The Designer's Guide Community Forum

The Designer's Guide Community Forum
https://designers-guide.org/forum/YaBB.pl
Design >> Analog Design >> Does CMFB loop bandwidth has impact on signal path bandwidth?
https://designers-guide.org/forum/YaBB.pl?num=1232441332

Message started by jlee on Jan 20^th, 2009, 12:48am

Title: Does CMFB loop bandwidth has impact on signal path bandwidth?
Post by jlee on Jan 20^th, 2009, 12:48am

Hi guys,

I have a question as the subject. A fully differential folded opamp is used with negative feedback to amplify the input signal with bandwidth of 50M. A CMFB is needed for this opamp. My question is how to confirm the bandwidth for this CMFB? Does this CMFB loop bandwidth has impact on signal path bandwidth?

Regards

Title: Re: Does CMFB loop bandwidth has impact on signal path bandwidth?
Post by Berti on Jan 20^th, 2009, 3:25am

Whether the CMFB has an impact on the differential-mode bandwidth depends on its implementation. In a SC-CMFB for instance, high common-mode bandwidth comes at the price of reduced differential-mode bandwidth. There is a trade-off. However, you should give more details.

Cheers

Title: Re: Does CMFB loop bandwidth has impact on signal path bandwidth?
Post by jlee on Jan 20^th, 2009, 4:52am

Thanks for your input.

The CMFB I used is sensing common mode voltage by two large resistors parelleled with fully differential output nodes and an opamp is used to force the sensed common mode voltage to a referced voltage. This opamp output is connected to the gate of tail current source of fully differential opamp.

Regards

Title: Re: Does CMFB loop bandwidth has impact on signal path bandwidth?
Post by thechopper on Jan 20^th, 2009, 6:22pm

It looks to me you are using a high swing CMFB circuit. IF that is the case actually is the differential signal path the one mainly setting the CMFB BW.
In my statement I'm assuming your CMFB signal is fed back at some intermediate node in your differential path beore the differential compensation stage (for example a Miller stage)

Regards
Tosei

Title: Re: Does CMFB loop bandwidth has impact on signal path bandwidth?
Post by jlee on Jan 21^st, 2009, 12:02am

Hi Tosei, thanks for your comment.

Yes, the CMFB feedback path is really the same as your statement.

However, I cannot understand why "the differential signal path is the one mainly setting the CMFB BW". ::)

Regards

Title: Re: Does CMFB loop bandwidth has impact on signal path bandwidth?
Post by vivkr on Jan 21^st, 2009, 1:37am

Berti wrote on Jan 20^th, 2009, 3:25am:

Hi Berti,

What do you mean when you say that high common-mode bandwidth comes at the price of reduced differential-mode bandwidth with an SC-CMFB? Could you elaborate please?

Regards,

Vivek

Title: Re: Does CMFB loop bandwidth has impact on signal path bandwidth?
Post by thechopper on Jan 21^st, 2009, 6:38pm

jlee wrote on Jan 21^st, 2009, 12:02am:

Hi Tosei, thanks for your comment.

However, I cannot understand why "the differential signal path is the one mainly setting the CMFB BW". ::)

As I tried to describe before, I'm assuming your CMFB signal is fed back to some intermediate point in your differential signal path, for example somewhere in the first stage of a two-stage Miller compensated opamp. Therefore such CM fed-back signal will have to go through the differential compensation stages, therefore it will be limited by the differential path speed.

Hope my explanation is clearer now.
Regards
Tosei

Title: Re: Does CMFB loop bandwidth has impact on signal path bandwidth?
Post by Berti on Jan 21^st, 2009, 10:10pm

Quote:

What do you mean when you say that high common-mode bandwidth comes at the price of reduced differential-mode bandwidth with an SC-CMFB? Could you elaborate please?

Vivek,

The bandwidth of a SC-CMFB depends on the size of the two capacitors connected between the outputs and the controlling node. Larger capacitors will give larger common-mode bandwidth but increases the loading of the amplifier, therefore decreasing the differential-mode bandwidth.

The following paper is worth reading:

Ojas Choksi and L. Richard Carley , Analysis of Switched-Capacitor Common-Mode Feedback Circuit ; IEEE Transactions on Circuits and Systems II ; 2003 .

Regards

Title: Re: Does CMFB loop bandwidth has impact on signal path bandwidth?
Post by vivkr on Jan 21^st, 2009, 11:31pm

Berti wrote on Jan 21^st, 2009, 10:10pm:

The bandwidth of a SC-CMFB depends on the size of the two capacitors connected between the outputs and the controlling node. Larger capacitors will give larger common-mode bandwidth but increases the loading of the amplifier, therefore decreasing the differential-mode bandwidth.

The following paper is worth reading:

Ojas Choksi and L. Richard Carley , Analysis of Switched-Capacitor Common-Mode Feedback Circuit ; IEEE Transactions on Circuits and Systems II ; 2003 .

Regards

Hi Berti,

Thanks for the link. I will check it out. But I think you are speaking not about the size of the capacitors used in the SC-CMFB but the ratio of the sampling to the integrating caps (it is an integrator). Am I right? So, if the smaller of the 2 caps is constrained to a certain size, then increasing CMFB loop bandwidth implies increasing the loop feedback factor (larger cap).

Regards,

Vivek

Title: Re: Does CMFB loop bandwidth has impact on signal path bandwidth?
Post by Berti on Jan 22^nd, 2009, 7:44am

Actually no. I mean the capacitor of the SC-CMFB which is connect to the output of the amplifier.

Cheers

Title: Re: Does CMFB loop bandwidth has impact on signal path bandwidth?
Post by vivkr on Jan 22^nd, 2009, 11:34pm

Berti wrote on Jan 22^nd, 2009, 7:44am:

Actually no. I mean the capacitor of the SC-CMFB which is connect to the output of the amplifier.

Cheers

Yes! I mean the same one, the one which is connected permanently. That is the feedback cap for the integrator formed in the CMFB loop. If i make that one larger, I increase the feedback factor for that integrator (=> larger loop BW), but at the cost of larger load. Maybe I will try to get hold of the paper and read it first.

Thanks for the link.

Regards,

Vivek