[GreenKeys] Re: Model 28 keyboards

[jhhaynes at earthlink.net]

The ordinary, not gold-plated contacts, are definitely rated to work in
a 60 ma 120 volt loop.  Just look at the schematics, where they are shown
in the local test loop.

I sorta remember reading somewhere about gold plating over the regular
contacts and how if they were used at low level they worked, and if
used at high levels they would blast off the gold and then would work
as high level contacts.  Don't remember if this was in some Teletype
publication or in Western Union Technical Review.  There is a very
thorough article about contacts in W.U.T.R. October 1961.  

Meanwhile, an anecdote.  The Model 35 keyboard is basically the same
as Model 28, and with speed increased to 110 baud there were some problems
with distortion of the signals.  It's hard to move parts that fast and
have the timing be consistently accurate.  Designed into the Model 35
was provision for a transmitting signal regenerator.  This was a little
circuit using an SCR, and using a separate contact on a flutter cam to
time the pulses more accurately.  In all the early production units the
SCR circuit was replaced by a circuit card that simply jumpered the
signal generator contacts to key the transmit line directly.

The Bell System decreed that as of a certain date all the Model 35
equipment they bought was to have the regenerator installed.  So
Teletype began building the machines that way, and (surprise) they
started failing tests.  The regenerator was converting timing errors
into bit errors; instead of a jittery signal you got one that had
completely wrong bits.  So a hold was placed on shipping the machines,
but the factory continued building them, and they were being stacked
up in the aisles and wherever space could be found while R&D worked
on the problem.  Presumably the regenerator was not adequately tested
when it was first designed, and the problem went unnoticed so long as
none of the machines being shipped had the regenerator cards installed.

The problem turned out to be contact failure in the flutter contact
because of oily films forming on the contacts.  I don't remember
whether the solution was to gold plate the contacts or to increase
the voltage and current to where the oil film would be burned off.



-- 

jhhaynes at earthlink dot net