openinverter forum

I was having GREAT difficulties getting the current sensors to read accurately, lots of signal "noise" and jumping about +/-2A or more. With all the on-off-on I did troubleshooting my errors (open circuit to one inverter input and failing to set "FWD" when calibrating), I suspected that better wires with shielding might help. I re-wired all the VCU to Inverter wiring, separated the MG2 enable from those wires off to the logic power circuit and all is well!

For about 10A of input, I have il1 = 9.7A and il2 = 10.1 and significantly less apparent signal noise

Nice one. I'm back to basics and manual run mode tomorrow with delta light bulb setup..

Just an update on MY project and the overview....

I am building the VOLTswagen Squareback not because I want an EV, but because I want to LEARN more about what it takes to make one. If it were simply a matter of desire to have, I could by a used LEAF and be "off-the-shelf", or a Prius and weld the planetary and replace the pack, for about the same money or less than this project will cost. But what would be learned? Little to nothing, both have been done and are well understood.

I chose the Openinverter VCU because of the work done by Damien and Johannes to enable folks to recycle parts from hybrids and EVs into their own projects. To that end, I neither want nor expect an out-of-the-box solution for any part of my EV system; I WANT the challenge of figuring out how to integrate the parts and get them working for my application.

If one component or another is not the best suited, I will figure that out and make the necessary adaptions.

What I do not expect, is to experience condescending attitudes and for the most part that has been a success, with some exceptions.

The openinverter forum members who are doing conversions, do a great job of helping each other out and sharing the knowledge

LOOK OUT MGR! I almost have your (syncoffset) number!

Fianally, bi-directional control and no rotation OR current draw at 0% throttle, either direction

What did I learn yesterday?

- The MGR REALLY LIKES volts!!!!
- If the Sin & Cos are incorrectly wired (as in electrically backwards), there is no "correct" syncoff setting as you cannot get rotation at modest voltages, but it will pull amps!
- Syncoff (FOC) set-up is more effective @ higher voltages, but I am not sure if the final results differ.

Not claiming to have the FOC tuning right, but DAMN that sounds nice!!!

ZooKeeper wrote: ↑Sun Aug 23, 2020 11:09 pm Not claiming to have the FOC tuning right, but DAMN that sound nice!!!

Fabulous

Awesome, awesome !

Looking good. Like the 'no spaghetti' look to your wiring set up .

RetroZero wrote: ↑Thu Aug 27, 2020 1:20 am Looking good. Like the 'no spaghetti' look to your wiring set up .

I cannot work produce decent results in chaos.... Thanks for the kind words

Someone suggested, based on "the numbers" an MGR would not have enough RPM to work as an effective EV motor.

Well, in THEORY there is no difference between Theory and Practice, in Practice there is

I am pleased to report than an MGR will happily spin at 4000 RPM on well fewer than 200v and yes, there is some proof of this

ZooKeeper wrote: ↑Thu Aug 27, 2020 12:43 pm there is some proof of this

This is good news

Please note that you can embed your YouTube videos as follows;

[media]your_youtube_url[/media]

Looking good

Would you be able to post the parameters you are currently using? I have exactly the same setup that I’m about to start testing so would be good to compare.

Per your request

Good to know, want to place it in wiki? https://openinverter.org/wiki/Configuration_Files

I'm surprised about respolepairs=2 and polepairs=4. How did you arrive at that?

johu wrote: ↑Fri Aug 28, 2020 1:31 pm
I'm surprised about respolepairs=2 and polepairs=4. How did you arrive at that?

Good question, I found the data somewhere, I will find and post it.

<<MAJOR EDIT, I found it, it was Paul M Holmes who found it, linked video below. Other links removed as they conflicted each other.>>

<EDIT: I set the resolver to r=1/p=2 as a test and the indicated motor RPM was 2x faster than measured output shaft x gear ratio>

VERY close to 7000 RPM!!!! So much for needing hundreds of volts to get an MGR to road speed, I guess that is now completely DISproved

ZooKeeper wrote: ↑Fri Aug 28, 2020 4:04 pm So much for needing hundreds of volts to get an MGR to road speed, I guess that is now completely DISproved

Probably need a load to be sure but that's fantastic progress

Kevin Sharpe wrote: ↑Fri Aug 28, 2020 6:17 pm Probably need a load to be sure but that's fantastic progress

Definitely! But if V provides the speed, A provides the Tq, I am on my way

Now if the damned car would get here

To answer Herr Huebner's question, it was Mr. Holmes, whom I consider an "authoritative" source as well as a number folks on this forum

One thing I noticed was I do not have the smooth "pulseless" rotation Konstantin or Paul do
They seem to have more "whine" than I am getting, although during one test over a week ago I think I did get similar. Mostly I get some growling, almost like the phases are not producing equal Tq or something.

Any ideas???

I did some testing over the extended weekend. Somehow the potmin became slightly less than pot but only in reverse??? Took hours of testing and re-testing to determine why I was getting pwm immediately in reverse even with no throttle input and normal function in forward.

After that, some very good testing on the Tq of the motor, around 160v and <10Adc I was able to break a string wrapped around one of the output flanges, so 2" or 50mm from the center. Same string can lift the 100# / 45kg MGR! Discounting the tensile strength 50%, that suggests that the motor output was over 100lb-ft / 130nm

Given the SB was born with a 1600 rated about 40kW & 110nm, this is looking VERY promising

Some clickbait

Well I got the math wrong in the video. The real question is how much Tq @ 12" is needed to achieve 100# @ 2"??? The answer is 1/6th or 100#, or 16.67 lb-ft.

Oh well, still cool.


<Edited becose I cant spel>

According to the gear ratios and Tq numbers I can find for a 1971 VW (72lb-ft/98NM @ 2k RPM w/4.1x or 4.3x RAR (call it 4.25)), the MGR will fare VERY well.

Wheel Tq for the MGR is about 655 lb-ft or 890NM

Wheel Tq for the VW transaxle is about:
1st = 1163 / 1583NM
2nd = 630 / 858
3rd = 386 / 525
4th = 270 / 367

Yea, that'll work just fine

I can hear it now.... That will NEVER work, like starting off in 2nd all the time. Duh, NO. The motor makes peak Tq @ 0 RPM (aka: at launch), while the aircooled needs to spin up to 2000 rpm to generate peak Tq, they make PITIFUL small Tq numbers off-idle

I thought I had good torque. 2 litre bus engine (i.e. extra torquey) is 111 Nm @ 3000 RPM and that seems like plenty.