An optimal parametric design to improve chip cooling

The thermal problems of most electronic components have been solved in recent years by installing various cooling equipments. The passive cooling, which has no need of the fluid-driving devices, can be more reliable and achieved with less cost. In this paper, the Taguchi's method is applied on the optimization of the passive cooling for electronic systems, and a representative CFD (computational fluid dynamic) model is selectively implemented for statistical analysis. The selected parameters in this study are the power density, mother board orientation, chip geometry, opening between chips and the flow pattern. The optimal combination for thermal cooling is obtained by using a two-level statistical approach. Analysis results indicate that about 50% of the effort in performing experiments and simulations can be saved. Further, the results confirmed that the most important parameters affecting the thermal behaviors are the openings in the mother board, power density, and the flow pattern in sequence. Finally, the concept of opening increases the reliability, reduces the manufacturing cost, and simplifies the assembly procedures. (C) 2007 Elsevier Ltd. All rights reserved.