[R-390] Synthesizers, was 390() vs 390(A)

[John KA1XC]

Well, a communications frequency synthesizer must not only generate the
desired frequency but must do it cleanly, meaning without adding noise
sidebands AND reference sidebands. In a typical synthesizer the PLL
bandwidth is usually made sufficiently high to knock down the VCO phase
noise, but this is exactly the opposite of what is needed to attenuate the
reference from frequency modulating the VCO; this would require using a loop
bandwidth MUCH smaller than the reference frequency. It's the classic
engineering example of  conflicting requirements which in turn leads to
increasing design complexity. Good modern synthesizers must use two and even
three loops to achieve small tuning increments without generating all kinds
of garbage off the side.

Sounds like the guys who designed the crystal synthesizer were able to make
good reference frequencies, but they were probably full of mixing products
and thus unusable directly. But by locking a high Q (thus low noise) LC VFO
(not VCO) to them with an electro-mechanical feedback loop, it completely
isolated the VFO from being modulated by the reference frequency. I bet the
whole thing probably had an effective loop bandwidth of just a few hertz!

Electro-mechanical PLL's and FLL's are classic 1950's and 60's technology,
just like the R-39x mechanical RF deck. Other related gear that made use of
electro-mechanical tuning loops were the CV-116 FSK converter and CV-157 SSB
converter.

John


----- Original Message ----- 
From: "Dave Maples" <dsmaples at comcast.net>
To: "Michael Murphy" <mjmurphy45 at comcast.net>; <paul at pdq.com>; "Barry
Hauser" <barry at hausernet.com>
Cc: <r-390 at mailman.qth.net>
Sent: Tuesday, December 28, 2004 9:30 PM
Subject: RE: [R-390] 390() vs 390(A)


> All: I once had an ARC-38A and associated 180L-3 automatic tuner.
Autotune
> mechanisms in both of those were typical Collins.  The most bizarre part
of
> the operation was the frequency control for the ARC-38A.  They used an
> electromechanical phase-locked loop.  It was strange: the VFO frequency
was
> compared to a crystal-synthesized reference switched by the autotune
> mechanism, and the VFO shaft was driven with a motor to match the
reference.
> The control loop was a proportional / integral loop!  Very strange...they
> had crystals to synthesize the reference in 500 Hz steps...why not just
use
> that to control frequency???
>
> Dave Maples
>