Finite element analyses (FEA) have emerged as a process for assessing stresses and strains in electronic equipment in order to compute the expected structural life. However, potential pitfalls may compromise accuracy. Guidelines have been established to improve the accuracy of these results. A method has been outlined that allows simplified linear FEAs to be used instead of the more complex elastic-plastic nonlinear FEA. Guidelines for mesh generation have been established to eliminate arithmetic errors caused when materials with large stiffness differences are adjacent to each other. The accuracy of FEAs when dealing with very small dimensions has been verified. Procedures for combining various loadings in order to predict life have been established for materials that exhibit stress relaxation and for those that do not. With these guidelines, FEAs can be an effective tool to predict the structural life of electronic equipment.
Predicting the reliability of electronics is as much of an art as an exact science. There are too many factors, both controllable and uncontrollable, which can not be readily identified, quantified, or combined into an all-inclusive and deterministic equation. However, progress has been made to simplify the prediction process. McDonnell Aircraft Company (MCAIR) has developed a software tool (E-Life) which will predict the fatigue life of solder joints, component leads, and plates through holes. E-Life is also capable of identifying the critical components that require analysis, greatly simplifying life predictions. The results of complex finite element analyses of printed circuit assemblies (PCA) are an integral part of the software, but are transparent to the user. Thus, the time to perform an analysis is reduced and the probability of error is decreased. Experimental tests have been used to improve the accuracy of E-Life. Thermal, mechanical, and dynamic test results were used to obtain material properties previously unavailable. Test results were also used to validate the analytical models. When combined, E-Life and existing prediction techniques provide a more comprehensive reliability prediction process than was available in the past.