Think I can make it? (electric 9-pole radial engine replica)

[metrologist]

Gregg, thanks for the idea. That would work too. A lot of small parts I'm finding it hard to work with.

Another try was to make a flat hub, more like a gear, so the magnets could be oriented with N-S facing poles and have some separation. My tool was a little small so the magnets will not quite fit into the slots. It's also kind of sloppy so I'll make another using a rotary table.

I'm adding a pic ture with a US dime for size

[Gregg]

I just reviewed a recent AvE video where he takes apart a Makita brushless drill.  About halfway into the video he has a nice view of the rotor and stator.  The rotor has 4 flat neodymium magnets inserted from the side.  I thought it may give you additional inspiration.  He showed less poles in the rotor than windings in the stator.
You are making it smaller than I imagined.  The thin brass hobby tubing may be useful.  That small I would try making the rotor two piece and glue the magnets in with epoxy and machine the OD down after the epoxy has set.  Maybe tap the holes in the non magnetic ring part way to help the epoxy bite, especially if using plastic.
I don't envy you winding the coils; I don't see that well anymore.  I used to fix clocks without magnification, but not anymore. 
Good luck on your project and post some pictures or a video of it running.

[metrologist]

I don't do much without magnification anymore either, but it is proving harder to hold things than see what I'm doing, because I have some optical stuff laying around. Those little magnets are about the size of broken off pencil lead.

I got inspired by small things when a lady at work gave me a toothpick that she had whittled into a giraffe - with a tail and 4 legs standing on grass... I still have it somewhere.

[metrologist]

I got all the holes milled to depth and may make an aluminum ring to press around the circumference to hold the magnets, or I may just super glue them.

I have a good idea for the body now too, since the RT is set up...

A question or two:

I was thinking of using 4 transistors to swap voltage polarity across the windings (12 transistors total). I was going to wire them as Wye, but without the return lines connected together as neutral. The return lines would feed through transistors to ground. I would have pulsed AC square wave, each phase 120° degrees apart. This seems like it would work, but not sure. I attached one phase driver for reference where V2, V3 are out of square wave from mcu. I used an inductor for the winding with 0.1L and 2Ω series resistance. LTspice shows ~1.2A through the coil.

If that would work, I have not yet seen how the motor could be wound Delta using the above driver. I'd need 6 connection points so separating each leg looks the same as above.

I also have a regular RC speed controller with three wires. How would I convert my driver so it only has three connection points instead of 6?

[metrologist]

Someone thought it might not work. I decided to use steel pegs shaped like a T for the cylinders, which will be pressed into the body and come closer to the rotor, then wrapped on the outside with motor wire. Then I drew the magnetic fields as I think they are at the start condition, with #1 at top going clockwise.

[Gregg]

Those are very nice drawings; keep up the good work!
I know this is just a novelty and basically just needs to turn and start by itself; maybe turn a small propeller for effect and some cooling. 
For possible inspirational purposes, I don't pretend to be an expert: If you could couple the outer coil cores together with magnetic material it will provide a better circuit for the magnetic flux.  Even if the connecting material was looped toward the axis for appearance.

[metrologist]

Now, that is a good idea to use small machine screws. Would it be better to make the body from steel instead of aluminum?

I had considered it would be good to make the body from many layers of steel sheet, as pictured, with plates or endbells on each end.

I loathe sheetmetal work

[Gregg]

Your magnetic field of the sheetmetal as shown would only work for a motor with permanent magnets on the outside.  The center ring will conduct the magnetic flux mostly between coils, leaving very little if any to turn the rotor.  If you invert the design, it would have an outer ring that would take away the esthetics of a 9 cylinder radial engine.
I guess it is back to the drawing board.  I thought if the outer ring of magnetic materials was formed along the lines of the radial coils but far enough to not interfere with the coil magnetics too much, it would make your motor look more like an engine. 
(Motor and engine are often interchanged and therefore seldom used correctly on the internet)

[metrologist]

I think it will be as such.

(I added the wiring)

[Gregg]

I may be wrong, it certainly wouldn’t be the first time, but I think you should re-think the ways magnetic flux work in a motor vs. a single magnet in free air.  I believe that it is the flux between poles that gets the job done via the electromagnets pushing and pulling on the permanent magnetic fields.  You certainly have the 3ph concept down. 
By the way, what are you using for CAD?  It looks interesting

[metrologist]

I am using Adobe Illustrator.

Do you know if the flux would be the same as you've drawn it if there was no structure? Magnets and coils floating in space. I was considering it that way because I am using aluminum.

[Gregg]

If all the coils are mounted independently it would be much better than the ring of magnetic material near the rotor.  Air has some magnetic permeability; about the same as aluminum, wood or many plastics.  Iron has many orders of magnitude better permeability which would cause the magnetic flux to follow along the path of greater permeability rather than interact with other magnetic fields like the permanent magnets.  A quick Google search for magnetic permeability will show some interesting results. 
The problem with aluminum in close proximity to changing free magnetic flux (flux not being directed by a medium of much greater permeability) is that aluminum is such a good electrical conductor that eddy currents will be generated in the aluminum by the changing fields of the electromagnets which will cause heating because they are essentially short circuit and will need more current to overcome this loss.  To demonstrate this, try dropping one of your small magnets through a similar ID copper tube and note how long it takes to fall.
Plastic would be a better material; using something like Delrin you wouldn’t need shaft bearings.

[metrologist]

I have black Delrin. I think I'll still give it a whirl with the aluminum body and later compare with Delrin. The aluminum is already chucked and the roughing started a couple days ago. Hope I have some suitable screws.

[Gregg]

Nails cut to length may make good cores for the coil and they are cheap, easy to chuck up and power wind the wire prior to cutting them to length.  The sheetmetal laminate plates you mentioned earlier may be a little more efficient, but what a pain for winding coils

[metrologist]

I was thinking 16 turns. But I would need a continuous wind between 3 cylinders. I was just going to wrap the single wire manually, but I think to get the wires to lay down nice, I'd need to wind all three runs at the same time, from cyl to cyl...

[Gregg]

I probably thought you were using more than 16 turns since the magnetic field is proportional to ampere-turns.  More turns at lower current equals the same magnetic flux and usually less heat.
But just getting it to work is the first step, take some measurements and modify if necessary. The magnetic field doesn’t care if the wiring is neatly wound. You do need to make all the coils wound the same direction.
I believe you have a lathe.  You could wind the first coil, put some heat shrink or tape around to keep it from unravelling; hold the first wound nail against a chuck jaw with tape while you wind the second, repeat for the third.  You probably don’t want to wind it very fast so you can count turns.
This is a great project; you do all the work and I just observe. 

[metrologist]

Not much else to do but put it together now. I'll see how many turns feels write when I start rapping.

[metrologist]

I got as far as winding the first set of 22 turns. I decided to test for a magnetic field by running an amp through the circuit, and could not detect any field with ferrous objects or magnets. I set a small magnet on top of one of the nails and reversing current had no effect on the magnet.

I recall an elementary school exercise involving a 6V alkaline lantern battery, some speaker wire, and a nail. I should go back and study my notes.

This is 28ga wire and can only handle about 1.5A, so I'm sure this design will not work.

[jpanhalt]

Your magnet wire looks used and could have multiple shorted turns.   Try the test with ordinary, vinyl insulated hook-up wire (stranded or solid).

[Gregg]

I don’t envy your task of winding wire in such a tight configuration.  Maybe it is time to bench test your theory.  Maybe three nails through a block of wood to try more windings, different voltages etc. and test the magnetic fields with a lot less hassle and if a meltdown occurs it won’t ruin all your hard work.  For a bench test you could wire coils with separate lengths of wire to make things easier and possibly test three phases.
I’d start by drilling holes in the wood so that the nails are a tight fit but not split the wood and grind off the nail points flush with the back.  Then during your tests you can put a suitable steel washer or sheet of steel across the bottom of the nails and see how much it enhances the magnetic fields and what it does to the current draw of the coils.