AN EXPERIMENTAL STUDY OF COUNTING LOSSES DUE TO DEAD-TIME EFFECTS

Abstract

When using counting methods for studying instantaneous reactions produced with pulsed accelerators such as synchrotrons or synchrocyclotrons, the systematic errors in counting due to dead-time effects may be considerable. One of us has already published a discussion of the simpler cases of pulsed single-channel counting; the present paper describes experimental investigations designed to extend our knowledge of these corrections, especially to cases of coincidence counting. In Part I of the present paper, the method and apparatus are described, together with calibration procedure; in the course of checking results already known from computation, defects in the functioning of the equipment were uncovered and the correction of these defects is described.Part II of the paper deals with the problems of coincidence counting at high speed but without the complication of a pulsed source. Difficulties arise in this case unless the rate of occurrence of coincidences is much smaller than the counting rates in the separate counters, a condition which may not be satisfied in many cases in the study of high energy reactions. The corrections for this case have therefore been investigated experimentally and a relatively simple correction formula obtained when chance coincidences are negligible. The dependence of the number of additional chance coincidences on coincidence gate-width is determined both theoretically and experimentally, with satisfactory agreement.In Part III the investigation is extended to pulsed coincidence counting. A theoretical formula is developed neglecting chance coincidences, for pulse-lengths less than twice the dead-time, and experiments confirm the results and indicate how to extrapolate for longer pulse-durations. A first-order formula for the chance coincidences is also obtained and confirmed.In all cases near-rectangular pulse-forms are used, but the equipment can readily be modified to study other forms. For the guidance of the user, the principal formulae adopted for determining counting corrections are collected in Appendix III at the end of this paper.