Files in this item

FilesDescriptionFormat

application/pdf

application/pdf8409776.pdf (7MB)Restricted to U of Illinois
(no description provided)PDF

Description

Title:A Comprehensive Fault Model for Concurrent Error Detection in Mos Circuits
Author(s):Halperin, Daniel Lee
Department / Program:Electrical Engineering
Discipline:Electrical Engineering
Degree Granting Institution:University of Illinois at Urbana-Champaign
Degree:Ph.D.
Genre:Dissertation
Subject(s):Engineering, Electronics and Electrical
Abstract:A comprehensive fault model is developed for concurrent error detection in MOS integrated circuits. This fault model is based on a thorough examination of physical failures in MOS integrated circuits. Models of MOS circuits are also developed which are used to determine the behavior of these circuits under failure. It is found from this analysis that many types of physical failures may result in logic signals that are not well-defined. In particular, it is shown that physical failures may lead to constant values that are neither logic 0 nor logic 1, timing failures, or oscillation. The concept of indeterminate faults is developed to describe the behavior of such failures. It is shown that most traditional fault models are unable to model the behavior of a circuit with an indeterminate fault correctly.
Ternary algebra is used to facilitate the analysis of circuits which receive indeterminate value inputs. Using ternary algebra, necessary conditions are developed for the propagation of indeterminate values through a circuit. It is shown that in many cases, an indeterminate value can propagate through a circuit even when a Boolean value cannot propagate.
The methodology of totally self-checking systems is used to provide concurrent error detection. It is shown that the traditional definitions of the totally self-checking property are inappropriate for failures which include indeterminate faults. A new definition of the totally self-checking property is developed which is compatible with indeterminate faults. It is shown that under our fault models, duplication may be used to provide a totally self-checking implementation for any function. Procedures are developed to determine if a function has an implementation using a separable code which may provide concurrent error detection at a lower cost than duplication. Issues involved in the interconnection of several totally self-checking circuits are considered, as well as the requirements for checkers in systems which may experience indeterminate failures.
Issue Date:1984
Type:Text
Description:214 p.
Thesis (Ph.D.)--University of Illinois at Urbana-Champaign, 1984.
URI:http://hdl.handle.net/2142/69264
Other Identifier(s):(UMI)AAI8409776
Date Available in IDEALS:2014-12-15
Date Deposited:1984


This item appears in the following Collection(s)

Item Statistics