Message-ID: <000b01c4d3fc$f0716880$6507a8c0@Main>
From: "Alan Melia" <Alan.Melia@btinternet.com>
To: "LF-Group" <rsgb_lf_group@blacksheep.org>
Date: Fri, 26 Nov 2004 21:14:30 -0000
MIME-Version: 1.0
Subject: LF: MSK etc and stability
Content-Type: text/plain; charset=iso-8859-1; format=flowed
Content-Transfer-Encoding: 8bit
Sender: owner-rsgb_lf_group@blacksheep.org
Precedence: bulk
Reply-To: rsgb_lf_group@blacksheep.org

Hi Bill and Alberto, I feel I should wear a silly smile and pair of glasses
(is it still true you can avoid getting thumped if you are wearing glasses
?) before saying this.....

About 16 years ago just as DSP was begining to make itself felt in the
professional modem field I was involved in an abortive project for a
multitone HF modem. It was to be centred on an ASIC design, and was
squelched by the DSP developments. I saw a bit of DSP coding done by a
research student that produced software decoder, but it also had a vital bit
of coding for commercial HF use... it had "AFC". Now I dont know whether
building in some sort of two speed AFC is possible, but I am sure it would
be an advantage for audible (or visible) signals.

The problems occurs for the operator at LF where we want to set up a
receiving facility "blind". I know the difficulty of this as I tried to
receive Larry for about 4 months on AFRICAM and an S-D unit, without
success. I am sure this was because although my receiver was stable enough I
was not able at that time to be totally sure of the frequency the computer
thought it was tuned to. (I did not have the fine resolution synthesiser I
have now) When Larry finally gave up with BPSK and went QRSS he was received
almost the first night.....but I was "listening" 0.8 Hz away and rejected
the broken line as a weak Loran line, knowing Larry had GPS locked
standards.....he was off-frequency as he later admitted.   The other way of
doing this might be to accept another local signal input...maybe the other
of the stereo channels.... which would be an accurately defined frequency,
rather like the ARGO calibration routine. It would then be subject to all
the soundcard errors, but not the receiver mistuning. The only way I could
think of around this would be to couple an accurate weak signal into the
aerial and have the decoder lock to it for calibration.

Because of the difficulties with fading one would need a "search" facility
for "optimising the tuning" of the signal over a known small range and then
when copy was received, a "hold" facility for locking to it even when it was
faded into the noise, and was not being decoded. The software "overhead" for
this might be too high.....I have no way of knowing what degree of mistuning
could readily be corrected. However if we work to 0.1Hz as someone suggested
would it be reasonable to have at least a similar degree of mistuning
tolerance??

I dont yet have a GPS locked standard, but as Crown-Castle and the NPL dont
seem to be interested in the reported instabilities on 198kHz Droitwich, I
will soon have to get moving on that project.

Cheers de Alan G3NYK
alan.melia(at)btinternet.com