Electrolytic tank analog computer for IF design
 

This actually pertains to receiver design rather than broadcast gear, but it seemed too general and theoretical to go in the black and white TV section.

In Donald Fink's Television Engineering Handbook, 1957, in the section of I-F amplifiers and detectors by H. H. Wilson, Jr., a curious analog computer for designing AC circuits, and IF amplifiers in particular, is described.

The linked file shows this section of the book and three references to the technique. A shallow tray of saline solution is used, representing the complex frequency plane.

Theory:
Mathematically, calculation of the steady-state (sine wave) response of an AC circuit is done using complex frequencies and "poles" and "zeros." The poles and zeros are points on the complex plane, that is, complex frequencies have a real and imaginary part. The complex plane has axes of real and imaginary frequency, so a complex pole or zero is specified by its location on the plane.

The response of the circuit to a real frequency (with only a real, sine wave part) depends on the distance of the real frequency from the poles and zeros of the circuit. The real frequencies have no imaginary part, so they lie on a single straight line separating the plane into half with positive imaginary part, and half with negative imaginary part.

The illustrations of IF response on page 16-51 show one half of the complex plane along with the resulting IF response to real frequencies, which are on the horizontal axis. The reason for using only half of the plane is that this is the half where the poles and zeros of "minimum phase" networks always appear. This includes all common AC circuit elements like R-C circuits, tuned circuits, etc., but does not include lattice circuits, which have phase splitters or inverters.

The electrolytic tank analog computer was apparently developed at Harvard and Imperial College, and originally used to calculate electric fields between electrical conductors. Apparently, the similarity of the equations to those for calculating AC circuit response was noted later. The references state that accuracy of 0.1 percent was obtained, and various improvements were made by adjusting the shape of the outer ground ring (I think to better emulate an infinite tank).

I haven't found any reference to indicate that a TV maker actually built one of these to help with circuit design. I could imagine RCA, MIT or Bell Labs doing so, but don't know.

Of course, digital computing can do these computations much more easily, quickly, and accurately today, but I thought it was interesting to see how the problem could be tackled in the late 1940s.

http://www.bretl.com/tvarticles/docu...ctankforIF.pdf