Gathering the quality evidence

E-mail

Creating a workable, economical test and inspection strategy often seems like a bit of a black art. We apply all of our experience, yet we wonder if we are looking for the defects that the boards actually contain. Gathering good defect data is expensive and time-consuming, so we often don’t have a lot to work with.

Eight years ago, when Stig Oresjo, then with Agilent Technologies, analyzed his original data on more than a billion solder-joint defects from 15 companies—including both OEMs and contract manufacturers—he shed some light on the situation. His results showed that the vast majority of defects consisted of shorts (22%), opens (46%), and insufficient solder (17%). More significantly, he discovered that the average number of defects per million opportunities (DPMO) vastly exceeded manufacturers’ goals.

Oresjo has now released a new study that shows progress achieving better quality, despite increased product complexity. Test strategies have changed somewhat, and because this time he included companies from Asia, they proved less consistent. The article on p. 68 in this issue examines some of his results.

Much analysis remains to be done—3.7 billion solder joints do not surrender their secrets easily. It will be interesting to see what new insights they will reveal—and how the next study eight years from now will compare.

Contact Steve Scheiber at sscheiber@aol.com