Ultimate lightweight wheel-motor concept

[madin88]

Fall from what?

At 400rpm: 132W copper losses, 13.4W iron losses, efficiency 90.5%

At 800rpm: 132W copper losses, 34W iron losses, efficiency 94.3%

nice to see efficiency increases with rpm at same torque output
800rpm means about 80volt
x 40amp are 3kW output and still >94% efficient which is awesome

with some bigger stator this motor would be a monster, but keep make it lightweight for legal bicycle commuting

 

[Falco]

Fall from what?

At 400rpm: 132W copper losses, 13.4W iron losses, efficiency 90.5%

At 800rpm: 132W copper losses, 34W iron losses, efficiency 94.3%

This is a typo correct?
With 2x the power but spinning twice as fast it is more efficent??

 

[Miles]

With 2x the power but spinning twice as fast it is more efficent??

Yes. It's more efficient, at that torque level, because the copper losses are the same and the iron losses are still a minor factor but in order to double the speed, the voltage has doubled. So, losses have gone up a relatively small amount for the doubling of the input power.

 

[Falco]

With 2x the power but spinning twice as fast it is more efficent??

Yes. It's more efficient, at that torque level, because the copper losses are the same and the iron losses are still a minor factor but in order to double the speed, the voltage has doubled. So, losses have gone up a relatively small amount for the doubling of the input power.

Thank you for explaining. I think i confused myself thinking km/h would double and be more efficent.
This concept is very exciting.
1kw light weight motor like this was just what i was looking for. I have an idea for the torque arms for this that would work universally on bicycles. Just need to make a sketch.

 

[speedmd]

Simulation at 40 amps:
X-76t-68p-40a-100degc.pdf

Hi Miles

Glade to hear your feeling better. Great work on the design concept. Looks like it will handle much more power efficiently before it fully saturates. Looks like it will run very efficient/ cool at the 500 watt output level. Just perfect for a light weight commuter. Look forward to building one. Local machine shop I use often is happy to quote the case prototype/production when we are ready. My lathe is a bit too small to turn the outer diameter details like cooling fins and such, but I can make the inners on a few for certain. What do you think we will need for plate thickness to hog a stator housing out. I have some drops of 1 inch plate big enough for one set aside.

 

[speedmd]

1kw light weight motor like this was just what i was looking for. I have an idea for the torque arms for this that would work universally on bicycles. Just need to make a sketch.

Hi Falco

Perfect timing, look forward to seeing the sketches.

 

[Miles]

What do you think we will need for plate thickness to hog a stator housing out. I have some drops of 1 inch plate big enough for one set aside.

That might be enough if the housing only covers to the far edge of the stator core, or less.

 

[speedmd]

I will keep my eyes open for a thicker plate to mess with.

 

[Miles]

I guess it's not such a bad idea to use the stator core to register a joint between the two case halves.......?

 

[madin88]

Yes. It's more efficient, at that torque level, because the copper losses are the same and the iron losses are still a minor factor but in order to double the speed, the voltage has doubled. So, losses have gone up a relatively small amount for the doubling of the input power.

what would be the highest rpm for this motor before iron losses get on top? Or do you think it will be limited by controllers ERPM due to the very high pole count?

 

[Miles]

The latter.

 

[Falco]

Now this is a really rough sketch as all i had was a pen and paper and no drawing skills and im at work in a very remote place
The idea being that there are a few rails up the side of the case that these torque arms slid on. These clamp to the swingarm. I say a few rails as it would be ideal to spread the force along the bike frams so its not in putting all the force in one spot but we might only need one if we place them as far from the center of the casing as possilbe.
Hope this sketch makes sense to you. I would make it all out of aluminum.

 

[Miles]

A two part case, along these lines, is a possibility. It would give fixing points for the reaction arm, too.

 

[speedmd]

A two part case, along these lines, is a possibility. It would give fixing points for the reaction arm, too.

Looks like a multi pc design can take advantage of using thick aluminum pipe and thin water jetted or laser cut plate. Looks near impossible to overheat the way you conceptualized it. My head is spinning with Ideas/ possibilities!

 

[Science Faction]

Absolutely loving this Thread.
Miles you are really inspiring me to learn emetor and better understand the intricacies of magnet machines.
One question though, as you guys are seeking the ultimate lightweight motor wheel concept:
Is 3 phase the ultimate configuration?
How about 5 phase like the Falco?
Or 7 phase like the wavecrest?
Or 23 or more phase like the Axiflux?
ref: viewtopic.php?f=34&t=48058
Obviously I'm not sure what controller can run it.
Maybe Lebowski's sine wave controler can be extended to more than 3 phases.
Cheers
Paul
Ps I know a fantastic laser cutter in Murwillambah, Northern NSW Australia, Michael Dowling.
He has one of the fastest cutters in the world, 100 made by Mitsubishi, Boeing have the other 99.

 

[speedmd]

Thanks Paul

You bring up good points on control side of motor design. I am not qualified to answer, but feel also we should air out challenges on that front before heading into a build. For a light weight build, we may just want to make it as simple, reliable, efficient and practical as possible this round. Certain we can wind / wire it differently if the protos prove out that it is close to simulations for various more advanced control topologies, at least for a conventional style BLDC motor. We are open to anyone with access to advanced simulators that can model and share results of other control scenarios and help advise on that front. Hoping next month some time we can at least start on some prototype planning for the first generation ULW motor build.

 

[Joe90]

Hi,

Here is my idea.

Hollow center that bolt on brake + freewheel adapter for single speed.
Or use an adapter for cassette

We can lace a front diske brake hub to suit this.

Side are black G10
Stator shrink fit
Overall diameter 329mm
thickness : 35mm

Motor
90S 100P
95% eff at 400 rpm and 38 km/h load of 550 Watts
Active weight 2200g
N35 magnets
Regular electrical steel.

 

[madin88]

Here is my idea.

Hollow center that bolt on brake + freewheel adapter for single speed.
Or use an adapter for cassette

We can lace a front diske brake hub to suit this.

this is it! one of the best ideas posted here in this thread. it makes the motor as universally usable as it can be on a bicycle. GREAT!!

a second, internal freewheel only would be needed on cassette side. on brake side we want to have the regen function.. or not?

cassette side: freewheel with thin, large diamter bearings. mount the rotor and the bearings on the sprocket part of the freewheel and the motor case on the bearings. fix the adapter for cassette mount on the inner thread of the freewheel (M35 thread i think?). maybe there are some buyable freewheels we can machine work to make them fit.

brake side: only the bearings between the case and the adapter for centerlock / brake rotor flange.

we need to figure out how much space there is left on brake side. as far as i know on the front there are normally only about 10mm left between disc flange and fork which is not much.
on the rear the motor has to be as close as possible to the wheel hub or/and the spokes - otherwise it will touch the swing arm.

place this nice drawing to a CAD frame with hub and wheel and figure out how it will fit.

 

[liveforphysics]

Using the sim to estimate the efficiency assumes a perfect drive with no induced harmonics.

You will get that performance if you are able to tune out your harmonic content and ringing with your drive. Extremely fortunate for us, Lebowski's controller might enable you to have super low ringing with about any BEMF shape.

 

[Joe90]

yes, this is simulation not real world.
I dont know what is ringing but I guess that is mismatch between bemf and drive current.
From what I understand this cause torque ripple.
The bemf look sinusoidal but it s not. no more flat top.
There is also the power factor that ask the drive to work more for a given motor output.
This is inerent to motor configuration. short motor seem more problematic for this.

On simulation it s possible to use square or sine drive current. it make a big difference.
The torque ripple is much lower.
I guess that what people speak about when the say that a Sine drive is silent and ultra smooth.
For efficiency it makes no significant difference. This is surprising since mismatch between drive and motor should show more.

I know that after the fact I may get many point lower efficiency but I feel that we should try to built something that show good potential to start with. I know people that were computing 98% on paper and ended up with 96% mesured on the bench.

Based on this after hand building and winding and drive mismatch I would be happy with anything over 90%
With careful atteytion to all details and more experience I beleive that some people can push this higher but at greater expense like everything else.

From what I read a given package that is designed for more efficiency will be less powerful than the same designed to maximise torque. I dont think we are making a neck snapping wheelee motor here. it will push gradually without much heat and add maybe 10% range from efficiency alone. A recumbent versus a mountain bike would beat that any day with a cheap 9c or golden motor.

 

[speedmd]

Great work Joe90. Like the fully enclosed motor also. Changeable center hub that could adapt it to either side would be excellent. I would prefer a perfect freewheeling bike, but the regen capability is a big deal to lots of potential customers. For regen and freewheel capabilities, thinking a eddy current clutch would be killer. With it you could run the motor at efficient rpms independent of wheel speed and ride the clutch when in burn up load situations. Burn outs, wheelies anyone? 8)

For added room on the left side, thinking the simplest way would be, to build a multiple part hub with very close spoke flanges. Even if they were 25 -35mm apart, the wheel would still build up strong and have plenty of lateral strength. No added structural issues as long as the axle was strong enough to handle the load in the center of the span.

Looks like we mainly agree on size, pole and tooth counts, so we could be close to potential core details/ build.

 

[madin88]

i have an idea for cassette side mout:

for brake side we can machine turn a part with SAME SIZE as the freewheel. a centerlock flange will easy fit and i also think the bolt circle for normal brake rotor mount is smaller than the bearings inner diameter. this will make fabrication and assembly easy.
i think it will not be a problem to clamp this freewheel in a turning machine and turn a fit on it for the rotor and bearings.
what do the others say?

 

[speedmd]

Hi madin88

The freewheel would be good for motor center IMO. It would allow freewheeling of motor when you have no battery left and allow you to coast down hill like a bowling ball when you just want to hear the air whistling past. I posted on page 10 of this thread the modified ( broached ) center that fits freehubs already available by es members. A double row bearing model in steel would turn down nicely for outer stator support bearings. Should easily fit a adjacent 3 speed as miles pictured on that same page.

 

[madin88]

I posted on page 10 of this thread the modified ( broached ) center that fits freehubs already available by es members.

i do not completely know what yours idea was, but i think of mounting the rotor and the case-bearings directly to this buyable freewheel. we only need to turn two flange for the bearings and one for rotor to the sprocket part and modify the center for freehubs.
for brake side use we can make a machine turned AL part instead of the freewheel with same flange diameter for rotor and bearings. on brakeside we want to use the regen function - or not?

A double row bearing model in steel would turn down nicely for outer stator support bearings.

one double row bearing is not sufficient IMHO. two normal bearings with the rotor running between them will be more long lasting.
the bearings in the freewheel will not move if the motors is spinning. they only move if the bike is pedaled without motor.
the flux and fly freewheels have top-notch quality with very good sealed bearings.

Should easily fit a adjacent 3 speed as miles pictured on that same page.

normally you will be able to install more sprockets on the cassette beside the motor (or beside the motors freewheel), because the smallest sprockets of a big cassette (11t, 12t) are located beside the freehub and not on it.

 

[speedmd]

Double row bearings inside the freewheel would be my preference as it would be a much higher quality support. Agree on the outside shoulder being a great area for single row stator support bearings (one on each side).

Without a free wheel on the pedal sprockets, motor would want to turn crank full time. Having a freehub with two freewheels on board would free you to either pedal or motor alone.