Multiplicity counters: utilisation test of an exponential modelization on a detector chain for which this hypothesis is little realistic

Shift register method is very used for Pu mass determination using neutrons counting. It is based on the fact: when counting neutrons in randomly triggered gates and neutrons triggered gates, counting difference which is the real coincidences number of order 2 has no sensitivity to the presence of any (α, n) parasite neutron source. The SCRD device produces additional informations, called as 'multiplicity results': these informations are coincidences number of order 1, 2, 7, > 7 in both randomly and neutrons triggered gates. It is expected that multiplicity results can be used for recognizing the isotopic composition of the spontaneously fissile sample. We describe a measurement in which we attempted to recognize a Cf 252 sample with presence of a strengther Am.B sample. Results are wrong. Apparent reasons for misfits are: - Use of an exponential model: the actual response function of the used detection head is not an exponential one. - Other effects neglected in the model: for example SCRD has a 500 ns dead-time corrected with a FIFO device. - Accuracy of used values for emitting probabilities of Cf 252. (authors). 2 figs., 2 tabs., 2 refs