Statistics of a two-parameter analyzer with associative memory

Abstract

This paper deals with pulse-height analyzers with associative memories of M words, where M is very much smaller than the total number of possible different channels. We show that if the first M channels to record one event each are stored, the probability that any given channel is included in the store is simply proportional to its counting rate. The counts per channel in a region of the spectrum are therefore indicated not only by the contents of the recorded channels, but also by the density of recorded channels. In general, if the first M channels to record n events each are stored, the inclusion probability is proportional to the channel counting rate raised to the n^th power.