Bypass Effect in High Performance Heat Sinks

Studies using commercial computational fluid dynamics software, running on a supercomputer, were carried out to investigate the effects of fin density, inlet duct velocity, and clearance area ratio, on the extent of flow bypass and its impact on the thermal performance of the heat sink. Flow bypass was found to increase with increasing fin density and clearance, while remaining relatively insensitive to inlet duct velocity. An optimum geometry, for a fixed inlet duct velocity, bypass clearance, fixed heat sink volume, and constant thickness, was determined. INTRODUCTION Continuing Moore's Law improvements in semiconductor technology are leading to larger, more complex integrated circuits with power dissipations and heat fluxes anticipated to reach 170W and 30W/cm2, respectively, by the middle of the current decade. Despite its relatively poor thermal properties, air continues to be the coolant of choice for many high-end electronic applications, necessitating the development of highperformance heat sinks, composed of thin, closely-spaced fins. The performance of such heat sinks is adversely affected by flow "by-pass," resulting from planned or inadvertent clearance gaps around the heat sink. Early investigations of flow bypass phenomenon in longitudinal plate fin heat sink configurations include Sparrow et al [1978] and Sparrow and Kadle [1986], who studied the effect of tip clearance on thermal performance. These studies showed the ratio of heat transfer coefficients, with and without clearance, to be strongly influenced by the tip clearance to fin height ratio, and to be independent of air flow rate and fin height. Experimental results in Butterbaugh and Kang [1995] indicated the thermal resistance to correlate exclusively with the pressure drop in the duct across the heat sink, while appearing to be relatively independent of the amount of bypass. Analytical network flow based models using laminar flow in parallel plate channels have been presented in studies by Butterbaugh and Kang [1995], Lee [1995], and more recently by Reis and Altemani [1999]. Wirtz et al [1994] used analytical solutions for the heat transfer coefficient for developing flow between parallel plates, in conjunction with experimentally obtained data, to construct correlations for the prediction of heat sink thermal performance. This study also showed the existence of an optimum array design, for a given inlet flow condition and shroud configuration. The experimental data collected for the various studies presented in literature, utilized heat sinks with relatively thick fins (≥ 1.27 mm) and moderate fin spacing (≥ 2.4 mm). The studies did not elaborate on the loss of airflow from the top of the heat sink (“leakage”), and the distribution of airflow through the various regions of the duct. The objective of the current work is to examine these phenomena in detail, using numerical CFD modeling. The present paper begins with a discussion of the analytical heat transfer and fluid mechanics relations that govern the behavior of compact heat sinks and the geometries needed to achieve the desired cooling capability. Next, the laboratory apparatus used to experimentally characterize these compact heat sinks will be described, along with the results of experiments with fullyand partially-shrouded compact heat sinks. For validation purposes a comparison of these results, with values obtained using analytical and numerical (CFD) techniques, have been presented. Attention is then turned to the use of commercial computational fluid dynamics (CFD) software to explore the effect of bypass on the thermal capability of three candidate heat sink designs, over a range of clearance area ratios and air mass flow rates. The thermal resistances and distribution pattern of these three heat sink designs are compared. To demonstrate the existence of an optimum geometry for a fixed inlet duct velocity and bypass clearance, the array geometry was optimized for a fixed heat sink volume, by varying the fin spacing while using fins of constant thickness. ANALYTICAL MODELING OF SHROUDED HEAT SINKS The analytical model developed by Holahan et al in [1996] for calculating the thermal and pressure drop performance in fully shrouded, laminar, parallel plate heat sinks has been utilized to characterize the thermofluid performance of parallel plate heat sinks. The model represents the flow field as a Hele Shaw flow and fin conduction is characterized using the superposition of a kernel function determined from the method of images. A 2-D flow field was assumed in the parallel plate channel between adjacent fins with uniform heating in the heat sink base and was shown to give good agreement with the 1-D results of Iwasaki et al [1994]. The side-inlet-side-exit (SISE) configuration considered in the current study is depicted in Figure 1, showing the nomenclature of the array geometry, including the fin height, H, fin thickness, t, inter-fin spacing, s, width of base, W, and the length of the heat sink base, L.