Common mode feedback in comparators design

[yvu10545]

Why do comparators sometimes need common mode feedback and sometimes not?

I have not seen any good explanation.

 

[Tahmid]

The output common mode (CM) is ill-defined for differential signals. Adding common-mode feedback allows you to have a well-defined CM level. You've probably seen them in comparators with differential outputs and haven't seen them in single-ended output comparators.

 

[dick_freebird]

Differential outputs like LVDS can get by without CMFB and still meet the offset specs.

 

[D.A.(Tony)Stewart]

Why do comparators sometimes need common mode feedback and sometimes not?

I have not seen any good explanation.

Where have you heard or seen this? CM feedback is only used on analog designs like INAs using RLD feedback.
Comparators have a much wider CM range than INAs so as long as it is within a wide DC range all you need is a clean differential signal.
To have CM feedback you must have a CM source which comparators do not have.

Whereas, INAs ( differential INstrument Amps) with multiple sensors for EEG or ECG can extract a CM signal two diff. sources & invert and drive a CM source to negate much of the induced grid e-field that passes through the body to improve low-level signal/noise ratio.

Is it possible you misread a schematic? If not pls show a presumptive example.

Other methods of reducing CM noise are : Ferrite Balun, balanced shielded twisted pair cable (STP), AC or DC grounding one source. "ground = 0 V"


Cheers,
Tony

(greetings Tahmid)

 

[nitishn5]

Is the CMFB in the Pre-Amplifier stage of the Comparator, which could be a linear(ish) stage?
A pre-amplifier output which has a properly defined output common mode could be used to drive the final latch/comparator. This could be helpful in reducing the delays since the differential output of the preamplifier will be at 0Vdiff and known Vcm and on application of the inputs, it does not have to swing wildy.

This is my guess. A circuit diagram would be great...!