Comparing antenna using WSPR
As some of you may be aware I’ve been comparing antennas by WSPRing from two sites simultaneously one band at a time, using the same callsign but different QRA locators. One antenna was a 33m long doublet N-S at 10m, the other was a 12.5m long loft mounted doublet N-S at 10m.
I have discovered several things from this experience
Comparisons on the LF bands are relatively straightforward as most antennas are relatively omni-directional. If lobes exist they tend to be very broad. On 80m and 40m the loft mounted antenna has approximately 6dB less gain. I could only rely on a small proportion of all the reported spots, as noise (which I assume was local to the receiving stations) varied considerably during the period I conducted the tests. This resulted in some cases 10 to 12dB worse S/N ratios for periods of time when other stations in the same geographical area were reporting normal levels. As a check, I could receive my own signals at each other site and found that the S/N reported by the loft mounted antenna was 6dB worse than that reported by the external antenna, this is the first time I have been properly able to observe non-reciprocity between TX and RX in WSPR reporting and it is purely due to a higher noise floor at one site. This meant that I could only calculate the differences between my two signals by using C/N figures as reported by others.
On the HF bands the lobes of the two antennas were not quite in alignment, however for stations in Europe the two antennas were within a few dB of each other. It was very gratifying to find a consistent difference of 2dB reported by all German stations and a difference of about 16dB from North American stations.
On 20m I also observed a huge difference between the RX performance at the two sites. My spots as reported by DH5RAE showed an average antenna gain difference of 2.3dB, but on receive I noticed a difference of 13dB. Some detective work has traced the problem to broadband noise being radiated from my PC power supply. This doesn’t show up on the waterfall, but it does severely affect the number of stations received.
I also tried varying my transmitter power from 5 Watts to 50 Watts at one stage. This seemed to improve consistency, changing the ratio of internal receiver noise to external noise received along with the wanted signal.
I think WSPR somehow needs to monitor or indicate changes in the level of receiver noise floor, or some method needs to be devised in order to better characterise receiver / external noise contribution. It’s easy to believe that a signal level has increased due to improvement in propagation condition, when it could actually be simply due to the neighbours turning their TV or PC off and removing a source of noise.
As an aid to measuring antenna performance WSPR is useful, but it’s very easy to misinterpret what is being observed.