[BC] Phasitron...

[RichardBJohnson at comcast.net]

From: "PeterH" <peterh5322 at rattlebrain.com>

>My own institution had an equally old exciter, a Serrasoid.

A major problem with Serrasoid and other phase modulation schemes is that generating a linear ramp does not guarantee an ultimate linear frequency deviation! A Serrasoid exciter, freshly tuned and tweaked will work okay because the phase-shifting is "jammed against" nonlinear tubes which help compensate. Once the tubes start to age, the distortion sets in. This paper http://www.urel.feec.vutbr.cz/ra2008/archive/ra2006/abstracts/018.pdf shows a linearity of about five percent.

Western Electric came up with "direct on-carrier frequency modulation," well before the term was applied to solid-state exciters. They used reactance tube modulators and a "Frequency Watchman" phase locked loop! It is surprising the technology developed in the '30s! Since the resonant frequency of an L-C circuit varies inversely as the square root of the L C product, varying the capacity using a reactance tube, will generate considerable FM distortion. Much later, solid-state exciters varied only the capacity with a varicap, but used only a small, almost linear, segment of a larger nonlinear range.

Western Electric's solution was to use two reactance tubes, one emulating a variable capacitance, and the other emulating a variable inductance! This required a push/pull oscillator circuit so that the required excitation phases were available. Since both the L and the C varied with modulating voltage, the square-root curve provided a linear voltage to frequency conversion!

The PLL comprised of a number of "regenerative dividers" that divided down the operating frequency  and the frequency of a quartz crystal to audio-frequency "tones" that were applied to a two-pole motor. If the frequencies were different, the motor would turn and adjust a capacitor to bring them together. Since the motor could not turn fast enough to cancel any AF, the frequency control was distortionless. The motor would turn until the average phase of the two signals was the same, preventing any further rotation.

When the FM band was moved up to 88-108 MHz, Western Electric just added a doubler stage so the scheme was no longer "direct-on-carrier..."

Cheers,
Richard B. Johnson
Book: http://www.AbominableFirebug.com/