Marcus wrote:

> The equispaced lines are presumably from Loran-C Anthorn,
> their frequencies are exact multiples of 50000 / 6731 Hz.

Thanks, I was wondering about those.  The tx is 155km from
here.  Those sidebands might turn out to be useful markers.

Domenico wrote:
> Initial transmission at 22.00 [CET!] will contains 22
> characters at 1 s/symbol

Decoded easily:  Eb/N0=7.5 ber=0.19 decode time 22 seconds

Quite a strong signal, the BPSK was actually visible on the
spectrogram!

I got no decode of the following hourly transmission with
6 chars, 3 second symbols.   I did not try any time or
frequency offsets, or any impulse blanking.  Seems to be
quite a lot of impulse noise, appear not related to sferics.
I suppose local interference, switching transients.

Carrier measurement:

 http://abelian.org/tmp/lf151023c.gif

21dB S/N in 3.33 mHz.
Spectrum peaks at 137502.241338 Hz +/- 1uHz.

The 20:00 UT 22 char message doesn't decode when I use
137502.241338 for the frequency, so spectrum peak is
not a good guide.

Plugging 21dB and 300 seconds into the carrier measurement
on the signal calculator page


http://abelian.org/ebnaut/calc.php?sndb=22&snbws=&snmps=300&code=8K19&sp=1&L=1&nc=22&submit=Calculate

predicts Eb/N0 = 7.48 for the 22 char 1 second symbols, which
is close to the Eb/N0 = 7.5 of the actual decode.

Dropping the power by 10dB and changing to 3 seconds, 6 chars,


http://abelian.org/ebnaut/calc.php?sndb=12&snbws=&snmps=300&code=8K19&sp=3&L=1&nc=6&submit=Calculate

gives Eb/N0 of 4.14 dB which should easily decode.

So, I have some investigation to do!

--
Paul Nicholson
--