Ac analysis of switched capacitor circuit

simetrix simplis download full version

I have problems to simulate swiched capacitor circuits with or cad pspice. I have realized a low pass switched capacitor filter with ideal elements (ideal switch and opamp found in analog.lib). If i make an ac simulation of this circuit,i obtain wrong results (because if i do a transint with sinusoidal voltage source, i obtain different results) ...in allen's book i've found a description about models used to simultate ac time variant circuits,but in .cir file nothing i've found...
1)how can i do a correct ac simulation in pspice?
2)what is the cause of the results that i've obtained?
thanks in advance and excuse me for all these questions

capacitor circuit noise analysis

1.) AC simulation is a small signal analysis and works only for linear and time-invariant circuits.
2.) Therefore, when s/C circuits are going to be simulated in the frequency domain, it is necessary to replace each switched cap with a time continuous element before.
3.) The most popular way is to use a combination of delay elements (artificial with ABM parts or delay lines) together with positive resp. negative resistances.
These combinations are called "storistor".
4.) This leads not to a kind of approximation but it is rather a correct method to apply ac simulations to S/C circuits.
5.) The method is described in detail by NELIN (IEEE Trans circuits and systems, CAS-30, 1983, pp. 43-48) and by GHAUSI and LAKER: Modern Filter Design, Prentice Hall.

I have used this method very often with good success.
Regards

switched cap ac simulation

Thanks for the infos,
i wrote .cir file with equivalent models that you have described. I've obtained good results (low pass filter has ac respnse equal that i've expected). But i have some doubts:
1)I have considered equivalent conductance equal to 1/(wC) where w=1rad/sec: is this ok?Why this value doesn't change with frequency?
2)if i do ac response with equivalent models, i obtain a gain value not equal with that obtained by transient with a Vsin voltage source with a fixed frequency: do you know why? (in ac response i obtain a gain value equal to my hand calculations, in transient i have a better value)
3)If i do an ac sweep of my swhitched filter (without any equivalent model), what i obtain?
Thanks

capacitor ac response

lionelgreenstreet said:

Thanks for the infos,
i wrote .cir file with equivalent models that you have described. I've obtained good results (low pass filter has ac respnse equal that i've expected). But i have some doubts:
1)if i do ac response with equivalent models, i obtain a gain value not equal with that obtained by transient with a Vsin voltage source with a fixed frequency: do you know why? (in ac response i obtain a gain value equal to my hand calculations, in transient i have a better value)
2)If i do an ac sweep of my swhitched filter (without any equivalent model), what i obtain?
Thanks

Click to expand...

To 1): What means "better" value ? (Larger or smaller ?) Is your clock rate high enough ? On the other hand, is it low enough to allow capacitor loading with parasitic resistances and a finite time constant ?
T0 2): I suppose you will get "garbage" (meaningless results).

ac analysis

LvW said:

1
3.) The most popular way is to use a combination of delay elements (artificial with ABM parts or delay lines) together with positive resp. negative resistances.
These combinations are called "storistor".

Click to expand...

Does it give good results for noise analysis also?

ac analysis of low pass filter

pixel said:

Does it give good results for noise analysis also?

Click to expand...

No, of course not, as the switching action is not included in this kind of analysis.

pspice switched capacitor ac sweep

ok, than spectreRF

ac analysis in simplis

Should run PSS and PAC for SC circuit, isn't it?

ac simulation switched capacitor

Try Simetrix/Simplis for SC frequency-domain analysis.

Thanks for your reply....

What means "better" value ? (Larger or smaller ?) Is your clock rate high enough ? On the other hand, is it low enough to allow capacitor loading with parasitic resistances and a finite time constant ?

Click to expand...

Larger value....I think my clock rate is larger enough (100 times greater than -3db frequency); i used in my simulation ideal switch (in analog.lib)
thanks

SIMetrix sounds a good software.
Where can i download a full version ?