This reminds me of a claw-pole laid out flat.

Speaking of which wonder if you could do the same with an old 12si - rewind the big coil in the centre of the claws with thicker wire, then put a ring of big mags around the outside where the stator used to be. Clamp the 'rotor' and rotate the ring of magnets.

Wow - I might have to give that one a try! The claws should pretty much look after cogging too. If I can think of a way to do a sim in FEMM I will. But that might lead to building one!

Keep up the great work!

Also, if the outer bends of the plates were extended farther out, wouldn't this reduce the shorted out flux path between plates?  Shorted out N plate to S plate, from the outer bends to each other.

Did either of those make any sense?

How about taking the plates out much farther radius, double bend them back so the overlap on differant planes but the N-S poles overlap.  The magnets on a non-flux conductive plane between the 2.  All the flux, from both sides, of every magnet, must go through the coil, no choice.
Meaning the flux now being conducted in the magnet retaining plate is being used in one of the coil plates.
Both sides get used.

O Boy.  I know what I mean.  I could email a drawing.

Positioning the magnets outside the rotor, rather than beside it, so the flux enters the edges of the plate, might help.  (Think of it as a laminated pole piece with a single lamination.  B-)  )  That would also get the magnetic field attraction force off the thrust bearing just as a dual-rotor would.

You could also use a stack of rotor "flowers" to make a wider laminated structure.

But you'd have similar eddy-current problems in the center.  I don't see a simple solution for that at this time.  (Maybe a bundle of laminations through the center?  Metal wires?  Or concentric cylinders?  Meanwhile the center of each "flower" in the stack would need a different-sized hole.)

The rotors of auto alternators don't have this issue because they are the field coil, not the output coil.  The mag field through them changes very slowly as the regulator adjusts it to control the output voltage.

Instead of having the coils
outside the rotor, the coil(s)
are inside the rotor.

It is easier to get speed
for the magnets this way.

Hugh's 'brakedrum alternator'
is based on quite similar idea.

I think that this design
is a good one, but probably
needs a laminated core for
the coil(s)?