Convergence of Least-Squares Multi-Harmonic Fitting Algorithm

Digital signal processing algorithms are becoming increasingly attractive for instrumentation systems. Their robustness combined with increased processing capabilities of DSPs will enable the development of a new brand of low cost flexible instrumentation. These instruments will implement different algorithms for each measurement need. Sine-fitting algorithms estimate the sinewave parameters of a digitized signal. They can be used to characterize analog to digital converters, measure impedances and test the frequency response of linear systems. In this paper, the convergence of the least-squares multi-harmonic fitting algorithm is analyzed. Simple improvements are presented to increase convergence and minimize the number of iterations