[GreenKeys] Mechanical and electronic RTTY

[Jim Haynes]

I love the old mechanical Teletypes - it's a joy to watch them in action
and to contemplate the beauty of the designs.  And with good signals
such as you get from ITTY they perform perfectly.

However

for on-the-air amateur RTTY they are not really competitive, especially
in contesting.

The greatest problem with mechanical TTYs is the mechanical selector. The 
selector cam shaft has to rotate a complete rotation once its clutch is 
unlatched.  If you get a false start pulse, however brief, or if the stop 
pulse is not there at the right time, the selector cam makes several 
rotations out of sync with the incoming signal and you get several errored 
characters.  The character error rate as related to signal strength or SNR 
is something like 15dB worse than it would be in a synchronous situation.

A second weakness is present if you don't have automatic carriage return
and line feed parts installed.  An errored carriage return turns a whole
line of print into a black square at the end of the line, and an errored
line feed causes a line to be printed on top of the previous line.

I have spent a lot of time over many years in trying to get the most out
of FSK RTTY.  There is the radio receiver, the demodulator (which we
used to call a terminal unit) and the printer or terminal.  In earlier
times we just fed the receiver audio into the demodulator.  You'd really
like the receiver to have bandwidth narrow enough to just contain the
FSK signal, so that signals outside that range don't affect the AVC
and depress the desired signal.

The demodulators have ranged from that single-tube single-tone 5763
design, through FM limiter-discriminator designs, and then to non-
limiter two-tone designs.  The simple ones are worthless unless signals
are very good, such as you get at VHF.  Practically all amateur-designed
limiter-discriminator designs are faulty.  The faults are inadequate
bandwidth restriction ahead of the limiter, discriminators peaked at the
mark and space frequencies rather than being linear far beyond them,
and inadequate low pass filtering after the discriminator.  The two-tone
demodulators became the accepted standard around about 1960.  Amateurs
have it harder than commercial/military users in that in contests there
are almost always interfering signals right on top of the desired signals.

The problems with the mechanical printers I have already mentioned.  One
partial solution is to employ an electronic regenerative repeater ahead
of the printer.  This helps overcome the mechanical selector problem
because the electronic selector can stop on a dime, so to speak, and
overcome sometimes the false starts and missing stops.  And it can
deliver perfectly timed signals to the printer, so that even an
out-of-adjustment or worn printer gets good enough signals to work well.
By reducing the character error rate the regenerator also cuts down on
the errored carriage return and line feed problems.

An alternatave to the regenerative repeater is a PC with software to
turn it into a Baudot terminal.  The UART chip on the COM port can
do just about what a regenerator does.

A lot has been learned about two-tone FSK reception over the years.
Probably the best source for up-to-date knowledge is the W7AY website of 
Kok Chen.  The other big development since early RTTY days is digital
signal processing, and especially DSP based on the IBM PC.  This makes
it fairly easy for people who know how to implement and experiment with
FSK demodulation to try all the latest ideas without having to build
hardware and tear it up again.

At one time I imagined that synchronous operation would lead to better
demodulation.  In due time I came up with the idea of a "compatible
synchronous" transmission by transmitting 7.00 unit code at 45.45
baud.  Since this included start and stop pulses it could be copied
by mechanical printers, but the idea was to implement a demodulator
that would take advantage of the synchronous nature of the signal,
which would insert a fill character whenever there was no message
character to transmit.  I built a transmitter circuit to produce this
kind of signal, but never got around to working with the receiver.
I did learn that Stromberg-Carlson had worked on a similar-sounding
scheme for the Army, but it took me a long time to get a copy of their
report.

In the meantime K6STI developed his RITTY program, which ran on a
PC with a Pentium CPU and which used what he called a "digital flywheel"
to synchronize the detector to a suitably repetitive signal regardless
of whether it was 7.00 or 7.50 or anything else.  By then electronic
keyboards were routinely generating that kind of synchronous signal, 
calling it "diddle".  I was quick to acquire a PC and a copy of his
software - it was what I had always dreamed of doing.  And it worked
very well, although not as well as I expected.

I think I realized why when I got the Stromberg Carlson report.  They
had also used a slightly different way of generating a compatible
synchronous RTTY signal.  They found that a true synchronous detector
didn't help much because of the phase perturbations inherent in HF
propagation via the ionosphere.  K6STI made many improvements to RITTY
over the years as a result of his operation on the air in contests.
But ultimately he discontinued distributing the program because of
software piracy (RITTY was not free) and because it required a PC
running DOS and a specific sound card.  I believe the requirement
for DOS is that the software engine must be non-interruptible.
So if you could dedicate a PC running DOS to the application you had
a fine RTTY machine, but otherwise you had to make use of software
developed later by others that would run under Windows or Linux.

One thing K6STI did for me was to provide a cleaned-up Baudot output
from the received signal, so I could drive a Real Teletype printer.
I wish the other RTTY software makers would do the same.

Currently the top-rated RTTY demodulator software seems to be 2Tone
by G3YYD.  This is free software but it is not simple to run, as it
has to run under N1MM logger.

Disclaimer: there's a lot going on in this world that I don't know about.

Jim W6JVE
 	---

 	"Ya can argue all ya wanna, but it's dif'rent than it was."
 	"No it ain't! No it ain't!  But ya gotta know the territory."
 		Meredith Willson, The Music Man