ES Motor Project?

[John in CR]

Could handle a 10-12kw input in the 2000-2500rpm range?

That's around 45Nm of torque. I'd guess not.

Really? I must be missing something in the sizing relationship. I run my motor with the 150mm dia 50mm long stator at 6kw input limits with max rpm on the road of 1045rpm at voltage sagging to 75V. That's a sealed motor, plus I'm running it conservatively. It has plenty more to give, since it's never gotten remotely hot, so at higher voltage and maybe ventilation if necessary, I'm confident that 12-15kw input is a piece of cake at 2-2.5krpm.

I don't understand why a 112mm diameter 100mm long stator isn't more motor. It's roughly 50% more magnet area working at only 35% less radius and a better winding factor due to the longer stator. If necessary, a simple radial fan blade goes on the end for far better cooling than is possible with my more closed motor. Of course I'd like to see a larger diameter, but my plate is so full I have to be mostly a follower on this one.

John

 

[Miles]

It's best to avoid using "input power" as a criteria....

Continuous output torque is the only thing I judge by.

How did you get "50% more magnet area"????

 

[Thud]

This is a take off with some 10mm magnets paired to make a "pole"
iIn this 5" flux can it leaves a lot of "open space"....I know, more room for copper.

But there has to be a ballance point between the perminate magnet & copper density per tooth dosen't there?
it visualy scales better if we reduce the flux can to 4.5" in dia.
View attachment 18T 16P.dxf

 

[Miles]

But there has to be a ballance point between the perminate magnet & copper density per tooth dosen't there?

Yes. More directly, the iron to copper ratio, which is what Eric and I spent ages playing with on my last design.

 

[liveforphysics]

Just as some design info, the Sevcon is limited to ~500hz commutation freq (if I remember correctly).

The Kelly Ultra-High-Speed option is limited to 1100hz (in my testing, and it was getting glitchy past 1000hz).


At 18poles, this means the Sevcon would only be able to drive (500hz*60rpm/min)/18 = 1666rpm maximum

At 18poles, the kelly would be limited to about 3,666rpm maximium


I think the "infinions" can do something around 2000hz (but there have been so many versions and chips used by this point, I have no idea what current ones do), so they could spin it ~6,666rpm (possibly, but someone needs to confirm this with whatever chip they run these days.)

 

[Miles]

Just as some design info, the Sevcon is limited to ~500hz commutation freq (if I remember correctly).

The Kelly Ultra-High-Speed option is limited to 1100hz (in my testing, and it was getting glitchy past 1000hz).


At 18poles, this means the Sevcon would only be able to drive (500hz*60rpm/min)/18 = 1666rpm maximum

At 18poles, the kelly would be limited to about 3,666rpm maximium


I think the "infinions" can do something around 2000hz (but there have been so many versions and chips used by this point, I have no idea what current ones do), so they could spin it ~6,666rpm (possibly, but someone needs to confirm this with whatever chip they run these days.)

Thanks for the info Luke.

Don't you divide by the pole pairs, though?

 

[Miles]

It's worth reading this, Todd:

 

[John in CR]

magnet area comparison
150mm X 3.14159 X 50mm = 23,562mm2
112mm x 3.14159 x 100mm = 35,185mm2 49% more

Regarding power comparisons, the only hard numbers I have is input. It seems reasonable that the new motor would be of comparable efficiency, and I can reasonably estimate output in the range above peak power. That's where I operate after launch acceleration, so that's what I have to go by since I lack the tools for computing or measuring torque. Similar input with similar efficiency and rpm means similar torque doesn't it?

I'm sorry that I haven't been able to convert myself to the torque machine view, but that's because I can't can reasonably estimate torque directly.

John

 

[Miles]

John,

Sorry. You're comparison is fine. It just seemed wrong and I never checked properly (it's the middle of the night, here...).

I'm sorry that I haven't been able to convert myself to the torque machine view, but that's because I can't can reasonably estimate torque directly.

You know the velocity constant of your motor, right? So, you know the torque constant. If you know amps in, you know torque out (@ WOT).

 

[John in CR]

I'm sorry that I haven't been able to convert myself to the torque machine view, but that's because I can't can reasonably estimate torque directly.

You know the velocity constant of your motor, right? So, you know the torque constant. If you know amps in, you know torque out (@ WOT).

Homework time, so I can get this stuff till it's 2nd nature, which should be a nice addition to my subjective understanding of the practical limits of hubmotor performance and the influences of current limiting (eg we can't and wouldn't want to run our motors with controller settings to achieve max power because the efficiency is too low and they'd burn up).

Knowing the Torque constant just from Kv isn't intuitive to me, because that means motors with the same Kv have the same torque constant. I guess current handling is what makes the difference.

John

 

[Miles]

Knowing the Torque constant just from Kv isn't intuitive to me, because that means motors with the same Kv have the same torque constant.

Yes, if a motor has a certain Kv it has the corresponding Kt. [Kt = 1/Kv (SI units)]

 

[Miles]

With 10mm width magnet pairs (as in Thud's DXF, above)

 

[Miles]

Any volunteers to do the FEMM analysis?

 

[John in CR]

We wouldn't really go with something with gaps in between the magnets would we? Shouldn't we start with magnet size? Also, I hope someone has a wholesale connection for best pricing on magnets, because it's looking like cannibalizing an existing motor may be the way to go, especially for Miles' application with only 1200rpm, so premium laminating steel isn't as important. A couple of MAC geared hubbies from Cellman may be the ticket for the size Miles motor. For those who can tolerate a 6.25" diameter, the high efficiency motors I can get are probably a better choice, though it would be a lot less making than modding.

 

[Miles]

We wouldn't really go with something with gaps in between the magnets would we?

 

[Miles]

The 10mm magnet pairs give 90% coverage.

 

[ElliottE]

I may be able to provide some help with this project, I've just acquired SolidWorks Professional (there's a great special on this month, buy SolidWorks standard and get a free upgrade to Professional) and I'm learning it now. Once I get some proficiency with model building I'd like to try my hand at putting together some 3D models of motors, I'm thinking I might be able to start on this in the next couple of weeks?

 

[John in CR]

The 10mm magnet pairs give 90% coverage.

Then a 10mm paired with a 12mm would be perfect. . Gotta love those easy solutions.

 

[liveforphysics]

The field distorts locally when the opposite orientations touch side by side. The way hub motors do it costs them performance.

I've often heard the minimum placement rule of thumb is 1/4-1/2 magnet thickness spacing between opposite polarities minimum.

Once you have the magnet width, placement, and strength determined, you can begin the tooth head design. The shape of the tooth determines how the BEMF will look, and being just a half mm thicker or thinner or wider or skinnier in different places makes the difference between a motor that rips and stays cool and one that is a melty scrap heap.

Member Biff here is the only person I know of who can do that work. His wife just had a baby a couple days ago, so I assume he has his hands full at the moment.

 

[Miles]

The field distorts locally when the opposite orientations touch side by side. The way hub motors do it costs them performance.

I've often heard the minimum placement rule of thumb is 1/4-1/2 magnet thickness spacing between opposite polarities minimum.

Right. Around 80-85% coverage seems to be the rule of thumb in the RC world, 1/4-1/2 of magnet thickness seems a better guideline, though. Look at the picture of the Joby motor that I posted above.

 

[Miles]

I may be able to provide some help with this project, I've just acquired SolidWorks Professional (there's a great special on this month, buy SolidWorks standard and get a free upgrade to Professional) and I'm learning it now. Once I get some proficiency with model building I'd like to try my hand at putting together some 3D models of motors, I'm thinking I might be able to start on this in the next couple of weeks?

Hi Elliott,
Thanks - that's great. In fact, modelling a motor would be a good project to learn on. They are relatively easy to model - well, apart from the coils... I've been using Alibre for the last 7 years. Anybody else who wants to have a go, Solid Edge is free for non-commercial use: http://www.deelip.com/?p=6517

 

[Arlo1]

This is a cool project. I choose 16 magnets for my build 8 poll pairs so it could be spun to ~13000 rpm with a 12 fet china controller. At the time marko give me the max erpm the controller was good for I will go find it.
If I was to get one of these motors I would want to be able to built it for very Hi rpm. So thin lams and a carefull selection of how many magnet pairs is needed. I think the dspic30f4011 chip lebowski and I are useing is capable of very hi clock speed. I will have to go read my notes on that too. After this the last peice to the puzzle is a DIY battery...

 

[John in CR]

The field distorts locally when the opposite orientations touch side by side. The way hub motors do it costs them performance.

I've often heard the minimum placement rule of thumb is 1/4-1/2 magnet thickness spacing between opposite polarities minimum.

Right. Around 80-85% coverage seems to be the rule of thumb in the RC world, 1/4-1/2 of magnet thickness seems a better guideline, though. Look at the picture of the Joby motor that I posted above.

Interesting. Hubmotor manufacturers must use essentially 100% coverage only because alignment is easier. Still, I can't imagine them having an 10-15% in magnets for no reason with the Chinese approach seeming to be minimize copper and magnet even if it means more steel and aluminum.

John

 

[John in CR]

After this the last peice to the puzzle is a DIY battery...

We start building our own batteries then we stockpile plenty of raw materials and start our own country. 8)

John

 

[Miles]

Re. magnet size availability:

I compared the price for a stock magnet with that of the same specification created as a custom magnet using the HKCM calculator. For 200 off it's 228 euros for the stock item and 260 euros as a custom run. Not a huge difference.....

Ref:
https://www.hkcm.de/product_info.php?products_id=53836&dna=3

https://www.hkcm.de/hkcm.php?osCsid=1562cf60bc3705a1a35c1743dfa9046b&Design=Design&dna=3&des=on&mA=20&mB=10&mH=3&bo=F4