Using the Allan variance and power spectral density to characterize DC nanovoltmeters

When analyzing nanovoltmeter measurements, stochastic serial correlations a re often ignored and the experimental standard deviation of the mean is ass umed to be the experimental standard deviation of a single observation divi ded by the square root of the number of observations, This is justified onl y for white noise. This paper demonstrates the use of the power spectrum an d the Allan variance to analyze data, identify the regimes of white noise, and characterize the performance of digital and analog de nanovoltmeters. L imits imposed by temperature variations, 1/f noise and source resistance ar e investigated.