Been working on a SC BPF for a while, simulating with spectreRF.  I use current mode integration sampling at the input to a SC biquad, and find that noise aliasing at the input is described by a transient noise derivation I came up with involving finite integration length.  When simulating with pnoise I see the expected noise in baseband (including the folding predicted by the transient derivation), but I also see up-conversion of strong low frequency noise to my clock freq and 10x clock freq.  I recognize that the spectrum of the noise currents should have up-conversion of the low-f noise (I essentially chop them with my sampling phase), yet I feel that the noise at the output of my filter due to these low-f components should all be at baseband, and I bang my head against the wall trying to find a physical explanation.

Could these be sim artifacts from trying to do a periodic noise simulation on a transient noise phenomenon?  Or maybe I should bang my head more?  Thanks!!

Thanks for the reply.

Found my issue, I had a poorly chosen Pnoise frequency sweep and had points right on fclk and 10*fclk.  I seem to have minor up-conversion of low-f noise that was being greatly exaggerated.

pancho_hideboo:  I'm not trying to explain my situation anymore because it is solved.  Please see my post above. Thank you.