Time distortion of a Poisson process and its effect on experimental uncertainty

By means of computer simulations of spectrometry counting, the effect of pulse pileup and system dead-time on the experimental uncertainty is investigated and compared with theoretical expectations. Different counting statistics are observed depending on the type of count loss, i.e. due to pulse pileup, extending or non-extending system dead-time or a combination of these mechanisms. Simulations were also applied to test the ability of “loss-free counting” to deal with the different event loss mechanisms, and to assess the corresponding counting uncertainty. Some findings have been tested experimentally by comparison with the statistical behaviour observed in real γ-spectrometry at high incoming pulse rate, including the counting uncertainties in loss corrected spectra (LFC with virtual pulser).