problems about chopper amplifier analysis with pss+pnoise

[lily1981216]

Re: problems about chopper amplifier analysis with pss+pnois

hello everyone!
i do noise analysis on chopper amplifier with pss+pnoise.in my circuit i have three
sources,two pulse as clocks of chopper ,one sin source as input singnal of amplifier.the pulse's frequency is 100khz and the sin's is 10khz.and the setting of pss+pnoise is like this:

Added after 25 minutes:

then is the setting right? should the sweeptype be choose relative? should the input sin source be put?and if i put the input singnal and i choose "relative" ,i have to choose 10khz as beat frequency and i wonder if i can only see the noise with frequency aroud 10khz or not.where could i see the 1/f noise.
can anyone help me?i look forward to your advice!

 

[steer]

Re: problems about chopper amplifier analysis with pss+pnois

In pnoise setup you should chose absolute and chose the range of interest. I would chose more sidebands, say 100 or so.
In pss setup I would chose more harmonics, may be 100.
Other than that, the setup is ok.

How do you plan to observe the simulation result? This is not a trivial thing as well.

 

[lily1981216]

Re: problems about chopper amplifier analysis with pss+pnois

yes,it's really a problem,i choose absolute and the plot is like this:the first one is the whole spectrum,then the last one is the spectrum below 10khz.
i want to observe the spectrum below 10khz .but can i belive the result of cadence
whose spectrum is below 10khz?

Added after 3 minutes:

oh,steer,thanks to your advice!!

lily

 

[steer]

Re: problems about chopper amplifier analysis with pss+pnois

You should look on the results in a different way.
There is no point to look on spectrum above pss nyquist frequency, at least in your case. Since pss frequency is 10kHz, the nyquist is 5kHz. So, your second graph shows the correct noise, except you should limit it by 5kHz. Also, you might want to chose time domain sampling point in pnoise setup panel to ensure the noise is calculated at right time - in case your amplifier noise properties vary in time, not always a case.

 

[lily1981216]

Re: problems about chopper amplifier analysis with pss+pnois

hi,steer
now i think i'm puzzled .if my signal frequency is 10khz .shouldn't the nyquist frequency be 10khz?and the effective sampling frequency should be 20khz,shouldn't it? what does the pss nyquist frequency refer to?why should it be 5khz?
thanks to your advice!
lily

 

[steer]

Re: problems about chopper amplifier analysis with pss+pnois

I thought 10kHz is your clock frequency. Sorry for the confusion.
In general, the nyquist frequency in pnoise analysis result plot panel should be half the clock frequency.

But wait, most noise simulations in amplifiers are done with input signal equal to 0. Is there any specific reason you apply 10kHz input signal in your noise simulation?

 

[gingerjiang]

Re: problems about chopper amplifier analysis with pss+pnois

i think the beat frequency is your pulse frequency 100khz
input signal may use a dc voltage with ac=1

 

[pseudockb]

Re: problems about chopper amplifier analysis with pss+pnois

Just curious about the pnoise plot. From papers I read about chopper stabilization method, the 1/f noise of the PSD of chopper amplifier should be shifted to odd harmonics of chopping frequency. But in the PSD plot, I could still see the 1/f phemomenon. Is this correct?

 

[jc2]

Why does sweeptype option have to be absolute?? what is the difference between absolute and relative option??

thank you