how to analyze transient response of chopper stabilized bandgap reference ciruit?

[Kit_Leong]

how to analyze transient response of chopper stabilized bandgap reference circuit?

Hello,

I am currently trying to design a chopper stabilized bandgap reference circuit using tsmc 0.18u technology in mentor graphics.Attached is a journal that shows a full schematic of what I am designing(figure 4) as well as a more basic bandgap without chopper stabilization (figure 3) in which I have already managed to successfully implement figure 3.

The problem I have now is moving on to figure 4.For the modulator and demodulator,I am using MOSFET's as my switches(as depicted in the Paint screen shot entitled '4 transistors as switches',these transistors will be driven by a square wave pulse alternating between -1 and +1 to perform the modulation and demodulation of the signal and to remove the 1/f noise with the help of a low pass filter)

I tried using NMOS for my switches but it is always in the subthreshold region,even after modifying the width and length ratio.How do I get it into linear or saturation?I tried using PMOS instead of NMOS instead,and I managed to get it to always be in a linear region at the default W/L ratio of 5/2.Does a PMOS function well as switches at linear region?Also,the bulk of a NMOS is usually grounded,while for the PMOS the bulk is connected to a VDD,does that mean I have to connect my bulk to a VDD as well for this 4 PMOS configuration for switching?

Also I would like to ask what are the expected transient waveform would be like to show that my bandgap is actually chopper stabilized.I have also attached the expected waveforms of chopper stabilized circuits and some explanation on it,but it is in the frequency domain.I would like to know its transient response instead as I do not know how to plot it in the frequency domain,as well as my lecturer has advised me to do it in the transient domain as he also said I can't do it in the frequency domain.I'm not sure why exactly he said I can't do it in the frequency domain,probably due to some limitation in mentor graphics I suppose.But as long as I can prove it is chopper stabilized,I am fine with either transient or frequency domain.Please help.

Thanks.

 

[dick_freebird]

If you don't need a lot of current then subthreshold
operation might be tolerable. I don't know why you
are using unrealistic gate drive voltages, why not
realistic supplies and CMOS inverters as commutation
drivers?

You can run a series of transient analyses with small
signal scale stimuli and record gain / phase information
to compute Bode points if you want.

 

[eladla]

It is usually better to use a transmission gate as a switch.
Transmission gate - Wikipedia, the free encyclopedia

Unlike PMOS only or NMOS only switches, it can pass both a low and high voltage with out degrading it. (as well as negative and positive)
The NMOS passes the positive and PMOS the negative.

This is more or less what you wanted to do, but with both used at the same time
Hope this helps.

 

[Kit_Leong]

Hm,this is interesting,I will give this a shot and see what I get,thanks alot man!

 

[eladla]

You were thinking in the right direction.
These TG devices are so simple, yet can be used for many applications.
They can be used as stand-alone switches, like what you want.
You can also create whole logic gates with them.
And their original application - they were used in ethernet drivers to disconnect the TX driver from the line when recieving and to disconnect the RX driver when transmitting.
Hence the name, transmission gate