Measuring IPM Motor Parameters from Plots

[johnspark]

I agree with aiming for 80% with only 20% effort .
Well here is some info on harmonics to help understand this stuff a bit better.

So there are no rotation forces on the motor either positive or negative, but we do get the center of the 3 windings wanting to do something, so if the neutral is isolated, which it is, there will be 3rd harmonic voltages on this center point.

For 5th harmonics we get torque going against our fundamental frequency torque. Mind you, we hope 5th harmonic is very low so this effect is very low.

7th harmonic rotates WITH the fundamental! At last! we have something nice!

[johnspark]

EMC with very sharp voltage source waveforms.

For Toyota motors, each individual phase wire is shielded, so no voltage EMC from the 3 phase supply wires.
As far as current EMC, if the motor has high L, current EMC will be low...

[Pete9008]

Thanks to @FFMan for taking some logs we now have data for the Outlander rear motor

Lq=0.36mH, Ld=0.20mH, Flux Linkage=60.8mWb, Rs 27mR

Also added to summary of motor data post here : viewtopic.php?p=52763#p52763

[Romale]

Thanks to @FFMan for taking some logs we now have data for the Outlander rear motor

Lq=0.36mH, Ld=0.20mH, Flux Linkage=60.8mWb, Rs 27mR

Also added to summary of motor data post here : viewtopic.php?p=52763#p52763

and where can I see the settings file of this motor and the pinout drawing of the resolver socket? I recently bought such a motor, but I haven't figured out the connection yet

[Pete9008]

and where can I see the settings file of this motor and the pinout drawing of the resolver socket? I recently bought such a motor, but I haven't figured out the connection yet

Best data I'm aware of is herehttps://openinverter.org/wiki/Mitsubish ... Drive_Unit

[Romale]

Best data I'm aware of is herehttps://openinverter.org/wiki/Mitsubish ... Drive_Unit

excellent. thanks a lot, that's what I was looking for in the wiki, for some reason I couldn't find it myself. It remains to find an excellent settings file for it, so as not to look for the correct sensor angle)

[nickyivyca]

Seemed to work for getting params for my mystery motor.

Since I am running what the inverter thinks is reverse to go forwards, I had to flip the sign on iq and uq as they are read from the file, otherwise I got nonsense values.

What could the error spikes in Vd mean? Are those just from when flux weakening kicks in and throws off id from the normal MTPA calculations?

[Pete9008]

Looks good, that's a pretty high value for flux linkage which explains why the motor doesn't like to spin too fast. It should make good torque but it's going to need a lot of field weakening current!

Not 100% sure about the spikes but think it's due to the flux moving up the B/H curve, not quite saturating but getting there, which reduces the apparent inductance and so increases the error. I did try using different error terms in the calculator to bias it more towards higher currents, it improved the spikes but overall made things worse. The plan, once I have a motor in the car, is to make the inverter dynamically measure the inductance as the current changes to allow a full motor model to be built up. No idea whether it will work, or whether the improved characterisation will allow improved motor control, but curious enough to give it a try. In theory it should be possible to do it with logs and the calculator but it would need much larger data sets.

Good point on the sign reversal, I should really add a checkbox to cater for that!

Edit - I produced this the other day https://openinverter.org/forum/vi ... 79#p59079 . Given that your motor has much higher flux linkage than all the others used when producing it (and it is only calculated using Ld) I'd be interested to know how your kp and ki compare to it once you have finished tuning?

[nickyivyca]

I retuned my gains the other night and ended up with iqkp of 250, idkp of 125, and curki of 22000. The previous ones I was using were not too different. Not sure what the metric for 'best' gains is - I did my tuning with a current clamp probe and tried to tune in the best step response as possible, using d current so the rotor would stay mostly stationary. Then doubled the gain for iq because of higher q inductance, just as a rough guess. This made the motor sound pretty similar to the original inverter which makes me think I'm in the right ballpark - with lower gains it would sound more 'rumbly'. And with these updated gains I seem to maybe be getting slightly better regen performance at higher RPMs.

I did then re-run the parameter estimator, and got slightly different results, though I was driving in a different area this time, still mostly flat though with no braking.

Still seems like I need more flux weakening current - I was inching it up and getting better performance until I was approaching low battery today, I think I need to give it even more but that's an adventure for another day.

[Vita1ii]

Hello friends,
I've been working on building an electric car using parts from Opel for some time now.

I've created a non-standard combination of a Prius Gen2 inverter and a Leaf EM61 motor with the original Nissan gearbox. Currently, in my tests, the battery voltage is at 190V.

I would like to request assistance with calculations and calibration settings for my inverter and motor setup. How can I gather the specifications?

I've taken a few screenshots of my testing while climbing a slight incline and driving on a straight road.

[johu]

The motor parameters are the same regardless of the inverter. For the EM57 we found Lq=0.426mH, Ld=0.164mH (lqminusld=262 uH), Flux Linkage=73mWb
viewtopic.php?p=52763#p52763

[Vita1ii]

Thank you, I'll make changes today.

(lqminusld=262 uH), Flux Linkage=73mWb
viewtopic.php?p=52763#p52763

Johannes, You mentioned "lqminusld" in uH units; should I record it as 0.262 mH ?

[johu]

Thank you, I'll make changes today.

Johannes, You mentioned "lqminusld" in uH units; should I record it as 0.262 mH ?

Oh yes indeed 0.262 mH

[Zieg]

Does anyone have the software up and running and a couple minutes to check some logs? When I press auto tune the numbers come in pretty low, and pressing it again just makes the numbers drop lower and lower. Not sure if the logs are bad or if I did something wrong, but even after the first round the numbers come in low compared to other Leaf motors. Does it matter that it's running in reverse?

Log 2 is a stationary locked rotor test with a small accelerator blip, the others were started and stopped while moving on flat ground at close to constant speed.

Vehicle weight 750 kg, reducer ratio 8.2, tire radius 0.3m, motor poles 4