Abstract
Acceleration factors (AF) are key to designing an effective accelerated life test (ALT). They represent the ratio of the time in field to the time in test for a particular event to occur (typically a failure event related to a specific failure mechanism). Time to failure for a device generally correlates with the amount of stress applied (the higher the stress, the quicker the device will fail), and failure models exist to mathematically define that correlation for various failure mechanisms. This allows for use of a higher stress in test than in the field, thereby providing an acceleration factor that shortens the time in test to demonstrate a failure-free time period.
ALT can take the form of qualitative or quantitative testing. The latter is used to determine the life characteristics of the device with some reliability and confidence level. Usage rate acceleration and higher stress acceleration can be used. It is important to consider the design limits of the device based on its specification and material properties, and limit the stress levels in test so as not to induce failure mechanisms that the device would not otherwise have experienced in the field.
ALT results are used to make life predictions for the device tested. With no failures, the test results demonstrate the required reliability and confidence level metrics for the failure mechanism of interest. With several failures, a reliability software tool can be used with the appropriate analysis method, rank method, and confidence bounds method chosen in order to extrapolate to an expected life in test. The equivalent field life is based on multiplying the expected life in test by the AF. If the field stress and/or test stress are not constant, there are multiple acceleration factors to utilize. As a result, an equivalent acceleration factor needs to be calculated and used as the AF when predicting equivalent field life.