Hello Andre, I suspect three things being at work here:
1. Spurs from the mixer / synthesizer: Plot the |FFT|² (i.e. simply play back the file through a Qt GUI frequency sink) with a high FFT length. Do you see peaks in there? With the PSD (assuming the signal is WSS) being both the Magnitude square of the Fourier transform of the signal AND the value of the Fourier transform of the autocorrelation function, peaks here indicate a sinusoidal correlation, as you might be observing. 2. The anti-aliasing filters of course convert a perfectly white signal to a bandlimited signal: you should see "roundish" rolloff at the edges of the nyquist zone. That of course (through the inverse FT) corresponds to a non-dirac autocorrelation function, i.e. correlation. 3. By overdriving the analog receive chain, you might be bringing it to behave nonlinearly; a simple thought experiment shows that this might be ruining the whiteness of your noise (i.e. the uncorrelatedness): uncorrelated noise must have a zero mean. Squaring zero-mean noise of any variance > 0 must lead to non-zero mean noise. Zero-mean noise means a dirac impulse in the PSD at f=0. That means correlatedness. Also, nonlinearity will lead to increased spurs and spur intermodulation (see 1.). So, make sure you're not overdriving your receive chain – it's a delicate flower! Best regards, Marcus On Tue, 2019-05-14 at 13:38 -0600, Andre Rosete via USRP-users wrote: > Hello > > I captured noise samples (from an attached noise diode) with the > B200mini using GNU Radio Companion. I set the gain to 90 dB with the > intention of minimizing the noise figure. However, this caused the > noise samples to show correlation - as if they were samples of a > signal, rather than AWGN, which should be uncorrelated. I used 40 MHz > of bandwidth and 40 Megasamples per second, with center frequency at > 5800 MHz. > > Is this a known issue, and are there optimal settings that I can use > to minimize noise sample correlation? For example, with respect to > gain? > > _______________________________________________ > USRP-users mailing list > USRP-users@lists.ettus.com > http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com _______________________________________________ USRP-users mailing list USRP-users@lists.ettus.com http://lists.ettus.com/mailman/listinfo/usrp-users_lists.ettus.com
