pooh_bear
Newbie level 5
Hi, I was wondering if there is a significant difference to using a single ended vs a fully differential amplifier as the gain boosting amplifier.
What I know:
Choice I:
Gain Boosted w/ Singled ended auxiliary:
+Needs 4 auxiliary amplifiers (2 pairs for tand 2 pairs for bottom current source)
+auxiliary amplifiers have Pole zero cancellation due to being singled ended so very slight reduction in BW
Choice II:
Gain Boosted w/ Fully differential auxiliary:
+Need 2 auxiliary amplifiers (1 for top and 1 for bottom current source)
+needs 2 CMFB (1 for each of the auxiliary amplifiers)
+Larger Swing (But I don't think this will make much of a difference? The cascode fets' gate source volage shouldn't change a lot, it's being regulated by current source load)
Choice III:
Also I saw another way to add CMFB for fully differentail auxiliary amplifier shown in a couple of papers :
"Optimization of Sample/Hold Circuit for High-Speed and High-Resolution ADCs, Junxiao CHEN, Lu ZHANG, Lenian HE Institute of VLSI Design, Zhejiang University"
and
"A New Modeling and Optimization of Gain-Boosted Cascode Amplifier for High-Speed and Low-Voltage Applications Mohammad Mahdi Ahmadi, Student Member, IEEE"
They added a dummy branch to mimic the input differential pair to set the CMFB. So there is an extra current branch running to the source node of the input pair. Is this one prone to offsets? It seems nice, simple and compact for a CMFB.
I'm trying to make this amplifier for a pipelined adc but I am not sure which one would give the best result. Do the differences matter much or am I just wasting my time? Any help or comments are appreciated.
What I know:
Choice I:
Gain Boosted w/ Singled ended auxiliary:
+Needs 4 auxiliary amplifiers (2 pairs for tand 2 pairs for bottom current source)
+auxiliary amplifiers have Pole zero cancellation due to being singled ended so very slight reduction in BW
Choice II:
Gain Boosted w/ Fully differential auxiliary:
+Need 2 auxiliary amplifiers (1 for top and 1 for bottom current source)
+needs 2 CMFB (1 for each of the auxiliary amplifiers)
+Larger Swing (But I don't think this will make much of a difference? The cascode fets' gate source volage shouldn't change a lot, it's being regulated by current source load)
Choice III:
Also I saw another way to add CMFB for fully differentail auxiliary amplifier shown in a couple of papers :
"Optimization of Sample/Hold Circuit for High-Speed and High-Resolution ADCs, Junxiao CHEN, Lu ZHANG, Lenian HE Institute of VLSI Design, Zhejiang University"
and
"A New Modeling and Optimization of Gain-Boosted Cascode Amplifier for High-Speed and Low-Voltage Applications Mohammad Mahdi Ahmadi, Student Member, IEEE"
They added a dummy branch to mimic the input differential pair to set the CMFB. So there is an extra current branch running to the source node of the input pair. Is this one prone to offsets? It seems nice, simple and compact for a CMFB.
I'm trying to make this amplifier for a pipelined adc but I am not sure which one would give the best result. Do the differences matter much or am I just wasting my time? Any help or comments are appreciated.
