Development of Reliability Allocation and Assessment Algorithms for Designing Multilevel Microelectronic Systems

Abstract

Design-for-reliability of complex systems involves top-down reliability allocation approaches, reliability prediction of both random and wear-out failures, and bottom-up reliability assessment approaches to provide more insight into the system-level reliability. Designing complex microelectronic systems, while considering reliability in the early design stages, is a challenge because these systems have multilevel structure and logical groups, and numerous components are associated with failure modes and mechanisms. To address these difficulties and to design reliable systems in a systematic way, reliability allocation and reliability assessment algorithms and associated reliability predictions methodologies are presented in this paper in the context of a System-Design-for-Reliability (SDfR) framework. Reliability allocation algorithms are presented for both parallel and series systems that calculate the target reliability of subsystems from the given target reliability of their parent systems. The reliability allocation algorithm is demonstrated for random failures in a video broadcasting system that consists of a four-level packaging structure. The reliability assessment algorithm is demonstrated for wear-out failures in a USB board system that consists of multiple logical groups and various failure modes and mechanisms. The reliability assessment algorithms also demonstrate the use of physics-based reliability prediction of each logical group before assessing the system reliability. The demonstrated results show that the algorithms are useful for determining system configurations and design parameters. Such design changes will reduce the burden of downstream reliability activities.