I am seeking help. I have trouble with the simulation of my CMOS chopper amplifer.  And thanks for help in advance.
----------------------------------------
The structure of circuit (standard chopper amplifier):
Switched-capacitor ckt to get the orginal signal, then go throuhg 1st chopper, open-loop high-gain amplifier(120dB at dc, but gain falls down fast, at 1.5KHz, the gain is reduced to 65dB already, unity gain bandwith 5MHz), 2nd chopper, low-pass filter. After the switched-capacitor ckt, and before the low pass filter, there is another switch and capacitor in parallel to form a closed loop feedback W.R.T. the switched-capacitor ckt at the front of this structure.  

--------------------------------------
Simulation Set-up:
Clocks:the freq for switched-capacitor ckt (fswitch) is 100KHz (T=10us), the freq for chopper i used are 6.25KHz(T=160us), 25Khz(T=40us), 50Khz(T=20us),  input signal is less than 1KHz, i used 1Khz as input signal freq. For example, i select beat period =20us when i run PSS with chopper freq =50KHz. Both choppers are using the same clock.

Select the output port, gnd, leaving input port not selected (so all the noise plot is measured at the output node of the whole circuit, it is not input-referred). Noise Type:sources:SSB noise analysis, Noise sideband is 6 or 10. And the reference side-band is 0. Spectre Gear2, liberal is used for all the simulations.  

-----------------------------------------
Simulation result:
I have the correct gain from transit analysis, for example, i used gain =2 for the whole circuit. But at PSS+Pnoise, i barely have any 1/f noise reduction at low freq range.Between the 0 and chopper freq, there is still large 1/f noise sitting overthere in the Pnoise plot. Noise measurement ranges from -89dB to -127dB between 0Hz to 500Hz. And the chopper freq selection did not change noise performance.

----------------------------------------------
Questions:
(1). Is my simulation set-up correct?

(2). Why the 1/f noise is not reduced? Even I do see the flipped 1/f pattern at chop freq in the Pnoise result plot? But between 0 and chopper freq, 1/f noise shape is still dominant.

(3). What would be a good choice for chopper freq? It seems that the PSS + Pnoise is very sensitive to the choice of chop freq and sometimes doesn't? When i run with fswich = 200KHz, chopper freq = 100Khz, PSS will converge, but Pnoise give bad result --- they go 300dB!!!?

(4). If i use chop freq = 1/2 fswitch = 50KHz, do i still need enough ope-loop gain at 50KHz?



Thanks very much.

Tony