Design of Self-Diagnosis System for Nuclear Instrumentation

The use of non-randomness measurements for self-diagnosis of a pulse-counting system is described. Characteristic patterns of interval distributions for nuclear pulse trains have been measured under various types of malfunctions. They can be fundamental data in designing a stochastic indicator for malfunctions. Discussions on possible applications of interval distributions are made. An example of diagnosis system is shown which indicates the presence of malfunction by data processing of time interval distributions.