Issue No.05 - September-October (2007 vol.24)
Published by the IEEE Computer Society
Scott Davidson , Sun Microsystems
DOI Bookmark: http://doi.ieeecomputersociety.org/10.1109/MDT.2007.154
This is a review of Advances in Electronic Testing: Challenges and Methodologies (edited by Dimitris Gizopoulos), which is part of a series called Frontiers in Electronic Testing. This book can be thought of as a set of advanced tutorials in book form. One of the best things about this book is that several chapters go beyond the what of the subject and into the why. The book is also well integrated, with pointers to other chapters when necessary and with a minimum overlap between chapters.
Attendees at many Test Technology Technical Council (TTTC) conferences have the opportunity to attend tutorials on various test-related subjects. Those interested in an introduction to a new topic area or an advanced survey can take a full- or half-day class with an expert, and come away with a valuable set of notes. Advances in Electronic Testing: Challenges and Methodologies, edited by Dimitris Gizopoulos, can be thought of as a set of advanced tutorials in book form. In fact, many of the authors in this book have been tutorial presenters. I think that's one of the reasons it's so successful at achieving its stated purpose.
When I read a book for review, I watch out for passages that leave room for questions, and for sections where the exposition could be clearer. I found almost none of these in this book. Perhaps the authors got questions on unclear points during the tutorials and fixed them in the text, or perhaps the editor found them all. In any case, I wrote very few notes in the margins as I was reading.
One of the best things about this book is that several chapters go beyond the what of the subject and into the why. For example, the first chapter, by Rob Aitken, describes how defect-oriented testing is different from structural or functional test, and how defect-dependent variation must be distinguished from circuit- and environment-dependent variation. Aitken starts with physical-defect mechanisms, both classical and those for newer technologies, then relates them to faults and fault models, provides some experimental studies on faults and defects, and concludes with new test methods such as statistical postprocessing.
Another example is Chapter 9, "Mixed-Signal Testing and DFT," by Stephen Sunter. Besides providing an excellent introduction to the topic, Sunter explains which DFT techniques haven't been accepted and why. He also helpfully provides a set of evaluation criteria for mixed-signal DFT. Someone reading this chapter carefully will emerge with not only basic background knowledge but also with some hints as to how to read papers and press releases on this topic. In fact, I consider this to be a model of how a tutorial chapter should be written.
Other chapters don't take the standard approach to their subject matter. Chapter 11, "Loaded Board Testing," by Ken Parker, spends relatively little time on boundary scan, a subject on which quite a lot has been written. Instead, he begins with the PCOLA-SOQ model for board-level faults, continues with in-circuit tests for components, and finishes with inspection systems. This is another excellent tutorial. I would have liked to have seen something about board functional test, though, since it still plays an important role in the complete board and system-test process.
The other chapters are outstanding as well. Chapter 2 describes details of the failure mechanisms we are now seeing, and why they are getting worse. There is a bit of overlap with Chapter 1, but it is kept to a minimum. Chapter 3 gives a much-needed tutorial on silicon debugging, including a case study. Ineffective debugging can cause major delays in time to market, and I know many people who believe scan's greatest value is in debugging rather than test generation. Chapter 4 gives a good introduction to delay test, and goes beyond the usual content of such chapters and into I/O delay characterization and the philosophy of various methods of delay test application.
Chapter 5 covers another important topic not often seen in this type of book: the interface between the device under test and the tester. Low-cost, DFT testers have made some inroads into the ATE market. Chapter 6 gives not only the architecture of this type of tester but also its philosophy and the design decisions used in the development of low-cost ATE. This chapter illustrates one of the things I find most valuable in this book: Much of it is reality based. There is a frank discussion of what this class of tester can and cannot do, and why reducing the cost of test is not as easy as marketers, and even eager researchers, claim it is. I also appreciate that this chapter gives specific solutions to issues arising in low-cost testers without appealing to a particular product as an example. Chapter 10, on embedded-memory testing, is another excellent chapter. It takes the position that, to generate a good test, the test engineer must understand a memory's physical implementation. This is important to keep in mind as defect types get more arcane but the time we have for testing them does not increase. We can't afford to test for defects that won't appear, and we can't afford not to test for defects that will.
I found the chapter on SoC testing (Chapter 7) to be a bit heavy on generic DFT methods and a bit light on the problems of SoC designs in general. Only a short section describes functional test in this environment, and someone receiving an IP core without DFT to be integrated into an SoC won't find a lot of help here. There is more on BIST than on test access mechanisms, for instance. I would have preferred to see less on the basics of DFT and more on SoCs.
Finally, the chapter on RF testing takes the approach of describing a low-cost RF tester built at IBM. This chapter, starting at a higher level, was out of line with the rest of the book. Many terms used are not explained, making me think the target audience should be skilled in RF test already. There is a bit too much of the flavor of a paper describing an implementation here, as opposed to the tutorial nature of the other chapters. I thought Chapter 6 did a much better job at handling the problem of justifying a tester architecture.
Given the diversity and depth of this book's subject matter and the clarity of so many of its chapters, I can't imagine anyone who won't get something out of it. I'd especially recommend it at the graduate level as a method of illustrating the real issues students will run into when they work in the industry. The book is also well integrated, with pointers to other chapters when necessary and with a minimum overlap between chapters. This volume is part of a series called "Frontiers in Electronic Testing." I fear that some might think the frontier of test research consists of far-out research that will never be relevant. Quite the contrary, this book covers the frontier where all of us now live—a frontier we need to know more about.