Re-wiring induction motors for more power

[lompilomp]

So I've been trying to delve into the world of EVs, and theres a lot of questions.

I was toying with the idea of making an electric motorcycle, since I managed to get my hands on two very cheap Tesla 5.3kwh modules.


So after been doing a bit of research online and on Alibaba, it seems like the prices for new motors is a bit extreme. As in $800 for a 10kw motor (just the damn motor).

I was wondering if anyone had any experience re-wiring/re-winding a normal induction motor for much more power. I ask this because such motors are readily available locally for super cheap, and they are pretty stout compared to what's coming out of China these days.

I don't have much experience in this field, but looking at the size of the zero motorcycle motor, which is maybe 8 inches in diameter and 4 inches deep, weighing around 35lb, puts out over 60hp

In comparison, it seems that 1hp in an induction motor is around 25lb.(this goes down as they scale up slightly)

Some quick maths reveal that it produces less than 1/40 the power per weight.

So, looking online, one can find much "beefier" 3 480v phase motors for near nothing, compared to other alternatives. But obviously, their power output is 6 fold less than those used in the Zeros.

While such motors were obviously designed for their specific operation, if you were to cut and put whatever coils can go in parallel so (maybe even without having to rewind the stator, just cut and reconfigure). And maybe even lower the rotor diameter a bit on a lathe and put some neo magnets in the proper configuration. To turn it into a pmsm. (Which supposedly output almost 2x per weight unit compared to straight up induction motors)

This should allow the motor to produce much more power at low voltages, and at high voltages should be a beast.
Obviously heat comes into play, but that's something that would need to be looked at in operation.

Again, if this is impossible and I'm just spitting out gibberish correct me, I've never done such a thing before, but I was just thinking and I guess you never know what you can expect unless someone does it, so... Has anyone done something like this?

[slow67]

The hp rating on the nameplate of those motors is typically continuous. The rating on the zero motor is far from continuous, it is only for short bursts.

One simple way you can get more power out of the motors is to take a Wye wound motor. You then re-terminate the ends to Delta wound. This effectively makes the motor make 1.73x less torque per amp, BUT it can take 1.73x more amps for the same amount of heat. This lets the motor spin 1.73x faster for the same voltage, and hence a motor that once spun too slow (say an 1800 rpm 24VAC motor) can now spin much faster for the same voltage.

The other thing you can do is to simply push more amps to the motor than listed on the nameplate (obviously only for short durations because of heating).

Look up the Buick LaCrosse hybrid generator/starter assist. Its rated at 15-18kw, but for short durations they got ~40hp out of it. Maybe more with a beefier controller (they were only using off the shelf smallish controllers).

[lompilomp]

That looks interesting, if only the operating voltage was able to be lowered. It seems like the guy got 36hp out of it at 96v. After doing the delta mod. I mean like I might be able to get another module for 18s but hell, that's a lot of battery and it would only be 66v nominal, though at 1000A peak so producing the power shouldn't be much of an issue, it's just getting the motor to accept it.

Obviously that voltage is less than ideal, it would be MUCH easier to handle if it were 12s37p per pack rather than 6s74p, I'd be able to get over 130v at 500a, which would be pretty much perfect, I would say extreme overkill as it would have more capacity than the longest range Zero there is, but the packs aren't all that big.

Alas, since I cannot easily modify these modules to any stretch of impermanence (it involves a lot of cutting of the buses, and the original Tesla BMS wouldn't work with it anymore, I guess I'll stay looking to see what I can cook up. I'm not one to look for perfection in anything but rather, efficiency. Like what can I do with X amount of money and a reasonable amount of time, and sometimes they're not even on the same boat. (When there's no money there's time, but when you get money, there's no time).

Obviously I wouldn't expect the performance of a zero motorcycle that's been specialty designed and built electric from the ground up, but I can do whatever I can to a reasonable extent to make sure I get as darn close as possible.


Do you know of other similar motors that are maybe slightly larger than that? I remember reading in one of those forums that it weighs 18lb, which makes me think its rather small still, compared to the zero motor.

[Stephen Darbey]

Thought rewinding was dead.

Last year I completed a rewind of a originally 7KW ACIM. Still in my shed untested but when I complete a power section for a Rev2 inverter I will use it and test accordingly. Wired in delta mode and using 8 x 27 metres of 0.91 gauge cable approx per phase. The process itself is almost therapeutic, until you make mistake of course. I will post results if you like. There would be a serious financial advantage if successful as a typical Leaf type motor fetches anywhere from 700-1000 euros on this side of the pond. Of course you don't have to worry about resolver as 'an almost homemade encoder' should do the trick. The investment for me was a large roll of enameled copper, some insulation paper and a load of reading. If it works the next five motor rewinds will be free. As you say, the motors are cheap as chips and plentiful. I am confident that it will pull an old motorbike along with loads of excess. That's the only guinea pig in my hideout at the moment. Proof of concept will lead to bigger and better things, not to mention the learning curve.

However PMSM motors as we all know are lighter and more efficient, hence the price. But as a DIY project I am hoping the above will bring its reward. Save the real money for a battery pack.

Will find out soon enough, I hope.

Regards
Stephen

[arber333]

I had my ACIM motor revound by a professional who also recalculated winding distribution and adjusted parameters to what my transmission needs.
ACIM was 75Vrms and i used it with 120Vdc voltage. After new winding i got 180Vrms calculated and motor should be used with 320Vdc, but i use 385Vdc. I am not sure it gives more power, but motor runs much cooler at higher voltage. Ergo you can sustain more continuous power. Also i had a feel it wanted to run on the high side of RPM. There is definitely better cooling at 4000rpm since its built in axle fan.

Now enter Johannes. We discussed small phase voltage adjustment because of winter and generaly you need some voltage overhead because battery sag across SOC. I had my motor wound too low because of that overhead and at high SOC it was quite jumpy, but sluggish at lower end of SOC.
Now that software edition was Nobel prize achievement. After carefull tuning motor with voltage correction runs smooth across whole SOC range as it did before at only a narrow voltage.

[MattsAwesomeStuff]

I don't really have a good grasp of why or when to rewind a motor, and I'm a bit scared my motor won't work in the application I'm using it.

I have a 48v forklift ACIM motor. Rated for I think ~1000rom.

I intend to run it on, I dunno, haven't decided, 200-600v perhaps. And need it to spin at 5000rpm to hit highway speeds.

In what ways will me running it at higher voltage than intended (without rewinding) have consequences? I don't even really grasp the big picture tradeoffs of why you'd do A vs. B.

[Isaac96]

Well the first problem is the strength of the insulation. Running 48v insulation at 480v will most likely cause quite the sparks. So chances are you'd want to rewind the motor just so that it won't short immediately, using wire rated for the right voltage and temperature.

Now on this part I'm less sure.
-Rewinding might change the V/F curve, basically the point at which field weakening becomes necessary/power is maximum.
-It could also change the current handling capacity; fewer thicker turns would give more current capacity, but might increase RPM (??not sure about this, I have only rewound R/C BLDC motors??)

As far as I know, rewinding mostly just changes the 'rated' RPM and maybe the available torque -- of course you can't improve both, but you can make tradeoffs. And you do need proper high-voltage wire.

BTW most of the inverters we are all using are only good to 400-450v; the Leaf can do 420 afaik, Tesla is okay to 450; Chevy Volt inverter has 650V IGBTs, but who knows about the rest of it (drivers etc.)

So there's my 2 cents; don't run a motor on 10x its rated voltage. Not sure about the rest. Maybe choose an inverter while you think about rewinding that motor?

-Isaac

[arber333]

Well the first problem is the strength of the insulation. Running 48v insulation at 480v will most likely cause quite the sparks. So chances are you'd want to rewind the motor just so that it won't short immediately, using wire rated for the right voltage and temperature.

This is not debatable. You have no excuse not to go rewind it. You may select phase voltage somewhere in the range of 180V - 200Vrms so that when you multiply with sqrt3 you get somewhere 320Vdc - 360Vdc which is your battery nominal voltage. I recommend it since your industrial stuff like power supplies and chargers are made for 1phase 230Vac or rectified 320Vdc. They can accept up to 400Vdc input and makes your life much easier.

Now on this part I'm less sure.
-Rewinding might change the V/F curve, basically the point at which field weakening becomes necessary/power is maximum.
-It could also change the current handling capacity; fewer thicker turns would give more current capacity, but might increase RPM (??not sure about this, I have only rewound R/C BLDC motors??)

Yes it does. I actually have much sloapier curve now that FW region moved from 128Hz to 148Hz. But i like to run the motor at 4000 - 5000rpm now.
I recommend you dissassemble it and inspect its bearings and rotor construction. My rotor was designed up to 7000rpm max so i cant run it beyond that without risking obliteration . And if you are so far aas to pull the rotor off motor, then just get new bearings that are designed beyond your new speed. That is peanuts compared to other expenses.

BTW most of the inverters we are all using are only good to 400-450v; the Leaf can do 420 afaik, Tesla is okay to 450; Chevy Volt inverter has 650V IGBTs, but who knows about the rest of it (drivers etc.)

Dont mistake IGBT rating with actual capability. You need to consider voltage spikes. I would say Volt IGBTs are good to 420Vdc and not beyond.

[MattsAwesomeStuff]

This is not debatable. You have no excuse not to go rewind it.

Bah. I have lots of excuses!

"I don't want to"
"I'm too lazy"
"That costs money"
"It'll take too long"
"I don't know how"
"I'll probably screw it up"

See?

Some of those excuses even survive the validity test to becoming actual reasons.

You may select phase voltage somewhere in the range of 180V - 200Vrms so that when you multiply with sqrt3 you get somewhere 320Vdc - 360Vdc which is your battery nominal voltage. I recommend it since your industrial stuff like power supplies and chargers are made for 1phase 230Vac or rectified 320Vdc. They can accept up to 400Vdc input and makes your life much easier.

Temperature is handled, I forget the rating but I don't think I've ever seen a forklift motor that wasn't rated for the highest temps. "H" class?

And, is the wiring they use really only rated for 48v, or is it like everything else on industrial equipment, it's good for far higher? Would they even use special wire with thinner insulation than they would've in industrial equipment rated for 600v?

Also, guys used to run 48v DC forklift motors to 144v (or higher, 400v in one popular example) all the time. Is mine likely to be any different?

I.E. Is there anything to gain by swapping the wire itself? I'd certainly feel foolish taking the time to re-wind it and end up using the same wire because it was fine to begin with.

And if you are so far as to pull the rotor off motor, then just get new bearings that are designed beyond your new speed. That is peanuts compared to other expenses.

I suspect the original bearings are going to be better quality than whatever I buy at a supply place. They seem perfect, so, other than "things to watch out for" I'll probably leave it as is.

If I'm going to rebuild a motor, I might as well just swap to using a Toyota MGR or Prius or the LS450H gearbox instead.

I'm more worried about the number of turns in the coils and things like that. I guess, I don't understand the impact of building a motor for one voltage versus another, and what the practical differences would be if I had two otherwise equal motors side by side, one wound for low voltage one wound for high.

[arber333]

Hah! Mere assertions !
I know you want to mess with that motor.

Well with 48V you will have difficulty finding good inverter power stage to handle enough power.
But If you use more voltage and extend FW knee you will either have a lame motor or you will run hot all the time. I tried to runmy car on 72V motor with 300Vdc and it was strange.

[MattsAwesomeStuff]

Hah! Mere assertions !
I know you want to mess with that motor.

Shut up. Get out of my head. Feature creep kills completion. I want a car I can drive. Stop taunting me with optimizing :p

Well with 48V you will have difficulty finding good inverter power stage to handle enough power.
But If you use more voltage and extend FW knee you will either have a lame motor or you will run hot all the time. I tried to runmy car on 72V motor with 300Vdc and it was strange.

Hmm. My Prius Gen2 MG2 can handle 350A. 48V lead acid is actually 60V when full. So, that's 28hp. So, technically, I dunno, maybe I could, but I doubt I'd get enough RPM at 60v.

What is it, mechanically, that winding for a higher voltage accomplishes?

Higher RPM per volt? (or lower?)

I have no intention of running it at 48v. I intended to run it at 200-600v (don't know where in that range yet).

I mostly have no intention of rewinding the motor unless it's necessary.

It's spec'd for 48v and about 1000 RPM.
I'm giving it, say, 250v and wanting about 5000 RPM.

I don't really know what it means to overvolt a motor using common terms I could understand (beyond insulation rating, which I think is probably fine).

[SciroccoEV]

There's a disadvantage to Delta windings. With a Y connection you have a floating neutral point and any imbalance in your winding characteristics mostly just makes the neutral potential vary more. With Delta the imbalance currents cause extra heat in the windings.

To take the example of the 1hp industrial ACIM, it's going to produce full rated continuous torque at 50 or 60Hz, which for a 4 pole motor is approx. 1400rpm. To generate the same torque at 100Hz - 2800rpm (and 2hp), you'd need to double the supply voltage. Or, you could rewind the motor to run on half voltage at 50Hz.

The motor in an AC G-Wiz has a nameplate rating of something like 28V for use on a 48V system.

[MattsAwesomeStuff]

There's a disadvantage to Delta windings. [...] With Delta the imbalance currents cause extra heat in the windings.

I don't know if mine was wound Delta or Y. I haven't cracked it open. Spec's don't tell me.

However, mine was not an industrial ACIM connected to the power grid, it was used in a forklift.

Here is my nameplate:


(I was wrong, it's Class F insulation (155'c) not Class H (180'C).

It says it was for 3-35 V ac, 50hz. 961 RPM... but meant to run on a 36/48v battery.

It's a 6-pole motor.

Despite Arber's attempt to make sure I never finish my car, I still have no intention of rewinding a motor with different wire. But, since Delta/Y swapping is just changing some endpoints of the coils, I might crack it open to do that.

To generate the same torque at 100Hz - 2800rpm (and 2hp), you'd need to double the supply voltage. Or, you could rewind the motor to run on half voltage at 50Hz.

If I understand this correctly, in order to get my motor to spin 5x as fast (I desire 5000 RPM for highway speed), I need to give it 5x the voltage anyway (5x... 35? 36? 48? 60?). So, my motor might already be correctly wound for the voltage I was hoping to use anyway? (5x voltage, 5x RPM?).

[arber333]

Well yes and no. You can use higher voltage with this motor, but at existing wiring it has very small iR (wiring resistance). This means if you give it higher voltage its winding will need the same ammount of boost as before. But you have more voltage available and inverter will wind up V/Hz faster. So motor will be jumpy and it will overheat etc...
You can solve this also, but it means to compromise the start by drawing the FW point further from the optimal. Motor will feel sluggish, but it will be drivable... I dont recommend it though, because such a car will not be fun to drive.

I see your motor is rated for 18kW for an hour. This means it will probably give 40kW out in short bursts, maybe more. That depends on how you handle the heat.
My opinion would be either you rewind it and use original transmission or use a Lexus transaxle. This would be even more fun.

EDIT: And you drive ACIM motor by frequency, so you can develop 3000rpm even with 48Vdc. Maybe even 5000 if you use more FW. I think you could get away with 80Vdc and motor would not feel so much different. But 500A with 48V and 80V is a lot more power. Might be good to try.