Abstract
While processing the signals from radiation detectors, for finding the true mean-count-rate, algorithms with hybrid pulse collection methodology have been proposed and used over the years. An algorithm based on this technique with a new methodology of adoption and implementation including spurious rejection is proposed here. It enables a specified and controllable error when the mean-count-rate remains within certain predefined limits from its true value. Effort is made to optimize the response time of prediction at low count rates preserving the optimum possible relative-standard-deviation (RSD). Chi-squared test is utilized for verifying the counting system to check if the observed fluctuations are consistent with the expected statistical fluctuations. A C-program code has been developed to test the algorithm. An observed set of detector outputs are given as input to the program and the corresponding Output is analyzed. A comparative study between the proposed method and floating-mean method is presented for the same set of observations. A typical short-lived high voltage (HV) induced spurious noise pattern is fed as input to the program verifying limited-spurious rejection capability of the algorithm. An embedded C program was written for microcontroller implementation of the algorithm. Case-study of a neutron roentgen equivalent man (REM) counter is presented for evaluating response time for various ranges of operation with calculation of RSD at these ranges. This general-purpose algorithm can enhance the read-out accuracy of radiation monitors used for radiation safety applications.
