[GreenKeys] Teletype Synchronizer

[jhaynes@alumni.uark.edu]

I can perhaps say a few things about this.

Teletype built some time-division multiplex equipment for the military,
first using vacuum tubes (AN/FGC-5) and then using transistors (AN/UGC-1,
AN/UGC-3).  The multiplex distributor ran about 2% faster than the
start-stop speed, which meant that the characters delivered by the mux
had shorter than usual stop pulses; and every 50 characters or so there
was a pause while the mux sent a non-character blank.  For some reason
this made the crypto equipment unhappy, so there was developed a thing
called a signal normalizer that received start-stop signals from the mux
receiver and stretched out the stop pulses to standard length. This
equipment is documented in Western Union Technical Review, January 1957;
and I'm sure there is also an article with the title "Printing Telegraph
Signal Normalizer" and probably an author name of Dingley and probably in
IRE Transactions on Communication Systems from about that period.  It was
called the "dingbat" in honor of Dingley.  See also U.S. Patent 2,721,230.

A good many years ago I bought some surplus stuff, which was a
regenerative repeater and T.D. stepping facility.  This had crystal
oscillators and frequency dividers, in duplicate, for each of the standard
operating speeds.  The T.D. steppers further divided the frequency to the
character rate and had power transistor drivers for the operating magnets
of transmitter-distributors.  It was also possible to operate keyboards
this way.  See a late Model 28 parts book for parts to add "synchronous
pulsed operation" to the keyboard.  And I'm somewhat aware of military
patching systems where they carried a wire for the transmitted signal and
a second wire for stepping the T.D. or keyboard.

What I haven't understood is why the crypto equipment required this kind
of treatment; why it couldn't just do buffering and then send slightly
fast like the multiplex did.