which makes me wonder what I've done wrong.

The e-Assist is an ACIM, and uses a resolver for the position sensor. I've been playing with the gain on the resolver excitation to get a good signal back from the resolver. Here's what the signals look like right now - upper trace (yellow) is the sine signal at 1 V/div, lower trace (blue) is excitation at 2 V/div. I rotated the motor by hand to get the largest amplitude on the sine signal - about 1.5 Vpp.



I plotted the ANGLE spot value received back - here's what I get with the motor stationary. Looks to me like the calculated angle is just spinning:



Thoughts? I've experimented with the resolver excitation amplitude, to give me a larger signal back, but with even up to 2.5 Vpp max on the sine/cosine signals, I see the same behavior. Firmware is ver 4.87R-Sine.

[kiwifiat]

It is not possible to say if your sine and cosine feedback coils are connected properly because we don't know what angle the motor shaft is at and we can only see the sine and drive signal. The attached images may help to confirm you have the feedback signals wired correctly. Note that the second image is at zero electrical degrees. Notice that between 0 and 180 degrees electrical that the sine feedback signal is in phase with the drive signal. One thing that is assured is if the feedback signals are not wired correctly the algorithm that calculates rotor electrical angle will not be correct.

I would not think that reversed feedback signals would cause jitter like you are seeing. The common neutral point voltage for the feedback coils looks ok.

Have you checked the cosine feedback signal?

sine cosine signals

Sin_Cos_Resolver_waveforms.jpg (109.82 KiB) Viewed 1936 times

Resolver signals at zero electrical degrees.

[johu]

I assume you're using the sine firmware as it's an async motor. And I further assume that what you're seing is the slip frequency So temporarily set fslipmin to 0 and re-observe.
You resolver signals looks good to me

[ChazFisher]

I assume you're using the sine firmware as it's an async motor. And I further assume that what you're seing is the slip frequency So temporarily set fslipmin to 0 and re-observe.
You resolver signals looks good to me

All good assumptions, Johannes. The firmware doesn't let me set fslipmin to 0, but I changed it to the minimum 0.28 Hz from the 2Hz I had previously selected, and the slope of the logged angle showed the much slower slip frequency. I could also reverse the direction of the rotation by changing from FWD to REV. So clearly a misunderstanding by me of what that ANGLE parameter was meant to indicate - position plus angle due to the slip frequency.

Thank you - solved!