Counting statistics for nonnegligible dead time corrections
Robert L. Lucke · Review of Scientific Instruments · 1976
If a detecting system records N counts in T seconds, with each recorded count followed by a dead time τ, the statistical variation in N is ±N1/2/(1+Rτ) (not just ±N1/2), where R is the true rate of occurrence of events. The value of R inferred from this experiment is subject to a statistical error of ΔR/R=±N1/2. This equation is accurate as long as N1/2/Rτ≫1.