This paper was produced for the 2019 NAFEMS World Congress in Quebec Canada

Resource Abstract

One of the most fundamental issues in electronics design is to properly cool the Printed Circuit Board (PCB) assembly. Miniaturization of modern electronics and ever-increasing complexity requires thermal management to improve reliability and maximize electrical performance. To ensure proper heat dissipation, electrical and mechanical engineers must collaborate to optimize heat transfer at the component, PCB and system level.



The PCB layout designed within the Electrical CAD software needs to be exchanged with the Mechanical CAD application, where it translates into a 3D assembly model. Board geometry, component geometries and positions as well as the PCB stack-up layers are required to construct a model on which different heat transfer mechanisms may need to be considered: conduction, convection or radiation. Based on the environment conditions surrounding the PCB, constraints or boundary conditions are applied to the model before solving the model solution.



A PCB can contain multiple layers that each includes different distribution of traces, vias and dielectrics. For thermal analysis where high accuracy is required, board features can be extruded and meshed in 3D. This approach is very precise as it can take into account small gaps between copper features and minimize modeling assumptions. Therefore, this process is time-consuming and can affect computational performance significantly.



Using Siemens Simcenter 3D, an automated modeling approach can be selected instead, based on a multi-layer shell PCB construction, where the board orthotropic thermal conductivity is calculated using a discretized algorithm. The resulting thermal conductivity of copper and dielectric is then interpolated to the mesh.



In this session, we will evaluate and compare the precision & performance of the 3D approach and the multi-layer shell approach, using a thermal management use case from the medical industry. The objective is to provide qualitative and quantitative metrics in terms of accuracy, pre-processing as well as computational time for a complete typical workflow encountered in the industry. We will demonstrate an efficient exchange process between ECAD/MCAD and show how PCB modeling techniques affect thermal analysis results.