SiPM signal processing via multiple linear regression

Abstract This paper presents a novel approach using multiple linear regression to process transient signals from silicon photomultipliers. The method provides excellent noise suppression and pulse detection in scenarios with a high pulse count rate and superimposed pulses. Insights into its implementation and benchmark results are presented. We also show how this approach can be used to automatically detect the pulse shape from a given transient signal, providing good detection for count rates up to 90 MHz. Experimental data are used to present an application where this algorithm improves charge spectrum resolution by an order of magnitude.