Why not more axial flux DIY projects ?

qwerkus

10 kW
Joined
Jul 22, 2017
Messages
785
I've been wondering why there are not more axial flux motor ebike projects around. I understand high developing and manufacturing costs hamper large scale production of those motors, but at the DIY level, where people spend countless hours on awesome stuff, there seems to be no reason no to try it. AFPM motor promise more torque, less noise and less vibration for less weight - especially at low RPM. Or am I missing something here...

https://cafe.foundation/v2/pdf_tech/MPG.engines/HE_HP_electric_motors_Long_20090929.pdf
 
If I was building a non-hub motor from scratch, I think Lebowski's axial-flux is the best option right now. However, most people don't build their own motors, and the axial-flux motors that are for sale have a limited selection in size. Plus...the existing axials for sale are a bit pricey (Mars, Etek, etc)

"Building a DIY triple stator axial flux motor" (Lebowski, 8 pages)
https://endless-sphere.com/forums/viewtopic.php?f=30&t=46476

Controller
https://endless-sphere.com/forums/viewtopic.php?f=30&t=36602
 
spinningmagnets said:
If I was building a non-hub motor from scratch, I think Lebowski's axial-flux is the best option right now. However, most people don't build their own motors, and the axial-flux motors that are for sale have a limited selection in size. Plus...the existing axials for sale are a bit pricey (Mars, Etek, etc)

"Building a DIY triple stator axial flux motor" (Lebowski, 8 pages)
https://endless-sphere.com/forums/viewtopic.php?f=30&t=46476

Controller
https://endless-sphere.com/forums/viewtopic.php?f=30&t=36602

Thanks for the very informative links - great read! Though I wonder why go through all this trouble to build a non hub motor: Placing the motor outside the wheel = better gearing options = plenty of readily available motors out there (astro for an instance). I'd consider building an AFPM Motor as a better hub motor, not a mid drive. I keep thinking that this is the way lightweight DD motor development would have gone, if bionX would have survived; the D Serie is probably the lightest Radial Flux PM Motor for Ebike you can build. Only AFPM left, if you want to go lighter. Perhaps if I come up with a good concept, I'd apply for a grant and work with an established company to build the ultimate ligthweight AFPM DD motor. Developing one from scratch seems a very complicated task, at least for a noob like myself used to working in his garage...
 
Here's one of the few small axials I've found:

https://endless-sphere.com/forums/viewtopic.php?t=85996

file.php


xxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxxx

And here's the "ebike sized" axial, from Golden Motor

https://endless-sphere.com/forums/viewtopic.php?t=69633

file.php
 
spinningmagnets said:
Here's one of the few small axials I've found:

https://endless-sphere.com/forums/viewtopic.php?t=85996

file.php

Did those guys really just PRINT the stator :shock: Now I understand better your interest into multi-stator designs.
This guy used a similar type of winding (though with proper litz wire), claiming it's more efficient at low RPM based on a master thesis which link is dead. Do you know what he is referring to ? Certainly a neat way to wind a motor in terms of raw copper density ... or just aesthetics !
 
Toorbough ULL-Zeveigh said:
in a hubmotor you want the fluxgap diameter as large as possible to maximize torque.
for a given o/a housing diameter this means the spm radial outrunner owns the high ground (slightly) compared to afpm.

Interesting reply, thank you for the input. Do you have numbers proving that ? Most studies I read about afpm claim exactly the opposite. Max Flux dia might be larger in rlpm, but the average flux dia is larger in afpm, because the magnetic surface ist larger. Or did I misunderstand something ?
 
perhaps i am misunderstang something but wouldn't afpm flux diameter average at the face centre?
in which case that's a somewhat shorter radius than the outrunner gap maxed out at the periphery.
i am strictly referring in a hub motor & of course all things being equal (in particular acerage & mass) which i realize may not be entirely possible.
 
in a hubmotor you want the fluxgap diameter as large as possible to maximize torque.
for a given o/a housing diameter this means the spm radial outrunner owns the high ground (slightly) compared to afpm.

IF...a common radial outrunner (such as a DD hubmotor) has the same diameter and thickness as an axial-flux...the magnets near the periphery would have near the same amount of leverage. The major benefit of the axial flux (IMHO) is that all the "wasted space" near the center of a radial...can be filled with magnets and copper if designing an axial. The portion of the motor that is closer to the axle (as you have pointed out) exerts less leverage onto the rotor, however...the vast center of a large DD hubmotor has no magnets and copper coils.

One of the benefits of the large DD hubmotor in the radial flux configutration is that you can very discretely change the width just one lamination at a time. Adding width to an axial-flux means adding an entire stator and rotor to the assembly. There are axials with a single stator and a single rotor. A more efficient design is dual rotors working with a central single stator. The next step up is dual stators and triple rotors.

If I was designing a motor for high power to fit in a limited amount of space, my first option would be to explore the best possible axial. A radial would be a secondary option.
 
here just one of the many many articles comparing BLDCs (or PMSMs) in axial and radia flux modes. here another one.

The bottom line is always the same:
- Torque depends mainly on motor volume, though AFPM have a slight edge because of their higher efficiency in Low rpm mode (like an ebike)
- AFPM are around 4% more efficient if properly designed, and up to 17% lighter
(mostly due to shorter wiring as spinningmagnets said, and the possibility to reduce mag weight with halbach arrays and working coreless)
- the inherent design of AFPM solves cogging issues, which translates into very low drag
- AFPM are WAY more expensive then RFPM motors <- That's the main reason why we don't see many.

I contacted a manufacturer and asked if'd be interested into working on a high end ebike motor. He said the main problem would be cooling: there is not enough thermal mass in a coreless AFPM to suck up the starting heat in high torque modes, which means a special cooling design has to be added. Also DD mode seems hopeless, so it would habe to be a geared hub motor.

If cooling adds too much mass, I honestly don't see a reason why spend big bucks on a new motor, when there are good ones available with similar weight. Though the absence of cogging would be a huge plus (NO DRAG) in a AFPM motor. I seems to me the big difference in weight is achieved by removing iron core. So if coreless is the way to go (always in low rpm mode), there seem to be cheaper coreless RFPM available: https://www.thingap.com/
 
https://www.thingap.com

Not having the patience to study , for many hours, this kind of motor I am guessing it turns the outer ring ?

If so then a couple of things I can think of needs to be done / I would like to be see done .

1 ) The outer ring machined into a rim , preferably Tubeless Rim.
700c / wider 29 er / 27.5 / 26 inch and down to 16 inch .

2 ) If using that motor as a mid drive , there would have to be a spyder made from the outer ring to down near the axle so that a small tooth count sprocket can be attached , or , are these motors very low RPM/KV motors ?
 
Or would the TG23xx series motors work better / more easily made adaptable to E-bikes ?
At least for a mid drive for a E-Bike ?

https://www.thingap.com/standard-products/

How easy would it be for someone with a machine shop to make a TG23xx type of motor with the necessary spyder or axle for using on a E-Bike ?
 
ScooterMan101 said:
How easy would it be for someone with a machine shop to make a TG23xx type of motor with the necessary spyder or axle for using on a E-Bike ?

Careful not to get excited too fast. Making a PMmotor is easy. Making a highly efficient one is another story. I'd first ask as meany details as possible from existing ones. Maybe just buy one to compare. Than you need a good simulation suite (mathlab + simulink works well). And only after you have a good model, you can start with the manufacturing process.

Those thingap motors seem interesting, but in an outrunner config, they will be plagued by the same overheating problem as the AFPM discussed earlier. With a plastic stator and no thermal mass, how is all that heat supposed to escape ? Perhaps some ferrofluid magic could help - still a tricky problem. There seem to always be a tradoff between weight and cooling ability. No such thing as a superlight hub motor that will run at 3000W.
 
Just hoping someone with access to a machine shop can get excited about making one that is designed / parts machined to work in a e-bike application

I am more interested right now in Hobby RC motors, I was told recently by a Hobby Flyer that Outrunner motors are better, they have better torque ?
Is that right ? for E-Bikes we need lower Kv's which means higher torque I believe .

For Cooling people are starting to experiment with pushing or rather pulling air through a open air RC motor, just using off the shelf brushless small DC Fans.
 
spinningmagnets said:
in a hubmotor you want the fluxgap diameter as large as possible to maximize torque.
for a given o/a housing diameter this means the spm radial outrunner owns the high ground (slightly) compared to afpm.

IF...a common radial outrunner (such as a DD hubmotor) has the same diameter and thickness as an axial-flux...the magnets near the periphery would have near the same amount of leverage.

how do you figure the af magnet periphery to have much leverage when the flux concentrates at the magnets face centre & where it will effectively torque at.
that's where the axial flux-gap diameter is defined.
so you're down by half a magnet length squared in torque production, as i said is slight but it's there.

One of the benefits of the large DD hubmotor in the radial flux configutration is that you can very discretely change the width just one lamination at a time. Adding width to an axial-flux means adding an entire stator and rotor to the assembly.
double the stators means double the end-turn losses... if you believe that... which i know you do.
If I was designing a motor for high power to fit in a limited amount of space, my first option would be to explore the best possible axial. A radial would be a secondary option
ok, take the bionx-d housing & stuff an af rotor-stator in there.
is it even physically doable in af given that width is the dimension that's at a premium?
maybe.
and the bionx d is ported from an old, old 10 year old wavecrest design that came along for the ride with couple of engineers when magna bought bluwav.
now extend the flux-gap all the way out to the rim like this fake version an italien newspaper used to bait mechanical doperz (good fact checking karlG).
no way i can see an af flux-gap get out there but hey, there's lots more peopl clever than i.

mechanical-doping-electromagnetic.jpg


qwerkus said:
Torque depends mainly on motor volume,
hokay, call it thrust then if you want 2b specfic which is good
the possibility to reduce mag weight with halbach arrays and working coreless)
that's not keeping thingz equal or is there some reason a radial can't be halva?
 
Toorbough ULL-Zeveigh said:
ok, take the bionx-d housing & stuff an af rotor-stator in there.
is it even physically doable in af given that width is the dimension that's at a premium?
maybe.
and the bionx d is ported from an old, old 10 year old wavecrest design that came along for the ride with couple of engineers when magna bought bluwav.
now extend the flux-gap all the way out to the rim like this fake version an italien newspaper used to bait mechanical doperz (good fact checking karlG).
no way i can see an af flux-gap get out there but hey, there's lots more peopl clever than i.

mechanical-doping-electromagnetic.jpg


(...)

that's not keeping thingz equal or is there some reason a radial can't be halva?

Ha: this was exactly my first idea: use the entire rim length as rotor. Problem is: you need a stator of the same length, and a bearing in between. If you use an axle based structure, you will need 2 complete housing over the surface of the entire wheel + some solid way of supporting the stator = a very heavy wheel ? A solution to that problem would be going hubless, but that translates into a bearing of the size of the rim, which is just impossible to seal properly, being exposed to all shocks of the wheel, plus water and dirt. I cantacted 5 manufacturers: all of them told me: no doable.

So I'm back to axle-held motors, where AFPM seems to offer the highest power density. If I manage to brush up enough of my integral math skills, I'll show you why the magnetic force in AF is stronger than in RF, albeight by a small margin.

RF halbach array are doable, but the non planar orientation of the working mag field makes it less effective, hence nobody uses it. I saw some papers about halbach arrays successfully implemented in an RF design, using slotted magnets. But that's just as expensive as an AF motor...
 
qwerkus said:
A solution to that problem would be going hubless, but that translates into a bearing of the size of the rim, which is just impossible to seal properly, being exposed to all shocks of the wheel, plus water and dirt.
that's how i see the end game for an in-wheel motor but it likely will require graphene and/or some other material breakthrough to realize it.
in the meantime ima thinking reduced size "semi-hubless" or large-hollowtech (whatever you wanna call it) centreless axle.
still have some isolation with shorter spokes-to-centre(less) while reducing the size/weight of the 2nd housing.
with the "wasted space" in the hollow centre mounting an electric fan, possibly a radiator if you want.
I'll show you why the magnetic force in AF is stronger than in RF, albeight by a small margin.
i accept that AF's stronger but does it make up for RF's small added lever moment (that i guess i'm the only one that's hung up on) & will a full wheel AF fit in anything less than a fatbike rim is what i want to see.
 
Back
Top