A simplified theoretical model is developed to predict counting statistics for a stationary Poisson process passing through a spectrometer with pulse-pileup rejection. The model is applicable to digital counters used for spectrometry as well as set-ups utilising analogue electronics for pulse shaping and pileup rejection. In comparison with an existing model for a perfect pileup rejector, the new model addresses the common imperfection of having a finite time resolution of the fast channel, allowing quasi-coincident signals to pass through the pile-up rejector. From the Laplace transform of a simplified interval-density distribution, approximate expressions are derived for the throughput factor and the variance of the number of counted events. The results are compared with computer simulations of a cascade of extending dead time and subsequent pileup rejection. In addition, a rigorous throughput factor is derived from probabilistic reasoning, as well as an effective throughput factor for singular and coincident events.
Read full abstract