Can anybody explain me what is the need of using differential amplifier in signal conditioning circuit?

 

Mainly to eliminate errors due to the ground of the signal source not being at absolutely the same voltage as the ground of the signal processor, and also to remove any common-mode interference picked up in the cable to the signal processor.

 

Ian is correct! and my more detailed explanation includes an additional possibility that the source of the "signal" to be conditioned may not be referenced to any particular "ground" at all. That could include a strain gage bridge in a pressure or force sensor, or a magnetic field sensor coil, as common examples. Those are totally IN ADDITION to the explanations from Ian0, rather than alternative.

 

If someone asks you to connect a signal using a single-ended input, your response should be "Do I have to?"
Although, for very small signals (in the microvolt region) a single-ended input is quieter by a factor of √2, but you should consider if the increase in interference would outweigh the extra noise of the differential input circuitry.

 

This could lead to a very extended, and possibly useful, discussion about signal inputs.
In reality, all inputs are differential, but not all of them are balanced and symetrical.
By "BALANCED", I mean as referenced to a common point, often called "ground", both connections having a similar impedance to that common, and often, opposite phase polarities.

 

I have always used the term "balanced" to indicate that both parts of the signal had the same source impedance, and "differential" (as in "differential output") to mean that there were two signals of opposite polarity.
Much of professional audio consists of a pair of wires with one wire on which there is a signal and another wire which is connected to ground via an impedance chosen to match the impedance of the driven wire; and this is connected to a fully balanced differential input at the receiving end.
Its performance is as good as having two driven wires of opposite polarity. I would refer to this as a "balanced but not differential output".
. . . "When I use a word, it means just what I choose it to mean – neither more nor less " Lewis Carroll.

 

ALL of the "professional " microphone audio stuff I have used has two conductors plus a shield from the XLR connector at the microphone to the XLR connector at the other end. NO GROUND REFERENCE for the audio signal, direct or otherwise. When a non-balanced microphone must be used, one of the pair does get tied to the microphone "ground", and sometimes also to the shield.

 

ALL of the "professional " microphone audio stuff I have used has two conductors plus a shield from the XLR connector at the microphone to the XLR connector at the other end. NO GROUND REFERENCE for the audio signal, direct or otherwise. When a non-balanced microphone must be used, one of the pair does get tied to the microphone "ground", and sometimes also to the shield.

The screened-twisted-pair balanced-impedance but with signal only on one wire is used mainly at line level. Microphones are connect with the case grounded but neither wire from the capsule connected to the case, as you correctly state.

 

The screened-twisted-pair balanced-impedance but with signal only on one wire is used mainly at line level. Microphones are connect with the case grounded but neither wire from the capsule connected to the case, as you correctly state.

The shield would be used as the power return conductor for "phantom power" applications. But power is a separate and much different topic.