Continue to Site

# Fully Differential OP AMP with CMFB and Autozeroing

Status
Not open for further replies.

#### AlexanderTheGreat

##### Newbie level 4
cmfb amplifier

Hello,

for my Diploma Thesis i have to design a low noise NF amplifier in CMOS Technology.
I chose a folded cascode OTA with autozeroing and fully differential design.
I read the paper "circuit techniques for reducing the effects of op-amp imperfections.." from Enz and Temes.

I wonder about a few things:

1. Both, the main amplifier and the CMFB amplifier have an offset, that's clear. do I have to autozero both amplifiers seperately or in one step?

2. My first suggestion was to use "closed loop offset cancellation". in the opinion of my tutor i have to choose an additional offset nulling input.
Do I have a biasing problem in this case because I have two control circuits, one for the offset and noise and one for the common mode feedback?

3. I wasn't able to find any paper about autozeroing in a fully differential op amp? does anybody habe such a paper for me?

4. In the first chapter of the paper mentioned above, calculation of the noise and offset are presented. Can somebody help me with an example for calculation noise and offset in an autozeroing amp?

The design target is an offset and noise (refered to the input) less than 120µV because I need a SNR of 74dB. This constraing comes from the ADC resolution.
The Signal is an low frequency signal (f = 10 kHz)

In my suggestion I can calculate the basic band noise like the following:

2 B Ts So + foldover component and the offet is the residual offset.
The total Input refered "error" voltage is then the Vrms of the noise + the residual offset? ( B = bandwith, Ts = 1/autozeroingfrequency and So = Thermal Noise Spectral Density)

Thank you all very much for you help! javascript:emoticon('')

Alex

#### sudheerkm

##### Member level 2
fully differential op amp

yours could be aproblem of competing current source that is why the pmos current source and nmos current source which ever wwins make the other in tride bias it properly derive all the current from singls source

best of luck

bynonsun

Points: 2