The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl Design >> Analog Design >> PSS for SC-CMFB again https://designers-guide.org/forum/YaBB.pl?num=1120551427 Message started by ethan on Jul 5th, 2005, 1:17am |
Title: PSS for SC-CMFB again Post by ethan on Jul 5th, 2005, 1:17am Hello Everybody, For my fully differential fold-cascode OTA with SC-CMFB, according to the Ken's paper "Simulating Switched-Capacitor Filters with SpectreRF" and previous posts in 2003 and 2004, I use PSS and can achieve good common mode output voltage results when I let two input signals are disabled. I have used two "vsource" as my two differential inputs, and actually I set type=dc and leave others all blank. But how can I get large signal transient plot from PSS, which just looks like the transient plot through conventional "tran" analysis. For example, from conventional "tran", normally if we set input with peak-to-peak 1mv sine wave, we can see large signal level in output sine wave, say peak-to-peak 2v~3v in some cases. But with PSS, right now, I only can get differential gain with 4 in rough from the PSS transient plot. What I tried was to set vsource type=sine, and fill "frequency name 1", "amplitude Vpk", "phase for sinusoid 1", and"sine dc level"=2.5v (which is common-mode input). But It is expected to get 2000~3000 voltage gain. So what I did wrong? Do I need to fill these specifications to get large signal transient plot in PSS? Anybody can share his or her experience? From then on, for PAC, after I set vsource "type=sine" with "display small signal params" and "PAC magnitude", "PAC phase", I got differential gain with 4 and about 10dB. Any suggestion? Or my circuit indeed has only 10dB gain? I have replaced SC-CMFB with equavilent resistors, its gain goes up to 70dB. Appreciate your help. part of the netlist is attached: // Cell name: fold_cascode_new_lowpower_sc_TG_PNOISE_invclk // View name: schematic //two input sources V4 (Vin net082) vsource type=sine freq=30K ampl=500.0u sinephase=180 \ sinedc=2.5 pacmag=500.0m pacphase=180 fundname="fin" V1 (Vip net082) vsource dc=0 type=sine freq=30K ampl=500.0u sinedc=2.5 \ pacmag=500.0m fundname="fin" // two pairs of clock for SC_CMFB V10 (ph2 0) vsource type=pulse val0=0 val1=5 period=5u delay=200.0n \ rise=500p fall=500p width=2.2u fundname="clk2" V36 (0 ph1_b) vsource type=pulse val0=5 val1=0 period=5u delay=0 rise=500p \ fall=500p width=2.7u fundname="clk1_b" V37 (0 ph2_b) vsource type=pulse val0=0 val1=5 period=5u delay=200.0n \ rise=500p fall=500p width=2.2u fundname="clk2_b" V8 (ph1 0) vsource type=pulse val0=5 val1=0 period=5u delay=0 rise=500p \ fall=500p width=2.7u fundname="clk1" ....... //PSS and PACsimulation setup simulatorOptions options reltol=1e-3 vabstol=1e-6 iabstol=1e-12 temp=27 \ tnom=27 scalem=1.0 scale=1.0 gmin=1e-12 rforce=1 maxnotes=5 maxwarns=5 \ digits=5 cols=80 pivrel=1e-3 ckptclock=1800 \ sensfile="../psf/sens.output" pss pss fund=10K harms=0 errpreset=moderate tstab=1.5m + method=gear2only maxacfreq=20M annotate=status pac pac start=1 stop=300M maxsideband=0 annotate=status modelParameter info what=models where=rawfile element info what=inst where=rawfile outputParameter info what=output where=rawfile designParamVals info what=parameters where=rawfile saveOptions options save=allpub |
Title: Re: PSS for SC-CMFB again Post by Ken Kundert on Jul 5th, 2005, 2:19pm Why don't you leave the large signal inputs disabled and simply use PAC? It will be much faster. -Ken |
Title: Re: PSS for SC-CMFB again Post by ethan on Jul 5th, 2005, 9:14pm Ken Kundert wrote on Jul 5th, 2005, 2:19pm:
Thank you for your suggestion, Dr. Kundert, The reason that why I set the large signal inputs is that I would like to see the large signal transient plot through PSS. If I disable the large signal inputs and use PAC mag only, I only can get common-mode output voltage level for PSS as before. Can you comment why I couldn't get large signal transient plot for output voltage in PSS even though I set large transient signal inputs? Thanks a lot. by the way, I still have a quesiton regarding PAC. I read a post in Mixed-signal Design forum that http://www.designers-guide.org/Forum/?board=ms_design;action=display;num=1101544899 basically, he said " PAC is considering the whole signal (over a complete clock cycle) when computing the gain. If your signal spends half its time at zero due to a reset phase, that will effectivly reduce your gain by half (6 dB). " Is that the reason why I only got 10 dB AC gain with PAC, instead of expected 60~70dB gain? Thank you for your time and help. ethan |
Title: Re: PSS for SC-CMFB again Post by Ken Kundert on Jul 6th, 2005, 12:38am Currently you are running a PSS analysis with both the clock and a "large" input signal present. This is not a transient per see, but rather a large signal steady-state analysis. Applying the input tone during PSS is expensive, and should only be done if you are interested in large signal effects that are present as a result of the large input signal, effects such as distortion. If you are only interested in gain, you can save yourself a lot of time by applying the input signal as a small signal during a PAC. Thus, typically, if you were interested in the gain of your clocked amplifier, you would apply the clock only and then perform a PAC analysis to compute the gain. There is no reason why you should be confined to only a common-mode response with PAC. The Fourier analysis in PSS does operate over the entire signal, and so if your signal has a 50% duty cycle, it will appear to have half the energy, or 6 dB less that a signal with 100% duty cycle. However, that cannot explain a difference of 60 dB. -Ken |
The Designer's Guide Community Forum » Powered by YaBB 2.2.2! YaBB © 2000-2008. All Rights Reserved. |