The forum is read-only currently.

Distortion & Noise Measurement of a new KiwiSDR, June 2018

I recently received a KiwiSDR from Mass Drop and made some measurements of distortion and noise of the unit. These measurements and their results are detailed in the attached file.

Glenn n6gn

Attachments:
http://forum.kiwisdr.com/uploads/Uploader/75/8d169145aa57b3ba76cf6219fb2f8e.pdf
WA2ZKDfractional_nrrobinetjksHB9TMCrz3dvp

Comments

  • Glenn,
    I've been experimenting with amplifiers for loop antennas for the KiwiSDR. 3 of the schemes are using Norton amplifiers and one is using a differential amplifier described by LZ1AQ. I lack the technical education to know what I'm doing, but doing it anyway, and having some success. One of the better amps, in my opinion is based on the Clifton Labs Z10040B using 14:1 transformers. It proved to be to much gain for the KiwiSDR and while it works with output attenuation, I should try one with 7:1 transformers. You can review some of these attempts here: https://easyeda.com/doxnairobi/

    If you have the Elmer tendency, I'm your guy. Hi Hi
    73,
    Ron Morell
    KA7U
  • Ron,
    Thanks for sharing interest in the Kiwi and for the encouragement but since posting my initial measurements a year ago I’ve discovered what I think significant errors and don’t want to let them stand. My posting was well-intended but I am now of the opinion that it was also misguided. The reported measurements were as stated but in the posting I implied that TOI, SHI and noise figure were suitable or even useful metrics for a KiwiSDR and that they could help to set expectations or help in comparisons with other receivers. I have come to believe that this is not correct and may be misleading and un-helpful. I’d like to recant.

    In the first posting I suggested that an ADC/SDR might be effectively without IMD. This simply isn’t true. Even a perfect ADC has quantization error due to a finite number of bits. The kiwi’s ADC may be limited by its SNR which can be due to other imperfections. The spurious responses created by these generally do not behave as they did for analog environments, where metric created from a single measurement allowed extrapolation to a reasonable expectation for performance elsewhere. Both distortion and noise impairments can vary as a function of input level. I now feel I made at least several errors:
    • I shouldn’t have implied that SHI or TOI were useful metrics for any kiwi.
    • Based on more measurements noise figure seems to be about 20 dB’ for small inputs on most kiwis, but can increase ~2 dB for large inputs.
    • Overload level seems to be nearer -14 or -15 dBm on most kiwis.

    I’m presently of the opinion that conventional metrics for describing distortion and noise may not be completely useful and might be misleading. Even alternate techniques, such as noise loading or interference-free range may not serve well to compare and predict aactual user experience of receiver performance. Coming up with a set of metrics that does provide a user with good advice is not easy and I’m still trying to learn about it.

    Meanwhile, the KiwiSDR has proven to be a very interesting receiver and tool for developing active antennas and other devices, as I originally hoped.

    So thanks Ron for the support but I think it safer to interpret what I wrote as “He didn’t know what he was talking about” and presume that this continues to be the case!

    Glenn n6gn
    June 23, 2019
  • Glen, I reread your original paper to remind myself of your setup and assumptions. I don't have access to equipment such as you use nor do I have the understanding to effectively use it without some expert training. Still, I have a general idea of what you were trying to measure in respect to signal input, noise floor, and distortion. In my use of the KiwiSDR and antenna inputs, the overload from strong signals, harmonics and spurious signals, and high level noise floors, are my concern and how to mitigate these problems is my interest. I could detail this further as to some specific problems I have, but I won't go into that.

    I do think you know what your are talking about. Empirical evidence often leads to incorrect assumptions, even though the observed evidence is correct. Without the effort, nothing would be ever be resolved, so don't be discouraged. I manage to get reasonable results in my efforts, and I certainly don't know what I'm talking about most of the time. LOL
    Ron
    KA7U
  • Ron,
    You say "In my use of the KiwiSDR and antenna inputs, the overload from strong signals, harmonics and spurious signals, and high level noise floors, are my concern and how to mitigate these problems is my interest." I actually think that this is precisely the correct approach. While getting a metric to capture it all may not even be possible, understanding the mechanisms and techniques to mitigate these are exactly my interest as well. Fortunately for us both, this can be done without having perfect a perfect metric.
    Best,
    Glenn n6gn
Sign In or Register to comment.