Browse Prior Art Database

# Cyclic-Short Detection

IP.com Disclosure Number: IPCOM000049492D
Original Publication Date: 1982-Jun-01
Included in the Prior Art Database: 2005-Feb-09
Document File: 3 page(s) / 59K

IBM

## Related People

Hsieh, HH: AUTHOR [+2]

## Abstract

A common failure in large-scale integration (LSI) and very large scale integration (VLSI) circuits is the short. Methods for their detection in the event that they do not introduce feedback (cyclic case) are well known. The following applies the D-calculus to compute tests for the cyclic case for a general technology including MOSFET and bipolar logic. The running time is comparable to that of the D-algorithm.

This text was extracted from a PDF file.
At least one non-text object (such as an image or picture) has been suppressed.
This is the abbreviated version, containing approximately 63% of the total text.

Page 1 of 3

Cyclic-Short Detection

A common failure in large-scale integration (LSI) and very large scale integration (VLSI) circuits is the short. Methods for their detection in the event that they do not introduce feedback (cyclic case) are well known. The following applies the D-calculus to compute tests for the cyclic case for a general technology including MOSFET and bipolar logic. The running time is comparable to that of the D-algorithm.

Assume in the diagram of Fig. 1 that in the correct circuit, e is a function of variables a, b, c and d alone. Under failure of shorting at s, e, because of the technology contraints, becomes a function of d alone. This is due to the fact that in MOSFET technologies the value of a variable is denoted by the 'voltage' of a line, and around a loop the voltage drop, with no voltage sources is zero, and the value of c will be the OR of d and the product of the voltage drops around the loop.

It is obvious that the voltage of e is that of d in the case of a short. Thus, we can easily obtain a test for this short by means of the D-calculus (*). For example, using cubical covers for a portion of the logic we have (see original)

Yielding the test a=l, b=0, c=0 regardless of the value of GOOD CIRCUIT SHORT CIRCUIT

a b c d a b c d e

1 1 1 1 1 x x x 1 1

0 x x x 0 x x x 0 0

x 0 x x 0 x x x 0 0

x x 0 x 0

x x x 0 0

Test obtained from D-calculus a b c d e

0 x x 1 I D

If a=0 and d=1, the good circuit is 0, and the bad circuit is 1.

This method may be extended to move general configurations b...