Ultimate lightweight wheel-motor concept

[madin88]

Double row bearings inside the freewheel would be my preference as it would be a much higher quality support.

a double row bearing would give better support for sure, but since the freewheels bearings only spinning when pedaling without motor assistance, i think the two internal sealed single row bearings at each side of the freewheel will be enough.
The freewheel on the picture above is "flux and fly" brand. it will be hard to find one with better quality on the market.
i have to look if i can find some norm bearings for stator case support with an inner diameter as close as possible to the freewheels flange diameter (1mm smaller would be perfect). not much material there to machine turn off. at home i have an FAG roller bearing catalogue here (german company). i got it in technical school.

 

[Miles]

Ref: http://www.endless-sphere.com/forums/viewtopic.php?p=867754#p867754

 

[madin88]

Ref: http://www.endless-sphere.com/forums/viewtopic.php?p=867754#p867754

this freewheel looks very similar to the flux and fly. they also have two sealed bearings, but made of AL.

miles, do you think it is a good idea to fit the rotor and case bearings to this freewheel?
do you think the inner flange will still be strong enough if we machine some notches in there to fit on the cassette?

 

[Miles]

do you think the inner flange will still be strong enough if we machine some notches in there to fit on the cassette?

The notches for the removal tool, on the inner part, appear to be within the thread core diameter. You could bore out to the minor diameter of the cassette driver and then cut the spline slots. If it's hardened Chro Mo., you might need to ask for Thud's help but it's possible that, like the ENO, only the outer is hardened. I doubt the alu. alloy version would strong enough...

 

[PRW]

what happened to this thread - it has all gone very quiet?

 

[speedmd]

Sorry to all for dropping this. Not forgotten. Playing catch up at my day job.

Some items we were waiting for were suitable / possibly stock magnets. Last I remember Luke had some new sourcing possibilities to report on if he had time. Once we have that component selected, we may be close to finishing prototype core details.

 

[speedmd]

Some thoughts I have while laying out the various options on the center bearing /freewheel and power durability issues have me concerned.

If we start putting the ponies to it, we will need a much more rugged clutch. Something like a automatic tranny sprag looks like it may be a suitable candidate. Not sure if they will run long without issues with only light grease or oil. Should have no issues with max power and would eliminate the need for double row bearings.
http://www.summitracing.com/parts/tci-373601/applications/

 

[johnrobholmes]

Bravo fellas! Very great thread here.


I have some solid magnet vendors available and would be happy to pass them on or tack on to our regular orders. Would be about a buck per magnet in the size and grade you are looking at, qty over 300 and nickle plated. Typical time from order to arrival via air is 4 weeks.

 

[speedmd]

Nickle plate should bond well, but not sure if adhesion will be as good as bonding to the virgin surface. Imagine there is a spec. surface that is ideal for this. If your in contact with them it may be something they can advise on.

 

[Miles]

Some thoughts I have while laying out the various options on the center bearing /freewheel and power durability issues have me concerned.

If we start putting the ponies to it, we will need a much more rugged clutch. Something like a automatic tranny sprag looks like it may be a suitable candidate. Not sure if they will run long without issues with only light grease or oil. Should have no issues with max power and would eliminate the need for double row bearings.

I fear the weight would be prohibitive.... We're not yet talking about a torque level that's beyond the capacity of a good bicycle freewheel?

 

[Miles]

Nickle plate should bond well, but not sure if adhesion will be as good as bonding to the virgin surface. Imagine there is a spec. surface that is ideal for this.

Yes, I believe the weakness is in the bond between the nickel plating and the magnet. I think epoxy coating might be better and would also give electrical isolation. It would also be easier to remove, if you wanted to bond direct to the base surface The specification of the coating (temperature rating) would need to be as high as that of the epoxy used to attach the magnets, of course.....

 

[Miles]

Bravo fellas! Very great thread here.


I have some solid magnet vendors available and would be happy to pass them on or tack on to our regular orders. Would be about a buck per magnet in the size and grade you are looking at, qty over 300 and nickle plated. Typical time from order to arrival via air is 4 weeks.

Thanks John!

 

[johnrobholmes]

From personal experience, raw neo magnets are extremely hard to bond without a closed environment. They surface rust almost immediately. Simple sanding and degreasing doesn't cut it. Cobalt magnets are easy to work with, but doesn't seem like the proper avenue for this motor.


Other treatments are available. Epoxy, nickle, zinc, gold, and tin. Nickle is cheap and pretty easy to work with, I've had no problems of bonding with proper surface treatment, even motors that get bashed on rocks don't pop mags. Zinc is really really easy to bond to, just not as good for corrosion resistance. Epoxy's downfall is cost and thickness. I would vote zinc or nickle from the manufacturing standpoint. They have the best tolerance and ease of surface prep.

 

[Punx0r]

That sprag clutch is elegant and cery robust-looking, but as with all freewheeling clutches it's added weight that isn't active material. When people are considering parts CF to save, say, 100gm it seems contrary to put back in <500gm of heavy duty clutch. That weight could be used for more copper

 

[madin88]

why not do some machine work on the flux and fly freewheel and mount the motor parts to it?

if the airgap is big enough, we could wrap some kevlar bondage around the magnets and rotor part. glue with epoxy. this is done so also in very high rpm rc model inrunner motors.

 

[speedmd]

There are some nice light weight sprag bearings, but with this diameter we may be out of luck finding a stock one that is light. Simple freewheel will last a bit, but it will depend much on the gorilla at the throttle. :lol: Direct mount most likely the lightest way to go as long as you don't mind spinning the motor when pedaling off throttle. May be a good way to re-gen on the slight down hills! 8)

Slight phosphoric acid etch should aid bonding the raw metal. Either zinc or nickle should be fine if just a thin flash coat.

 

[flathill]

Expensive and flux inefficient but this how Honda encapsulates their magnets with lams to make it bulletproof

(at least in the motor, it is easy to break the lams while removing unless you use plastic shims sliders, I have two these motors)

where it goes (flat wire for the win)

 

[johnrobholmes]

Slight phosphoric acid etch should aid bonding the raw metal. Either zinc or nickle should be fine if just a thin flash coat.

That's certainly an option, although I won't volunteer to be the worker sitting in front of the acid bath. I don't see much reason to over complicate what can be done with sanding/ sandblasting and regular degreasing. Since the design is inner rotor, it wouldn't cost much more to cut back iron lams so the magnets are held. It costs nothing to alter the magnet shape slightly to key into the back iron securely. Added benefit of such a design is jigless assembly.



If a run of these get made, count me in for one. I would love a superlight road bike commuter.

 

[Miles]

It costs nothing to alter the magnet shape slightly to key into the back iron securely. Added benefit of such a design is jigless assembly.

Are you saying that there's no premium for non-rectangular shapes?

 

[madin88]

It costs nothing to alter the magnet shape slightly to key into the back iron securely.

is this doable with this design? i mean there will be a gap between the magnets.

 

[speedmd]

I do also like the dovetail magnet fixturing method if there is enough room between to do it and the magnets can be made with a undercut without costing too much or loosing too much flux/power.

Bonding will work as long as you use a good high temp epoxy or make sure you don't get the motor too hot. Looked at several data sheets of high temp metal filled epoxy that have great heat transfer also. Maybe some upside to bonding in that respect. Best to price magnets both ways and see if it may be worth looking at this further. Hopefully johnrobholmes, you can get some more definitive numbers.

Phosphoric acid wash is common, safe, and used in most any quality auto body shop that deals with rusty metal. Low concentrations 10 -15%, dissolves /neutralizes/converts rust and gives a great surface to bond to and will not burn the skin on direct contact. I use it all the time and have for decades. Perfect for a porous irregular surfaces with rust that can not be fully removed with conventional means as on old cast iron. It is worth wiping some on/ rinsing it off, if surfaces are rusted at all or the rust just keeps growing and growing and eventually lead to things like bond failure in much shorter time frames.

 

[johnrobholmes]

It costs nothing to alter the magnet shape slightly to key into the back iron securely. Added benefit of such a design is jigless assembly.

Are you saying that there's no premium for non-rectangular shapes?

Low qty runs are made with EDM, the cost increase is very marginal unless waste material and tool path length are significantly affected. We are talking pennies per part, maybe 5 cents if you widen the base just a smidge. Don't design it for 5 axis machining :wink: Above around 5k pieces in a run and a molded magnet becomes more cost effective.


If we were using an off the shelf magnet there would be a cost difference. But we want curved , maybe bread loaf shape on a custom diameter. Optimize the motor, don't try to save a few bucks and screw up those last few point of efficiency.

Feel free to make them any shape. Typical tolerance with edm is +\- 0.002". Laser cut lams are +\- 0.005" if I recall correctly.

 

[johnrobholmes]

I do also like the dovetail magnet fixturing method if there is enough room between to do it and the magnets can be made with a undercut without costing too much or loosing too much flux/power.

Bonding will work as long as you use a good high temp epoxy or make sure you don't get the motor too hot. Looked at several data sheets of high temp metal filled epoxy that have great heat transfer also. Maybe some upside to bonding in that respect. Best to price magnets both ways and see if it may be worth looking at this further. Hopefully johnrobholmes, you can get some more definitive numbers.

Phosphoric acid wash is common, safe, and used in most any quality auto body shop that deals with rusty metal. Low concentrations 10 -15%, dissolves /neutralizes/converts rust and gives a great surface to bond to and will not burn the skin on direct contact. I use it all the time and have for decades. Perfect for a porous irregular surfaces with rust that can not be fully removed with conventional means as on old cast iron. It is worth wiping some on/ rinsing it off, if surfaces are rusted at all or the rust just keeps growing and growing and eventually lead to things like bond failure in much shorter time frames.

I've been buying 1-2k runs of magnets for a few years now and slight profile changes are the least of cost worries. The biggest worry IMO is the assembly time and lamination cost. Miles has already gotten switching frequency low enough for cheaper lams. As long as winding isn't a million turns you just need a decent assembly method and the motor is ready for low volume or high volume production. Design for manufacture is almost already done. A multi part case is about the only thing left to minimize material cost, but a full run would be needed to investigate labor vs material trade offs.


Depending on stator lamination costs, I would suggest the outer alloy housing allow for rebuilds. At least on the first few. It would allow for rewinds until the right turns are nailed down.

 

[Miles]

It costs nothing to alter the magnet shape slightly to key into the back iron securely. Added benefit of such a design is jigless assembly.

Are you saying that there's no premium for non-rectangular shapes?

But we want curved , maybe bread loaf shape on a custom diameter. Optimize the motor, don't try to save a few bucks and screw up those last few point of efficiency.

I understand now. You were assuming a non-rectangular shape in any case. On my previous simulations, I could achieve a slightly lower torque ripple with flat magnets, hence my use of them in the simulations above.

 

[johnrobholmes]

Yes, I would have assumed a curved face to offer better overall performance. Have you tested a bread loaf shape? It could offer a tighter average airgap than flat without the high detent and ripple typically gotten with simple airgap radius faces. Depends on magnet coverage and stator shape, so running a sim would tell the tale.

A quick way I have found to estimate breadloaf shape is to peak the middle at min airgap, and radius the top so the edges are at max gap found on other design iterations. From there, the leading/ trailing edges can be curved or chamfered to further soften magnetic transition from pole to pole. It's basically the inverse of a flat top, but with tighter coupling and more magnetic material.