[GreenKeys] HF radio relating to DSP TU Update

[Harold Hallikainen]

On Wed, July 6, 2022 6:46 pm, Jim Haynes wrote:
>
> HF radio is all we've got, but it just isn't very nice.  There is flat
> fading, meaning the whole bandwidth of the incoming signal fades together.
> Then there is multipath or frequency-selective fading, where multiple
> reflections from the ionosphere create creative and destructive
> interference based on slightly different path lengths.  So a particular
> frequency get nulled out, and that frequence keeps changing as the
> reflective paths keep moving around up there.  And there are multipath
> echoes which are the same thing except the path length differences are
> in the range of milliseconds, so it's like receiving several copies of
> the signal displaced in time.  The Dovetron demodulators were called
> "multipath corrector" but apparently it didn't work very well as the
> circuits were redesigned several times.  I think the most successful
> attack on the multipath problem was a system developed at MIT called
> RAKE, which put the signal through a delay line and then tried to
> take signals from several taps on the line and combine them
> constructively.  I'm told the RAKE technology is now used in things
> like cell phones because the signals get multipath reflections from
> buildings and things in the urban environment.
>
> And then in the amateur radio realm we have to contend with interfering
> signals, especially in things like contests and DX chasing.  And the
> ever-increasing urban noise floor.
>

I have not studied the Dovetron multipath corrector closely. I did look at
the one in the HAL ST-8000 (my comments at
https://w6iwi.org/rtty/TuComparison/ ).

So, we have a less than perfect channel! There are several ways around
this. These include:

1. More power to get through fades. Fades are generally caused by
reflections, and increased transmit power increases both the direct and
reflected signal, possibly causing no improvement. But the main problem
with a fade is that it fades below the noise. More power makes it so the
signal has to fade farther to get below the noise.

2. Signal redundancy. We have this with FSK, but with 170 Hz shift, both
signals tend to fade together (see
https://w6iwi.org/rtty/TuNotes/#SelectiveFading ). Some day I'll try some
850 Hz experiments. With dynamic threshold control, it might be more
reliable than 170 Hz. Other signal redundancy approaches include forward
error correction, sending the message multiple times (time interleve),
etc.

3. It would be interesting to try OFDM. It uses many narrow bandwidth
carriers and survives interference and selective fading taking out some
carriers.

4. Another approach would be something like JS8CALL, which uses something
similar to FT8 modulation (slow 8 level FSK).

In general, I've had some very solid RTTY QSOs running 170 Hz shift with
the Flesher TU-170. It is pretty simple and uses a limiter with no dynamic
threshold control.

I've added more scope images at https://w6iwi.org/rtty/DspTU/ . Next I'll
add some schematics.

I bounce between several projects, so it will probably be a few weeks
before I add more to this project.

Thanks for the comments!

Harold
https://w6iwi.org