Fully Differential Opamp and CMFB

• Dec 20, 2004

• #1

C

ccw27

Full Member level 5

Joined

Oct 13, 2004

Messages

267

Helped

14

Reputation

28

Reaction score

6

Trophy points

1,298

Activity points

2,558

differential output op amp

For simulation analysis only is it necessary to build a CMFB for a differential amplifier? It seems hard to bias a telescopic or folded cascode on a 1.8V supply voltage without CMFB.

Thanks

 

• Dec 20, 2004

• #2

C

cristianb

Member level 1

Joined

Sep 9, 2003

Messages

33

Helped

3

Reputation

6

Reaction score

1

Trophy points

1,288

Activity points

287

ideal cmfb

U need to used a CMFB to set the output ot the fully-differential op-amp, otherwise is is even harder to bias it. And u can do this evenwith lower voltage supply than 1.8V

 

• Dec 20, 2004

• #3

Y

yeewong_su

Advanced Member level 4

Joined

Oct 12, 2004

Messages

119

Helped

5

Reputation

10

Reaction score

3

Trophy points

1,298

Activity points

803

fully differential folded cascode

Full differential output opamp need the cmfb, you can use the sc ckt, and if your supply is 1.8v, you'd better using the foldcascode and gain boost, may be it's diffcult if you select the telscopic

 

• Dec 20, 2004

• #4

M

markty

Member level 3

Joined

Dec 15, 2004

Messages

59

Helped

2

Reputation

4

Reaction score

0

Trophy points

1,286

Activity points

493

how to design cmfb differential amp

CMFB is necessary. Otherwise the output common voltage may float to the supply in the real ic.

 

• Dec 21, 2004

• #5

B

bear7679

Member level 2

Joined

May 18, 2004

Messages

51

Helped

1

Reputation

2

Reaction score

1

Trophy points

1,288

Activity points

412

fully differential opamp

I think you must use CMFB circuit, you can use SC type or RC type. For detail you can see razavi's book.

 

• Dec 21, 2004

• #6

A

arsenal

Full Member level 2

Joined

Oct 17, 2004

Messages

143

Helped

15

Reputation

30

Reaction score

4

Trophy points

1,298

Activity points

1,103

fully differential amplifier with opamp

yeewong_su said:

Full differential output opamp need the cmfb, you can use the sc ckt, and if your supply is 1.8v, you'd better using the foldcascode and gain boost, may be it's diffcult if you select the telscopic

Click to expand...

u a right,and sc is more preferrable as compared to resistors arctec

 

• Dec 21, 2004

• #7

C

ccw27

Full Member level 5

Joined

Oct 13, 2004

Messages

267

Helped

14

Reputation

28

Reaction score

6

Trophy points

1,298

Activity points

2,558

differential output operational amplifier

How can I design an ideal CMFB ckt first?

 

• Dec 21, 2004

• #8

C

cristianb

Member level 1

Joined

Sep 9, 2003

Messages

33

Helped

3

Reputation

6

Reaction score

1

Trophy points

1,288

Activity points

287

gain of fully differntial opamp

you can simulate an ideal CMFB by using ideal op-amp but the topology for the real circuit is dependent on the maximum voltage swing of the output stage of the main fully differential amp and its current capability. you will find a good analisys of fully differential amps and CMBF topologies (including compensationi) in the last chaper of Gray & Meyer last edition (I think its the 4th) of Analogue Integrated Circuits.

 

• Dec 21, 2004

• #9

C

ccw27

Full Member level 5

Joined

Oct 13, 2004

Messages

267

Helped

14

Reputation

28

Reaction score

6

Trophy points

1,298

Activity points

2,558

fully differential folded-cascode

This is my ideal CMFB circuit, the amplifier is taken from ahdlLib. For initial design of low voltage fully differential amplifier this can be quite useful as it can help you focus on the design of differential amplifier rather than the whole thing. Of course, later you should substitute this with an actual CMFB circuit found in textbooks.

Does anyone know how I can implement the amplifier in the ideal CMFB circuit with finite gain and GBW? This can help the designer decide the minimum spec he needs to design the actual CMFB circuit.

Thanks

 

• Dec 22, 2004

• #10

C

cristianb

Member level 1

Joined

Sep 9, 2003

Messages

33

Helped

3

Reputation

6

Reaction score

1

Trophy points

1,288

Activity points

287

problems with fully differential amplifier

First of all I there's a problem with the way u read the input signal (voc) for the error amplifier in the CMFB. voc=von+vop, but this is equal to 2*VCM so u need (vop+von)/2. for this use a simple resistive voltage divider. As for your question, the gain for the error amplifier is direcly linked to the precision u want to control the common mode at the ouput related to the voltage swing. so lets say if u need a 10mV precision for a output voltage |Vo|=1V this leads to a gain of 40dB. Now having and open loop amplifier in silicon the gain will vary so u need to go even a little higher. if u need more precision u cane use a simple VGA for gain control and the main limitation is given by the offset of the error ampliefier.

 

• Dec 22, 2004

• #11

C

ccw27

Full Member level 5

Joined

Oct 13, 2004

Messages

267

Helped

14

Reputation

28

Reaction score

6

Trophy points

1,298

Activity points

2,558

common mode variation fully differential

Actually the VCVS have gain of 0.5 each so voc=(vop+von)/2, forgot to point that out.

Also if I have a two stage amplifier, folded cascode followed by common source. Is it ok to put the CMFB to track the output of the folded cascode instead of the output of the common source? From my simulations I needed a much higher gain amplifier if the CMFB input is taken at the output of the common source and its output drives the PMOS cascode mirror in the folded cascode.

Thanks

 

• Dec 23, 2004

• #12

C

cristianb

Member level 1

Joined

Sep 9, 2003

Messages

33

Helped

3

Reputation

6

Reaction score

1

Trophy points

1,288

Activity points

287

differential op amp cmfb gain

This depends on the aplication of your amp. If is just stand alone there is no problem. if is part of a bigger block you'll have to take into consideration the behavior of the follwing stage. If this stage does not tolerate common mode variations (for example a high order adc) then you to take the common mode signal at the output of the common source stage, otherwise if u use the signal from the output of the folded cascode its ok. in the first case u should take in consideration that higher gain could splell in stability problems of the common mode loop and compensaiting this is tricky.

 

• Mar 16, 2006

• #13

wjxcom

Full Member level 5

Joined

Sep 7, 2005

Messages

281

Helped

4

Reputation

8

Reaction score

3

Trophy points

1,298

Activity points

3,840

schema ideale cmfb

Dear ccw27：do you know how to simulate a full differential OPamp? for example we want to simulate ac character, PSRR character and so on, what circuit can we use?

 

• Aug 31, 2006

• #14

G

gte582w

Member level 5

Joined

Aug 31, 2006

Messages

94

Helped

7

Reputation

14

Reaction score

1

Trophy points

1,286

Activity points

1,700

fully differential folded cascode amplifier

I personally prefer to use RC to sense the CM signal if the swing is large.
The down side is it takes quite a bit area. Source degenerated xtor in triode
is good but has limited sensing range.

 

• May 20, 2007

• #15

ahmad_abdulghany

Advanced Member level 4

Joined

Apr 12, 2005

Messages

1,206

Helped

102

Reputation

206

Reaction score

22

Trophy points

1,318

Location

San Jose, California, USA

Activity points

11,769

fully differential folded cascode opamp

wjxcom said:

Dear ccw27：do you know how to simulate a full differential OPamp? for example we want to simulate ac character, PSRR character and so on, what circuit can we use?

Click to expand...

A simple OCEAN script can do it easily.

 

• Oct 14, 2010

• #16

R

raditsme

Junior Member level 1

Joined

Aug 26, 2010

Messages

18

Helped

0

Reputation

0

Reaction score

0

Trophy points

1,281

Location

india

Activity points

1,413

hello everyone... i have to design a telescopic amplifier on 0.18um Spectre. my ugb requirement is > 300MHz and gain > 80db .Although power diss. limit is given i decided my current be 320uA to be ok. I calculated all W,Ls according to the basic equation for SINGLE STAGE TELESCOPIC AMPLIFIER topology as given in Razavi on Oage no.299-300,, of course giving enough margins of Vod for each transistor. But when i simulated my design on cadence tool it is becoming impossible for me to bias all transistor in saturation for given current.i am considering only single stage in initial exercise. I have spent a lot of time on the the exercise but i got no sucees .....please help and suggest some idea.