trying to simulate a chopper amplifier. For testing it is configured as a voltage follower. There is only one clock all the non overlapping clocks ect are derived from it. The problem is that the flicker noise is not being reduced on the input noise plot when I use the following pnoise, pss settings:
pss pss fund=100k harms=3 errpreset=conservative+ annotate=statuspnoise ( out 0 ) pnoise sweeptype=relative relharmnum=1+ start=1 stop=50k log=20 maxsideband=20 iprobe=V2 refsideband=0+ annotate=status
I have tried various settings and cannot reduce the flicker noise. I have verified that the chopper is working and it does remove the dc offset when inserting a non zero vdc at the inputs in the time simulation. I have also tried following the chopper opamp with a with a sample and hold and taking the output for pnoise simulation from the sample and hold but this still does not reduce the flicker noise seen on the direct plot.
I've read Ken Kundert's sc-filters.pdf from www.designers-guide.org/Analysis/ and have successfully used pss and pnoise to simulate an oscillator in the past following the spectreRF tutorial.
There has been a few posts on this site as well as others but the only advise given is to eternally point people to Ken Kundert's sc-filters.pdf or the cadence manual.
Any ideas where I'm going wrong with the simulation setup?
BenMartin said:I have the same question. I don't see 1/f noise being modulated away from the origin. Did you ever figure it out?
It should be set to the zero harmonic ie relharmnum=0 not relharmnum=1 as that starts the sweep around the clock (see my original spectre settings above) which was not the intention. After that the settings are fine and the chopper works as desired regarding 1/f noise.
The error was obvious but no thanks to this forum though.