Field Weakening for synchronous motors

[johu]

As I mentioned elsewhere the current firmware doesn't explicitly support field weakening for sync motors i.e. running beyond base speed.

Now, I don't think it's hard to implement the latter, I'd just shift "syncofs" closer to the magnet to oppose and thereby weaken its field. The higher the speed, the closer. I'm not quite sure how to control torque there but I don't even think that will be hard to figure out.

What I am a bit afraid of is uncontrolled regen. So say the motor has a specified (unweakened) base speed of 3000rpm @ 350V then would expect it to actually generate 350V when you spin it externally to 3000rpm. Should you spin it to 6000rpm it would output 700V, right?

So, if the inverter shuts down at high speed for some reason, you get 700V on the DC bus. Not a problem for the IGBTs, as there are 2 600V ones in series. The cap might not like it though?

Alternatively you could keep the DC switch closed but I imagine insane currents flowing back into the battery through the freewheel diodes, right?

How do OEMs handle this? What if 12V power fails at high speed? That sort of incident surely doesn't blow up a Nissan Leafs drive train, would it? Am I seeing a problem that does not exist?

[arber333]

Hi Johannes

I see you came to the same conclusion i did when i blew my power stage.... repeatedly.
1. To control BLDC motors unprotected drivers must go!
2. You have to use a sort of UVLO and FAULT protection to shutoff PWM and release motor from battery in case of 12V brownout.
But with Leaf OEM inverter you already have this protection.

I can imagine that IGBTs would close and reverse diodes would block current after some time. In that time main cap is used to absorb any transients. That is why all OEMs use self healing film DClink cap.
Also i imagine OEMs use mainly 8pole motors that can be run with cca 60V at 1000rpm and 400V at 7000rpm such as i imagine is the case with Leaf motor. I tried to use one chineese bike FOC inverter and it worked with Leaf motor on 60V giving 1020RPM. I have a video somewhere.

I think i will ask Arlo how much RPM he can get directly from his inverter without weakening. I remember he said he uses 460Vdc so he can get 8000RPM and then 25% weakening to 11K RPM.

Here i found this document about regen
http://irtfweb.ifa.hawaii.edu/~tcs3/tcs ... ration.pdf

EDIT: I found a discussion about regen and BLDC a while back. It didnt seem significant at the time...
https://www.edaboard.com/showthread.php ... lled-Regen
https://www.edaboard.com/showthread.php ... -the-motor

[johu]

Hi Arber,

thanks for the input. So 8000rpm with no weakening sounds very promising. No 3000rpm limit after all?
I don't even want to dive into the hardware here, as I expect PMAC motor applications to use their original inverter.
So I think before writing any code I will await the results of the first tests.

[arber333]

I found Arlos comment on weakening here:
https://endless-sphere.com/forums/viewt ... g#p1303708

and here:
https://endless-sphere.com/forums/viewt ... g#p1271932

So he can rev the motor to 14Krpm with some 450Vdc and use 450A of current while performing weakening.

[johu]

Sounds like fun. Can't wait to do the testing!
Though I don't think VW transmission will like 14k rpm

[D@F]

I don't think VW transmission will like 14k rpm

14000 rpm on mine in 6th gear is somthing around 700 km/h !!!!

[Jack Bauer]

This might be of interest :

[johu]

Yes, interesting. Without field weakening 1080rpm@60V. That is 6500rpm@360V. Actually quite a usable speed.
The extra current for field weakening is really high, I'd rather shift gear

[arber333]

Yes, interesting. Without field weakening 1080rpm@60V. That is 6500rpm@360V. Actually quite a usable speed.
The extra current for field weakening is really high, I'd rather shift gear

Hi Johannes

What about those who run 200Vdc systems? Dont write them off just yet. Also we are using 600A switches with protection circuit so we can deal with more current.
Also remember true speed comes 10% or more lower than base speed due to dV and all sorts of lag in the system. I am not sure 3500rpm would be enough to run a good car on 150V or 200V.

[D@F]

yes and those who use the complete drive unit with only one gear.

[arber333]

yes and those who use the complete drive unit with only one gear.

Here here! I intend to use Leaf motor with original diff and 360Vdc system. I must get to 10Krpm at least to have 140km/h speed. If you will not support this i will be forced to use lebowski board. Not ideal but at least it can go 20% into weakening.
I rather be using your board since i have all test system prewired now. Can you make a parallel software for me to test when i have the car on the road? Like i said i have driver protection incorporated.

tnx

A

[johu]

Can you test with the current software first? I would really like to know what the ideal syncofs is and what rpm can be reached. I'm not quite ready to test my self so it will be sort of hard to know what I'm doing. However I could make a parameter like degree/Hz that tells the controller how much to advance the field per Hz above base speed.

[zippy500]

yes and those who use the complete drive unit with only one gear.

Here here! I intend to use Leaf motor with original diff and 360Vdc system. I must get to 10Krpm at least to have 140km/h speed.

I am thinking with your idea of leaf motor/gearbox fitment.

What would a 360v pack with motor rpm pf 6000 achieve as a top speed ( with 14in wheels for inistance)

what sort of RPM and top speed does a standard leaf make ?

[zippy500]

....so I found a online speed calculator

RPM-6000
Tire size 14"
Gear Ratio 8:1 ( leaf gear box ?)

Potential Calculated Vehicle Speed: MPH 79.24 or KM/H 127.52

These figures don't seem unreasonable if accurate

[doobedoobedo]

I think your calculations are off. Using this calculator and assuming leaf reduction gear is 7.92:1

225/45 R18s (standard tyre size for my car) are 778 revolutions per mile - make it easy say I'm doing 60mph, 778rpm X 7.92 = 6161.76 motor rpm to do 60mph. The leaf itself, if you take it's official top speed and standard tyre size, spins the motor up to 10,300 rpm.

Most cars without ridiculous wheels seem to do somewhere around 800 wheel revolutions per mile.

[zippy500]

OK, thanks

60mph isnt a really a good top speed then then , what is the maximum RPM thats J huebner controller will go up to ?

[arber333]

Can you test with the current software first? I would really like to know what the ideal syncofs is and what rpm can be reached. I'm not quite ready to test my self so it will be sort of hard to know what I'm doing. However I could make a parameter like degree/Hz that tells the controller how much to advance the field per Hz above base speed.

I will as soon as i manage to test the new resolver board i made (regular birds nest). Motor is actually mounted in the car now and it has manual 5speed transmission. But i have inverter at home connected to my motor for testing. The car is some 25km from me. I hope at this weekend i will see the wheels spinning.

Hm... it seems every Leaf motor has its own offset angle. It is marked under the motor in hex value or some other code. Manual says if you change motor you have to reprogram inverter with the new value.

[arber333]

OK, thanks

60mph isnt a really a good top speed then then , what is the maximum RPM thats J huebner controller will go up to ?

Sure, but you have to consider that at 4000rpm motor produces 200Nm and at 10000rpm it can produce only be 75Nm! That shouldnt demand too much current then? I imagine some 80A to maintain speed and another 30A to keep field reduced? Would my estimate be correct? Now with mazda i run in 4th and 6000rpm at 130km/h i need some 100A to maintain speed. Johannes what do you think?

EDIT: I calculate from T=9550*P/n and i get 78kW for 10000rpm where T is torque and P is power. I am sure Nissan designed graph around inverter not the motor though.

I would recommend you put a limiter in code how much (by amps or power) you can weaken field.