Tests Compare Forced Convection Heat Transfer

Because junction temperature is the main factor influencing semiconductor reliability and performance, semiconductor manufacturers specify the maximum allowable junction temperature. Maintaining junction temperatures within the prescribed limits usually involves the use of heatsinks, particularly for power and microprocessors at the p.c. board level. The present trend of decreasing package size and increasing heat dissipation, which causes heat flux to increase further complicated heat removal. Consequently, to improve a heatsink's ability to remove heat, thermal designers are looking at the various manufacturing processes and materials used for heatsinks.

Because of their cost effectiveness, extruded heatsinks are usually used for low power dissipation and low heat flux applications. However, when producing high aspect ratio fins, the extrusion die breaks more readily as the fin thickness and fin spacing decreases. For high-volume applications, the die-casting manufacturing technique is an alternative, due to its low average cost. However, high porosity and low purity alloys result in lower thermal conductivity products.

In heatsinks with bonded fins, the base is extruded with slots to allow the insertion of plates or extruded fins. You can attach the fins to the base using thermal epoxy, brazing, or swaging. Thermal epoxy is commonly used to bond high aspect ratio heatsinks, but epoxy has a very low thermal conductivity compared with aluminum, so its thickness should be lessened to minimize thermal impedance.

Brazing is a subgroup of welding that takes place at temperatures above the liquid state of a filler material (450°C) and below the solid state of the base materials. Capillary action plays a major role in filler flow through the joints.

The swaging process (Fig. 1) is a cold forming process used in the fabrication of high fin density heatsinks. Currently, this process involves the placement of fins with a tapered base into a slotted base plate and then the application of a rolling pressure on the opposite sides of each fin. This results in vertical and lateral pressure of the base unit material, which tends to push the fin toward the bottom of the groove in the base. This secure connection provides very good thermal contact between the fins and base, and also prevents air and moisture from entering the grooves, thereby preventing corrosion and allowing the heatsink to be anodized.

Heatsink Tests

To determine the best heatsink design, we compared the thermal performance of four heatsink combinations using forced convection: aluminum, copper, copper base/aluminum fin, and aluminum base/copper fin, as you can see in Fig. 2. We placed each heatsink within a vertical wind tunnel of Plexiglas walls so the fins were positioned vertically and parallel to the airflow inside the tunnel. We then attached a 585W block heater covering 5% of the base to each heatsink. Next, we performed tests for an approach velocity ranging from 2 m/s to 8 m/s.

Generally, the heatsink base plate area, fin height, and fin-center-to center distance were the same for all heatsinks, as seen in Fig. 3 and the table. The aluminum-serrated fins were extruded with an overall average thickness of 1.2 mm and an average base thickness of 1.33 mm. The thicker fin base helps secure the connection between the fins and base, resulting in good thermal contact through the swaging process. The extrusion process used to produce the aluminum-serrated fins is flexible enough to allow different fin body and base geometries (Fig. 4). For the swaging process involving the flat copper fins, the (constant) copper fin thickness was selected to be equal to the aluminum fin base thickness of 1.33 mm. Copper fins were sheared from 1.33 mm thick rolled flat sheets. The rolling process used to produce copper plates only allowed fixed flat sheets (Fig. 4).

To facilitate the single heatsink testing, we modified a section of the wind tunnel door to allow mounting a heatsink so the base was flush to the interior wall of the tunnel with the fins extended into the air stream. We constructed large Plexiglas baffles and flow diverters, and positioned them in the wind tunnel to ensure a smooth transition of the air into the duct upstream of the heatsink (Fig. 5). Following a brief test to ensure correct operation of all system components, we turned on the blower, adjusted it to the desired target velocity, and waited for steady state data acquisition.

The data represents performance of all the bonded heatsinks (Fig. 6, on page 54). The all-copper heatsinks showed the lowest thermal resistance while the all-aluminum showed the highest. The all-copper and all-aluminum sinks represented the limiting cases of performance. We can explain this behavior by the higher thermal conductivity of the copper compared with aluminum, resulting in lower thermal spreading resistance in the base, better heat conduction through the fins, and higher overall heat transfer to the ambient air.

With the copper base-aluminum fin heatsink, the copper base decreases the thermal spreading resistance in the base, which in turn helps decrease its overall thermal resistance. With the aluminum base-copper fin heatsink, its overall thermal resistance decreased because the copper used in the fins helps increase fin efficiency.

As seen in the table, on page 52, the average thickness of the flat copper fins is 1.33 mm: It's 1.2 mm for the serrated aluminum fins. Due to this difference in fin thickness, the fin spacing, S, of the copper fins is 2.1mm and the serrated aluminum fins is 2.23 mm. Fig. 7 shows you can expect the pressure drop of the air going through the copper fin heatsinks (aluminum base-copper fin and copper base-copper fin) to be higher than when going through the serrated aluminum fin heatsinks (aluminum base-aluminum fin and copper base-aluminum fin).

For heat source covering 5% of the base plate area, the all-copper heatsink had the lowest thermal resistance, with up to a 28.7% reduction in source thermal resistance as compared with the all-aluminum sink. Yet, the copper heatsink was 3.5 times the weight of the aluminum heatsink. Up to 15% reduction in thermal resistance was achieved by using a copper base-aluminum fin or aluminum base/copper fin sink. The copper base-aluminum fin heatsink had the lowest weight increase: It was only twice as heavy as the all-aluminum heatsink.

For more information on this article, CIRCLE 334 on Reader Service Card