Computing Reviews
Today's Issue Hot Topics Search Browse Recommended My Account Log In
Review Help
Search
Accumulator-Based Compaction of Test Responses
Rajski J., Tyszer J. IEEE Transactions on Computers42 (6):643-650,1993.Type:Article
Date Reviewed: Nov 1 1994

A technique of testing VLSI circuits using a signature method is described and analyzed. The technique is best suited for circuits that may have an adder available as part of the circuit. The basic idea is to use the adder as an accumulator to collect the responses of the circuit being tested. The authors carry out a detailed analysis of the method to determine the probability that a faulty device would pass the test if this technique were used to test the circuit.

In the second and third sections of the paper, the authors establish the superiority of the 1’s complement adder over a simple binary adder. Their reasoning is well founded. The nice thing about these two sections is that, after reading them, a reader can find many intuitive reasons to believe in the result. For example, a 1’s complement adder has a feedback, like a linear feedback shift register, whereas a binary adder does not have such a property, which ought to make the 1’s complement adder superior. These two sections and section 4 carry out theoretical analysis, whereas section5 uses simulation to arrive at a model and then analyzes the model. My only concern is with a part of section 4 of the paper in which two error models are analyzed.

The first model is single error pattern. The testing world contains many skeptics about the use of this model, yet it has become acceptable. It is simple and easy to deal with. The second is error pattern &dgr; and −&dgr;. The credibility of this error model is questionable because it seems to imply that faults are detected on two disjoint sets of outputs of a circuit. I do not find this model to be of much practical value, although the authors argue that it provides a worst-case scenario.

The paper is well written and easy to read. VLSI designers and test engineers are the audience for this topic. Although the paper is theoretical, readers without any strong mathematical background should also be able to understand and follow the paper.

Reviewer:  K. Saluja Review #: CR117897
Bookmark and Share
 
Built-In Tests (B.6.2 ... )
 
 
Parallel Circuits (B.6.1 ... )
 
Would you recommend this review?
yes
no

E-Mail This Printer-Friendly
Send Your Comments
Contact Us
Reproduction in whole or in part without permission is prohibited.   Copyright 1999-2024 ThinkLoud®
Terms of Use
| Privacy Policy