There were two versions of the C-S AP system. The dual stick which was the first to market, and the two axis single stick about 6 or 8 months later. The same stick assembly was then used in the early DP-5 radios.
The point of my previous post is about the encoding scheme for their so-called analog radio. In effect it was really an extension of a galloping ghost (pulse proportional) radio, operating at a much higher pulse rate. Changing the throttle was with the RF either full on, or full off to move the throttle to high or low, with changes in pulse width for rudder and pulse rate for elevator. While the throttle was moving, I believe that the other two servos tended to move back to their fail safe position (to neutral). When C-S came out with their GG system later (NPT for nervous proportional transmitter), all they had to do was change the pulse rates down to match up with the Rands systems. At any rate, it was a clever way to avoid having to build frequency discriminators ( a device that turns a change in frequency into a change in voltage) with the usual heavy toroids in other true analog designs.
The pulsed RF was also modulated with a tone at about 3700 Hz, which helped prevent CB activity from shooting you down (most voice information over a phone line is under 3000 Hz. It's also why it can be difficult to tell the voices apart of siblings, because all the higher harmonics are gone.). Anyway I think the filter is constructed from the transformer in the receiver after the signal is demodulated.
While the included schematics showed two sets of batteries, there were some circuits around for using a single set with the addition of noise suppression components. The receiver may not have been very good in this design, or may have been designed to ignore all but strong signals, since they recommend a vertical whip for the airborne radio. But if you flew upwind, the plane should get close enough to regain control if it went into fail-safe.
All in all, an interesting design in the development of modern RC systems.