[RTTY] RTTY newcomer looking for some information

[Kok Chen]

On Sep 22, 2008, at 9/22    9:21 AM, John Becker wrote:

> I have a couple of books on RTTY dating back to the 1960's. They talk
> about using TU's that can decode with only the mark or only the space
> signal.

Unless you make modifications to it, the KAM that you are using is  
not capable of Mark-Only (MO) or Space-Only (SO) copy.

> This seems pretty straightforward, since the mark and space
> together provide 100% redundancy.

The principal reason that FSK (mark and space) is better than Mark- 
only or Space-only (also called On-Off Keying or OOK) is that it will  
decode through selective fading (pretty much behaving as an automatic  
MO or SO decoder when called for) if the "slicer" is properly designed.

Those who have a crossed ellipse display will be able to tell you  
that it is quite of often that we see one of the two (Mark or Space)  
tones vanish into the noise while the other one survives, or does not  
fade out at the same time (frequency diversity). When selective  
fading takes one tone away, you can still copy with the other tone  
(albeit with a worse error rate than if you had both tones).

So why include MO and SO in a modem?  MO and SO can be used to copy  
RTTY when there is adjacent QRM that is interfering with one of the  
tones.  So you just copy him using only his "other" tone while  
ignoring the tone that has QRM.  You often encounter this during  
contests.  Better modems than the KAM, such as the HAL ST-8000 has  
this capability.  So do some software modems.

> I haven't figured out how to do that
> with what I'm running, and on weaker signals I'm finding a tuning  
> error
> of only 10Hz can make the difference between copy and no copy.

My two cents...

IMHO, you really can only tune strong signals using the bar  
indicators that you find in the old TNC (KAM, PK-232) and rigs today  
with built in decoders and tuning tools such as the Elecraft K3.  You  
could spend more time to eventually get a copy, but tuning bars don't  
give you accuracy and speed at the same time.  One of the better bar  
indicators I have used is on the PK-900.

For good weak signal copy, I recommend a "crossed ellipse" (or as  
some old hand RTTY people call "crossed bananas") display for  
critical tuning (if you cannot click-tune with a mouse, read further  
down).

Tuning indicators are probably a personal and subjective thing.   
Other RTTY ops may have their own preferences.  You really need to  
try them out to find which you prefer.  For example, some people like  
the DXP-38 crossed LED, but I found it completely useless when  
searching for the QSX in a split pileup or for rapid S&P during an  
RTTY contest.  For myself, a crossed ellipse is accurate while  
fast.   There is a slight ambiquity (am I too low or am I too high)  
when you are far off-tuned but that is where your ears come in HI.   
After a while, your eyes and ears cooperate automatically to control  
the fingers that are on the VFO knob without you even thinking about  
it (OK, so I spend too many hours on RTTY :-).

The KAM has provisions for you to bring out the filtered Mark and  
Space signals to use with an X-Y scope for use as a crossed ellipse  
tuning indicator.  The signal path is shown in the KAM Plus  
schematics, but not wired.  You will have to wire it yourself.  The  
PC board solder pads are already there and you can use two unused  
pins of the DB-25 connector to lead the scope signals out.  Many  
folks (including myself) had done that in the past.

If you use a Mac, I have an old twin-crossed ellipse program that I  
would be glad to send you.  (The "twin" is so you can send the stereo  
output from a two receiver rig such as your MkV to allow you to tune  
the A and B VFO to two separate signals.)  Feed the rig's stereo  
signal into the Line input of your Mac, and you are off to the races.

However, if you are going to go as far as to interface the audio of  
the rig for the computer to use, I would highly recommend that you  
use a software modem instead of the KAM.  Good software modems today  
can wipe the floor with the gold standard of hardware TU (the HAL  
ST-8000) from yesteryear.

At the least, you should be able to set the software up as a tuning  
display, even if you don't use it for decoding and transmitting.  Not  
all software programs come with good cross ellipse displays, so you  
need to find one that will make you happy.

Some software will allow you to tune an RTTY signal in a spectrogram  
(hams call them "waterfalls") accurately enough with a single mouse  
click (that is how cocoaModem does it, together with a crossed  
ellipse display that you use for fine tuning with a mouse wheel or  
arrow keys -- but when you are used to where to click, you almost  
never need to fine tune).

On top of providing better demodulators for less expense, the  
software route will also allow you to operate other more popular ham  
digital modes.  They are really win-win when compared to a TNC.  When  
there are free software modems about, TNCs are IMHO a thoroughly  
false economy.

BTW, John, if you want to keep using the KAM, you might want to check  
that you are selecting 170 Hz shift and not the TNC's default of 200  
Hz shift (the KAM protocol allows you to choose the mark and Space  
frequencies).  I have had no problem copying most RTTY signals  
through a cascade of two Inrad 250 Hz IF filters.  It is tight and  
tuning is critical, but not impossible -- but there is no need to use  
such a narrow filter when there is no QRM --  350 Hz to 400 Hz  
filters are more than sufficiently narrow for normal operation.

Absent of QRM, a very sharp 250 Hz is poorer than a 350 Hz filter on  
paper --  most of the time, you are cutting off signal energy from  
the FSK keying sidebands when they are still above the noise, i.e.,  
you are actually losing SNR by using a sharp 250 Hz filter,  
especially if the demodulator is using a matched filter that is  
matched to 45.45 baud.  When there is QRM, there are cases where a  
narrower filter will present better SNR to the demodulator.  The QRM  
does not need to come from the principal tones of the interfering  
station, but it can also be the result from FSK keyclicks from an  
RTTY signal hundreds of Hz away.

73
Chen, W7AY

P.S.  Crossed ellipses filters are not hard to design.  For a  
perfectly orthogonal (the mark and space ellipses are perfectly at  
right angle to one another and spaced 170 Hz) display, you just have  
to make the mark filter put out a 90 degree phase shift when  
responding to a space signal.  And vice versa.  (This is why a  
crossed ellipse display is skewed when you tune a 200 Hz signal in --  
the 90 degree phase shift is maintained at 170 Hz but not at 200 Hz  
-- think Lissajous curves).  In software, they are even easier to  
design.  I didn't even bother with designing perfectly orthogonal  
filters in cocoaModem, but simple use graphical transforms to  
orthogonalize the ellipses.