Evaluating the Dynamic Behavior and Analytically Predicted Displacements of Printed Circuit Boards (PCBs) Using the “Smeared-Mass” & Fine Mesh Approach

Abstract

I Abstract ue to the high cost of component failure during launch, component level vibration testing is used to verify electronic component design and certify electronic boxes and assemblies as acceptable for flight in aerospace applications. Prior to testing, Finite Element Analysis (FEA) is typically used to verify electronic assembly designs by simulating and affirming acceptable component dynamic responses and minimal electronic component deflection. Random vibration induces deflection of the mounted chips relative to the printed circuit boards (PCBs), which can cause fatigue-type failures. While a deflection-based failure mode can be evaluated using complex and detailed finite element modeling of the boards, components, and lead wires, such efforts assume a large resource cost, both in model generation analysis run-time. Thus, the need for a simplified modeling technique that accurately simulates the board dynamics is required. Using finite element models and random vibration test data for an electronic assembly, the authors will attempt to quantify the quality of the analytical predictions between multiple coarse modeling approaches. The predictions between a “traditional” coarse model approach (or one in which the mass of the mounted components is evenly distributed on the board using non-structural mass) and more detailed modeling approaches (local mass and stiffness patch-based modeling) will be explored. An approach using concentrated mass elements with the component properties (mass and assumed inertia) on rigid elements and approximation of chips using uniform generic solid elements will also be explored. The rigid elements (in a “spider web” configuration) and solid elements will introduce local stiffening of the board. A total of five (5) analytical approaches will be compared to each other and to actual test results in terms of modal frequency response and static PCB displacement. The goal of this paper is to demonstrate the quality of the test predictions of the various coarse modeling approaches and provide recommendations that will lead to better finite element model predictions in dynamic environments.