Resolution of silicon drift detectors with sampled waveform method

Summary form only given. The processing of the information from thousand of channels, as required in modern high-energy physics experiments, especially using semiconductor detectors, leads to the use of dedicated integrated electronics, signal sampling, and their analog or digital processing. The authors discuss the instrumentation used for the sampling of the signal pulses and the noise. The optimum processing of the sampled data, based on the maximum likelihood method, is developed for the signals known a priori in shape but not in time and amplitude. It is shown that undersampling does not introduce systematic errors but only an increasing loss of resolution, slightly sensitive on the pulse position with respect to the sampling comb. This makes it possible to choose the best compromise between the achievable resolution and computational requirements.<>