Reliability Evaluation Method for Electronic Device BGA Package Considering the Interaction Between Design Factors

摘要

The development competitions of electronic products are intensifying, since shortening the design period and the decreasing the development cost are important problems in recent years. Moreover, the miniaturization and the high integration of electronic device parts were progressed by the advance in technology. Furthermore, the reliability of fatigue life has been prioritized as an important concern, because the thermal expansion difference between a package and printed circuit board causes thermal fatigue. The reliability engineers have to understand the reasons of this problem and improve the design of package in very early stage. Though the packages structure become very complex, it is considered that the design factors have strong interaction. As a result, it is very important to solve the reliability problem with considering interaction of each design factor in short time period. However, it is very difficult and needs much cost. In this study, the method of understanding the relation between design factors and thermal fatigue life has been established, and an application of BGA (Ball Grid Alloy) package was examined. As a result, the interaction between design factors was clarified. Furthermore, the simple evaluation technique of the thermal fatigue reliability in early design stage was examined. By now, it was clarified that the thermal fatigue life of the solder joint is able to evaluate by the total equivalent inelastic strain range. In order to help the design engineers to improve the reliability, it is need to examine the relation between each design factor of BGA package and the total equivalent inelastic strain range generated in the solder joint when the thermal load was subjected. To clarify this relation, sensitivity analysis was executed by using the FEM analysis and the cluster analysis. First of all, FEM analysis models with various design condition were made based upon orthogonal table, and the total equivalent inelastic strain range of the solder joint has been calculated as a characteristic value. Next, the cluster analysis was applied to these analytical results. The analysis results were arranged in order of total equivalent inelastic strain range, and the close models have been clustered to several clusters by Euclid distance. Then each design factor was averaged in a cluster, and by comparing the averages of the design factors, whole relation between design factors and the total equivalent inelastic strain range was clarified. Furthermore, analysis cases where the certain one factor takes value of maximum or minimum were extracted from analytical results of various design conditions. Then those results were arranged in order of the characteristic value, and by applying cluster analysis the influence of the interaction on other factors when certain one design factor was changed has been clarified. As a result, all interaction relations between all design factors were clarified. By performing response surface method with considering the interaction relations, a characteristic value can be expressed by an estimated equation with high accuracy. And the influence of the design factor to the total equivalent inelastic strain range was calculated. This result would lead to the understanding of the relation between each design factor and the thermal fatigue life. And it is possible to evaluate the thermal fatigue reliability simply at the early stage of the design development by using the estimated equation. Therefore, this result will help for an adequate design.