Just got back from a week's camping in Wales on Shell Island.
Chatted with a radio ham who was camped nearby, saw his antenna
poking up above the sand dunes.

My rx runs happily while I'm away - my hobby is almost fully
automated, it doesn't need me any more!

So now there are all these messages to decode... Ah yes I am
needed - but only to read the emails and run the decoder.

Stefan wrote:

> f = 8270.005000 Hz
> Start time: 22.September.2017  06:00:00 UTC
> Symbol period: 24 s
> Characters: 12
> CRC bits: 16
> Coding 16K21A

E-field is recording from the new rx but is un-calibrated so I
can't use it for decodes just now.  So the following is just
using the loops:

 2017-09-22: +0.4 dB phase 126.0 list rank 8225;
 2017-09-23: +1.6 dB phase 131.0 list rank 0;
 2017-09-24: +0.3 dB phase 134.0 list rank 242;
 2017-09-25: +2.2 dB phase 127.2 list rank 0;

All decoded an obviously correct message with uniform reference
phase, using the optimum list length for uniform phase;

Decoder memory for this one is 1.9 GByte when using optimum
list length of 325k for uniform phase search.  Not too bad.

Stacking all four gives Eb/N0 = +7.2 dB.

--
Paul Nicholson
--