Roelof Bakker wrote:
> I presume that the antenna is directly connected to your
> soundcard.
No, there is a pre-amplifier using LT1028 op-amps. Input
transformers match the 20 square metre loops to the op-amps
for optimum noise performance. The 'receiver' is calibrated
using the thermal noise from a 1 ohm resistor.
> Which program do you use for these graphs?
For the graphs I use gnuplot. All the signal processing
software is homebrew - C/Linux.
A little more system description...
After the LT1028 front end, a couple of further gain stages
raise the signal to line level. Ground-loop isolating
transformers and a 200 metre run of cat5 bring the signal
back to the near the house, through another pair of isolating
transformers and into a PC located outside of the house
(installed in a spare dog kennel!). There the signals are
digitised with an M-Audio 192 sound card running at 192k
samples/sec, 24 bits and recorded raw to a large hard disk.
From the outdoor PC, the data is sent over mains ethernet
into the shack where another PC does some signal processing
and mains hum removal, SID detection, spectrum plotting etc.
From there the data is encoded into mp3 streams and sent
to a pair of servers located in St. Louis, Missouri where
they can be combined with signals from other VLF receivers.
The two servers run software which analyse the signals for
whistlers, risers, and other interesting natural radio events.
They also provide some audio and video stream serving.
Raw data is kept for about 10 days so that it can be
reprocessed if necessary to examine anything interesting.
I will switch the test signal off later tonight because I want
the laptop back :) Maybe the next signal to appear on the
spectrogram will be from a radio amateur in Germany?? It will
be very, very weak, but the background noise at VLF varies
quite a bit, some days are better than others. With luck and
a low background... well, it is worth a try.... Viel Glück!
--
Paul Nicholson
http://abelian.org/vlf
--
|
