Determine and design the best ALT

Abstract

Over the many years of development concerning accelerated life testing (ALT), our peers have found many ways to take advantage of the interaction of stress and failure mechanisms [1-15]. In an ideal situation, the reliability engineer will have ample time, samples, test resources and knowledge to conduct an ALT. This is often not the case. Trading off the risks in conducting the ALT and fitting within the myriad of constraints and expectations is a challenge. Understanding the basics of ALT approaches and associated assumptions, permits one to select the right ALT. `Right' being the ALT that provides meaningful results in time for technical and business decisions, plus meets the budget and risk tolerance limits. There is no one-way to design an ALT that will meet the specific set of conditions presented to the test designer. Being able to clearly articulate the tradeoffs involved permits the entire design team to fully understand the results when produced. The `best' ALT is one that adds value to the design process. The most accurate results involve testing all of the production units in actual customer application or use until they all have failed. While this is clearly not practical, neither is the simple-minded approach guessing at the results. In between these two extremes lies an optimal value: being the most efficient ALT that provides meaningful results. When the results provide information to make design or program decisions, the ALT adds value. Reducing ALT costs by reducing sample size or test duration is possible, yet may significantly increase uncertainly around the results. Running the test longer to achieve more accurate results is often constrained by the timeline to make decisions. It is this and similar tradeoffs that force us to carefully design each ALT and determine the best path forward.