SDR comparison test at KFS shows only a small difference in WSPR reception performance between RX888, KiwiSDR and Airspy HF+

[Rob Robinett]

Since KA9Q-radio (and Phil and Franco's support) was essential to getting this test running, I thought the results might be of interest to this group.

For the past several days I have been running a SDR comparison test at KFS, the top West Coast spotting station.  There RF from the TCI-530 antenna is fed through a 20 dB gain LNA, a high pass shelving filter, an 8-way passive splitter to six SDRs.  The net RF gain is about 8 dB on 10m going down to below unity gain below 30m. This 'top spotters' report for the last 24 hours shows that on 20m the KFS/OMNI (RX888) reports slightly more total spots and slightly more unique beacons than the /K (KiwiSDR), /A (Airspy HF+), and the /S (SdrPlay RP1a).   All four SDRs are being demodulated by KA9q-radio and decoded by the 'wsprd'  binary extracted from the WSJT-x 2.6.1 package.  So there is no analog stage in the RF signal processing path after the LM2208 ADC and the only difference is the SDRs.

Without the low noise gain stage, the high 15-20 dB noise figure of the Kiwi and RX888 would probably reduce their performance to below the Airspy and RP1a.  Without the high pass filter that additional gain AM and SW broadcast signals would overload the RX888 and Kiwi.  There are many good LNAs (but avoid the DXE Clifton Z10046), but Paul WB6CXC is the only commercially available high pass shelving filter: https://turnislandsystems.com/sdr-front-end-filter/

In my experience such a LNA+filter is essential for the best performance of an SDR system at any reasonably quiet QTH.

Rob

=============