Why are Ultra-low Noise Pre-Amplifiers so key?

My experience teaching and working with PhD students has shown me that many measurements are very far from understood and almost never optimized.  Many students will simply connect their samples to the front-end of a sampling voltmeter, or a standard laboratory lock-in amplifier, and have no idea what determines the noise, resolution, or measurement bandwidth.  Without this understanding, they cannot possibly know if the measurement can be improved.

For instance, take the Stanford Research model SR830 lock-in amplifier.  The noise on the input is specified to be 6nV/rtHz.  Using a Sierra Amps pre-amplifiers with noise levels of 270pV/rtHz, the measurement resolution can be improved by a factor of  22 times, or with the speed-up by a factor 500 times.  The simple use of a ultra-low noise pre-amplifier can make a signal which could not be observed, observable, or a measurement which took 24 hours, take only 3 minutes (!).  That is the profound effect an optimized measurement can have.

Another example is the Signal Recovery 5113 voltage pre-amplifier.  The input voltage noise of this amplifier is specified to be 4nV/rtHz and an optimal source impedance of 100k ohms. The noise on Sierra Amps SA1 is 400pV/rtHz (with the same source impedance) and can lead to either x10 higher resolution or 100x faster measurement.

Noise from the pre-amp is a big deal and can determine success or failure of a measurement.