Hello
Stefan,
“One
could try to track HWU and its phase and then correct the phase of my
signal by the HWU phase”
Question:
After first subtracting HWU transmitter phase variations (based on
HWU-signal phase variations seen at the grabber closest to HWU)?
In
other words, subtract the HWU phase variations seen at the grabber
closest to HWU, from the HWU phase variations seen at Edgar’s grabber,
and use the result to correct the phase of the DK7FC signal seen at
Edgar’s grabber?
73,
Jim
AA5BW
VLF,
Another file from Edgar. Now we can already compare two individual days
as well as stacked days. As expected, the phase seems to be quite
dynamic making it hard to pass a message to down under.
I improved the plot a bit by reducing the measurement interval to 5
minutes. I also added a trace for a 30 minute carrier which reaches up
to 7.55 dB in 556 uHz!
3 plots are available now.
15 March: http://www.iup.uni-heidelberg.de/schaefer_vlf/VLF/VK7_15MAR.png
16 March: http://www.iup.uni-heidelberg.de/schaefer_vlf/VLF/VK7_16MAR.png
15+16 March, unweighted stack: http://www.iup.uni-heidelberg.de/schaefer_vlf/VLF/VK7_1516MAR.png
Unfortunately the stacking does not improve the SNR, it reduces it!
This can be expected from an extremely dynamic propagation as it seems
to be.
We can now continue to watch a few more individual days and the stacked
days. If the phase is not reproducable at all, the stacked SNR will
become lower each day. If it is reproducable, there will be a peak
building up.
Another idea is to correct the phase by using other signals within the
band. HWU on 18.3 kHz is not so far away, regarding location and
frequency. One could try to track HWU and its phase and then correct
the phase of my signal by the HWU phase. Meanwhile i would call this
valid :-) Edgar is already preparing for HWU measurements!
73, Stefan
Am 16.03.2018 17:52, schrieb DK7FC:
VLF,
Edgar has sent me a txt file which i converted into a wav file using
Markus' tools. The file starts 15:11 UTC and covers nearly 9 hours, so
it covers the complete time where i transmitted on 17470.1 Hz.
I tried to decode a '*' EbNaut message taking 8 hours, without sucess.
Then i moved the file on the Linux PC and tried to decode an '*' EbNaut
message again, which is equal than analysing a carrier.
BTW my antenna current was just 900 mA at times. I could run 1.5 A, so
i'm missing 4.4 dB! This is not acceptable at all. I need to build a
better PA quickly.
I quickly wrote a script to generate the SNR of a carrier with variable
time offset (in 10 minute steps; relative to the file start) and
variable length. This way we can see how dynamic the propagation
(phase!) changes.
Working together with vlfrx tools and Microsoft EXCEL!!! is a thing
that some people don't like :-)
But this is how i can do it as fast as possible at the moment.
Attached you see a mst confusing series of graphs. Selfeexplaining for
some, but, well you can for example see that a 2 hour long carrier at 6.3
dB SNR 138.9 uHz can be extracted when applying a time offset
of 4.5 hours to the file start.
You can also see that a 4 hour long carrier will only reach about 3 dB
SNR in 69.4 uHz!
BTW i ignored the 20 minute transmission break from last night, which
will cause a small error.
Now we could generate another set of graphs showing how it looks when
applying the brand new iqnorm script by Paul.
And we can see how the curves develop when stacking two or more files!
This will tell something about the reproducability of the obviously
quite dynamic phase...
So far so good. The TX is running again, same procedure than
yesterday...
73, Stefan
