[GreenKeys] Polar Relay

[Jim Haynes]

On Mon, 6 Feb 2023, Tom Hunter wrote:

> Eugene,
> the polar relay. I also believe a properly adjusted polar relay is bistable
> (which would not matter in this situation.)
> 
Yes.  One of my first projects at Teletype was to work on, or at least
think about, an electronic replacement for the polar relay.  That led to
the low-voltage selector magnets of Model 32/33/35, but those were the
solution to the wrong problem.  A real polar relay is terribly difficult
to model electronically, especially in those days before we had opto-
isolators.

In a real polar relay the windings are electrically isolated from the
contacts and from each other.

A real polar relay is "side stable" meaning with no current in the
windings the contacts will remain in whichever position they were
before current was removed.

I think there is or was a paragraph in MIL-STD-188 which stated these
were requirements for a replacement for a polar relay.

So actually I worked on a different problem, and came up with what I
considered an excellent concept but was unable to sell it in the
company.  The concept was that every selector magnet should have a
selector magnet driver with an input something like RS-232.  RS-232
had not been created yet, but what I meant was with voltages, not current
loops, inside the set.  So every keyboard or tape transmitter should
similarly have a circuit for generating polar voltage signals.  What
was different from RS-232 was that any signal source should be able to
drive multiple loads.  What's beautiful about this is that it gets rid
of the current loops inside the set, replacing them with voltage sources
and sinks.  If you want a reperforator to copy what you are receiving
you don't have to break open the printer current loop and insert the
reperf selector; you just connect the reperf selector input to the
"hub" that is driving the printer.  And there are things you can do
with diode gates to simplify other switching internal to the set.
And you are all set up to add transistorized accessories such as a
regenerative repeater to a set without having to go through current
loop to voltage conversion and then back to current loop.  Then if the
input to the set is a neutral or polar current loop go ahead and use
the polar relay; it won't wear out with low voltages and currents on
the contacts.

We did have unattractive ways of isolating circuits in the days before
optoisolators.  The AN/FGC-5 time division multiplex, using tubes,
used a gas diode enclosed in a solenoid as a relay.  With no current
in the solenoid the gas tube would fire and conduct at the voltages
applied to it from the loop.  With current in the solenoid the gas
tube was unable to fire and made an open circuit.