[RTTY] fluttered signals

[Kok Chen]

On Apr 18, 2004, at 1:34 PM, Jacques RAMBAUD wrote:
> -what causes the signal to be so difficult to copy ? multipath ?....?
> -Why is MMTTY unable to deal with such signal?
> -Is there any other way , or RTTY decoder that are better in this 
> respect , how old TU with discrete filters (88mh) works ?
> -Are the KAM or PK232 better ?
> any litterature on the subject?

If it sounds odd and fluttery and difficult to print even when the 
signal is loud, it is pretty likely to be multipath.  It can happen not 
just from transpolar circuits but trans-equatorial as well.  I get a 
lot of flutter across the pole from Europe, but I have also noted some 
pretty bad flutter from across the equator with loud S9 signals in 
Portland from FO5, for example.  One time when JA1ELY was at PY0FF, his 
PSK31 signals were fluttered beyond copy (but his RTTY signal was still 
barely printable).

My opinion is that simple modems such as the KAM are not going to 
behave better than the software modem you used.

What makes it so hard to decode?  The HAL ST-8000 Technical Manual 
says: "Multi-path propagation of an RTTY signal may cause differential 
fading or severe pulse distortion of the signal."

I think the "differential fading" is a specific instance of the 
selective fading that we all have learned to hate.

The "pulse distortion" is understandable from observing that any square 
pulse signal is made up of different Fourier components.  If all 
components do not arrive at the same time instance, the result is no 
longer be square. When it is severe enough, this can cause the Mark 
detector output and the Space detector output to actually overlap 
(usually, when the Mark tank detects a signal, the Space tank circuit 
will not see anything since you do not transmit both Mark and Space 
tones simultaneously -- which you can in some other modes, but not in 
Binary FSK).

The ST-8000 has a very simple circuit (just a handful of gates) which 
takes the detected mark signal and the detected space signal and forms 
an output signal from the combination to estimate a signal whose mark 
and space do not overlap; pretty much an exclusive-OR type gate 
configuration.  (The ST-8000 has detector settings for normal, mark 
only, space only, and multipath.)

RITTY has selectable matched filters that are designed for different 
propagation conditions, but I don't know any more than that since I 
don't run that software.

I have been planning for some time to investigate DSP counter-flutter 
implementations, but have never gotten around to it.  I have even 
bought a used ST-8000 for the time I have more free time to investigate 
it and would need a reference modem to compare against.  For now 
however, except for some recorded sound files, this is as much as I 
know.

73
Chen, W7AY