Second build has shipped. Message will appear here when store is ready for third build ordering.

Waterfall appearance versus SNR

I'm feeding a new KiwiSDR with a wire dipole which is resonant at about 3.4 MHz (about 138 ft. total length), about 15 feet above ground, in a rural area of Michigan. At night I'm typically seeing SNR 34:30 values. Replacing the SMPS should improve things, since there is a visible noise pattern it adds, though less obvious at night. The SNR values being reported seem to be higher than many of the online KiwiSDRs that I have looked at....however I need some seasoned KiwiSDR owners/users to take a look at the attached spectrum/waterfall to assess whether SNR 34:30 is a "real" value, i.e. does it actually reflect an above average signal environment?


  • Could be. The issue with judging a full-span waterfall for SNR is trying to see the entire coverage but also the fine detail. If you zoomed into say 20m, the noise floor is probably incredibly low, but if there are some signals there, they probably benefit from the very low noise floor rather than any antenna gain.

    The fact the strong broadcast signals are not overwhelming the radio is a real bonus too, in Europe and many other places, a large antenna normally means having to selectively reduce those broadcast signals during darkness.

    Very nice.

    It will be interesting to see how well it works when the higher bands are active.


  • Hi

    SNR reports are often exaggerated, and in my location where background noise is very annoying, I used multiple antennas. The best result was on a Delta antenna suspended at 10m, where after using bandpass filters and a 3-10MHz preamplifier, cutting everything above and below, my receivers were regularly in the top 10 of the linkfanel with the results 47:46 and sometimes higher with the background ~ s8 but 70% of the result was given by AM stations transmitting with a sick power of several hundred kW in the 31.41.49 m band, causing Kiwi overload in the evenings. Since yesterday I am using a double H / V magnetic loop from a converted MLA-30 at a height of 2m without the filter attached and the result does not exceed 30:30 in the evenings and during the day it is around 18:18 despite the fact that I receive again quite well stations 160m to 15m with background ~ s5. So, in fact, if the SDR receives less signals than, say, 15MHz as on your waterfall, the result will be higher, which does not really translate into the quality, quantity and comfort of receiving stations. Of course, there are a few receivers on the list where the background noise level is a different story and deservedly have such high SNR results. This rat race for a place in the statistics does not make sense :)


  • I am definitely beginning to appreciate how misleading the SNR values can be without viewing the spectrum detail. Based on viewing a variety of publicly shared Kiwi's it looks like quite a few of the higher values are situations where the spectrum is filled with SMPS harmonic "carriers" the raw SNR values do not reflect that. And I can't think of any alternative way to arrive at a more "true" SNR result, unless there were a sophisticated algorithm to sort the junk signals from the "real" ones. Maybe there should be a "wall of shame" for the ugliest, noisiest waterfall available to the public(!). So my conclusion is that a high SNR score by itself only suggests a good environment, but a qualitative evaluation via viewing the fine detail and tuning in real signals is ALWAYS required. And based on that I am pleased with my current antenna/SDR configuration. Once I work out a safe way to expose the Kiwi to the WAN (give it a separate LAN segment?) I will add it to the publicly shared roster.

  • edited September 2021

    The cyclic nature of spurs and noise can often be a clue. But your eye is often the best way to determine that, unless some external AI was applied.

Sign In or Register to comment.