First-Hand:Seeing Was Believing
Contributed by: Thomas Cuthbert Jr., IEEE Life Member
I joined Collins radio with a B.S.E.E in 1959, after serving as a U.S. Navy pilot and electronics officer and attending M.I.T. and Georgia Tech. During my interview and thereafter, Collins designers talked about loaded Q in circuits, meaning the ratio of L or C reactance to the apparent resistance at a particular frequency; that is how much reactance power (vars) circulates relative to the real power (watts) passing through. In 1959, the main application of loaded Q was calculation of resistance transformations that are equal to 1+Q², "one plus Q squared" being the method's name. HF resonators usually are composed of adjacent L and C branches having equal reactances of opposite signs and operate at a chosen "parallel resistance" at the passband center frequency.
Arthur Collins' Preference
Arthur Collins always preferred top inductive or magnetic field coupling between parallel resonators because of the greater harmonic attenuation, which increased with loaded Q. Despite only two years of junior college, Arthur Collins was an intuitive genius, who invariably could find a superior qualitative circuit without a quantitative method to justify it. Fortunately, the famous Seymour Cohn published "Direct-Coupled Resonator Filters" in the 1957IRE Proceedings, and it later dawned on me that resonator loaded Q was simply related to the microwave designer's normalized lowpass (LP) prototype network element values, resonators, and the connecting inverters. The impedance inverters Cohn employed between resonators were just the capacitive or magnetic couplings that always concerned Arthur Collins
Working papers at Collins
Collins engineers could write "working papers" that were preserved in a cataloged system. I wrote WP-8014 in 1965, which related the immortal loaded Q to microwave filter design concepts. Fairly often in engineering meetings with Mr. Collins and others, I would remark that the loaded Q product they were considering would produce a harmonic attenuation of so many decibels and that the passband would be flat if the resonator loaded Qs were distributed in a certain pattern. Finally, Arthur Collins called me one day to say he had discovered my working paper and the source of my quantitative remarks. The title of this tale relates to Mr. Collins' doubt that Cohn's impedance inverters really existed. Inverters simplified filter design and tuning and produced voltages across resonators that are exactly 90 degrees out of phase at band center frequency. So Arthur Collins had my associate Dick Fenwick construct an elaborate three-resonator bandpass filter for the low HF band and tune it by the open-short-circuit method so familiar to microwave engineers. A signal at band center frequency was applied to this bench-top rig with vector voltmeters measuring the complex voltage across each resonator: All phases varied by 90 degrees and had the expected magnitudes. "Well, I'll be darned!" Mr. Collins said as he smiled at me and walked away. Seeing was believing!