Ah, the old memories come flooding back!
One of the hard tradeoffs in SRM design is that the torque is (almost) inversely proportional to the aligned airgap between rotor and stator. This is because the torque is proportional to the change in inductance between unaligned and aligned states, and because virtually all of the magnetic reluctance is in the airgap, cutting the airgap in half cuts the net aligned reluctance almost in half, which therefore almost doubles the aligned inductance, and hence the torque.
However, two bad things happen when you try to reduce the airgap too much, both of which Skogs alluded to. First, just stamping out rotor and stator pieces from lamination is no longer accurate enough, which starts eliminating the manufacturing cost advantages. Second, as the rotor's salient pole whips by the stator's salient pole, a virtual shock wave is generated in the air, making them very noisy.
Still, I am kind of disappointed that they haven't taken off. About 15 years ago, I did some playing with an SRM built by Warner Electric for the last generation of HP's vector pen plotters. These motors were a little smaller than a Coke can (~500W IIRC). I could run them to 30,000 rpm, and it was the controller that gave out, not the motor. I could also reverse the motor from +9000 rpm to -9000 rpm in 20 msec (unloaded). These were done under servo control with an optical encoder on the back.
jraef: Corvallis, OR is near me (LA, CA)? Only about an 18-hour drive. It's all relative, I guess...
Curt Wilson
Delta Tau Data Systems