MULE1.2 Axial flux test motor/bicycle specific

[Thud]

This will be the build log for the small "test Mule" Axial flux motor.

Objective- To learn as much about motor design & application as possible.

On my recent build I am quite pleased with the performance of my relitivly tiny RC outrunner motor. with the experiance gained I realize I will need much more power for my next build, Since I make things for a living, I thought, why not make the motor I want?
Here is a list of issues from my perspective:

1) RPM's- They are extream,freash charge voltage spins the 50mm outrunner at 12700.
This is an niose issue more than anything else. An open can whistling away,as well as the additional reduction
parts spinning at dizzying speeds-We need Tourque at a resonable rpm

2) Size- the outrunner I selected was size specific to cram into the rear triangle of a small bicycle frame. The
larger motors that have enough power create other issues, width moves most out of the crank circle.
I think thiner is a winner if we are not going to settle on a hub motor.

3) Controllability- As much as I love my sensorless RC controllers, I know they are stressed beyond there design intent. The smaller sensored controllers are going to provide a few more options regarding the performance envelop on this & future builds.

This motor is described as "Single rotor axial flux/split stator" There may be other topologies attempted latter, we will give them their own desin specifc attention when the time comes.

I wish I could list the expected performance numbers. There are too many variables inplay right now to speculate. The mule expectations are to match the 50mm outrunner for tourque(or at least 80% of it) & get the Kv as low as posible.

Progres:
rotor construction-
.5"dia x .25"t Neo-magnets n42 (or n48-i forget)12lbs pull force each. Aluminum hub & outer retainer ring

here it is with epoxy resin & fillers for securing the magnets

I Use my printer for a lot of layout work on most of my projects these days. print a 100% scale picture & rubber cement it to some plate & chop away.
Here is the rotor with the shaft pressed into it & the motor cases roughed out.

& here it is along side some other motors for scale.

After finishing the case (bearings & coil anchor points) its on to the stator teeth.
Dollar totals for the build thus far-$100.00 (magents & 12lbs of magnet wire)

 

[liveforphysics]

Here is a list of issues from my perspective:

1) RPM's- They are extream,freash charge voltage spins the 50mm outrunner at 12700.
This is an niose issue more than anything else. An open can whistling away,as well as the additional reduction
parts spinning at dizzying speeds-We need Tourque at a resonable rpm

2) Size- the outrunner I selected was size specific to cram into the rear triangle of a small bicycle frame. The
larger motors that have enough power create other issues, width moves most out of the crank circle.
I think thiner is a winner if we are not going to settle on a hub motor.

3) Controllability- As much as I love my sensorless RC controllers, I know they are stressed beyond there design intent. The smaller sensored controllers are going to provide a few more options regarding the performance envelop on this & future builds.

I agree firmly with all of this you stated above. I see some issues with plan A though It's ok, because we have lots of alphabet left, and you've got myself and other more smarty-pants guys to help us arive at the goal on one of those letters.

First, if you ask magnets to spin next to a conductor, you just made an eddy current brake/dyno. The magnets can be mounted on a disk made of a conductor, because it has no change in distance from the magnets to the conductor as it spins. They can't be asked to spin next to a big aluminum sheet though, or you are asking them to become a big inductive heater.

Second thing, with only coils on one side of the rotor, the magnetic field will not be captured at all. If I'm looking at things correctly, the flux will simply deform away from the rotor magnet. Am I not seeing it right?

 

[Thud]

Luke,
I am an electrical imbicial.
I assumed if I electricly isolated the winding cores from the plate, all eddy curents would be isolated to the coils, I know alumn is a conductor but never contemplated eddy curents in the flat plate of a nonferris material. Smarter now than a minute ago.I though I had end run inductance issues.
My only study of inductance is foundry related, & I know only enough to embarass myself. If that is the case I am already moving to plan "B" & will make the case from composits/non conductive materials.

I assume I can use condutive materials, out of the spinning flux pattern with reckless abandon?

The coils are actaualy split & straddle the magnets on either side. I realize I may need to add iron to the ends of the stator teeth to maintain flux densities. Here is a section view showing intent.

I know you are on the ironless track ATM but I would like any comments or suggetions that pertain to the desin intent of this configuration.
Thanks for the Catch & any additional input.(better late than never) please correct me at every slip.
Todd

 

[Miles]

I think it only needs to be a conductor, for eddy currents to occur. Presumably, the better the conductor, the worse the eddy currents....?

 

[rhitee05]

Yeah, eddy currents will be an issue. I'd suggest something like plywood for a first prototype, save yourself the effort of making something nice out of plastic/carbon fiber/whatever when you'll probably end up changing something in the design anyway!

As far as the coils, I don't see any reason you can't try it ironless first and then add some sort of core later if you choose. It would be a useful data point if you got performance for both versions.

 

[liveforphysics]

Iron cores to help direct flux into the magnet flux area will be critical for a design like this. I'm pretty sure that the way to get around only having 1 rotor disk and not sacrifice half or more of the torque is to make the iron run in horse-shoe like shapes, with the coil anywhere along them, and these would go radially around the single disk.

I'm no expert, so you guys with a better grip on things please feel free to chime in, but a coil on each side, with an independent iron core, would use twice the power to make less torque than a single core with a magnet on each side of it.

In an axial layout, the hot ticket (as well as key to low KV) is pinching the coil down sandwiched between magnets. We don't have to pay (in watts) for the magnetic field that the magnets create. We do have to pay for every bit of the flux we generate with our windings, which is why getting something efficient (as well as torquey) requires capturing every bit possible of that flux we generate.


When you make the end-housings, try to stay away from carbon based composites. It's not a great conductor, but it would still be like leaving your foot gently on the brake as you're trying to run your motor, as well as potentially be a huge portion of the total heating of the motor.

Axials are tough to make right. I think the results of your work will be so well worth it in the long run though You're awesome for being a driven do'er, and it's going to pay off big when we get this all nailed down.

Your craftsmenship is outstanding!
You are the man Thud! We're going to get nail this axial motor design!

 

[bigmoose]

Yes, sadly, the current design with aluminum side plates is in fact an eddy current brake. Here is a tutorial:
http://farside.ph.utexas.edu/teaching/316/lectures/node89.html The sideplates will need to be nonconductive. Plexi/Acrylic is machineable with a wood router... well sort of.

Study the "right hand rule" it is the basis for all the eddy current problems in the world!

Pix of small eddy current dyno:

Don't worry though this is how we all learn. ... I use this in my bio: "I've worked, I've played, I've served. I've laughed, I've cried, and I've picked up the pieces more times than I can count. Each day, taken one at a time, shaken not stirred. Life is kewl... Prov 20:29"

 

[rhitee05]

Iron cores to help direct flux into the magnet flux area will be critical for a design like this. I'm pretty sure that the way to get around only having 1 rotor disk and not sacrifice half or more of the torque is to make the iron run in horse-shoe like shapes, with the coil anywhere along them, and these would go radially around the single disk.

I'm no expert, so you guys with a better grip on things please feel free to chime in, but a coil on each side, with an independent iron core, would use twice the power to make less torque than a single core with a magnet on each side of it.

I'm sure you're correct, Luke, that iron would increase the torque from this motor design. I'm not sure if it'll be a 2x difference like you suggest, but it would be helpful. I think what would be needed is back iron like what would be done with external magnet rotor design. I pulled a snippet from a paper that illustrates the flux flow in some types of axial designs:

View attachment Axial Types.JPG

His design is similar to that shown in D. Providing some back iron would allow flux to flow from one pole to the next circumferentially. I think what you suggested would also work with U-shaped pieces of iron, but that's a slightly different design.

I'd still be curious to see what the actual difference would be between the cored/coreless versions. I might attempt to model that if I have the time, but that would be pretty involved to get an accurate model.

 

[etard]

Maybe you could pick up some Ipe wood commonly used for decking during your beta testing, it has the same fire rating as concrete and steel. Might bring a whole new smell to "magic smoke" coming off the motor. :wink:

12 pounds of magnetic wire? How much of that will this motor require?

This is an awesome project, I wish I had the time and the skills to tackle one of these motors.

It's funny, last year around this time everyone was hyped on the reduction designs for high rpm RC motors, now it's all about building your own motor. But we still don't have a good option for applying the power to the rear wheel. Gary's adapter comes closest to what we need, IF you have a three speed hub.

 

[liveforphysics]

What you've got there isn't a bad motor attempt, but a perfect motor load testing tool. Unlike friction brake dynos, eddy current dynos are extremely consistent. By mounting the case to a pivot arm with a spring scale supporting the other end and adding a reflective line on the shaft for a laser tach, you've got the ability to apply a load, and measure torque and speed. This is everything you need to know to dyno motors.

Make sure the pivot point for the case is shaft centric (like a large bearing that slips over the whole central shaft area). This way you won't get goofy numbers from tension on pulley belts and things when dyno'ing.

You've made a damn fine and very useful tool!

 

[enoob]

i dont know what you make for a living but looking good from here.

Seems to me you may have very usefull stuff on your hands here. if your gonna keep making motors this could be very good as liveforphyics has said.

those two plates making the case would be simple to transfer to some plywood or the fire retardant material already posted. may want to look at high grade marine or aircraft plywood they are very nice to work with as well with good structural integrity. lil fibreglass for extra strength if needed.

that all said . how much you want for one of your eddy current dyno's

id love one.

 

[Thud]

LOL, Thanks LFP. that is really turning lemons into lemonaid! dyno problem solved! (completely left handed )
Thanks guys,
Don't feel bad for me, as I am getting so much already from just hanging around all you "large brained" fella's
Also materials & applications won't be an issue, I used to be the guy in the small room ajacent to the engineering dept. in all major manufacturing facilities.(you know the guy all the deigners & engineers bounce ideas off from & have make their prototypes of stuff & what not) I have all sorts of goodies to build with & I even have prototype foundry capibilities.

I have some fabric made of spectra fibers I been looking to employ

I guess given the information presented here I may abandon the Mule1 & retitle it "Thuds Serendipitious Dyno"

Mentor's, please answer me this- If I were to make a laminated core like this:
HORSE SHOE CORE

It would be an improvment to mirrored coils by nearly 2x?
& would a laminante core direct the flux feilds with the Laser like acuracy I can get with a mirrored set up?

It Is my desire to reduce rotating mass, but I have no problem subordinating to the laws of physics. I am allways ready to meet in the middle

 

[Jonathan in Hiram]

Here's my take on it.. The purple is magnets, the white iron and the orange is copper..

The measurements are one inch cube magnets and the entire motor is a little over ten inches in diameter and three and a half thick. The torque arm length is about three and a quarter inches from the axis to the center of the magnets.

 

[liveforphysics]

Thermal epoxy a CPU heat sink on that eddy dyno, and drill some outside radial holes to keep the magnets from cooking. That thing should make a fantastic dyno. If you can get the aluminum to almost contact the magnets, it will lower the speed at which it provides a counter-torque. It's quite drastic how much it effects things, even 1mm less airgap could cut the RPM in half that it provides a given amount of counter torque. Of course this can all be easily adjusted by having a few pulley choices available to drive the dyno with on your test bench.



What Jonathan drew is exactly what I was saying about using horseshoe shaped iron, and then running your coil around it. I tried to communicate that in an earlier thread, but I seem to suck as verbally describing things to everyone except Jonathan, whom I apparently share a communication wavelength with (I likewise find your words/descriptions to always seem exceptionally clear in my mind for some reason).

The challenges it poses are of course that the stator encapsulates the rotor disk, so that building one means the stator must be multi-piece, or assembled and wired up around the stator. Not a big obstacle for custom one-off stuff.


I gotta sleep/rest now. Flu = kicking my ass.

-Luke

 

[Thud]

Etard,
Regarding getting power to the rear wheel, I think the freewheels to freehub option is going to be suprisingly usefull, asuming all works out on the Beta side. You can trust Luke to "test-to failure"-asuming it can fail, Luke is going to power this thing with his twin motor setup. Thats way more than any reasonable human would ever need.
That wire will last me awhile.....or will it?

I will definatly be making the dyno & see how it turns out. Seems much more repeatable than a prony brake.
(I will add it to the catolog when the time is right)
TOTALY OFF TOPIC-(my thread my rules )
What do I make? lately its furniture stuff for architechs:
Here is a Board room table I installed in Maryland last spring.

This section of the project, I can say is all my handy work. (I need to show the floor guys this desk Jocky still has some chops every now & again 8) )For referance it is 17-1/2 feet wide & 34 feet long.
22 pc of matched cherry veneer/ebonised inlay/stainless steel edge inlay/lights & glass in the center section
steel super stucture in the base. from the Architechs head to my hands. (wood working is easy) on a side note We did the manufaturing engineering for all the custom designs utilized in this install.
I was very lucky when I landed my 1st real job after HS. I was mentor'd by anchient craftsment in wood working & by tool makers in the boat factory. Combine that with a "why not?"attitude & you become ME

 

[Jonathan in Hiram]

What Jonathan drew is exactly what I was saying about using horseshoe shaped iron, and then running your coil around it. I tried to communicate that in an earlier thread, but I seem to suck as verbally describing things to everyone except Jonathan, whom I apparently share a communication wavelength with (I likewise find your words/descriptions to always seem exceptionally clear in my mind for some reason).

Eh, it's just that we share a special kind of crazy..

Your sig line and attitude reminds me Clarke's second law: "The only way of discovering the limits of the possible is to venture a little way past them into the impossible."

The first time I built a hot rod I made a killer engine and then proceeded to break everything in the driveline, one piece at a time. Clutch, transmission, u-joints, driveshaft, rear chunk, axles, rear suspension, lug nuts.

 

[etard]

Nice table Thud, it kinda looks like agiant surfboard. Very classy, and posh, not even gonna ask how you got it in there. Anyway, I don't think the weak link in LFP's test will be the tranny, I got my money on the Dual Drive for sure, either that or his domepiece. :lol:

So this Mott that Jonathan has illustrated here, I guess this is able to be layered in how ever much power you want? It would be truly a step in the right direction to have a motor that is modular and one could simply change out or add parts depending on how much power you needed. Talk about simplicity and ease of mass production!!

Question:

If you built a single layer motor such as the illustration in a 20" diameter or more and made it similar to a hub motor, what kind of performance would one expect? Is that what the csiro guys have basicly done?

 

[Miles]

Question:

If you built a single layer motor such as the illustration in a 20" diameter or more and made it similar to a hub motor, what kind of performance would one expect? Is that what the csiro guys have basicly done?

For a hub motor style, I think you'd use a twin rotor design (CSIRO/LaunchPoint et al), not a twin stator...

 

[Jonathan in Hiram]

If you built a single layer motor such as the illustration in a 20" diameter or more and made it similar to a hub motor, what kind of performance would one expect? Is that what the csiro guys have basicly done?

The larger the diameter the more torque and the lower the Kv and probably the more efficient the motor will be overall. The beauty of this design it that you can pack an awful lot of copper in there and keep your copper losses low by using large diameter wire. A motor is most efficient when copper losses and iron losses are equal which usually happens somewhere around 75% to 90% of maximum load. At higher loads copper losses predominate while at low loads iron losses are greater.

The OD of the motor could be cut significantly by shaping the iron somewhat differently where the coil wraps around it, I'm just a total n00b at 3D design in general so I couldn't get the more sophisticated shape I was looking for, or at least not in a reasonable length of time. The big advantage to the shape I posted is that it can be made with regular grain oriented transformer steel in rectangular shapes which makes it both efficient and easy and cheap to get. Cedric Lynch's interview clued me in to that idea.

If you had a source of old power transformers you could strip them down for the laminations and get started that way, they probably wouldn't be as thin as you would like but they would do for a first try on the cheap.

 

[Thud]

OK,
On the section view of my perceived "horse shoe" core, you guys are saying it would be of no value to make the final turn & point at the magnets like a fire hose?
compare this pic to the previous......now,

Indulge my ignorance for a moment, Given that the pole generated by the winding is parallel to the magnet poles I would assume polar alignmet would see some advantage.(unless magnetic flux is so mobile as to make the twist & not show any reductions)
what if... the core was horse shoe shaped (more race track now) & the coil was split to stradle the magnet like this?

it is essentaly the same winding, however many turns to get the desired effect at the selected voltage/current.
thanks for the indulgance. I hope I am not chaffing with my educational probes.
Edit: After re-reading the resonces I better understand the point. I seem to be hooked on the flux feild being strongest near the coil, core or no core.
Am I way off base in that assumption?

Bigmoose, Thanks for the link, very good read & clear language too. I will be refering to that often.

Regarding design intent:
Is the single rotor design worth pursuing?

I have no issue changing gears if we can reach a consensus. I am itchin to make something spin & can lap & fit parts to finite tolerances if required. I just want to be pointed in the right direction when the fuse is lit, its a long walk home :lol:

 

[Miles]

A big advantage of Jonathan's design is the ease of winding. 8)

Alignment is going to be tricky with these yoked core designs.........

 

[Miles]

Edit: After re-reading the resonces I better understand the point. I seem to be hooked on the flux feild being strongest near the coil, core or no core.
Am I way off base in that assumption?

I don't think it's that significant - the reluctance of the iron is so low (unless it's beginning to saturate, of course.....).

 

[enoob]

may have been said already in language im not getting but another identical rotor sandwiching something like the stator in this post from earlier ? some steel behind the magnets ?
http://www.youtube.com/watch?v=hPyl776j7no
may not be what you started after but it would spin


off topic . thud now you got me digging for shots of the board room table i did for a internet start up a few years ago .
i love nice inlays with different material .
very nice . cherry gives such a rich color

 

[Miles]

off topic . thud now you got me digging for shots of the board room table i did for a internet start up a few years ago .
i love nice inlays with different material .
very nice . cherry gives such a rich color

We should start a woodworking thread in General Discussion

 

[enoob]

We should start a woodworking thread in General Discussion

something tells me if these diy af motors are a hit that thread will be a must. AND i may have something tangible to contribute in there