Advanced algorithms for signal processing scintillation gamma ray detectors at high counting rates

Abstract Gamma ray spectrometry measurements at high detector counting rates (on the order of 107 s−1 and higher) are relevant for high-temperature plasma diagnostics for existing tokamaks and during the development of gamma ray diagnostic systems for the ITER tokamak under construction. At high loads, to obtain spectra without distortion and with a small amount of dead time, it is necessary to use advanced scintillation detector signal processing methods, which can resolve superimposed pulses. Two algorithms that can be used for digital signal processing of scintillation gamma ray detectors with many piled-up pulses are considered in this article, the fitting and deconvolution methods. These algorithms are compared with both one another and two less sophisticated pulse-height analysis algorithms (maximum height and total sum under the peak) in applications that process model and measured signals. Baseline detection algorithms are also considered, which are necessary when processing signals from detectors. The algorithms are applied to LaBr3(Ce) detector signal processing. For the modeled signals, the best results in terms of the number of resolved events and energy resolution at counting rates up to 2 x 107 s−1 of the LaBr3(Ce) detector are demonstrated by the fitting method. In the real gamma ray measurements at a loading of 5.1 106 s−1, the deconvolution method demonstrated the best energy resolution.