ABSTRACT
In nuclear applications related to radiation energy measurement, digital pulse processing for fast sampling facility has great advantages compared to the traditional analog pulse processing. A distortion in energy spectrum and degradation in timing resolution take place when pulse pile-up occurs. This in turn leads to degradations of the signals used in applications such as nuclear medicine. Pile-up pulses are commonly removed in most algorithms if they are detected, but in this case, getting a clean spectrum becomes a challenge due to data loss. A proposed pile-up detection and recovery algorithm based on fitting the preamplifier data and signal extrapolation is presented. The extrapolated pulse is used for recovering (extracting) the piled-up pulse to estimate its parameters. A smoothed first derivative algorithm is used for peak tracking inside the detected events. To evaluate the proposed pile-up recovery algorithm theoretically and experimentally, it is applied on both simulated input signals generated with different time displacements between the piled-up events, and experimental preamplifier radiation signals. The obtained results prove that the proposed algorithm achieves better utilization of power and can be used for high count rate (high activity elements) spectroscopy systems.
