Hi Paul,
Great, we've got it! That was harder than expected but as we can see,
it's only due to the very high QRN levels during the last days. More
than 10 dB difference between 28th and 30th.
A last transmission of the message is running right now! If you like to
include it to the stacking, maybe you will get the missing 0.25 dB.
But using the known phase from the previous carrier transmissions is a
legithime method as well, i would say.
Can you provide the phase and SNR for each day please?
Actually the start time of the transmission was a bit early. I expect
the terminator was passing the path during the transmission. Thus the
error map would be interesting as well :-) These days, the change of
daytime per day is already 3min:22sec here, so maybe there is a
significant time difference between the terminator passing between the
23rd and 31st, which could reduce the stacking gain a bit. All this is
very interesting i find :-)
73, Stefan
Am 01.09.2017 01:38, schrieb Paul Nicholson:
Success!
Daily Eb/N0:
23rd -6.2 dB
24th -16.8 dB
25th -9.2 dB
26th -10.3 dB
27th -9.8 dB
28th -15.8 dB
29th -8.0 dB
30th -4.9 dB
31st -10.2 dB
Leaving out 24th and 28th, I decode 'PEACE' at Eb/N0 = -0.8 dB,
BER 41.9% constant reference phase.
I used a 'known phase' decode for this, using the phase measured
from your 2 day carrier.
It does not decode with a uniform phase search or a full phase
search. (the extra gain of the known phase decode is about 0.25 dB).
Ok, you see why my 0.13 dB rx mystery is important to solve!
We cannot afford to discard even 0.1 dB with these experiments.
--
Paul Nicholson
--
|
