[GreenKeys] Range and Bias

[Jim Haynes]

On Sun, 21 Jul 2013, Jeffrey D Angus wrote:

> Depending on the amount of distortion, meaning the mark to space or
> space to mark transitions not being where they should be, the range
> finder will allow you to move the "look" slightly before or after the
> middle of the data bit to correct for the transitional errors.

It would be interesting to know, if we had some of the Teletype oldsters
still around, whether start-stop synchronization and point selection is
most useful for dealing with distorted signals for for dealing with
motor speed errors (in an era when all the motors were speed-governed).

> The reason for the extra 1/2 width stop pulse it to give the machine
> time to settle in case it's not quite running at the correct speed.
>
Welllll, maybe.

In the original Morkrum design, with rotating electrical distributors
for send and receive, the receiving distributor shaft turned faster than
the transmitting shaft.  This continued with the all-mechanical designs.
The result is that the receiving clutch is ready to be latched up by
one-half pulse width into the stop pulse.  The machine can receive 7.00
unit code with no difficulty.  Western Union standardized on that early
on, giving a rate (at 45.45 baud) of 6.5 chars/sec, or 65 wpm.

Western Electric, on the other hand, had developed some machines where
the transmitting and receiving distributors were on the same shaft, hence
running at the same speed.  For start-stop synchronization to work the
receiving shaft has to stop between characters.  Therefore the 
transmitting shaft, running at the same speed, had to stop between
characters.  They accomplished this with a relay controlling a clutch
on the transmitting shaft.  The relay was found empirically to introduce
a delay of about 9 milliseconds, so the stop pulse was elongated by that
much, leading to the common 7.42 unit code.  The Bell System insisted
that the Morkrum equipment interoperate with the Western Electric 
machinery, so the 7.42 unit code was carried into the Morkrum product
line, even long after the old Western Electric equipment was retired.

I know they like 7.5 unit code in Europe, and maybe there are mechanical
reasons why European equipment needs a longer stop pulse.  That is
certainly not the case with the Lorenz machines, since they are copies
of the Model 14 and 15 Teletype machines which have no trouble with
receiving 7.00 unit code.

Once we got to 100 wpm the Teletype machines did need a longer stop
pulse.  Models 33 and 35 doubled the length of the stop pulse, giving
11.0 unit code.  Model 32 at 45.45 or 50 baud may not have needed
more time, but the code was made 7.5 unit for interoperation with the
European machines in Telex service.  At 50 baud this works out to
66.6 wpm.

At some point the U.S. government decided there was a need to standardize
speeds.  100 wpm, 7.42 unit code works out to 74.2 baud.  This was rounded
up to 75 baud, and then each speed in the standard is twice the next lower
speed.  That leaves the speeds of 110 baud (Teletype) and 134.5 baud (IBM)
non-conforming.