SRM has to add and remove energy from a rapidly changing inductance.
Braindead sinewaves won't produce the smooth result you might expect.
There may not be any simple simple fix for silencing audible bumps that
are harmonics of any reasonable direct drive cog rate.
Now on the other (not direct drive) hand: Dyson has shown that SRM can
be silenced by cogging faster than you can hear it. The challenge then is
to reduce that without mechanical whining. I don't have a ready answer
for the mechanical challenge.
Assuming the rotating part does not endure flux reversal, other parts
that will have to endure it probably aren't going to get above audible
with simple laminated iron. Ferrite could, with greatly reduced flux.
Which requires higher RPM to make the same power. Ferrite seems
a bit brittle for motor use. Other sintered powders? Hiflux, Sendust?
Those same materials could apply to direct drive at low cogging rate
but very high switching rate with lots of opportunity to control those
bumps. But even the best powders carry far less flux than solid iron.
Without reduction, might be talking about a very wide and heavy
motor to have decent torque.
https://www.micrometalsarnoldpowdercore ... ls/hi-flux
Hiflux is gonna be spendy with 50% nickle...
There is another. I forget the name which makes hard to lookup a link.
But its some sort of powdered high silicon steel. Probably ferrosilicon.
The stuff is comparatively cheap. Strong? Used as a filler in JBWeld...
I've molded JBWeld into distributed gap cores that could handle 1MHz
Don't know how that changes when pressed and sintered for maximum
Doh! On the same site, in plain sight, I just missed it.
https://www.micrometalsarnoldpowdercore ... ls/fluxsan
----- edit to add later thought -----
Suppose we were to bury a hall sensor in the rotor.
I dunno how you connect it, rings and brushes maybe?
The object here, to learn a PWM curve that serves to
keep the rotor flux constant. Could that help shape a
smoother hand-off from one pole to the next?