Browse Prior Art Database

Fault Isolation Probe Algorithm

IP.com Disclosure Number: IPCOM000088632D
Original Publication Date: 1977-Jul-01
Included in the Prior Art Database: 2005-Mar-04
Document File: 2 page(s) / 13K

IBM

Related People

Herberg, RT: AUTHOR [+2]

Abstract

This diagnostic algorithm for testing logic cards and the like uses a binary search technique to isolate the fault with a minimum amount of probing. Logical data obtained from stuck-fault simulation is used if available, otherwise physical data is used.

This text was extracted from a PDF file.
This is the abbreviated version, containing approximately 64% of the total text.

Page 1 of 2

Fault Isolation Probe Algorithm

This diagnostic algorithm for testing logic cards and the like uses a binary search technique to isolate the fault with a minimum amount of probing. Logical data obtained from stuck-fault simulation is used if available, otherwise physical data is used.

The steps below demonstrate one way of implementing the technique for the proposed binary search fault isolation procedure. Step 1. Assign weights to possible failing components based on

precalculated diagnostics. Enter the components

mentioned above, with their respective weights, into a table

or "List A".

Step 2. For each failing output and shift register latch

construct a list of all nets controlling the failing output

or latch. These lists are referred to as lists B1, B2,

B3,...,Bn.

Step 3. For each of the Lists Bi generated in Step 2, do the

following:

a. For each net in List Bi, find the component(s) that

drive(s) this net. Assign a weight to the net based

on the component weight(s) from "List A". If "List

A" is a null list, then assign a weight of "1" to

the net, unless the net is a primary input in which

case a weight of "0" is assigned. If "List A" is

not null, each component not in the list will have a

weight of "0".

b. Sum the weights of the nets in "List Bi".

Step 4. Select "List Bj" from lists B1 to Bn, where Bj has the

lowest total weight, and refer to List Bj as "List C".

Step 5. For each net in "List C" of Step 4, find the SUMI,

referred to as SUM Ci, of the weights of each net

controlling that net (including that net). The net with the

maximum SUM Ci is designated Nf.

Step 6. Select the net from "List C" whose SUM Ci is closest to

the total weight of "List C" divided by two, and refer to

this net as net Np.

Step 7. Probe net Np.

Step 8. Construct a list, referred to as "l...