federico.butti
New Member
Offline
Posts: 2
|
Hi,
I have a question for the experts here! I am doing a SUBsampled PXF analysis (and a sampled PAC analysis) for a driven circuit (cml2cmos converter + inverters) to simulate supply noise. The thing that confuses me is the following: 1) fclk = 100 MHz 2) supply noise BW = 1k-300 MHz Therefore I set PSS fundamental as 100 MHz.
In my understanding, sampled PXF output is the contribution of each input sideband to the output frequency sweep chosen, which in this case should be Nyquist BW (50 MHz). Therefore, in the plot I see the first input sidebands which combined reaches 300 MHz (supply noise BW). Is this a correct way of seeing this thing? I suppose that if one wants to have the actual spectrum of the output sample sequence, one should sum up in the Nyquist BW all the input sidebands with the related folding, i.e. BW 0 between 0-50 MHz, sideband -1 between 50 and 100 MHz and so on, due to the subsampling.
In case noise BW would be smaller than 50 MHz, no subsampling would occur and therefore result would be correct without summing and folding, simply by taking sideband 0. Am I right?
|