Hmmm. Some food for thought here. Thanks.
Zubbly's ebay supplier (tropical_son for others who are interested) is supplying skewed magnets for induction motor conversions. Not quite what we need, but maybe he'll be willing to have the magnets we need machined for us. I'm still curious as to how the machining is done, but it looks as if it might not be necessary for us to do it, though it still seems to me that counterbored holes would be better than countersunk, since flathead screws exert some radial force on what seems to be a very brittle neo.
It also seems that the epoxy coating that's supplied on the magnets isn't very durable, suggesting a few coats of epoxy paint which could also be used to coat bare magnet surfaces. Epoxy automotive paint has a good track record for longevity and we'll probably be coating the rest of the machine with it (speaking of nasty chemicals!) so painting the hub plate and magnets before assembly and depending on the paint system to resist corrosion (both fretting and moisture) seems like an effective strategy.
Just FYI, the blades will be mounted on the windward side of the 1/2" steel plate and each blade will be "topped" with a 1/4"x8"x8" plate that covers the root end. Four 1/2" grade 8 bolts and nylocs hold each blade plate and blade to the steel hub plate. It could be more structurally elegant, but it's simple, proven to last and allows easy individual component replacement on the tower.
"If you still want screws, how about filling in between and around the magnets with a urethane epoxy and then drilling and screwing the epoxy to the plate (use non-magnetic stainless screws)?"
Maybe I'm misunderstanding, but I don't quite see the point of doing both adhesive and screws. Either method works by itself, why make it more complicated than necessary?
News flash- just found a hardness value of Vickers 500 for neos, which converts to Rockwell 49C, easily machinable with carbide tooling. (Don't think EDM is necessary.) Now how to get rid of those pesky chips?.
Will keep you posted.