3D thermal model to investigate component displacement phenomenon during reflow soldering

Abstract

This paper reports the development of a 3D component-level thermal model to investigate the displacement of the components during the reflow soldering process. One of the root causes of displacement is the temperature deviation between the contact surfaces of the components. Therefore, our model examines the temperature distribution at the level of discrete components and not at the level of whole assembly as is the case of general models used in reflow soldering [Whalley DC. A simplified reflow soldering process model. J Mater Process Technol 2004;150:134–44; Sarvar F, Conway PP. Effective modelling of the reflow soldering process: basis construction and operation of a process model. IEEE Trans Compon Pack Manuf Technol C: Manuf 1998;21(2):126–33; Inoue M, Koyanagawa T. Thermal simulation for predicting substrate temperature during reflow soldering process. In: IEEE Proceeding of the 55th ECTC; 2005. p. 1021–6]. Our model is based on the thermal or central node theory. This means that the investigated area is divided into thermal cells representing the thermal behaviour of the given material. During the model discretization a nonuniform cuboid cell grid is used, with this the resolution of the model can be refined to focus on areas of interest (e.g. the soldering pads) whilst reducing the involvement with the less important areas (such as the epoxy body of a component). This approach therefore allows to achieve a significant increase of spatial resolution of the calculated temperature distribution in some distinct areas with only a little increase of model complexity. The model is described by common heat conduction and convection equations and these are solved by the finite difference method in order to achieve high calculation speeds and easy implementation. The ability of our single purpose model is then compared with a general purpose FEM analyzer. It results that the calculation accuracy of our model is comparable with more detailed models. Also, its calculation speed and application is much faster and easier.