CMFB simulation setup steps

[Junus2012]

Hello

Please I want the simulation setup of testing the CMFB. Ifinished the design of a fully differential amplifier and my teacher is asking me to characterize it
so what should I do

I designed the full MOS CMFB

I would apriciate any help of you

 

[rakshitdatta]

Simulate the common mode feedback loop gain. This has to be done in the same way as you do it for the main differential loop. Unity loop gain frequency of this common mode feedback loop and the corresponding phase margin is what you should look for.

 

[jgk2004]

Additionally, if you do a time domain simulation and apply a common mode step up down to the input of your amplifier, you can monitor to see if your stable if no ringing occurs. Additionally, it might be interesting to see how well your CMFB regulates to your common mode level, AKA offset.

JGK

 

[Junus2012]

Dear rakshitdatta and jgk2004

Thank you for your reply to my thread

I have a little small problem, we have two outputs and two inputs , if I would apply the step input with the buffer connection, whom should I connect to the output to make the buffer and which one is my input

I will be great helpful to assist your answer with picture

thank you both
Regards

 

[jgk2004]

connect it as an inverting amplifier with a gain of 1 then step the input.

Jgk

 

[Junus2012]

even if I assume to connect it as a buffer, the probelm of me I didnt deal with fully differential amplifier so I dont know how the circuit will be

could you please submit me the connection with an image ?

 

[jgk2004]

Here is how a diff-diff inverting amplifier should look.. Sorry for the ruff drawing.

https://obrazki.elektroda.pl/6159537400_1357806313.png

JGK

 

[Junus2012]

thank you jgk2004, it is clear

I was thinking before you give me the circuit that we need two signal inputs,

now I would like to ask how to connect the circuit for the AC open loop simulation test please

thank you in advance

Here is how a diff-diff inverting amplifier should look.. Sorry for the ruff drawing.

https://obrazki.elektroda.pl/6159537400_1357806313.png

JGK

 

[prestonee]

JGK you should get a good grade
ac sims: dc source from ground to cm set to your cm voltage, then have two ac sources each set to .5, with 1 pos to pos input of amp and neg to cm, the other will be flipped and connect to the neg input. this will be your ac ref of 1v. you should runa trans and wait for your cmfb to settle, then save operating points and run your ac sim with these operating points. at my work we have a cool widget that lets us sim the loop gain without breaking the loop. but you will probably need to break your loop, just remember when you break the loop you are changing the loading of your amp, and so the results you get will be slightly off depending on the values of your feedback network(typically a broken loop ac gain sim will look better then it is, unless youre incorporating the feedback loading as part of your compensation, in that case you may see a decrease in pm).

-Pb

 

[Junus2012]

Kindly do you have any schematic of this please

JGK you should get a good grade
ac sims: dc source from ground to cm set to your cm voltage, then have two ac sources each set to .5, with 1 pos to pos input of amp and neg to cm, the other will be flipped and connect to the neg input. this will be your ac ref of 1v. you should runa trans and wait for your cmfb to settle, then save operating points and run your ac sim with these operating points. at my work we have a cool widget that lets us sim the loop gain without breaking the loop. but you will probably need to break your loop, just remember when you break the loop you are changing the loading of your amp, and so the results you get will be slightly off depending on the values of your feedback network(typically a broken loop ac gain sim will look better then it is, unless youre incorporating the feedback loading as part of your compensation, in that case you may see a decrease in pm).

-Pb

 

[jgk2004]

Once gain, here is a quick schematic. Sorry for Ruff draw. Additionally, as prestonee said, its better to look at loop gain, since when opening the loop you change the loading of your amplifier, but since it sounds like this is for a school project, maybe we should keep it college simple.... So thus you need to simulate your amplifier either with a switch which can change for each simulation, for example, Cadance has a switch, spst_switch, it can be closed for DC and open for AC by just going to properties. If you don't have this. You can place a extremely large low pass filter in your feedback. aka an R_C filter with R=1Gohm and C=100F..Therefore, DC will pass and AC will be open.
https://obrazki.elektroda.pl/3060075200_1358071996.png

hope this helps

JGK