Multiscale analysis of multilayer printed circuit board using mechanics of structure genome

Abstract

The structure of Printed Circuit board (PCB) is very complicated because it consists of woven composites and custom defined conducting layers. To improve the reliability of PCB, it is critical to predict the constitutive relations accurately. This study by implementing Mechanics of Structure Genome (MSG), homogenizes a multilayer PCB to determine the effective mechanical properties and coefficient of thermal expansions. The homogenization is divided into two steps. The first step is to obtain the yarn material properties. The second is to write python script in TexGen4SC to generate the woven model and combine it with conducting layers in ABAQUS to obtain the PCB effective material properties. A new tool, which is freely accessible at cdmHUB, was developed to integrate the two steps woven homogenization. The homogenized material properties were validated with representative volume element (RVE) analysis. A good agreement was observed between MSG and RVE analysis. Finally, both structural analysis and dehomogenization were done to predict a PCB plate global responses and local stress fields. MSG-based plate structural analysis and dehomogenization matched the direct numerical simulation (DNS) very well. However, MSG is significantly faster than DNS.