Hello Jim, LF,
Jim, thanks for that informative picture and the effort to make these tests
understandable! I am impressed that yesterday your noise level was so low, that
the SNR would have been so good!! With that SNR we could easily switch to
DFCW-200 or so! Normally my ERP should have been 3 dB higher, as told. But the
next time!!
OK, one can see that clipping makes absolute sense, especially in a low noise
region! 0 dB above average is clipped, does this mean, that each sample that is
above 0 dB is just left away out of the whole integration that gives one pixel
in the spectrogram?
This clipping is done by SpecLab but you also implemented a narrow band filter
in your preamp, that is switchable. You said you want to show a picture with
the filter switched on and off (and display 0...20kHz). That would really
interst me! Have you used compact inductors in a resistor housing or real wound
pot cores for high Q? And what is the increase in the SNR @ 8,97kHz?
BTW, i like to read your mails. Always well formulated and informative ;-)
Best 73 and i am looking forward to the next experiment :-)
73, Stefan/DK7FC
________________________________
Von: [email protected] im Auftrag von James Moritz
Gesendet: So 04.04.2010 16:52
An: [email protected]
Betreff: Re: VLF: 4th VLF experiment by DK7FC/p
Dear Stefan, LF Group,
Looking at screen captures from my portable RX yesterday, your signal is
hardly visible The first dash was obscured by a burst of 50Hz noise, the
subsequent shorter dot is not enough by itself to clearly identify as a
signal. So here's hoping for better conditions next time...
However, I did take the opportunity to do some simple tests of the noise
level, and the effectiveness of clipping in reducing the noise floor due to
QRN. I used the preamp circuit with bandpass filter shown in the previous
mail on 23/3/2010, and the "hard limiter" function in Spectrum Lab, with
the limiting level set to "0dB above average". The attachment shows a
spectrogram (5mHz FFT resolution) with a marker signal at 8969Hz. Initially
at the /P location, at the bottom of the waterfall, I set the marker
amplitude to about the same level as your signal in your previous test -
this gave a SNR of roughly 10dB, and would have been readable as a DFCW
signal. This was equivalent to a field strength of about 1.1uV/m.
Moving up the waterfall, the bright band of noise is the effect of turning
off the clipping - the rise in noise level obliterates the signal (part of
the noise was also a nearby electric fence, producing a loud clicking
noise). Turning clipping on again restores a readable signal. The black band
shows the effect of removing the antenna and replacing it with a dummy
antenna - the preamp noise is well below the band noise, by 10dB roughly.
Further up the spectrogram shows the noise at my home QTH. The wideband
noise level increases several dB, and there are also bright streaks of
narrow-band QRM. I had to increase the marker level by 5dB to restore a
"readable" level of signal, ignoring the much stronger narrow-band noise. At
the top of the waterfall, turning off the clipping again also causes a rise
in the wideband noise level - to produce a "readable" marker, the amplitude
was increased by a further 5dB to 3.8uV/m.
So, overall, the use of clipping produced about 10dB apparent reduction in
noise level in the relatively quiet /P location, which is a significant
difference - the signal would not be detected here without the use of
clipping. Also, the preamp/antenna system noise is well below the external
noise, so internal noise is not a limiting factor.
Looking forward to next time...
Cheers, Jim Moritz
73 de M0BMU
<<winmail.dat>>
|
