Cross-spectrum Analyzer for Low Frequency Noise Analysis

Abstract

The design and performance of a sensitive and reliable cross-correlation spectrum analyzer for studying low frequency transport noise is described in detail. The design makes use of common PC-based data acquisition hardware and preamplifiers to acquire time-based data, along with software we have developed to compute the cross-correlation and noise spectral density. The impedance of device under test may cover four decades from ${100\,{\Omega}}$ to ${1\,{\mathrm{M}\Omega}}$. By utilizing a custom developed signal processing program, this system is tested to be accurate and efficient for measuring voltage noise as low as $\sim\!10^{-19}\,\mathrm{V}^2/\mathrm{Hz}$ from ${0.001\,}$Hz to ${100\,}$kHz within one day's averaging time, comparable with more expensive hardware solutions (bandwidth in real measurements may be limited by the sample impedance and stray capacitance). The time dependence of measurement sensitivity is discussed theoretically and characterized experimentally to optimize between measuring time and accuracy. A routine for noise component analysis is introduced, and is applied for characterizing the noise spectra of metal and carbon film resistors, revealing an almost strict $1/$frequency dependence that may reflect an ensemble of random resistivity fluctuation processes with uniformly distributed activation energies. These results verify the general applicability of this analyzer for low level noise researches.