IMPROVING THE DIAGNOSABILITY OF MODULAR COMBINATIONAL LOGIC BY TEST POINT INSERTION,
Abstract
To improve module level diagnosability of modular combinational logic circuits, three are described for allocating test points. The first, analogous to the prime implicant covering problem of classical switching theory, requires a large amount of computation but produces a minimum test point set. The second, a graphical worst case analysis, is considerably more economical but produces non-optimum results. The third is a combination of the first two and lies between them in computation required and optimality of results. It is proved that the graphical test point specification algorithm is sufficient to guarantee complete module level diagnosability of the circuit if the test points are used to gain more information on the result of each test. A general graph-theoretic characterization of the non-optimum set of selected test points is given in terms of accessibility properties. A final theorem is proved to show that tests derived on the basis of fault detection only are sufficient for diagnosing the previously troublesome fault classes. An Appendix is given to present an efficient notation useful for analyzing the behavior of a combinational circuit under all allowed fault conditions. The procedure is algorithmic in nature and programmable, and flow charts are given to illustrate one possible implementation. (Author)
Document Details
- Document Type
- Technical Report
- Publication Date
- Mar 01, 1969
- Accession Number
- AD0686385
Entities
People
- Terry Gene Gaddess
Organizations
- University of Illinois Urbana–Champaign