Riccardo wrote:

> f = 8270.0033187866 Hz
> Start time: 20:00:00 UTC
> Symbol period: 60 s
> Characters: 2 (repea­ted)
> CRC bits: 10
> Coding 16K21A

After four repeats the results are now:

Todmorden 1413km
----------------

 17/18th Eb/N0 +4.0 dB, -31.5 dB in 1Hz, phase -117.5 deg *decoded*;
 18/19th Eb/N0 +6.0 dB, -29.5 dB in 1Hz, phase   +8.7 deg *decoded*;
 19/20th Eb/N0 +7.6 dB, -27.9 dB in 1Hz, phase  -35.7 deg *decoded*;
 20/21st Eb/N0 +4.6 dB, -30.9 dB in 1Hz, phase -164.7 deg *decoded*;

Bielefeld 845km
---------------

 17/18th Eb/N0 -3.1 dB, -38.6 dB in 1Hz, phase  -27.7 deg;
 18/19th Eb/N0 +0.2 dB, -35.3 dB in 1Hz, phase +111.2 deg *decoded*;
 19/20th Eb/N0 +1.4 dB, -34.1 dB in 1Hz, phase  +56.5 deg *decoded*;
 20/21st Eb/N0 +2.4 dB, -33.0 dB in 1Hz, phase  -80.9 deg *decoded*;

Cumiana 371km
-------------

 17/18th Eb/N0 +1.1 dB, -34.4 dB in 1Hz, phase +139.2 deg *decoded*;
 18/19th Eb/N0 +0.9 dB, -34.6 dB in 1Hz, phase  -96.4 deg *decoded*;
 19/20th Eb/N0 -8.3 dB, -43.8 dB in 1Hz (rx timing fault)
 20/21st Eb/N0 -5.4 dB, -40.9 dB in 1Hz (rx timing fault)

Warsaw 1035km
-------------

 17/18th Eb/N0  -5.8 dB, -41.3 dB in 1Hz, phase -28.3 deg;
 18/19th Eb/N0  -3.1 dB, -38.6 dB in 1Hz, phase +88.7 deg;
 19/20th Eb/N0  -1.8 dB, -37.3 dB in 1Hz, phase +97.1 deg;
 20/21st Eb/N0 -10.8 dB, -46.3 dB in 1Hz

Getting reliable decodes at Bielefeld.   Cumiana not timing the
signal very well, many thousand timing breaks each day
since something went wrong at 2017-12-19 06:42.

> A new transmission scheduled this night, trying a true
> "stacked decode" in Warsaw:

For unknown reason a poor S/N at Warsaw 20/21st.

Stacking 18/19th + 19/20th produces a false decode at +1.6 dB,
beating the correct decode at +0.5 dB.  Pure bad luck there.

Phase shifts from 18/19th to 19/20th:

 Todmorden:  -44.4
 Bielefeld:  -54.7
 Warsaw:      +8.4

Phase shifts from 19/20th to 20/21st:

 Todmorden: -129.0
 Bielefeld: -137.3

Still something is changing the phase from one night to the next.

--
Paul Nicholson
--