APL's DIY axial-flux motor

[Ianhill]

I totally disagree on the controller part. This motor should not be anything special in the ”difficult to control” department.

Get one of the bigger VESC if you want foc and programmability.

(And, if you want options.. If you can build a motor then you can also add halls. right?)

Trampa 75 300 on a premium
Flipsky 75200 on a budget https://flipsky.net/products/flipsky-fsesc75200-75v-high-current-200a-esc-base-on-vesc-with-aluminum-heatsink-for-e-foil-fighting-robot-surfboard-agv-robot-arm

Don’t make it harder than it is :thumb:

£300 posted is not a budget controller in my eyes, vesc foc implementation is not the best and theres problems with inductance on larger motors i couldn't run a dualtron hub for example kept stalling and losing sync, stayed well clear since then.

More metal in the motor more likely to have issues with vesc good for hoverboard motors and thats budget.

I cant really recommend a controller as such but all i can say is dont get a under designed piece of crap and pay big for it cuz vesc 6.6 is sat idol wont run anything bigger than an apple in anger.

 

[larsb]

Plenty of people would disagree with you. Why would a larger motor be problematic? 300 for a controller is not a lot. If you’re down in the chinese controller swamp and comparing then maybe it looks expensive but those, in my view, are not worth it.

 

[Ianhill]

I cant comment on the 75v version its the vesc 6 i had issues with and that was £145 down the drain when a £120 72v40amp bluetooth controller got the job done.

Ive used the vesc since but its on a hoverboard motor and for that it runs sweet got regen very adjustable but still get the occasional throttle error and the ride needs a hard reset.

I was gonna get a silixcon controller for my freerider before the motor overheated, shame as its the controller that helped kill the ride along with my lead hand, 12kw full speed for good 10 mins and it was cook an egg time a decent FOC with temp control id have gone into limp mode at worst plus there be less stray currents giving less heat and a quieter noise pity cause a id spent out loads on getting frame sweet the brakes upgraded to 220mm and vented calipers chamged the charger to a chargery and portsto anderson all it needed was a decent controller now i need a motor to and thats £1000 im tapped out its sat idol in shed wasting away.

Maybe ill put a qs motor in there one day and a silixcon controller or maybe we have to pay £5 for a loaf of bread nw.

 

[TorontoBuilder]

Yup, new avatar. Might as well get something out of V1,.. she was a beauty, if nothing else.

Well Mike, you got one hell of a great thread out of version 1 too, not just pic of a nice shiny looking motor. A thread I may add that has a great deal of solid information. I read the entire thread, and reviewed a lot of the materials that were linked to. I'm a research junky above all else.

I wanted to comment on dozens of posts but didn't know if later in the thread someone had already pointed out what I wished to comment on, but I will have a few comment posted and suggestions coming soon.

Again I loved your build and thread even if the results were not perfect.

 

[APL]

Well, "it ain't over till the fat motor spins", (or something like that),.. so they say. :lol:
I'll be back at it one of these days soon enough.

I won't be happy until there's a DIY motor in the bike, happily feeding on electrons and pulling me down the road!

Thanks for the thumbs up on this thread, :thumb:, and she's been a long road indeed, enough to give anybody a
headache, especially me. (They should hand out a certificate for reading through this thread!) :wink:

Probably should have posted this a long time ago, but I took a breather from the motor build and built a new bike
over in bike builds category, and now that has turned into a battery pack build, which will likely take a while.

Bike; https://endless-sphere.com/forums/viewtopic.php?f=6&t=111159

Battery; https://endless-sphere.com/forums/viewtopic.php?f=14&t=113512

But, I still have plans to build the next version motor, although it's still not clear which direction it's going to go,
axial, or radial. I have so much invested in axial that it would be a shame not to go that way,.. but another failure
would spank me pretty hard. Radial would guarantee no rotor rub, and more likely succeed, but not as exciting
as axial.

 

[Ianhill]

Personally i think you give it a good go and not to win first time is fine keep at it redesign and rebuild.

I think a yokeless design is where you can gain the required strength try and emulate magnax design.



Looks like ive got 2 more threads to read though keep up the good work well done

 

[TorontoBuilder]

Taking a look around, I find quite a bit of info on printed coils, mostly for small motors, but I think that the process
is 'etching' for removing copper from the board. I think that it can be used for thicker copper too, and may be an
alternative to water jetting, but I'm sure it has limits.

etched stator coil.png

Something that looks right about it. Interesting, as it lets you design optimal geometric coil patterns that you can't do
with wire.

I thought that maybe a 3D router table might be the perfect tool for the job too, and wouldn't require any chemicals
or mask's. Especially for thicker copper.

I found an article about printed coils for the archives, it gets long and technical as usual,.. but the first few paragraphs
are a good intro.

https://www.pcbway.com/blog/Engineering_Technical/Designing_Printed_Circuit_Stators_for_Brushless_Permanent.html

Important Tech Tip

Printed PCB coils lack density to attain high power & torque. However, Binder Jetting 3D printing copper powder is a far faster and superior manufacturing method to attain complex coil designs similar to those found in pcb designs. Binder jetting is similar to laser sintering machines, but 100 times faster, since it uses inkjet print heads to spray adhesive onto the powder bed. By not using heat to fuse the particles in the machine the machine does not require an interior nitrogen environment. Where required parts are later fused in a heat treatment oven.

The method is also far superior to CNC coils such as in the greenway motor prototype. Binder Jet is a 0% waste manufacturing method. CNC routed coils can produce 90% scrap material! The other benefit of binder jetting over CNC is multi-layer coil production and the ability to change the aspect ratio of parts.

BUT WAIT, there's more!

You can use Binder Jetting with many different types of metallic and non-metallic powders to make so many different parts.

For years I wanted to find a foundry to make me small cast iron parts for a 1/2 HP steam engine I designed. After years of searching I gave up and had decided to cast the parts myself in aluminum. When speaking with a local 3D printing service however they told me that I could skip the intermediary steps and have the metal parts printed directly and sold me on the benefits of binder jetting over laser sintering.

As a result I now have model steam engine parts manufactured in a steel alloy to my own design that are dimensionally more accurate than lower quality cast iron. The parts were affordable and took minimal time to produce (less than 2 weeks). I cant yet say how the alloy will machine but the parts are far smoother and dimensionally accurate than sand cast parts.

What's that got to do with motors dude? I'm glad that you asked...

There is a growing trend to use 3D printed metal parts in motor manufacturing.

I am going to try Binder jetting techniques for a couple of motors I wanted to make and for a wind turbine generator since I have used the process before and I have heat treatment oven to sinter. I was planning on using any steel alloys I could get my hands on if necessary but if you can get somaloy powder as someone suggested earlier in this thread that would be awesome.

You only need to find a 3d printing center near you with binder jetting or laser sintering equipment and ask them to print stator cores for you. I dont know how Somaloy's insulative coating reacts at the laser fusion temps which is another reason I prefer binder jetting and subsequent oven sintering. I know what temperatures oven sintering is used for die compacted parts that work well.

I personally will take the green non-sintered parts from the printer and sinter them myself in my heat treatment oven in a nitrogen atmosphere. There are dozens of build videos for small DIY heat treatment ovens on youtube. It is easy to add nitrogen and cheap if you use that gas already elsewhere.

Here is a great review of Additive manufacturing of soft magnets for electrical machines from Materials Today Physics
Volume 15, December 2020:
https://www.sciencedirect.com/science/article/pii/S2542529320300791

and here is a research paper on Hysteresis measurements and numerical losses segregation of additively manufactured
silicon steel for 3D printing electrical machines Published in Applied Sciences (Switzerland) 01/09/2020 that provides great comparison of properties of laser sintered parts:

https://acris.aalto.fi/ws/portalfil...ely_ApplSci_2020_10_finalpublishedversion.pdf

IMO the laser sintered parts have less desirable grain characteristics but I need to read both these papers much more deeply...

I have a couple more posts coming to pull everything together to help Mike and others build better motors easier and that could be group component buy and build opportunities if people wished.

 

[TorontoBuilder]

Personally i think you give it a good go and not to win first time is fine keep at it redesign and rebuild.

I think a yokeless design is where you can gain the required strength try and emulate magnax design.

magnax.png

Looks like ive got 2 more threads to read though keep up the good work well done

I second the opinion of Ian. Seldom does the first iteration achieve the desired goal. Redesign and build the next version. I will be sending my ideas for Mike to consider over the next day or two.

There are designs that use a yoke that are very viable and yokeless that are also viable... BUT IMO it must be an inrunner single stator – double rotor design for Mike's usage. it will make the build much easier.

 

[TorontoBuilder]

Thanks for the thumbs up on this thread, :thumb:, and she's been a long road indeed, enough to give anybody a
headache, especially me. (They should hand out a certificate for reading through this thread!) :wink:

But, I still have plans to build the next version motor, although it's still not clear which direction it's going to go,
axial, or radial. I have so much invested in axial that it would be a shame not to go that way,.. but another failure
would spank me pretty hard. Radial would guarantee no rotor rub, and more likely succeed, but not as exciting
as axial.

Damn I have to start to learn to insert emojis here. Definitely a huge

When can I expect my certificate to arrive

Dude, it must be axial. Too many benefits NOT to use axial, and you have learned so much and we can toss out fresh new thinking prior to the start of redesign to assure your success.

Teasers...

definitely use a single piece stator with single coil and flip your coils 90 degrees.

perhaps use square wire high density coils wound on a bobbin (perhaps mould the bobbins yourself)

Alternatively, 3D print coils if there was a group buy/build project. Can be potted using silicone moulds to make modules for assembly.

These coils can be made similar to edge wound coils illustrated below but with several benefits...

Yasa uses edge wound coils using a high speed CNC bending machine... you can see a it of it at about the 50 second mark of this video that was released just after Mercedes bought Yasa.

[youtube]https://youtu.be/YkldZQuzR3E[/youtube]

 

[TorontoBuilder]

Henery2000, thanks for the info., I wasn't aware of the Linear Labs HET motor per say, but I've seen others like it. Still
don't see how they mount the stator exactly, but it looks like they use brushes in the center.
Your right, they really have it closed up tight, with no cooling that I can see. Big power = big heat.

Mike picked a good topology in the single stator – double rotor design, but very early on in this thread I wanted to ask him one very important question.

Why did he choose an outrunner design over an inrunner design? An inrunner design would bring manufacturing simplicity.

The outrunner design forced him to choose a hard to implement stator design. The difficulty became compounded with the choice of split stator design and opposing coils.

Axial motors tend to generate enough force between the rotor and stator discs to deflect the rotor against the stator and cause a failure. I believe that despite the rotor design where the magnetic forces cancel each other out it was a combination of the stator design and flex and the built up sandwich that led to sufficient lack of rigidity causing the rotor to contact the stator disk at higher rpm.

IMO an inrunner design with the two rotors connected directly to one another via a very rigid shaft inside a motor case will solve that issue. Such a motor can be yoked or yokeless. In either case the stator windings can be placed into direct contact with the outer aluminum casing for better heat transfer.

It is preferable to use a single toroidal wound coil rather than the split design chosen, made so that there are no overhanging loops that don’t contribute to generating flux. in such coils 100% of the windings are active. The HET motor uses such a design as does the YASA and Magnax motors.

I'd consider rotating the coil 90 degrees so to make manufacturing of yoked design with far less material much easier. The HET motor does this and does the Lynch motor.

When Mike posted the pic of the HET motor I was sure that you had stumbled across what I wanted to suggest, but he seems to have missed what some of the best companies have done and why because he was too far down his design path to change.

Linear labs does not say how they fix their armature to the motor housing, but they do show a brilliant stator armature arrangement that assembles components around a tape wound split ring yoke and spider. This permits them to use discrete mass produced parts of excellent quality.

The armature is assembled from power modules that slide over a yoke comprised of a ring that separates into two half moon segments. The power modules are made up of two stators and a coil element.

Mass produced toroidal coils (with very high fill density due to the use of square wire) are wound onto plastic bobbins. The wires from the coils are soldered onto a narrow PCB board, which have long pre-bent coil outlet pins soldered onto them.

The coil output wires are soldered into a separate pcb board that bring all the coils together into the desired connection arrangement...

The coils are wound onto plastic bobbins that are sixed to fit snugged very tightly in between 2 stator poles with a hole in the center, all of which slide onto a split ring to hold everything together. Due to the design of the stator poles which virtually surround the coils they capture all of the magnetic field from the coils.

The coil design is similar to yokeless and segmented armature arrangement that I believe magnax use as well.. at least the mass produced square wire toroid's that insert into the segments.

In most other yasa designs the stators are set into the motor shell so as to cool to the exterior of the shell or have cooling oil injected into the armature portion to perform cooling in larger motors. Some experimental in wheel motor have cooling finds that penetrate the shell and extend into the stator assembly..

I'm not sure what HIT does, what with their hybrid axial and radial flux design, but I suspect its oil cooled.

If you want to attempt a yokeless design here is a paper on the Comparison and Analysis of Stator Plate Holder on Yokeless and
Segmented Armature Machine from the Journal of Physics: Conference Series:
A F Desanti et al 2020 J. Phys.: Conf. Ser. 1577 012039
https://iopscience.iop.org/article/10.1088/1742-6596/1577/1/012039/pdf

and here are the best images I have found on a YASA design

plan view of a motor shell lamination showing the stator holding arrangement.

Before the stators are put in place and potted with epoxy.

and here is the post potting pic

The above images are from the following paper:

Stator heat extraction system for axial flux yokeless and segmented armature machines
H. Vansompel, A. Hemeida, P. Sergeant
Published 2017
Materials Science
2017 IEEE International Electric Machines and Drives Conference (IEMDC)

pay per view paper.
https://www.semanticscholar.org/pap...eida/c35f1a09869dc7f65a6474fb848685dae1c5fa48

 

[TorontoBuilder]

Another route, and my own preferred model is just being developed for small ebike motors is the 3D printed coils that insert into an innovative lightweight armature...

Here is image of the coil and stator assemblies and how they insert into the plastic spider array.

The spider array assembly drops into a three piece stator block... which attaches to the outer shell

and finally a close up of the wiring connection.

They entire assembly can be 3D printed.... including the aluminum ring and the stator wedges...

Note too how that when you flip the coils 90 degrees the modules are comprised of 2 stators and a coil with no wasted wire loops. Just like the HIT and Lynch motors.

This coil inside a power module comprised of two stator pieces and a coil must be less than 3/4" wide... the entire motor with two rotors is about 2" wide

 

[HalbachHero]

APL,
Thanks for sharing those other links. I had not seen the battery build. very cool. I will likely be going down this road when I am happy with the motor. Also I am excited to hear that you plan to revisit your motor project in the future. Can't wait for that

TorontoBuilder,
Wow, love the collection of photos. I have seen a lot of those, but that last one is pretty neat. It looks a lot like the kind I am trying to make. I have to assume those are air cores. Do you have a link for where you found those. I would love to know more about this build.

 

[TorontoBuilder]

TorontoBuilder,
Wow, love the collection of photos. I have seen a lot of those, but that last one is pretty neat. It looks a lot like the kind I am trying to make. I have to assume those are air cores. Do you have a link for where you found those. I would love to know more about this build.

That last pic is a coil from the coil module from the motor in the proceeding group of pics. Those pics are of the unpotted parts of the oddly offset and grossly over-priced German manufactured Binova Flow motor. Their website is www.binova-flow.de

The pics though are from the German company Additive Drives. That is the company whose technology is being used to 3D print the coils. https://www.additive-drives.de/en/project/pedelecmotor/

I'm not exactly sure if the coils can be called coreless because the two outside pieces of the 3 piece modules are iron based stators. I am unsure if one, both or neither of the two parts has a tooth that would extend into the open space in the middle of the copper core. I'd suspect that the flux path would be improved with such a design.

IF anyone wants to contribute to a fund to buy a 2000 euro motor kit so we can cut the motor open to see lemme know :lol:

 

[HalbachHero]

The outer faces of the stator appear plastic to me, unless you mean the metal ring that goes radially around the stator. I am not sure how that would contribute to the flux path. Very neat though. I wish those sites went into a bit more detail of the specifications. thanks for sharing!

 

[APL]

Wow guys, I just noticed all this activity! Thanks for all the thumbs up! Guess I better pay more attention.

TorontoBuilder, you've given me brain-burn with all your posts! Lots to read and think about,.. unfortunately I've
got a busy day today and have to get going,.. I'll have to get back to all this tonight/tomorrow.

Awesome stuff! :thumb:

 

[TorontoBuilder]

The outer faces of the stator appear plastic to me, unless you mean the metal ring that goes radially around the stator. I am not sure how that would contribute to the flux path. Very neat though. I wish those sites went into a bit more detail of the specifications. thanks for sharing!

HH, I tooo wish those sites went into more detail... in fact I wish every company posted detailed info on how their products are designed and made so that consumers could assess the durability/longevity of products and also make our own :lol:

Until they do such things we have to resort to things like patent documents and our knowledge of the topic in general.

You are correct in that the outer faces of the stator are plastic, likely polypropylene. But those are just a "spider" to hold elements in place for the potting process. I was not referring to the ring that wraps around the stator. That ring is used to mount the motor to the bike.

I was referring to the wedge elements that surround the copper coils.

In the above pic I've highlighted two pieces that make up the split stator core on two separate coil assemblies. They must have teeth that penetrate the center of the coils.

HIT by the way calls these assemblies of their split stator and coils "power modules" in their motor design. In the pic below you can see how the stator yoke connects all the modules and completes the connection between the two stator halves.

From the Binova patent documents it is clear that they originally used a traditional coil wound onto a bobbin that then had a two piece stator with a core that passed thru the bobbin center. The new 3D printed coil omits the bobbin but must have one or both sections of the stator pass thru the core of the coil.

Description
Paragraph [0007]:

"Furthermore, the invention provides that the coil elements each have a coil carrier with a coil core around which a coil wire is wound, wherein the coil core at its opposite ends along the coil axis has two mutually opposite polar bodies, which are in the axial and / or radial direction relative to the axis of rotation of the rotor on both sides extend beyond the spool core. Thus, the coil wire is fixed by the polar body on the spool core, it can not slip due to the above polar body from the spool core. In cross section, the bobbin thus substantially in the form of an H. The coil element has a concentrated coil winding. By the coil carrier, the coil element to a compact unit, which allows easy production of the stator. The opposite polar bodies of a coil element always have an opposite polarity, since they are assigned to the opposite coil ends."


The following paragraph explains the plastic "spacers/spiders" on the outer stator faces. Sadly, here are no images of the final potted stator anywhere.

Paragraph [0015]:

"Under coil element is a magnetically isolated single element with independent coil, such as a wire winding, understood, i. the coil of a coil element is individually electrically controllable for generating an individual magnetic field of the coil element. In particular, an individual control of the coil element is possible independently of adjacent coil elements. Furthermore, the coil carriers of adjacent coil elements are expediently spaced-apart components, in particular by a plastic layer separated from each other and / or cast into a plastic mass components."

How is your German? The main patent documents are a mix of English and German, and explain the rationale and benefits of the design...

https://patentimages.storage.googleapis.com/ea/18/c7/92b72a70764308/EP2638618B1.pdf

and

https://patentimages.storage.googleapis.com/ef/b0/a9/b82c5f25cd30f9/WO2012062710A2.pdf

Here are two excerpts for example that explain well some benefits

Paragraph [0013]:

"The disc rotor motor according to the invention is characterized in that the pole bodies are radially aligned relative to the axis of rotation of the rotor, so that the distance between the two polar bodies of a coil element increases with increasing radial distance from the axis of rotation of the rotor. As a result, a uniform interaction of the pole faces with the permanent magnets of the rotors or of the rotors is ensured, which is important in particular with a large radial extension of the pole faces. In other words, this design of the pole body allows the provision of Pole faces with comparatively large radial extent, which leads to an increase in the performance of the disc rotor motor with a further compact dimensions with appropriate design of the permanent magnets of the rotors."

Paragraph 0014]:

"The disc rotor motor according to the invention is further characterized in that the coil core is a flat body, the radial extent of which exceeds its axial extent by a multiple relative to the axis of rotation of the rotor. This allows axially compact stators and thus compact disc rotor motors. At the same time, in particular, a comparatively large radial extent of the coil core and thus of the coil element leads to a correspondingly large radial extent of the pole body and thus of the pole faces, which likewise enables powerful compact motors."

 

[TorontoBuilder]

HH I found another incomplete patent document but completely in English (yay!) but lacking the images (Boo!)

https://patents.justia.com/patent/9438092

Between this new document and Figure 4, we can see definitely that the new design split stator pieces must have a core thru the coils. The old design must have used a one piece yokeless stator which a traditional coil was wrapped with normal magnet wire. I surmise that they may have had manufacturing issues that led to the upgraded design choices.

FIG. 4 furthermore shows that the coil core 19 has two mutually opposite pole bodies 21 at its opposite ends along the coil axis, which runs through the coil core 19, to be precise in the center of the winding from the coil wire 20. The pole bodies 21 extend in the axial and radial directions in relation to the axis of rotation 13 of the rotor 12 (not illustrated in FIG. 4; the alignment of the coil elements is shown in FIG. 5) beyond the coil core 19 on both sides. In other words: both pole bodies 21 of a coil element 17 protrude on all sides beyond the coil core in a plane perpendicular to the coil axis. The coil wire 20 therefore cannot slide from the coil core 19.

 

[APL]

Well that's too much stuff to comment on all at once, so I'll just hit a few things for now,.. first of all the Yasa and
Magnax motors are sweet tech units, and I struggled with seeing how they actually mounted the coils and cores.

Thanks to your posts and photos I can see it all a little better now. :thumb:

The yokeless design gives greater strength to the stator assembly for sure, because it holds them from the diameter
which is where my design was lacking. Trying to hold everything from the axle is very difficult, compounded by the
fact that in most every decent attempt you wind up encircling the coil with metal somehow, which leads to a flux
short.

Evidently those two companies are using potting and epoxy to hold things together along with the yokeless design,
which I was unwilling to do. I guess I feel that one of the greatest benefits of a DIY motor is the ability to take it apart
and fix it or modify it, and the easier the better. I certainly had this thing apart plenty of times.

In summary I guess my takeaways on the last attempt was that the rotor ring needs to be one piece, the segmented
magnet attempt was inadequate, and trying to run the flux path of the cores through a reduced central holder is
a definite no-no.

Bad design, but it did pull me down the road, so not a complete fail.

TorontoBuilder, your suggestion to go 'single sided' inrunner is a good one, and I'm definitely going to go inrunner next.
(forget radial,.. another day)

 

[APL]

I like the toroidal air core motors being shown,.. seems to be more than one , they have to be run sinusoidal since
all the coils are in series. Something I wasn't comfortable with in my build, since all the wiring diagrams were showing
coil gang configurations for balancing.

I like the flat wound coils a lot, compact, and you can pump a lot of current through them, which air core motors need
to get going, but it also needs to be liquid cooled too or all that plastic is going to melt down.

The split ring stator holder is cool, never thought of that. I assume that it's composite? I don't see any reference to SMC
powdered iron parts anywhere on these.

Also, 3D printing powdered copper coils is just too cool! I don't think the SMC can be done like that though, it only
works if it's compressed under 10 -20 tons of pressure, and the binders that coat and bind the particles are plastic too.

 

[TorontoBuilder]

The yokeless design gives greater strength to the stator assembly for sure, because it holds them from the diameter
which is where my design was lacking. Trying to hold everything from the axle is very difficult, compounded by the
fact that in most every decent attempt you wind up encircling the coil with metal somehow, which leads to a flux
short.

Evidently those two companies are using potting and epoxy to hold things together along with the yokeless design,
which I was unwilling to do. I guess I feel that one of the greatest benefits of a DIY motor is the ability to take it apart
and fix it or modify it, and the easier the better. I certainly had this thing apart plenty of times.

In summary I guess my takeaways on the last attempt was that the rotor ring needs to be one piece, the segmented
magnet attempt was inadequate, and trying to run the flux path of the cores through a reduced central holder is
a definite no-no.

Bad design, but it did pull me down the road, so not a complete fail.

TorontoBuilder, your suggestion to go 'single sided' inrunner is a good one, and I'm definitely going to go inrunner next.
(forget radial,.. another day)

In suggesting an inrunner design I did not mean to imply that you should use a single rotor. Far from it. I like the two exterior rotor with interior stator design for it's many superior qualities. Such a motor can be built as an inrunner or outrunner and it can also be built as a "caseless" inrunner motor. More on that later.

The twin rotor design gives you other options as well. Yours was a fractional slot winding using traditional stator arrangement as illustrated in the image on the left below.

But you could have rotated the coils 90 degrees and used the same sort of fractional slot winding scheme but in a torus configuration, or as I plan to use on my generator a simple integer slot torus wound stator. By using a Integer slot scheme I can mass produce connectors to connect the coil outputs to the next coil. All the pins will have exactly the same configuration.

Example... 36 slot 12 magnet integer winding connectors span 3 slots. The scheme goes AA|bb|CC|aa|BB|cc|AA|bb|CC|aa|BB|cc|AA|bb|CC|aa|BB|cc|AA|bb|CC|aa|BB|cc|.... to change from CW to CCW the coil modules merely need to be rotated 180 degrees if I recall. IF you recall the HIT motor photos, they use one pcb board to accomplish this, but normally insulated stiff wires can be bent to shape.

That said, there is a huge variance in fractional winding schemes that can be great or terrible, simple or complex. I'd select carefully. In your previous design I'd have gone with 21 slot 28 pole to get a simple concentrated winding scheme with ABCABCABCABCABCABCABC... the connection scheme could use mass produced connectors or a pcb.

Figure 1 below shows (a) ISTW motor and (b) a FSCW motor. I recommend either of these two choices. Really the best starting place is to determine the motor diameter and then engineer the largest stator and rotor that will fit within that diameter and with the best winding scheme by first going thru design iterations prior to building anything.

Binova actually constructed their middrive motor as a "caseless" inrunner motor to eliminate unnecessary weight of the exterior can, and to maximize the rotor radius in order to increase the lever effect to gain torque. To accomplish this they mount the motor using their stator disk. See the 40 second mark and the 1:20 mark of the video below to see how the motor operates in this configuration and how they mount the chain ring to the inside rotor disk. I think this design would have great promise for your bike...

[youtube]OdfaVZaRbBY[/youtube]

 

[TorontoBuilder]

I like the toroidal air core motors being shown,.. seems to be more than one , they have to be run sinusoidal since
all the coils are in series. Something I wasn't comfortable with in my build, since all the wiring diagrams were showing coil gang configurations for balancing.

I like the flat wound coils a lot, compact, and you can pump a lot of current through them, which air core motors need to get going, but it also needs to be liquid cooled too or all that plastic is going to melt down.

The split ring stator holder is cool, never thought of that. I assume that it's composite? I don't see any reference to SMC powdered iron parts anywhere on these.

Also, 3D printing powdered copper coils is just too cool! I don't think the SMC can be done like that though, it only
works if it's compressed under 10 -20 tons of pressure, and the binders that coat and bind the particles are plastic too.

Let me clarify, none of the toroidal coils I have shown are not really air cores. They're part of a power module made up of two outside stator parts and a toroid wound coil. The cores are all SMC or stamped laminates. The design does make manufacturing very easy.

I like flat "edge wound" coils of rectangular magnet wire and that had been what I was originally planning to use for my coils. I can design a simple winding jig with very smooth polished surfaces for my lathe so that the chance of damaging the insulation is lowered.

IF the insulation is damaged it is usually on the outside radius of the bend but that can be avoided by careful design of the bend radii, or using wire with high elasticity insulation. Additionally, you can more often repair cracked insulation on the outside bend radius using a bit of spray on insulation. The potting process will also likely fix this too.

But I may try the 3D printed copper, because I'm a sucker for such things.

In hindsight I would not likely try to get stator cores 3D printed and sintered. I know it will work to make the parts but I dont know how the sintering will affect performance of the SMC materials. I keep getting conflicting information when I research the product.

Some documents say SMC is used in highly compressed but unheated forms, yet other places say that parts may be sintered after die forming. I hate conflicting advice and haven't contacted manufacturers yet...I am saving that to beg for the ability to buy a small sample container of the material.

Instead I will mill up some dies out of steel in the shape of the two stator components. I'll make the cores by compressing them on our 12 ton shop press. I hope that will work to provide sufficient density for my needs. If not I can install my much larger hydraulic cylinder.

The beauty of potting the whole stator is that it becomes a structural element that also acts as a heat pathway to the motor exterior, and that protects the SMC material from shock and damage.

Now just consider the stator as a replaceable element in it's entirety... far less expensive to make than a rotor. You first test all the coil components prior to assembly into power modules. The modules get tested prior to and after connecting coils in series, only then do you pot up the entire stator assembly.

Transparent silicone based potting gel can also likely be used during the prototyping phase. Given all the interlocking parts I am confident that it will be rigid enough for testing purposes without a load.

[youtube]ogJbC9H7TS8[/youtube]

 

[Ianhill]

Whats strange from my observations is the stator teeth on the magnex picture i posted, they have a sarreted edge down the face reverse to what i have seen to date specially non axial any breakdown will occur at these points.

Got me thinking is this what they are doing to eek a bit more efficency out adding more force from the steel albeit for free as its hard not to power the steel if its present.

 

[APL]

Not sure what your referring to Ianhill,.. I think it's the core edges? The motor has both segmented magnets and
segmented cores, usually better for eddy current reductions, probably better for production as well.

From what I understand, the edges can have rough shapes and not diminish efficiency too much. I would think the
opposite, since this is where most of the torque flux is happening.

Not sure if they're using laminations for cores, but I would assume so, and this might be a way to reduce assembly.
But surely they must have to machine smoothness for the coil windings, so why not the edges? Perhaps to match the
magnets.

Like you say, most likely it has to do with efficiency, I'm pretty sure these guys aren't trying to save money anywhere.

I found a patent that shows the method a little more closely.

 

[TorontoBuilder]

Whats strange from my observations is the stator teeth on the magnex picture i posted, they have a sarreted edge down the face reverse to what i have seen to date specially non axial any breakdown will occur at these points.

Got me thinking is this what they are doing to eek a bit more efficency out adding more force from the steel albeit for free as its hard not to power the steel if its present.

Ian I too am not sure what you mean about serrations on the magnax stator face. The pics and patent info I have for the components show that while they appear to be made from laminations, they dont have serrations.

 

[Ianhill]

See how theres a jagged edge down the face without rounded edges, on traditional motors that would be hot spots on the steel.

Im aware the segments are built from laminations but they stack in a different oreantion theres a flat face of stator facing the magnets rather than edges face on.

These edges face the opposing coil im not expert but its 1 or a mixture of these 3 things help efficency, loeer high speed losses or smoother motor reduce torque ripple by sharing losses mutually coil to coil.