Hi Paul,
Wow, what a great result after such a lomng time :-) I'm very happy.
40 characters on 6470 Hz, some kind of a new dream :-)
Hopefully some newcomers read the message and think about it :-)
Thanks for your efforts, as usual, and the most informative set of data
each time.
Probably the decode could have happen much earlier if the phase
inversion would be faster. This will be done in the next experiment!
My tree is back, the accus are fully charged. There have been 4 sunny
days now. The Raspi didn't start up automatically when the accus were
reasonable charged (a hardware problem that i already know, have to fix
it soon), so i had to start it manually today again. It was a very nice
day in the clear cold frozen forest!
Now i have my local feedback channel again and can do a local EbNaut
experiment with 1 second symbols on 6470 Hz. This should work without
problems. Markus' new SpecLab config is running since a few days but i
wanted to continue with the old psk technique until that 40 char
experiment was finished. Now i can change it...
Tomorrow will be a sunny day again and i have some time. The plan is to
run 100 mA on 2970 Hz for the first time, including a carrier
transmission taking a few hours. So be ppepared, if you like. The ERP is
7 dB higher than in my 32 km QRB experiment.
73, Stefan
Am 03.12.2016 17:32, schrieb Paul Nicholson:
Stefan wrote:
> f = 6470.00000 Hz
> Start time: 23.11.2016 08:00:00 UTC (Daily repetition)
> Symbol length: 12 s
> Characters: 40
> CRC 16
> Coding 8K19A
A decode after 11 repeats:
Copied 'HI FOLKS, JOIN THE VLF RADIO EXPERIMENTS'
Eb/N0 = -0.3 dB using normalisation by the RMS amplitude.
Eb/N0 = +0.3 dB using normalisation by the power as suggested
by Markus.
Constant reference phase. S/N 23.84 dB in 38.0 uHz,
-54.3dB in audio bandwidth.
...
A slightly better decode, Eb/N0 = +0.7 dB if I use a 2 second
offset, ie the decoder starts work 2 seconds late. Maybe
a consequence of the leisurely phase switching.
--
Paul Nicholson
--
|
