Something puzzles me. When I shorted the 15 unused inputs of my 16 channel bio-amplifier with a signal generator connected to channel 1. The output not only is cleaner but increased in amplitude.

Do you have any theory why shorting the unused inputs and connecting all the grounds and references in the software make the amplitude bigger in the following? is there a principle that makes noises produce smaller amplitude or I wonder if the DSP just amplify the gain when the signal it detected is cleaner?

g USBamp Demo - YouTube

 

Leaving unused Inputs "Floating" will definitely cause Oscillations,
or, the amplification of external Noise.
This will be outlined in the Spec-Sheet,
and applies to ALL High-Impedance-Inputs on ALL Chips.

These Oscillations or Noise massively increase the Internal-Load drawn by the Chip,
and could actually cause it to overheat and fail under certain circumstances.
.
.
.

 

Leaving unused Inputs "Floating" will definitely cause Oscillations,
or, the amplification of external Noise.
This will be outlined in the Spec-Sheet,
and applies to ALL High-Impedance-Inputs on ALL Chips.

These Oscillations or Noise massively increase the Internal-Load drawn by the Chip,
and could actually cause it to overheat and fail under certain circumstances.
.
.
.

But why does the amplitude increase and become cleaner when the ground and references (V-) of the 4 banks were connected as shown in the video see below too. Without the ground and references of the 4 banks connected, the effect is like the unused inputs not shorted and signal smaller and dirty.

 

Noise. with lights on and off, within a house, near the work bench, surrounding your equipment all should be logged
and recorded, The copper wiring can be a source of harmonics that can constructively increase your reading.
once that is known the frequency range with all the cables connected should be as low as possible.
The negative impulses or harmonics are known to travel back on a cable sometimes canceling the positive ones.

 

A differential amplifier amplifies the difference between Reference (negative input) and positive input. If Reference is disconnected or floating, can it not only pick up of noise, but also cause a diminished signal? Can anyone give an example of diminished return if Reference (negative input) is floating?

 

The short answer to the original question is that in such a circuit all of the noise takes a bit of power to amplify, and that tends to reduce the gain available for the intended signal. This may also be compared to saturation of some intermediate stage resulting in gain compression.
So when all of the noise is removed there is a bit more gain for the intended signal. The complete detailed explanation of the exact mechanism is rather long and tedious.

 

Wait. Im using channel 1 of Group A below. Are you supposed to connect the Reference (In -) of Group A to the Reference (In-) of the other 3 banks where I shorted all Inputs (V+), Reference (V-) and Ground together? Isnt the effect like shorting the V- and ground of channel 1 while using channel 1 in Group A since the ground of all 4 Groups are also shorted??