Dear LF,

I could just barely see Jay's WSPR signal, but a little too weak to decode 
here. The top panel in the attached picture shows the
faint traces, as seen on the WSPR spectrogram at 23:10.

The periodic nature of the WSPR signals tempted me to try SpecLab's built in 
averaging function to stack a number of 2 minute
cycles. The software signal generator was set up to deliver a burst every 120 
seconds, which was used to trigger a new spectrogram
sweep each time. After gathering about eight averages at 23:13, the noise had 
leveled out a fair bit, and details of the modulation
started to become visible (middle). The start of the transmissions is on the 
first of the two minute markers. The bottom snapshot
was taken an hour later, including 38 averages.

This incoherent averaging is a slow process, as noise power is reduced only 
with the square root of the number of averages, ie.
1.5dB SNR gain for double observation time. The advantage is that it does not 
require long term phase stability of the propagation
path.

Stacking of a number of spectrogram captures could possibly be also applied to 
dig repeating QRSS beacons out of the noise, provided
the repetition period is precisely known a priori.

Kind regards,
Markus, DF6NM

we2xgr2_av_090131_2258-2313-0013.jpg
Description: JPEG image