Study On The Dependence Of Digital Filter Performances On Sampling Parameters For Random Pulse Measurements

Recently, radiation detection based on the digital signal processing has been widely applied in the fields as high energy physics, radiation imaging and so on. Compared with analog circuits, the digital filters have the advantages as flexibility stability and adaptivity, although its performance is limited by the Analog-to-Digital Converter (ADC). In random pulse measurement of radiation detection, there has been no systemic study in theory on the ADC sampling induced error analysis. The study on effect of ADC parameters to digital filtering performance is necessary for digital filter design. For example the power consumption becomes critical issue for signal processing circuit when the number of readout channels increases significantly. In such case, the quantitative study of the dependence of the filter performance on ADC parameters is necessary for reducing ADC power consumption.This thesis analyzes the dependence of digital filtering performance on ADC sampling frequency in random pulse measurement based on commonly applied digital signal processing circuit. The thesis includes the following parts. 1) Derive the signal spectrum and the noise power spectral density after sampling, analyze quantitatively the amplitude measurement performance on the sampling frequency, and then obtain the optimum signal-to-noise ratio under certain sampling frequency as well as noise index for several commonly applied digital filters. 2) Analyze the random arriving time induced measurement errors on sampling frequency, including the expected amplitude shift and fluctuation, which decrease significantly with increasing sampling frequency. When the sampling interval is less than 1/3 to 1/5 of the signal rising time, the measurement errors can be ignored. The timing estimation algorithm is proposed to amplitude measurement correction and analyzed preliminary.The dependence of amplitude measurement fluctuation on the sampling frequency is demonstrated by the numerical simulation and measurements with digital oscilloscope sampling. The dependence of experimental results on the sampling frequency is consistent with the theoretical results, which include the noise of ORTEC 142AH charge sensitive preamplifier and the ²⁴¹Am energy spectrum of CdTe detector. The theoretical analysis is demonstrated.