THE DECONVOLUTION METHOD OF FAST PULSE SHAPING AT HADRON COLLIDERS

We describe a new technique for front-end signal processing of signals from LHC or SSC detectors which precisely defines the origin of an event in time while maintaining amplitude measurement with an excellent signal to noise ratio. The method is designed for use with silicon detectors whose leakage currents may be substantially increased during the lifetime of an experiment by radiation damage, although it is likely to be applicable to other types of detector. It makes use of a shaping amplifier with a time constant of several beam crossing intervals and is particularly well matched to CMOS front ends where low power consumption and low noise is best achieved by utilising pulse shapes with time constants approximately 50 ns. It is based on discrete time filtering of data extracted from an analogue pipeline after a first level trigger. A finite impulse response type filter deconvolutes the sampled voltages of a shaped pulse to retrieve the original impulse signal with high precision. We describe the mathematical basis of the technique and its implications for timing and signal to noise. Measurements have been made on a CMOS amplifier intended as a prototype for readout of silicon microstrip detectors at LHC which demonstrates the power of this approach. A CMOS circuit emulating the filter is being built. It has been implemented with extremely low power consumption ( < 0.1 mW/channel) in a small silicon area to be utilised on every channel of the system.