NMX detector algorithms

A Gas Electron Multiplier (GEM) detector with a Gadolinium converter is being developed for the Neutron Macromolecular Diffractometer (NMX) at the European Spallation Source (ESS). Charge data from the orthogonal copper strips of the GEM detector will be collected by a number of VMM3 chips which each delivers a stream of data triplets from 64 copper strips. Each triplet consists of strip number, a charge maximum and the time when the maximum was recorded. This data has to be clustered in real time into X- and Y-cluster pairs each corresponding to a neutron event such that the entrance point location and time for the neutron event can be determined. This paper describes algorithms and their implementation for performing this task. The algorithms include an innovative streaming sorting, a novel clustering algorithm and a pairing algorithm for constructing the neutron event data. The implementation and algorithms have been tested extensively and yield a satisfactory performance without any optimization so far. The processing time is close to being constant per data triplet.