[Milsurplus] Mystery FSK/DFSK demodulator CV-2310/G

[Jim Haynes]

On Tue, 27 Nov 2012, David I. Emery wrote:

> 	I believe the US/NATO world developed digital logic based time
> division multiplex systems (lots of 4 channel ones) during about the
> same era... which mostly put faster bit streams on standard 2 channel
> FSK signals.
>
Yes.  Teletype Corp. had an electromechanical time-division multiplex
in its catalog back in the 1930s.  I imagine this predated the AT&T
purchase of the company, and stayed in the catalog because of sales to
certain private wire companies such as railroads.  In the late 1930s
Teletype hired an out-of-work radio engineer to work on an electronic
version of the time-division multiplex, but that work was delayed by
WW-II.

During WW-II the Navy made some use of Western Union time-division
multiplex, which was largely electromechanical with a few electronic
aids.  This may have been because a W.U. executive was commissioned
into the Navy to help get their communications in shape during the war.

After the war Teletype resumed work on the electronic multiplex and
came up with a working model which excited the Navy, so it went into
production as the AN/FGC-5, two seven-foot racks of equipment using
vacuum tubes and thyratrons.  Use of this spread to other government
services and agencies such as the C.A.A. (ancestor of F.A.A.)  Actual
manufacture was done by a subcontractor Power Equipment Co. which was
a big maker of power supplies and was later acquired by North Electric.

All this time Western Union was operating time-division multiplex on
its own wire circuits; but AT&T was not interested in that form of
transmission so Teletype's work was strictly at the request of the
government.  And the equipment being built to military standards was
probably unattractive in price to the private sector.

I remember hearing lots of FSK multiplex signals on the air in the
late 1950s.  Didn't know what they were until I had a summer job at
Teletype in 1958 and we had a radio receiver and FSK converter in
our lab and were copying some of that stuff on the multiplex equipment
in the lab.

At that time the company was developing a transistorized version of
the four-channel multiplex, which became the AN/UGC-1, occupying a
cabinet about three feet high.  This was interoperable with the FGC-5.
It may have been operable as a two-channel mux if radio conditions
did not support the shorter pulses of the four-channel operation; but
I just don't remember that and don't have the manual at hand right now.
This went into production about 1959.  As an aside, it originally used
some point-contact transistors made by Western Electric, which had
some unique useful properties in spite of their general crummy-ness.

Teletype also developed for the Navy the AN/UGC-3 sixteen-channel
multiplex.  This was operable also as 8-channel or 4-channel and in
4-channel could interoperate with the earlier products.  At the time
I assumed it was for use on radio also, in spite of the very narrow
pulses in 16-channel operation.  Much later I learned that the real
reason for it was that they had a crypto machine that would encrypt the
16-channel signal.  Then another Teletype machine the AN/UGA-1 would
demultiplex the signal into sixteen single synchronous signals which
were transmitted by frequency division multiplexing.  About the same
time the military was moving from time-division to frequency-division
multiplexing generally.  And Western Union phased out its time division
multiplex in favor of frequency division sometime in the 1950s.

W.U. made a frequency-division mux terminal for the military, which
was quite a bit larger than the time-division systems at the time.

Teletype was not the only maker of time-division multiplex for the
military; I've read of others.  RCA Communications had time-division
multiplexing back in the 1930s, of a different sort that was not
interoperable with the AN/FGC-5.  And we recently discussed the
cheapie time-division multiplex obtained by using a two-headed
transmitter distributor sending regular start-stop signals, with
the data from one tape in the early part of the pulse time and from
the other tape in the late part.  Then demultiplexing was just a
matter of adjusting the range finders in the receiving machines to
sample early or late.  There's mention in the manual for the TT-63
electronic regenerative repeater of using it to receive such signals
and regenerate them as ordinary start-stop signals to drive two
printers or other circuits.