Something just occurred to me. I did get to thinking.

My previous answers were wrong. Peter's spinning antenna wouldn't
produce a circularly polarized wave (as universally defined) even if it
was synchronous with the wave frequency. As I've said, a circularly
polarized wave has constant E field amplitude; Peter's wave would have a
time-varying amplitude. If it were synchronous, the nulls and peaks
would always occur at the same places in the rotation cycle, so they
would occur at fixed angles relative to a rotational reference point. If
non-synchronous, the nulls and peaks would rotate at the beat frequency.

It seems to me that the way to mechanically generate a circularly
polarized wave would be to rotate a source of *static* E field, for
example, a short dipole with constant applied DC voltage at the
feedpoint. That should produce a circularly polarized wave with the
frequency being the rotational frequency of the dipole. At any point in
space, the E field would change with time, and would propagate, and it
would look exactly like a circularly polarized wave broadside to the
rotation plane.

If the scheme works and radiation is occurring, then power must be going
into the antenna, which in turn means it's drawing current that's in
phase with the applied voltage. When stopped, no current will flow, but
when rotating, it does. So how does the antenna know it's rotating? How
about this -- if you instantaneously move the antenna into some
position, a static E field appears there, and propagates outward at the
speed of light. Closer in than the leading edge of the propagating wave,
the field is static. When we rotate the dipole to a new position, it
moves through the field from its previous position, which induces a
current in it. Hence the current. It's fundamentally a generator, with
the field being in the air.

I'd be willing to bet a moderate sum that if you did apply a DC voltage
to a dipole and rotated it, you'd see an alternating current with a
frequency equal to the frequency of rotation, and a circularly polarized
wave broadside to the antenna. I suspect that the current and the
radiated field increase in amplitude with rotational speed, so you might
have to get it going really fast before you can detect the effects.

Now there's some food for thought.

Roy Lewallen, W7EL