A technique is described whereby an estimate of the false coincidence signal, suitable for background correction of data acquired in a coincidence experiment, is obtained by using a pseudorandom pulser to generate a stream of “false” start events. The statistical properties of this simulated source are adjusted to mimic those of the real source of electron start events. False ion coincidences with the simulated starts are measured concurrently with the real coincidence signal, with the mean count rate of the pseudorandom pulse source automatically tracking that of the true electron start events. In this manner any long term instrumental drifts during the course of an extended experimental measurement will similarly affect both the real and simulated coincidence data. Subtraction of the simulated background of false coincidences from the real coincidence data then yields an improved estimate of the true coincidence signal. ©1998 American Institute of Physics.