WSPR checkout on new workstation

I have been building a new workstation and a simple test of its reliability for logging signals via the IC-7300 sound card is to run WSPR.

So, WSJT-X was configured for WSPR on 40m and run for 24h using the low G5RV inverted V dipole with tuned feeders (designed primarily as an NVIS antenna for local contacts).

Above is a map of the spots involving VK2OMD over the 24h survey.


Above is a list of heard and hearing for the survey period.

This was more a test of the receiver and computer than a propagation report. There are a bunch of issues with WSPR that question its use for propagation reporting.

Let’s look at just one, the dynamic range.

Above is a frequency distribution of data extracted from a month studied in 2011. There are almost half a million spots on 40m contributing to this analysis, so it covers a wide range of propagation conditions during the month, and includes all stations spotted by all stations.

Note that the distribution is almost linear from about -18dB to 13dB SNR, a range of just 31dB. If you read the description of the decoderĀ  you might at first expect that the probability would increase linearly down to a wall at -27dB SNR, but it gently slides away below -22dB SNR so although the decoder might perform very well to -27dB SNR on a constant signal, it doesn’t do so well with fading signals on 7MHz.

Given that fading exceeds 15dB over an hour on 40m, a dynamic range of around 30dB is very limited and is likely to distort results and any conclusions would be weak. Any study that includes observations below -15dB SNR or above 10dB SNR are likely to be tainted.

Measured ambient noise FS runs in -104dBm in 2500Hz (the noise bandwith used in WSPR SNR calcs) at this location, so a SNR=10dB represents S=-94dBm, or 21dB below S9, a quite modest signal at about S6 saturates the measurement range under these conditions.

Data mining of WSPR spots is made easy using SQL. The mysql SQL used for the summary follows.

use WSPR;

system rm "/home/owen/mysql/wspr/snrfreqdist.txt";

create table temp1 select band,snr from SPOTS
where band=7 and distance>0;

select snr,count(snr)/(select count(snr) from temp1) AS N
from temp1 group by snr
into outfile "/home/owen/mysql/wspr/snrfreqdist.txt";

drop table temp1;

You could use SQLite3, but some changes are needed due to SQLite3’s cutdown functionality.

The small output table is then graphed, I used Excel.