Return to search

Digital Moments Analyzer: Design and Error Characteristics

A portable special purpose computer (s.p.c.) is described which provides decimal readouts of the first four moments of a fluctuating voltage v on four separate registers. A fifth register provides a readout of the measuring time which can be within the range 10 ms. to
30 Hrs. The s.p.c. can be switched to another mode which provides a measure of the cumulative amplitude distribution of v within sixteen positive and negative levels. Salient characteristics of the s.p.c. are as follows:
(a) There are no low frequency limitations. The upper frequency limit, established by error considerations, is about 5 kHz with 99.73% confidence that the error is within 1%.
(b) At the end of the measuring time T, all the four moments are immediately available in magnitude and sign.
(c) The outputs can be available in any code, the only change necessary being in the code of the counting readout registers.
(d) All computations for a sample are completed before the next sample arrives so that programming and unnecessary storage facilities are eliminated. The voltage input v is rectified and sampled systematically by an equi-interval a.d. converter. The samples, together with the sign bit, are fed into special purpose digital multipliers based on a
"weighted feed" principle. The outputs from these multipliers, with the sign bit, arc fed to accumulators via parallel adders for each of the moments. The overflows of these accumulators are shown to be contributions to the various moments and are fed to the decimal display registers. Direct computation of the standard deviation (σ) of the input, from
measured first and second moments has also been investigated. A theoretical analysis of the various errors which occur in such an s.p.c. has been made. Results indicate that for most signals the overall error is within 1% for all four moments. Finally, the development of a universal arithmetic cell, for use in iterative and near-iterative arrays, is reported in this thesis. It is shown that use of such arrays in the arithmetic units of the
s.p.c. can lead to a considerably simplified design. / Thesis / Doctor of Philosophy (PhD)

Identiferoai:union.ndltd.org:mcmaster.ca/oai:macsphere.mcmaster.ca:11375/19597
Date03 1900
CreatorsMajithia, Jayantilal
ContributorsKitai, R., Electrical Engineering
Source SetsMcMaster University
Languageen_US
Detected LanguageEnglish
TypeThesis

Page generated in 0.0021 seconds