An Optimal Parametric Design to Improve Chip Cooling

In recent years, most electronic products were overcome the thermal problem by additional cooling equipments. These designs induce the manufacturing cost and the risk of the cooling failure problem. The present study utilizes Taguchi experimental design to speed up the investigation in the passive cooling design for electronic systems. A representative CFD (computational fluid dynamic) model has been selectively implemented to be a database for statistic analysis. Power density (A), orientation of mother board (B), chip geometry (C), opening between chips (D), and flow pattern (E) are the parameters selected in order to find out the optimal combination in thermal cooling with two-level statistical approach. The result indicates about 50% of effort in experiments or simulations could be saved. The results further verified that openings in the mother board, power density, and flow pattern are sequentially the important parameters and they do significantly influence the thermal behavior. Finally, the concept of opening not only improve the reliability but also reduce the manufacturing cost and simplify the assemble procedures