Search

An adaptive technique to restore scintigraphic images is presented. A few examples are discussed to show its performance when the procedure is implemented on a minicomputer with limited core memory.

Experimental outcomes are often described by a convolution integral whose kernel is the impulse response of the instrument. The program DEO, Deconvolution of Experimental Outcomes, corrects the effects of the instrument finite resolution on experimental data. Basic ideas and performances of DEO are here discussed.

Automatic peak locating techniques for γ-ray spectra are compared with reference to the probabilities of detecting fraudulent peaks and missing true peaks.

We present the content-addressable-memory feature of a new system designed in our laboratories for non-destructive testing of nuclear reactor pressure vessels based on acoustic emission analysis. The content addressable memory is divided into two parts: the first selects the most frequent events among incoming ones (FES: Frequent Event Selection memory), the second stores the frequent events singled out (FEM: Frequent Event Memory). The statistical behaviour of FES is analyzed, and experimental results are compared with theoretical ones; the model presented proved to be a useful tool in dimensioning the instrument store capacity.

A technique for the estimation of first and second order properties of a stationary counting process is presented here which uses standard instruments for analysis of a continuous stationary random signal.

A method has been introduced for the evaluation of mean and variances of estimators of autocorrelation parameters for Poisson doubly stochastic point processes. The behaviour of a particular digital instrument developed at Brookhaven National Laboratory for nuclear safeguard purposes has been analyzed by the suggested method.

Starting from the theory of linear time invariant filters for optimum processing of the amplitude of single pulses, an ideal time variant system has been synthetized for the analysis of a sequence of random pulses and its performance has been evaluated. Demonstration is given that the system is the optimum one under the hypothesis of gaussian noise and taking as criterion for the evaluation of amplitude resolution the rms deviation.

A very fast pulse amplitude-to-digital converter has been realized connecting a sampling height-to-time converter to a digital time sorter. Very good differential linearity and encoding accuracy have been obtained with an equivalent cloc frequency of some GHz.

In this paper, the pile-up amplitude distribution of pulses having the most common shape is studied (single and double RC differentation, single and double delay line differentiation). The pile-up effects on timing circuits are then analyzed, and experimental results are discussed. Finally the effeciency of pile-up rejection circuits is examined in order to design the most suitable ones.

The problem of achieving the highest resolution in the amplitude measurement of pulses randomly spaced in time is considered, taking into account both noise and pile-up fluctuations due to slow signal tails. Signals composed by a pulse of finite duration and a long tail, are considered: a measuring technique is introduced, based on a linear combination of sampling measurements on the pulse. It is shown that, by a suitable choice of weights, optimum compromise between noise and tail pile-up effects can be reached, allowing good resolution at high counting rates. Results of calculations are given for the well known finite width cusp pulse and exponential tails.

Equations for the effects of the quantization on the mean value and the variance of a random variable are presented for some usual channel profiles in nuclear instrumentation.

A method is presented to smooth a noisy waveform by introducing a correlation between adjacent channels dependent of the channel number.

A new procedure for evoked response evaluation is presented which uses an estimation of the original response power spectrum to compute an approximation of the optimum linear filter.