Making Products Die Faster—In the Lab, Not in the Field

A workshop discusses ways to test technologies for problems that will occur near the end of their life cycles

27 July 2011

Do the electronic industries need a new approach to qualifying their products? That’s the question surrounding the upcoming IEEE Accelerated Stress Testing & Reliability Workshop (ASTR), to be held from 28 to 30 September in San Francisco. Qualifying a product means evaluating it to ensure it meets all expectations for its application. That involves not just quality—or adherence to a set of requirements—but reliability, which is complying with those requirements over the product’s expected life. But how do you predict what the life of the product is going to be, based on test results? And is the testing the most effective available?

ASTR is co-sponsored by the University of California at Santa Cruz and the IEEE Components, Packaging & Manufacturing Society, with support from the International Conference on Prognostics and Health Management as well as the American Society for Quality’s Reliability Division.

The workshop will try to determine which types of qualification testing are best, a topic that’s attracting much discussion in industry, says IEEE Member Don Gerstle, the technical program co-chair.

Ensuring reliability includes stressing products often to death with thermal cycling, vibration, shock, voltage and current extremes, exposure to other environmental stresses, and highly accelerated life testing (HALT). The idea is to accelerate failure mechanisms that might occur during actual use in the field without inducing failures that otherwise would not occur.

Accelerated testing is especially important because today’s electronic systems often are expected to achieve mean-time-to-failure of thousands of hours. Testing in real time, or non-accelerated testing, would take too long. It might not reveal hidden flaws until the products were already superseded by new models; HALT can reveal defects before the products go into production. That’s closely tied to reliability physics—the analysis of failure modes and mechanisms—which can guide the design of tests as well as the products themselves. It also involves predictive modeling to lower the probability of field failure.

Even so, testing is more a matter of techniques and best practices than of theories. So ASTR is truly a workshop rather than a traditional conference. By popular demand, there’s only a single track of sessions so one can attend any or all of them. Included are keynote talks, panel discussions, tutorials, case studies, and sessions on equipment, reliability, reliability testing, and the physics of failure.

“I learn a lot about many real-life reliability problems at this workshop,” says Ephraim Suhir, the conference’s general chair. “It’s a forum for sharing novel and useful methodologies. We’re all equals here, and we learn from each other.”

Session titles include “The Role of Accelerated Testing in Attaining High Reliability in Small Production Quantity Products,” “The Design of Reliability Tests Using Physics of Failure—Bringing Math and Physics Together,” and “The Electronics Industry Needs a New Frame of Reference for Qualifying the Reliability of Products—Moving from Time Metrics to Stress-Limit Metrics.” Also covered are techniques for test design and such application areas as solar mirrors and the power grid.

Several sessions are devoted directly to the workshop’s theme, while others cover protecting customers from field failure, new technology, and assessing whether electronics reliability is slipping.

Panel moderators from around the world, including Finland, Israel, Japan, and the Netherlands, are scheduled to participate, representing industry, academia, government laboratories, and the U.S. military.

IEEE membership offers a wide range of benefits and opportunities for those who share a common interest in technology. If you are not already a member, consider joining IEEE and becoming part of a worldwide network of more than 400,000 students and professionals.

Learn More