[GreenKeys] How were teletypes connected across long distances?

[David I. Emery]

On Fri, Nov 24, 2023 at 05:00:28PM +0000, Sheldon Daitch via GreenKeys wrote:

> When the local radio station in my hometown in Georgia went on the air
> in 1960, it was an AP radiowire subscriber.?? M-15 RO printer fed via a
> DC loop from the Southern Bell telco central office about 2-3
> miles??away.?? ??How the AP circuit got to that telco office??? I really
> don't know, but I suspect there were both DC circuitsand tone on audio
> circuits between Atlanta and that my hometown in east central Georgia.

	My experience with a college radio station UPI circuit in the
late 60s spanned the conversion from a DC polar loop over leased copper
"telegraph" pairs (I think +- 20 ma, but definitely polar) to the tone
pack over conditioned voice grade analog data line.

	The polar relay telegraph circuit fed a traditional round gray
beer can plug in TTY polar relay which keyed a 100 volt or so DC power
supply and power resistor 60 Ma local loop with the model 15 selector
magnets in series.  I do remember the polar relays (not the later sealed
mercury kind but the adjustable kind that could be opened up with
contacts that could be burnished and cleaned) were a fairly regular
source of problems with copy quality and garbles... and needed to be
replaced and refurbished and readjusted fairly often  or the machine
would begin to show noticeable errors in copy.   

	The relay and power supply were located in the base of the table
the 15 sat on...

	The later conversion to the Lenkhurt carrier replaced the polar
relay and power supply with a box that sat under the machine on the
table..

	All the UPI redo wire machines in the city  were connected in
series on the same 20 MA polar loop and disconnecting any one of them
from the telephone line caused all of them to stop printing - something
that would happen from time to time as folks moved machines around for
cleaning or re-did telco cabling.. this of course resulted in lots of
calls from station to other nearby stations asking if the other guy had
done something to his machine or telephone wiring.... or was seeing the
same outage or not.

	I did get to actually visit the local CO and see the huge old
vacuum tube FSK modems in floor to ceiling racks used in the CO to drive
the wire service (and other long lines DC telegraph circuits) from tone
pack signals on FDM channels on the telco microwave and coax carrier
systems that brought them in from a serving CO near the state AP and UPI
bureaus.  They took up a couple of feet on the telco racks for just one
channel... and must have used a good many watts of power for each one
too.

	I understand these came into common use in telco plant for this
kind of teletype circuit  as early as the late 1930s or 1940s - with
those modems driving the machines in a local area over series DC polar
circuits well before the Lenkhurt solid state demods were developed.   I
don't think they got as many as 25 channels on one 3.2 KHz audio
circuit, the shift and tone separation was wider and it might have been
only something like 8 or 12 or 16 channels max.

> Sometime in the late 1960s, the AP in Georgia converted to a tone pack
> system.?? The DC circuit was replaced??with a voice grade audio circuit
> and a Lenkurt 25A single channel tone decoder 

	In our area the tone pack audio circuit was terminated in a
fairly large gray telco data circuit termination box that included the
ability to loop back on command so the quality of the channel could be
measured remotely from the CO on another pair.   This was considered
a big improvement in troubleshooting circuit problems as it could be
activated from a test board at the CO rather than requiring a tech 
visit the station.   Testing the DC loops was harder and IIRC involved
a loop shorting switch and current measuring jack (maybe meter ?).
The gray box also had isolation amplifiers so the audio circuit
was buffered and level adjusted (and with a couple of additional
boards could be equalized for frequency response).

> No doubt the main force in converting from a DC loop system between
> the local telco office and the??subscriber was cost driven.?? ?? It was
> cost effective to convert the entire AP distribution network from??DC
> loops to the telco, then whatever the telcos used to get the circuits to
> the telco office nearest the subscribers and then back to DC loops to the
> individual TTY machines.

	The justifications I was given were that adding additional
wires to site was just a matter of another machine and a modem tuned
to a different tone... rather than added DC telegraph circuits that
had to be installed by the telco (at some cost).   And switching around
the details of a particular wire's network was much easier to do by doing it
at the regional offices of the wire service than multiple telco COs... 

	But the major reason was that one multidrop voice grade data
network was a lot more reliable (or supposed to be) than multiple 
DC telegraph based nets with issues like other customers moving machines
for cleaning or office re-arrangement.   

	And many many multidrop voice grade data networks were emerging
in the mid to late sixties with more and more telco folks trained to
maintain and troubleshoot them - while DC telegraph stuff was a dieing 
antique even back then, so getting competent repair and installation
and maintenance was easier for more modern technology and reliable
into the future and supported with modern test equipment and tools
and techniques that made diagnosis of problems easier and quicker.

	And by having a capability to remotely measure noise and
distortion and level on an audio tonepack signal all the issues of
correctly diagnosing noise, bias distortion, wet cable leakage problems
and cross connections in cabling would largely go away.  Many of these
were very difficult to troubleshoot in networks of series connected
machines...  and of course it is also true a lot of cable problems with
audio data signals could be diagnosed by just listening to the mux
audio, not so easy to do with low speed polar TTY signals.

	Also (and used eventually) it was possible to add intermixed
higher speed circuits over these FDM muxes - essentially just by
changing the  Lenkhurt modems used for that particular signal - allowing
110 baud 150 baud or 300 baud circuits - intermixed with the
45/50/75 baud ones - all without adding any additional telco circuit
costs or delays in installation and setup.
	
	As things evolved, by the mid to late 70s many newspapers were
getting their AP and UPI news as 1200 baud FSK ASCII over the same kind
of voice grade data lines (in some cases the same one).   This quickly
became computer to computer with actual teletype or other printers
optional since news was directly ingested into  databases on editing and
composition computer systems (mostly minicomputers in that era).

	Later a lot of this telco transmission was replaced with
satellite signals delivering the data audio rather than telco
circuits... and eventually with various higher speed satellite delivery
systems in purely digital formats.

	More recently (some of) the satellite stuff has been phased out
and replaced with VPNs over the public Internet or various paywalled
web sites from which story files can be retrieved as needed.
> 
> I suspect other folks may have additional information.

	Not clear how much I add, but something...

-- 
  Dave Emery N1PRE/AE, die at dieconsulting.com  DIE Consulting, Weston, Mass 02493
"An empty zombie mind with a forlorn barely readable weatherbeaten
'For Rent' sign still vainly flapping outside on the weed encrusted pole - in 
celebration of what could have been, but wasn't and is not to be now either."