I wrote:
> http://abelian.org/vlf/tmp/160419a.gif
Another plot from the same pipeline, all same settings
but with the sferic blanker removed
http://abelian.org/vlf/tmp/160419b.gif
DF6NM is lost in the noise without sferic blanking.
Average noise per 139uHz bin: 0.02 fT with blanker
: 0.14 fT without blanker
Blanker is reducing the background by about 17dB in these
quiet afternoon conditions. Around 20dB is typical.
0.14 fT in 138uHz is 12 fT in 1Hz, reasonable for daytime
background.
I did some blanking trials in 2010 when Stefan first began
VLF transmissions, see the rough notes at
http://abelian.org/vlf/blanking.html
The trials examine clipping and blanking with a range of
thresholds, smoothing factors and dwell times.
The settings I have used since then correspond closely to
the last graph of Test 1. Typically between 30% and 40%
of the incoming signal is blanked when the threshold is
set to 1.2 times the moving mean amplitude.
Sometimes I can improve the S/N by raising the threshold
to 1.8 or 2.0.
Estimating the effect of the blanker on flux density
measurements is difficult so I just guess. I normally
use a factor of 1/0.7 amplitude gain to correct for the
blanking effect. If I want an accurate measurement
I have to mix in a simulated weak carrier before the
blanker and measure that as a reference.
--
Paul Nicholson
--
|
