The errors in the 0 0 0 0 error map look quite equally spread.
How much weaker is the average night signal relative to the day signal?
If the difference is not much, it would be worth to use the nights as well.
I got the feedback that the message was not decodable on the other
station, unfortunately. So maybe the plain carrier which is running
since 12 UTC is more helpful. It is already visible in Italy again:
Now, what was the message? :-)
My signal looks terrible on the tree grabber. I still don't know the
reason. Is it the TX or the RX? The TX PC is very old. I thought it
could be a temporary overloaded USB port (WLAN + USB soundcard), so i
tried a LAN connection instead, with no change. Another thing could be
partial discharges in the coil arrangement! So i reduced the antenna
current from 21 UTC to 21:30 UTC from 225 mA to 210 mA. So far there is
no significant change. At 21:30 i will go down to 180 mA. If this makes
no difference i will go back to 225 mA again at 22 UTC. I will run the
carrier until tomorrow evening. Then i may have a new idea what we could
do next, on 58 km.
The next 2.97 kHz experiment is a bit delayed but the coil is ready for
150 mA. They say the wx will be terrible in the next days but as soon as
there are a few good days, i will be back on 101 km...
Am 06.01.2017 12:33, schrieb Paul Nicholson:
> f = 5170.000000 Hz
> Start time: 04.Jan.2017 08:00:00 UTC
> Symbol length: 60 s
> Characters: 8
> CRC 16
> Coding 16K25A
> I will repeat the transmission starting [5th] 8 UTC...
Best decode with reference phase pattern: 30 30 0 0
Eb/N0 = +3.6 dB, BER 35.0%, list rank 0.
S/N 20.24 dB in 11.8 uHz, -63.0dB in 2.5kHz
The phase pattern is appropriate for an advance of the
signal phase during day relative to night.
Symbol error map (with 30 30 0 0 pattern):
and with constant reference (0 0 0 0 pattern)
Not much difference, 500/1408 symbol errors with constant
reference. Down to 493/1408 with 30 30 0 0 pattern.
I haven't mentioned the message text because it's possible someone
else may be trying a decode.
> that makes a fine record of 5.17 kHz successes
> between the two of you!
Indeed, 5170Hz seems to be quite a cooperative band to use.