Riccardo wrote:

> f = 8270.00332 Hz
> Start time: 21:30:00 UTC
> Symbol period: 30 s
> Characters: 3
> CRC bits: 24
> Coding 16K21A

Copied an obviously correct message in Todmorden, UK.

Eb/N0 = +1.8 dB, symbol error rate 40.5%,
S/N -30.6 dB in 1Hz.  Rank 150 in the list decoder
output.

Constant reference phase (meaning the frequency
is correct and the path was stable).

Symbol timing is correct: Eb/N0 peaks at timing
offset zero.

Noise was a little higher last night compared to
the previous night, so S/N is a bit less than
expected.

No decodes via the on-line natural radio streams,
but some Eb/N0 measurements by cross-correlation:

Bielefeld (DL4YHF) 845.3 km:

   Eb/N0 = -9.5 dB, -42dB in 1Hz;
   There was a one sample negative timing break in the
   Bielefeld recording, about half way through.  These
   cannot be corrected for, so the phase may have stepped
   by about 90 deg at the timing break and lost a few dB.

Cumiana (IK1QFK) 370.5 km:

   Eb/N0 = -5.2 dB, -37.6 dB in 1Hz;
   I expected a better result than that from Cumiana.
   Four timing breaks, but forward steps - easy to pad
   over;

Warsaw (SQ5BPF) 1035.0 km:

   Eb/N0 = -15.4 dB dB, -47.85 dB in 1Hz;
   This was a long shot but worth a look;

--
Paul Nicholson
--