Re: Quick look at RPI Kiwi noise

classic Classic list List threaded Threaded
1 message Options
Rob
Reply | Threaded
Open this post in threaded view
|

Re: Quick look at RPI Kiwi noise

Rob
Hi Glenn and Tom,

Thanks for the work.  
Holger's KIWI_1 SNR vs KIWI_4 SNR heartmap chart appears to me to show a 1-2 dB SNR advantage to theKiwiPi KIWI_4.
But that advantage is probably within measurement error limits.

Rob

On Sun, Sep 6, 2020 at 10:32 AM Glenn Elmore <[hidden email]> wrote:

Using his Rig Expert as a source, spec'd to put out +13 and adding 60 dB of attenuation the RPI/Kiwi indicated -43 dBm at 15.092 MHz, about 4 dB high of what it should.  I managed to sweep it to 30 MHz one time and got this data:

Because we don't yet know the error v frequency and only hope the Rig Expert + attenuation is correct at 15 MHz, it's too soon to draw conclusions but at least at 15 the noise (with the 3-4 dB correction) would indicate arounbd -157 to -158 dBm/1-Hz in the absence of strong signals which is exactly what the 'standard' Kiwi normally measures. We didn't test to see if there was also noise floor lift in the presence of a strong signal that exercised all of the ADC bits and thus might expose non-linearity.

The code in the Kiwi or WebRX  has issues that appear to relate to the 60 MHz span. I had to zoom to max and back to get the display to be right. 

I don't know what the spectrum of the Rig Expert, 2nd and third harmonics of what is probably an SiLabs synthesizer seem normal  but I am suspicious of the non-harmonic term shown at 49+ MHz, It may be aliasing. Maybe they didn't get the filtering right:


That's all for now, an attempt to measure noise to 60 MHz failed with kiwirecorder.py not getting response after the first success. Don't know why not.


Glenn n6gn



--
Rob Robinett
AI6VN
mobile: +1 650 218 8896