Ansaldo Motor

[Jack Bauer]

My new Ansaldo motor in the E39 hates me. Finally got the cooling done today and decided to go try some tuning. Not very successfully. Can't get more than about 30 amps of battery current into the inverter but motor current very high. Speed signal is clean and current sensors are reading correctly. From what I can tell by the rating plate and looking in the terminal box, this is a delta wound motor. I picked it as it was available and seems to be rated for a low voltage as the Landyacht used to be DC. Any ideas much appreciated:)

[Jack Bauer]

Can't seem to add picture of the motor plate. Error parsing server response.

[johu]

Uploaded the file for you. It should work now, permission problems - of course.

So its a 2-pole pair motor (9000 rpm @ 310Hz)
It is also a bit of a give-away for the slip frequency: sync frequency at 9000rpm would be 300Hz. But they specify 310Hz -> 10Hz slip frequency! wow. Maybe try a range 3-10Hz or so.
Concerning V/Hz: fweak=73Hz (2200/30) at udcnom=183V (sqrt(2)*130V)

Is that about what you configured?

[kiwifiat]

What type of position feedback does that motor have? If changing fslip makes no appreciable difference to the performance it might suggest that the firmware doesn't know the rotor position accurately.

[Jack Bauer]

It uses an AB encoder bearing with 75ppr. I did have a noise issue on the speed signal but it is now clean. I think Johannes has a good point about the slip. I'll give this a try.

[Ryp]

It uses an AB encoder bearing with 75ppr. I did have a noise issue on the speed signal but it is now clean. I think Johannes has a good point about the slip. I'll give this a try.

Hi!
Are you sure about the bearing pulses? I have almost the same ansaldo motor with SKF bearing and 80ppr output.

[Jack Bauer]

No I'm not sure:) I just measured the pulses on the scope but could easily have made an error. That could very well be the answer. Thanks very much:)

[Jack Bauer]

I went down to the car today and changed to the correct 80 ppr setting for the encoder. Sadly however this didn't fix the problem although the motor does run much smoother. The main problem seems to be very high motor amps from the start. With the second parameter set, basically changed fslipmin to 2 and flipmax to 10 and slipstart to 10%, I get a very good response for the first few mm of throttle travel. But as soon as I push the pedal the motor amps max out , current limit kicks in and the car creeps.

I have attached a graph of IL1 and IL2 vs Speed for first gear and gradual full throttle pedal travel. In neutral I can rev the motor to 6000rpm no problems and regen it to a stop. If I try and start in a higher gear the motor amps max out and the car does not move. I'm going to trailer the car back home in the next few days as it will be easier to do tests when it's here. Any help much appreciated:)

[Jack Bauer]

One other thing of note is I think this is a delta connected motor if that makes any difference?

[johu]

Only thing that jumps out is the high fweak. Which I calculated to be more like 80Hz. Maybe try the ppr part of the tuning script (only works with over uart)

[Jack Bauer]

I had set fweak high to try and reduce motor current. Almost like the motor is not making any back emf.

[Jack Bauer]

Found this interesting data sheet.

[johu]

Hmm, very inconsistent. So here we have udcnom=288V and fweak=100Hz.
Could there be some damage to the rotor?

[Jack Bauer]

I guess it's possible but I hope not. Going to get the car home tomorrow so hopefully will get to the bottom of it.

[Jack Bauer]

This is probably nuts but what if the encoder bearing data sheet is counting 80 pulses over 2 channels and its really a 40ppr per channel output?

[johu]

Sounds unlikely, but worth a gentle try.

Another thing that gets me: they state the "limiting" speed for the 80ppr bearing to be 5600 rpm but the motor is rated 9000rpm. The shaft dimension for the 80ppr bearing is 40-45 mm. Is that the diameter of your motors shaft?

[Jack Bauer]

Well, got the car home today and tested the spare motor. It is indeed 80ppr on the encoder. Did a little bit of playing and still pritty much the same. ie hits current limit when trying to accelerate. This gets me thinking about a possible hardware problem with my Enova board. What would happen if the phase order or sequence got messed up? Like high drive swapped for low or something like that?

[Jack Bauer]

Ok forget that as I doubt it would have ran my Toyota hybrid motor correctly in this video if it had a flaw :

[johu]

Could it be that you are only running on 2 phases?

[Jack Bauer]

Could it be that you are only running on 2 phases?

I think that is very likely. Or 2.5 phases. When I modified the 3 phase cables on the Enova inverter I had to remove the gate connection board. They use these silly push on connectors and you have to line up all 12 in one go. I bet one didn't seat properly and only 5 igbts are getting drive. Will check asap.

[Jack Bauer]

Well, so much for that theory. All 3 phases working fine. Running in neutral in closed loop mode showed 62,66 and 68Amps peak on each phase. In gear with handbrake on showed 289 , 176 and 92A per phase with no rotor rotation. It just will not make any power regardless of what parameter combo I try and just hits current limit. More investigation tomorrow.

[Jack Bauer]

Measured the stator inductances :

216uH , 217uH and 215uH

for comparison the Tesla front drive unit gives :

379uH , 380uH and 358uH.

Is there any differences in driving a delta wound motor as I think this is?

[Jack Bauer]

Well I think I found the problem. The low inductance makes motors REALLY particular about their parameters. Am now getting to the point the car is driving but still a way to go. More updates soon:)

[johu]

Yes, the values and step sizes you used for finding a good setup would be really valuable.

Have a good new year

[Jack Bauer]

Parameters attached. I'll do a video on the procedures in a few days. The lower the inductance the more particular the setup. Happy new year