WSPR performance disappointment
I just began testing a new KIwiSDR for WSPR decoding, and have been disappointed by the low WSPR spot counts versus other SDRs, when attached to the same antenna. In comparison to an Airspy HF+ the spot counts are roughly 30% lower with the Kiwi. Wondering if the Kiwi is simply less sensitive than other common SDRs, in which case I'll need to consider an LNA. The antenna is a random length dipole (138 foot total length), with a Cat 7 ethernet feedline, and a video balun (used for ethernet to video cable installations) at the antenna input. It works beautifully with my Airpsy, SDRPlay, and RTL SDRs. In addition, subjectively when listening to weak signals the KiwiSDR sounds significantly noisier than other SDRs.
On the other hand, I am not helping things by using a switching power supply to power the Kiwi. But with the antenna input terminated with a 50 ohm SMA load the noise floor is nice and quiet and flat, generally below -110 dBm. I would have expected that switching noise would be manifested as assorted noise spikes or blobs in the spectrum.
I have a lot more testing to do, but just wanted to post some early thoughts to tap into the community wisdom here.
Comments
Both WA2TP and KD2OM, who amongst the top wspr reporters in North America, do it all with KiwiSDR. They run wsprdaemon external decoder systems. The kiwi internal wspr decoder can't handle big spot counts, not enough processing power.
The internal KiwiSDR WSPR extension was written 5 (6?) years ago as the very first example of the extension concept. It was not designed to be high-performance and has not been updated to use later/better versions of the WSPR code. That's what wsprdaemon is all about.
Even if the code were updated the limited processing power of the Beagle (compared to modern PCs or the latest RPi's) means it will never match them. But that was never the point..
Thanks for the rapid replies, I am new to the kiwiSDR so I did not realize the built in decoding for WSPR is not recommended for general use. From what I have read so far it sounds like the wsprdaemon is a separate program running on another Linux system such as a Pi. So for now I should focus on optimizing the signal sensitivity and trying to understand if/why it SEEMS less so than my other SDRs.
Unlike other SDR that use USB to connect to your PC, the kiwi uses your LAN. Common Mode Interference (CMI) has been shown to be a noise source and can be mitigated by use of ferrite. There's bad and good Switching PS, but hard to tell sometimes.