Yesterday I read about a new way to send more than one signal on a frequency at a time. (Slashdot link below).
It took a few hours contemplation for me to finally get my mind wrapped around this one, here's my explanation.
Imagine you have an AC motor, with 2 poles. If driven with 60 hz power, it will spin at 3600 rpm (60 rotations per second).
The same 60 hz power can be fed into a 4 pole AC motor which will spin at 1800 rpm (30 rotations per second)
Generalizing, The same 60 hz power can be used for a 2n pole motor to get a spin of (7200 /.n) rotations per second.
If you were to hold a magnet in the center of any of these motors it would vibrate at 60 hz. - This is analogous to a normal antenna trying to pick on one of these signals, it would pick up the carrier frequency regardless of the spin.
It is only when you let the magnet spin that you can measure speed at which the field rotates - This is why it takes multiple antenna to discriminate between signals.
Now the real fun is when you imposed BOTH a 2 pole and a 4 pole signal in the same space... if you put a rotor in and let it spin, you might get either 3600 or 1800 rpm... hard to guess.... but if you spun it up to just under 3600 rpm, it would pull forward to 3600 rpm, with the same effect if you spun it at 1800 rpm. There would be a lot of energy lost in cross coupling of the winding coils, though.
If we went a bit further, and considered a motor which had a way to simultaneously be powered in 2, 4, 6, 8, and 10 pole modes, you'd be able to measure the voltage fed into each of them by rotating the motor up to the appropriate speed and measuring the torque on the rotor... you'd get 5 different numbers, each linearly related to the amount of excitation for that set of windings.
You lose the utility of a motor, but you get the ability to convey 5 different signal levels with 1 frequency. Cool stuff!