I notice these would use round coils .
I must not be thinking correctly, I was under the impression you wanted the magnets to pass over straight lines. Round coils would make the coils much easier to wind.
Wound tightly over pvc pipe the pipe section could be left inplace while casting ( this could also be done for ovals or triangular coils by heating the pipe and shaping it to suit)
There is also the point of surface area on the magnet its self , in which case a 1" cube would have more face surface and more total magnetic mass.
I'd have to see the magnet field for different shapes.
Thank you for simplifing the 30/40 degree I was looking at a 30/20 degree off set Mags to coils and using 18 coils in series of 6 it would have come out to two magents centered on the first and last coil in the series. My misguided thought being to double the magnetic flux intering the coils. The other portion of this thought was to make the coils half as thick to gain staturation. In my drawings there would be no need to overlap.(stone aged drafting compass and protractor)
The rotor discs arent a big problem I have a number of 6,7,8,10" steel sanding discs 1/4"-3/8" thick .