dineshbabumm
Advanced Member level 4
We did common mode feedback for an opamp using a switched capacitor structure.. The structure is as shown below..
There voutp and voutm are output terminals.. We are controlling one of the bias with this cmfb so that output node remains a constant..
Our opamp is 2 stage architecture and we used 2 cmfb to control both the stages.. Only then it worked in all corners..
Actually while trying to replace the ideal switches in that archi with cmos (transmission gates) we are facing many problems.. We dont know how to design it.. We used minimum sized transistors for switches..
But doing so, in the first stage,output resistance of opamp comes in parallel with the transistor's resistance and gain gets hit.. Also signal is getting degraded with too many spikes! Also we have clock feedthrough..
So can some one help us about how to design the switches and also some inputs on cmfb as to what are the other possible ways through which we can design without affecting the gain and bandwidth.. We tried using this cmfb(from carley paper of ieee) and also tried feldman(another ieee paper) method.. But in 2nd method too we need shifting and we invariably had to use swtiched caps which involves switches!
Please help us in designing the switches and also please suggest some good ieee papers or other literature where we can get to know how to properly design switches!
There voutp and voutm are output terminals.. We are controlling one of the bias with this cmfb so that output node remains a constant..
Our opamp is 2 stage architecture and we used 2 cmfb to control both the stages.. Only then it worked in all corners..
Actually while trying to replace the ideal switches in that archi with cmos (transmission gates) we are facing many problems.. We dont know how to design it.. We used minimum sized transistors for switches..
But doing so, in the first stage,output resistance of opamp comes in parallel with the transistor's resistance and gain gets hit.. Also signal is getting degraded with too many spikes! Also we have clock feedthrough..
So can some one help us about how to design the switches and also some inputs on cmfb as to what are the other possible ways through which we can design without affecting the gain and bandwidth.. We tried using this cmfb(from carley paper of ieee) and also tried feldman(another ieee paper) method.. But in 2nd method too we need shifting and we invariably had to use swtiched caps which involves switches!
Please help us in designing the switches and also please suggest some good ieee papers or other literature where we can get to know how to properly design switches!
