The PNOISE analysis in Spectre gives strange results of sampled thermal noise from a simple resistor-capacitor network. After SINC compensation, the noise PSD is flat, but the integrated noise from 0 to fs/2 varies depending on the resistance R instead of 64uV/1pF fixed. Why is this?

Thanks,

- Yun Chiu
10/11/05

Thanks for your suggestion.

I used a 1k resistor and a 1pF capacitor, sampled at 10MHz. The following is my PSS and PNOISE statement:

pss_x pss period=T tstab=0 maxacfreq=1000*fclk
swp_x sweep param=nres values=[100 1k 10k] {
   pns_x (vo vss) pnoise start=0 stop=0.5*fclk lin=1e2 maxsideband=1000 \
   noisetype=timedomain noisetimepoints=[0] numberofpoints=1
}

I got a total sampled noise 177, 56, and 18uV for the resistance 100, 1k, 10k ohm, respectively.
Any suggestions?