AbstractsEngineering

This thesis discusses the design and evaluation of an instrument that can estimate the probability density distributions of bandlimited, aperiodic electrical phenomena. This estimator can obtain the probability density estimates of signals that are contained within the frequency range of 100 cps. to 20 kcs. However, the basic design logic of the instrument can serve as a design foundation for a device that should be capable of estimating the distributions of signals that range in frequency from a fraction of a cycle to 50 kcs. Transistors and solid-state diodes are used in the instrument's circuitry. This device, however, is not a transistorized version of another instrument. Area error, form factor, and end-point error evaluation techniques are discussed, modified and then combined to develop a procedure which will determine an instrument's frequency-dependency characteristics. The author also introduces an evaluation technique (the pseudo impulse test) that involves the use of a test sgna1 that has a theoretical probability density distribution which contains a series of impulse functions. The pseudo impulse test is a powerful technique whir.h will reveal the operating capahllitie9 of an instrument. All the evaluation techniques discussed here may be used to evaluate any device that is designed to estimate probability density distributions. The appendix contains information that may be helpful to those who would like to duplicate or improve the instrument.