History of Design-for-Test: Structural Test
It still amazes me just how long DFT has been around. I suppose that time passes quicker than you think; turn around, and fifteen years have passed. It was about that long ago when I had my first exposure to scan and ATPG. And to me - it was brand new technology. So when I read about the concept of structural test and stuck-at fault models possibly being kicked around fifty years earlier… well, you see what I mean? Amazing…
Structural Test is Born…
“In order for the successful operation of a test routine to guarantee that a computing system has no faulty components, the test conditions imposed by the routine should devised at the level of the components themselves, rather than at the level of programmed orders”.
Richard D. Eldred realized fifty years ago that functional testing was not efficient - and documented it in his article entitled “Test Routines Based on Symbolic Logical Statements,” in the Journal of the ACM, Vol. 6, pp. 33-36, 1959 (as quoted above). And in those four pages, the future of electronics test was forever changed. Eldred wrote the article while working on the Honeywell/Datamatic-1000. The Datamatic was introduced by Honeywell and Raytheon in 1955, had a core memory of 24k decimal digits, and filled an entire room[ref.], if you can imagine that - if not, click here and check out the photo.
So this was the beginning of Test turning the corner; the paradigm shifted from one of, “let’s verify the function of our manufactured systems, again and again “- to “let’s make sure that our verified design has been put together correctly, by checking the structures“.
And almost 30 years later, again, to add to my amazement, I struggle to convince designers to this day that functional testing, on it’s own, is effective enough to screen defective product.
Who’s fault is that?
