Repository hosted by TU Delft Library

Home · Contact · About · Disclaimer ·
 

Novel reliability assessment concept based on an accelerated de-rated strength approach

Publication files not online:

Author: Veninga, E.P. · Kregting, R. · Waal, A. van der · Gielen, A.W.J.
Type:article
Date:2013
Source:14th International Conference on Thermal, Mechanical and Multi-Physics Simulation and Experiments in Microelectronics and Microsystems, EuroSimE 2013, 14-17 April 2013, Wroclaw, Poland
Identifier: 477630
ISBN: 9781467361385
Article number: 6529983
Keywords: Electronics · High Tech Systems & Materials · Industrial Innovation · Mechatronics, Mechanics & Materials · MIP - Materials for Integrated Products · TS - Technical Sciences

Abstract

The introduction of new materials or technologies can have an enormous impact on the Time to Market (TTM) of new products. Preferably, the performance of new materials or technologies is known before these are designed into a product. This paper presents a reliability assessment approach which has been developed with the aim to reduce the so called Time to Technology (TTT). The method which is based on an accelerated de-rated strength approach has been expanded to a concept which could also include health monitoring and prognostics during lifetime. In this part of the work a combination of modelling and statistical techniques was used to explore the feasibility and potential of the concept. Ball Grid Array (BGA) designs were used as a vehicle with solder fatigue as the selected failure mechanism. Finite Element Modelling (FEM) together with Design of Experiments (DoE) revealed that the (package) substrate thickness, stand-off, (package) substrate size and the final solder ball diameter are the statistical significant factors with respect to fatigue life of SnAgCu BGA balls. Simplified linear models obtained from regression analyses were used to design de-rated strength variants and estimate test times. Simulations using a strain based lifetime model of Engelmaier together with a Monte Carlo method were used to generate lifetime distributions based on induced variations. A statistical analysis showed a significant difference in lifetime performance between the simulated de-rated strength designs. © 2013 IEEE.