AnsweredAssumed Answered

testing for correlation of noise

Question asked by DJOHNSON on Nov 14, 2003
Latest reply on Dec 4, 2003 by mfredrik
I am trying to do a simulation to test for phasenoise correlation at the output of two doubler circuits in RFDE (posting in ADS due to the higher volume).  My line of thinking is to take two doublers driven by the same OSCwphnoise source and then different OSCwphnoise sources, thus having 100% correlation of input phase noise in one experiment and uncorrelated sources on the other experiment.  Then I will look at the individual output pnoise in dBc of the two doublers and the sum of their noise outputs.  If the sum is 6dB greater than the output of each individual doubler then the pnoise is correlated at the output, if the output is 3dB greater then the output pnoise was uncorrelated.  In the second experiment, the output noise should always be uncorrelated since the circuits were completely isolated from each other and input phase noise was uncorrelated, but I get a 6dB increase of the phasenoise at the summed output of the two doublers saying that the output noise was correlated.  I think the issue is that the noise information that ADS gives the user does not contain the statistical information of the noise so when I sum them I always get a doubling of noise voltage.  Does anyone have a way of testing the correlation of noise, specifically phase noise, of two outputs? ???  Is there a way of outputing the phasenoise in such a way that I can add the output spectrums and get a noise doubling/power doubling depending on the correlation???