DaffyDuck
New Member
Offline
Posts: 6
|
Hi,
Simulation set-up: (A) gnd -- res -- sw ----- (B) gnd -- res -- sw -------- sw -- res -- gnd (ck) | (ck) | (ckb) cap cap | | gnd gnd res = 500 Ohms (noise turned on) sw = verilog-A model, noiseless cap = 300f
For (A) and (B), I turned on the "trans noise" option in ADE and set noisefmax to 10G, and then ran a 10u transient sim. I then used the "rms" function in Calculator on the voltage across the cap.
For (A), the rms voltage is about 108uVrms, close to the theoretical value of 117uVrms ( sqrt(kT/C) of 1pF cap is 64uVrms ). For (B), I got the same number! I was expecting sqrt(2) * 117 uVrms, since there are 2 clock phases, 2 different resistors, and hence no correlation on the 2 noise generators from res.
Can I use the rms function this way to compute the sampled kT/C noise, or .... what's really the problem?
Daffy
|