attached is the output stage of a fully differential push-pull class AB amplifier,
in my previous design I used to control the output common mode voltage by using one CMFB connected to CM1 shown in the picture, the circuit was working fine,
it came to my mind why not using two CMFB amplifier instead of one, where the other one is added to control the upper current through CM2 as shown in the picture,
I have simulated the circuit by different type of simulation and I found that loop control is becoming better,
However, I didn't see before designers who used two CMFB, so regardless of the extra power paid for the new CMFB, is there other limitation ?
You don't have two controls and trying to make them match. You'd better have one free variable (PMOS current for example) and control the NMOS current to match the two. Somehow I think it is problematic to have two feedback loops to control the same thing, they may not agree. What if the two CMFB amps have very different offsets and one of them drives the other completely out of its good operating region?
You don't have two controls and trying to make them match. You'd better have one free variable (PMOS current for example) and control the NMOS current to match the two. Somehow I think it is problematic to have two feedback loops to control the same thing, they may not agree. What if the two CMFB amps have very different offsets and one of them drives the other completely out of its good operating region?
In the circuit I presented I dont have two CMFB controlling the same node as you might think from my post, I used two CMFB to control two different node, the first CMFB is controlling the upper current by CM2, I used NMOS amplifier for this node, the second CMFB is controlling the bottom node by CM1 and I used PMOS differential amplifier for it.
is that what you ment?
also I did't understand your alternative suggestion
Yes, that was what I was saying. You use two CM loops to control the same thing, which is the output common-mode voltage. The alternative is to leave one of the current sources and control the other.
So you are afraid that the two loops will conflict with each other,
By the way suta, I have simulated the loop gain of the CMFB, when I connect the two CMFB it gives me 0dB or unity for both, I think this because bothe work against each other to reach the same goal, the above CMFB is trying to increase the curerent while the down one is trying to decrease it