Ultimate lightweight wheel-motor concept

If I recall its 10mm between flats on the rear. Thinking we can go as large as we can fit in a standard bearing and still allow enough meat for a freewheel thread or similar size free hub attachment with its own outer bearing / dropout support. Originally I was thinking a outer bb bearing size as a possibility. Lots to choose from that are excellent quality. With a multi part axle we buy a bit more freedom of design in the center/ stator support as LFP has stated.
 
I'm developing a concept for a lighter, stronger, stiffer axle design that works with conventional drop-outs.


A cr-moly tube section, perhaps 3mm wall section and perhaps 20mm OD, with internally threaded ends, perhaps m18 female thread.

A pair of short high quality m18 bolts, get the area just below the head on the shank machined down to a 10mm radius for a section about the thickness of the drop-out.

You loosen the bolts to position the machined-10mm gaps at the point where they are aligned to slide into the drops, install, and torque the bolts.

This large clamping surface will be awesome for rigidity of the system, and the axle could be very thin and light this way.

Actually, since the threaded area itself would be so strong with the m18 sized thread, even if the center of the axle were aluminum or something, it would still be a stronger threaded joint than a steel nut on a steel axle of only 12mm diameter.

Anyone need a sketch? Someone want to attempt CADoodling it up?
 
How light do you want to go? :D

What if we use spokes to support the rotor and the stator and the rim?

Would it all fit? We can shield from rocks with other means.

1.GIF
2.GIF
3.GIF
 
Do you have more info on that motor?

It appears to be a rigidly chassis mounted traction motor from a solar car?

It looks like they went to pretty extreme levels to be very scary lightweight. Do you have a picture of the active parts by chance? Im curious if they use a few hundred tiny teeth as I think is the ultralight path forward.
 
aircraft

aircore so you dont have to worry about stator steel wanting to bend stator from reluctance force

this is how they get away with supporting the stator only on one side with spokes,
but I guess the steel core Joby inrunner and other steel core outrunners get away with singled sided support too

double sided halbach array, no teeth per se but high pole count (low rpm operation)

of course more poles are better for a wheel motor but you quickly reach a limit to to practical concerns (winding and so many mags) and other issues (slot fill factor and slot leakage)

only steel in the 4130 shaft
 
both

I am wondering if we could do something similar to the bionx motor but use bicycle spokes to support the stator and rotor

so six layers of spokes

good for diy kit :)

not good for mass production

but this is supposed to be the ultimate iron core non halbach motor ( we have to have limits)
 
Just a thought: If this thing fits inside the spokes then we have a triangular space in which to fit this motor limited by the wheel diameter. If the wheel is 900c and the motor diamter is 400mm then we already know the maximum width it can be. Hopefully there's enough headroom to allow use in a smaller wheel, like the common 26".
 
Toorbough ULL-Zeveigh said:
does ultimate differ from optimal?
other than being ridiculously pompous.
Click to expand...

Ultimate within reason is optimal

The ironless, carbon fiber, double halbach motor is the true ultimate but not practical

I wish we could buy an off the shelf frameless motor for a reasonable price.

If we were going this route on an unlimited budget here is what I would buy:

technology-1B.jpg


I think $999.99 is a better goal for this motor as a kit (in a small group buy)

I would not spend 2k on a 1kw motor
 
My whole carbon fiber dura-ace cannondale cost 3.5k (Ya a good deal) and thats stupid expensive compared to what one would expect to pay for a bike....even then thats a whole elite class bike and all the optimized components to make it work well together... not 2k for just a motor..sans controller or batteries.

Down this path if you succeed lies the bike Yasa. Are we going for the spaceship one of bike hub motors, cool tec for the one percent... or trying to move the rep of the cheap chinese hubby a step higher up the evolutionary ladder. If the latter, then I think we have to look for design solutions that allow us to improve motors without ending up with a solution only DARPA could love. http://www.google.com/url?sa=t&rct=...ekR2Zq_xjWo8Jdbzw&sig2=yK59pVj781eF7SNddZLzYw .
How can we keep costs down to $500.00 in parts and still improve the genre. Kit only, like Vans aircraft or pygmy boats. (we won't count time building in the value equation)
 
First you must make what you want at any price, and refine the design until it accomplishes the project goals. (~1kW continous, direct drive, only 3-4lbs over a normal bicycle wheel).

Once you've made something that works at any price level, then you can find ways to make it cheaper. Or just send one of them to a hungry-for-new-designs Chinese company and let them figure out how to make it for an economical price.

Trying to both push new performance concepts make the best, and make it cheap, is like setting out to fail.

The lessons learned from these no-expenses-spared motors will pave the way for the next generation to hopefully get 70-80% of the performance at 20-30% of the costs.
 
liveforphysics said:
First you must make what you want at any price, and refine the design until it accomplishes the project goals. (~1kW continous, direct drive, only 3-4lbs over a normal bicycle wheel).
Once you've made something that works at any price level, then you can find ways to make it cheaper. Or just send one of them to a hungry-for-new-designs Chinese company and let them figure out how to make it for an economical price.
Trying to both push new performance concepts make the best, and make it cheap, is like setting out to fail.
The lessons learned from these no-expenses-spared motors will pave the way for the next generation to hopefully get 70-80% of the performance at 20-30% of the costs.
Click to expand...

I'm not sure about that LP . It may have worked in computer chips, over 20 years and with production volume of hundreds of millions of units, and a market worth billlions, but even then the people who designed the raspberry pi http://www.zdnet.com/we-thought-wed-sell-1000-the-inside-story-of-the-raspberry-pi-7000009718/ weren't on the team that designed the quad core pentium. We already have cheap heavy hub motors and light frameless aircore skunkworks motors, we need both goals addressed at the same time. When the twin goals of performance and value are considered together issues take on importance that wouldn't be considered in a cost is no object project. Instead of how do we make it better, it becomes how are we making it, what materials, and processes can meet our seemingly conflicting goals. It brings up the question, what is better, is better faster and more powerful or easier to make and more affordable to a larger part of the world. this is The whole field of appropriate technology http://makezine.com/magazine/make-37/patdelany/
If bikes don't belong in the arena of appropriate tec... what does... Oh and I'm still waiting for my $20,000 flying car or my cheap luner lander :) .
 
Hal2000- I respect that you're interested in a mass-market-appeal low-cost designed for volume production motor. I hope you can take away as much design as possible from this non-volume-production exotic motor development to use in such a product.

I am looking for the best hubmotor I can make for my roadbike. I want it to be very lightweight, strong/durable through good design, and very high efficiency.

I'm willing to pay another $500 materials cost to have the motor end up a pound lighter (or whatever). I know most folks wouldn't notice at all if you added or removed a pound from there motor, so paying a premium to be as light as possible obviously isn't going to be good bang for the buck. I also know most people wouldn't spend a few hours a day for a week winding hundreds of tiny teeth, nor could most folks even get it wound correctly and without cutting the enamel in places and getting shorts etc.

If you want to start a thread and call it, "the every-man's cost effective hubmotor designed for economical volume production", please do. I'm sure it will appeal to a radically wider audience than the fringe <1% ebike group who is willing to pay with both time and money/resources to attempt to create the best performing super light hubmotor they can spend many many hours building/winding/gluing magnets, etc.

I kinda look at it like this. If I'm going to put perhaps 20hrs of my time into winding the motor, it's already >$1k of my time being spent on the motor, if the other parts end up costing $1500-2000 or whatever, I don't mind so much.
 
After some price searching, it seems Hyperco 50 goes for about $100-150/lbs. That's not as bad as I was expecting. If we could make the rotor back-iron lamination from the same chunk our stator is made from, we can likely get our 1-2lbs of lams from only using perhaps 6-7lbs of Hyperco 50.

I found this crude rule-of-thumb price comparison online:

MATERIAL TYPE RELATIVE COST
Low Carbon Steel 0.5
Silicon Steel 1.0
Thin Electric Steel 10.0
49% Nickel (~Carpenter 49) 12.0
80% Nickel (~HyMu 80) 15.0
Cobalt Alloy (Hiperco 50) 35.0


Miles- Were you thinking N45SH for the magnet material?

Perhaps ball-park of something like ~200-250 of these magnets?
http://www.magnet-shop.com/neodymium-magnets/blockmagnets/blockmagnet-100-x-55-x-20-mm-nickel-n45sh-holds-1-kg::1077.html

That would cost $125-150.

The copper will be cheap, sub $50, even if we do use some exotic stuff.


So, active materials cost is likely in the $1200-1400 range. Then the machined aluminum stator support and custom axle and laser cut CF side covers and custom spoke flanges may be an additional $500-ish.

It's going to end up cheaper than many lightweight pedal bicycle wheels that don't include a motor in them. :)
 
Along those same lines
Maybe we should save the scrap hiperco for a smaller motor
Such as one of the one miles is working on :D

An "all out" design is awesome but I think we should also design for optionality

Easy enough to offer a kit with cheaper standard lams instead of hyperco

I dont think your hiperco price estimate is realistic. Not sure you can go off price per pound. By all means that should not stop hou from getting a quote

A few small hi end rc motor use hiperco and there is a huge jump in cost

Same with carbon fiber side covers... Easy enough to offer a cheaper alumnium version that will also sink some heat

As far as the cost of hi end wheels go...not a fair comparison to compare raw material cost and retail
Usually the retail price is 5-15x material cost for hi end stuff
Also makes you sound pompous :D
 
I'm hoping the end design will allow people some flexibility to pick-and-chose materials to suit their budget and requirements. I.e. replacing CF parts for machined or spun aluminium alloy, or high-end laminations for electrical steel, or magnet type (within reason). All design is a compromise.

If the sidecovers can be laser cut from off-the-shelf CF sheet then that sounds reasonable. Getting CF parts moulded would be too much, I think.

I'd love to see an open-source design, with models and drawings that people all over the world could take to local suppliers and source, or get made, the components needed. Simple tweaks could be made (width as already discussed), but maybe you have a really good local source of magnets, but they're ~0.2mm thicker than the design - no problem, just adjust a handful of dimensions on the drawings for the stator ring and side covers.

Here's a basic point: Can we have the wiring exit concentric to the axle, not down the centre of it, using a large I.D. bearing? It would simplify the axle design, the wiring and allow flexibility to add water-cooling or monitoring sensors etc.
 
I think $999.99 is a better goal for this motor as a kit (in a small group buy)
Click to expand...

This is a good long term goal, but with great motors like offerings from Joby, Plettenberg and large Astros selling well above that it may be just wishful thinking to get something great without adding the better / more costly materials. Certainly a goal to design out the costs that add no or little value, but we need to explore the limits a bit here.

On racing/road bikes, a few of the weight weenie riders I know, value component upgrades (lighter weight)/ tradeoffs at $1.00 per gram. So if wheelset A is 100 grams lighter than wheelset B and less than $100 more it may be a suitable upgrade if all other things are relatively equal. Same with bars, stems, shifters, cranks, etc.. Certainly the lower the cost/gram the better. Given that this hub motors weight reduction is a design goal, we can at least review the tradeoffs in a similar manner and ignore the more costly options at least for now.
 
True, but again you are comparing the price of finished motors.

The JOBY JM1S peaks at 12.6kw (different application I know) for only $1200. Joby's were still being hand made (wound, assembled) in small batches in the USA (at least last year)

Still a hub motor is much more complex in some ways, although the JOBY bearing and support structure is not cheap

What peak torque value are we shooting for? What continuous torque value?

For a 700cc rim or a 26" rim I'm thinking somewhere around 80-100Nm peak, but I'm not sure how much continuous we need. If it can't wheelie a 29" bike from a stand still without pulling on the bars....weaksauce. The thing that sucks about hub motors and large wheels is they overheat going up long steep hills. 20" wheel and cromotor, no problems with hills. An ultralight motor is not going to help here. Low thermal mass and no active cooling means were going to end up with a very high peak to continuous ratio. We should definitely leave provisions for temp sensor(s). The motor is still going to weigh more than the frame on an ultralight road bike. I'm interested in something that better than a MAC/BMC motor on the hills with large wheels, WITHOUT the gears. I live in San Francisco where the hills can be steep and long.

I was also thinking it would be nice if we could have a connector at the axle, instead of wires running out of the axle. Make it super easy to fix a flat and also waterproof. This would require a large oversize bearing on one side. Not sure if it is worth the weight penalty.

So have we decided, the Bionx D-Series style motor is the way we want to go as far as topology. We need to start a spec list and add it to first posting as we go:

1) Fits in wheels 26" and up using Bionx D-series style "non structural" (as in the cover/rotor doesn't connect to the rim) side panels
2) 1000 kw continuous
3) ? torque continous
4) $2000 dollars for components on hi end version
6) 130 or 135mm dropouts?
7) Connectors
8) Magnet temp class: 150C?
9) Wire temp class: 180C?
10) Vented?
11) Cooling features
12) Torque arm
13) Also works on front fork (see Justin's new motor)?
 
An ultralight motor is not going to help here. Low thermal mass and no active cooling means were going to end up with a very high peak to continuous ratio. We should definitely leave provisions for temp sensor(s). The motor is still going to weigh more than the frame on an ultralight road bike. I'm interested in something that does as well as a MAC/BMC motor on the hills with large wheels, WITHOUT the gears. I live in San Francisco where the hills can be steep and long.

I was also thinking it would be nice if we could have a connector at the axle, instead of wires running out of the axle. Make it super easy to fix a flat and also waterproof. This would require a large oversize bearing on one side. Not sure if it is worth the weight penalty.
Click to expand...

I am not sure you can say that without exploring full use of large areas of passive cooling surfaces with proper conductive paths for heat flow. The Joby and plett are just motors. No controller or any required gear train to drive a wheel. Also not a simple direct drive complete wheel. Power I stated in the OP was unlimited with a realistic target of 1kw cont.. If we can get peeks of 5kw out of it for long enough to scream up some of your shorter (less than a mile) hills I think you would be happy to trade a 20-25 pound wheel for a 6-8 pound one. Lets see where we get with some more realistic numbers say by mid april. Miles estimate was around 25 NM which should be plenty of torque for a 700c wheel. What is the mac or bmc for comparison.

On the larger axle bearing, Wondering if three or four (sun-planet style) smaller bearings supporting a large diameter axle/torque arm mount/cable connection center hub block would make sense hear on the none sprocket side.
 
Outrunner don't have good heat flow unless you use oil to transfer heat to covers. Oil is messy, but if we want to design for oil cooling there is much to be gained. Trying to seal a hub motor not design for oil is hit or miss

Vents can work OK too but not as well at low speed

Yeah we need to revisit the torque spec as it seems like we have thrown the 20" rim version out the window

Do you guys agree it should wheelie in a standard geometry frame (not stretch)?

Here is what I mean by larger (ID) bearing courtesey of zappy:
zappy_pie4.JPG

zappy_pie1.JPG


JUSTIN MOTOR AXLE
file.php
:
file.php


OPTIONALITY:
"You can also see that the motor has a hollow aluminum axle. This is 20mm ID and allows the motor to be used with thru axle suspension forks, or with stub axles for supporting from a single side only. For normal forks, we just stick in a small insert to each end that makes it quick release compatible. And the hexagonal shape on the end of the axle is for an integrated torque arm. So there will be no spinout forces present on the fork dropouts."

IMG_4771-600x399.jpg
 
flathill said:
1) Fits in wheels 26" and up using Bionx D-series style "non structural" (as in the cover/rotor doesn't connect to the rim) side panels
2) 1000 kw continuous
3) ? torque continous
4) $2000 dollars for components on hi end version
6) 130 or 135mm dropouts?
7) Connectors
8) Magnet temp class: 150C?
9) Wire temp class: 180C?
10) Vented?
11) Cooling features
12) Torque arm
13) Also works on front fork (see Justin's new motor)?
Click to expand...
No digs, just curious... :)

Some comments:
#2: 1000 kw? Is that correct? If so - that's 1341 hp continuous... on a bicycle? or motorbike? It's a typo, right? :)
#6 & 13: OK, so 135mm is STD for rear, and 110 for front. This means you must design for 110 and allow for 130/135/150 widths - or have two models. Personally - I'd want 150mm version so I could mount a 3" wide tire.
#8: I fear this is unattainable. Most Chinese magnets are about N32-N35 range @ 80°C. Affordable epoxies begin to have problems above 120°C - plus there is the added expense of high-temp tolerant magnets. Have you spent any time bidding on magnets in the past year?
#9: Good choice.
#10 & 11: Your power rating & efficiency will dictate how to proceed.

Once you start down the road of designing all this stuff - you will discover natural constraints to your intentions, and that's the time to re-scope your dreams.

Here's to wishing 8)
~KF
 
Luke and Miles agreed on sh magnets which means 150C

You are right tho about price. I have bought large amounts of magnets in the last year and the price quotes jumps dramatically with increasing grade due to the monopoly on dysprosium. Overall prices are down in the last few years. The chinese suppliers sell into their domestic market for less than half of export prices. Trying to force manufactures to relocate to china. Puts everyone else at huge disadvantage. Cant compete on price.

120c mags may work but heat soak can kill up if you dont keep a hot motor moving. Oil cooling will also transfer much more heat to mags. Have to be careful there too. I think we shoukd stick to sh grade and bump down to n40
 
You must log in or register to reply here.

Similar threads

RustyKipper
    • Like
    • Cheers
I'm in the 40,000 mile club!
Replies
8
Views
1,935
RustyKipper
RustyKipper
G
CRF450R with LG washing machine stator wheel motor.
Replies
14
Views
2,959
larsb
L
T
Multistage axial flux motor design and prototyping
2
Replies
25
Views
6,505
larsb
L
R
    • Like
    • Brainy
Really good article comparing Bafang vs. Tongsheng vs. CYC
Replies
14
Views
26,851
Chalo
Chalo
Torxed
Jawa 350 type 639 Chopper - First time conversion
Replies
3
Views
2,803
Torxed
Torxed
Back
Top