I don't know how to post pictures here any more but with a bit of luck I have managed to upload two images to my files.
Try looking for Radial jpg and radial 1 jpg.
The first is a slotless alternator using an inner rotating magnet, the second is the winding for a drum type where the magnets rotate outside the stator.
You should be able to see the laminated core and the 2 shapes of coils needed to do the overlap. You can do it as a 2 layer diamond winding if you prefer or you can do it as a diamond mush winding and just squeeze the end connections down over the wend of the core out of the way somewhere.
It becomes tricky to wind if you get the coils much over 1/4" thick so it is best to work with air gaps of about 3/8" or less. With a radial with decent bearings you should be able to get 1/4" thick coils in a 5/16" air gap but if you go up to 3/8" you still can manage with fairly thin magnets.
I used curved neos about 1/4" thick, they were the wrong curvature for the stator but were cheap ( ex voice coil motors I believe).
Standard rectangular blocks would be fine and 3/8 to 1/2" would again do well. The normal 2 x 1 x 1/2 should make a nice machine.
There is no force tending to throw the magnets off with an outer drum type and the magnet drum gives a considerable measure of weather protection, i have not had the magnet corrosion problems I have had with axials.
The iron loss drag is negligible and it can't cog so start up is good. Cooling is way better than a dual rotor as you have direct thermal transfer to the core. With neo I am not sure slots are beneficial, you get tooth saturation with gap flux much over 600mT and you can get those sort of gap flux densities or more in this slotless design with very much lower iron loss. The things are resistance dominated with no signs of reactance limiting under normal working conditions.
Flux