Non-linear fin heat sink

1. Field of the Invention

The present invention relates to methods and apparatuses for heat transfer. More particularly, the invention relates to optimized extended surfaces used for cooling electronic components and other objects whereas such methods and apparatuses involve heat transfer, such as the removal, absorption and/or dissipation of heat.

2. Description of the Related Art

A “heat sink” (alternatively spelled “heatsink”) is a device used for removing, absorbing and/or dissipating heat from a thermal system. Generally speaking, conventional heat sinks are founded on well known physical principles pertaining to heat transference. Heat transference concerns the transfer of heat (thermal energy) via conduction, convection, radiation or some combination thereof. In general, heat transfer involves the movement of heat from one body (solid, liquid, gas or some combination thereof to another body (solid, liquid, gas or some combination thereof). In the present invention the term “heatsink” may also apply to heat exchangers, radiators, air and liquid-cooled coldplates, and other devices through which heat is transferred.

The term “conduction” (or “heat conduction” or “thermal conduction”) refers to the transmission of heat via (through) a medium, without movement of the medium itself, and normally from a region of higher temperature to a region of lower temperature. “Convection” (or “heat convection” or “thermal convection”) is distinguishable from conduction and refers to the transport of heat by a moving fluid which is in contact with a heated body. According to convection, heat is transferred, by movement of the fluid itself, from one part of a fluid to another part of the fluid. “Radiation” (or “heat radiation” or “thermal radiation”) refers to the emission and propagation of waves or particles of heat. The three heat transference mechanisms (conduction, convection and radiation) can be described by the relationships briefly discussed immediately hereinbelow.

Conductive heat transfer, which is based upon the ability of a solid material to conduct heat therethrough, is expressed by the equation q=kA_cΔT/L, wherein: q=rate of heat transfer (typically expressed in Watts) from a higher temperature region to a lower temperature region; k=thermal conductivity (W/m K), which is a characteristic of the material composition; Ac=cross-sectional area (m²) of the material (perpendicular to the direction of heat flow; ΔT=temperature difference (° C.), which is the amount of temperature drop between the higher temperature region and the lower temperature region; and L=length (m) of the thermal path through which the heat is to flow.

Convective heat transfer, which is based upon the ability of a fluid to transfer heat energy through intimate contact with a solid surface, is expressed by the equation q=h_cA_sΔT, wherein: h_c=fluid convection coefficient (W/m² K), wherein h_c is determined by factors including the fluid\'s composition, temperature, velocity, and turbulence; and, A_s=surface area (m²) which is in contact with the fluid.

Radiative heat transfer, which is based upon the ability of a solid material to emit or absorb energy waves or particles from a solid surface to fluid molecules or to different temperature solid surfaces, is expressed by the equation q=A_s∈σ(T_s⁴−T_a⁴), wherein: ∈=dimensionless emissivity coefficient of a solid surface, which is a characteristic of the material surface; σ=Stefan−Boltzmann constant; A_s=surface area (m²) which radiates heat; T_s=absolute temperature of the surface (K); and, T_a absolute temperature of the surrounding environment (K).

It is theoretically understood that, regardless of the heat transfer mechanism, heat transfer rate q can be increased by increasing one or more of the numerator factors on the right side of the equation.

In current practical contexts, heat sinks, coldplates and heat exchangers are generally designed with a view toward furthering the conductive properties of the heat sink by augmenting the thermal conductivity k, for conduction; the surface area, A_s, and heat transfer coefficient hc, for convection. In this regard, according to conventional practice, a heat sink structure is made of a highly thermally conductive solid material, thereby maximizing the conductivity k characteristic of the heat sink; an extended surface comprising a plurality of manufacturable fins or pins, thereby maximizing the surface area A_s; and a geometric shape in contact with the fluid medium, thereby maximizing the heat transfer coefficient h_c.

Following conventional design practice, the heat sink structure tends to be rendered large (e.g., bulky or voluminous), therefore heat sinks are often rated by a heat transfer efficiency, or thermal resistance θ, found by dividing the ΔT, temperature rise of the heat source by the power input, i.e., ° C./W, whereby a lower value for thermal resistance θ, equates to a more efficient design.

As surface area and volume is increased, ancillary issues such as flow resistance and mass must be minimized. In order to gauge these ancillary effects on the efficiency of a heat sink, pumping power P_p, (measured in W), and mass M, (in kg) can be weighted and added to the efficiency equation resulting in η=ΔTP_pM/W. Flow resistance can be particularly important because this resistance increases at the square power of coolant velocity. High flow resistance may require larger pumps or fans to generate additional pumping power, which may also require additional cooling capability.

Due to manufacturing costs, optimized heat sinks are usually limited to a linear array of identical fins having fixed spacing, which are intended to increase the surface area available for heat transfer and increase the heat transfer coefficient. These heat sinks are further compromised by containing simple fin shapes such as squares or rectangles, and occasionally round pins.

Several factors combine to reduce the effectiveness of these conventional heat sinks. One of the most common problems is that the heat absorbed by a coolant media results in a higher temperature media. Due to the temperature rise of the coolant and because a passive heat sink can not cool a heat source below the temperature of the coolant, the temperature of the last device in a row of equally powered components will be hotter than the upstream components. The temperature rise of the coolant is found by ΔT=q/{dot over (m)}c_p, where {dot over (m)}=mass flow rate (kg/s) and c_p=specific heat (J/kg K) of the coolant media. This effect can also greatly change the coolant properties. Therefore, a linear fin array which is optimized for a specific inlet coolant temperature will not provide optimum heat transfer for the coolant after heat is absorbed. An extreme aspect of coolant media property change is in high heat flux applications whereby a saturated liquid enters a heat exchanger, becomes a two-phase flow through nucleate boiling, and subsequent vapor flow.

In addition, heat is not usually spread evenly across the heat input surface of the heat sink. Common practice is to have a plurality of small heat sources share a common heat sink. In such cases, a linear array of fin protrusions will require the same amount of pumping power to flow through the unheated regions as the heated regions.

Thus, there are potential problems associated with conventional approaches to effectuating heat sink cooling of an entity behaving at a high power density. Firstly, prior art manufacturing approaches result in an array of fin protrusions that are more optimized for cost and not for heat transfer. Secondly, a low-cost prior art fin array, consisting of identical fins with identical spacing, will waste pumping power on unheated regions, usually resulting in the need for larger fans or pumps. Thirdly, prior art fin arrays have no provision to account for the temperature rise of a coolant media or the changes in physical properties of the coolant, resulting in decreased efficiency. Fourthly, prior art heat sinks are often grossly overweight, due to the limitations of the manufacturing approach.

Of interest in the art are several United States patents, each of which is hereby incorporated herein by reference Klein et al. U.S. Pat. No. 4,151,548 issued Apr. 24, 1979 teaches the use of square or diagonal cross-section pegs in a fluid flow whereby turbulence is created to enhance cooling. Klein also teaches that opposing inlet and outlet ports cause a higher velocity between the ports. Klein does not teach the use of efficient structures or the role of flow resistance.

Pellant et al. U.S. Pat. No. 4,188,996 issued Feb. 19, 1980 describes a device that contains a plurality of spaced parallel channels. The channels being divided by studs spaced longitudinally in an effort to promote more fluid turbulence. Pellant does not teach the use of efficient structures or the role of flow resistance.

Iversen U.S. Pat. No. 4,712,609 issued Dec. 18, 1987 discloses a roughened heat exchanger surface with a coolant flow heated to boiling and producing pressure gradients to remove nucleate bubbles. Although Iversen teaches that low flow resistance is important, Iversen does not teach, and makes no provision for the fact that the optimum heat transfer surface for liquid flow is very different than the optimum for two-phase and gaseous flow.

Steffen et al. U.S. Pat. No. 4,997,034 issued Mar. 5, 1991 teaches a heat transfer surface consisting of diamond-shaped protrusions on a pie-shaped plate and recognition of manufacturing ease and flow resistance. Steffen does not teach that different aspect ratios will produce different heat transfer and flow resistance results, nor does Steffen teach the use of mixed shapes and heights of protrusions.

Wolgemuth et al. U.S. Pat. No. 5,453,911 issued Sep. 26, 1995 discloses the use of nozzles to cause impingement of a coolant onto the baseplate of an insulated gate bipolar transistor (IGBT) or silicon-controlled rectifier (SCR), and deflectors to cause greater a greater heat transfer coefficient at hot spots. Wolgemuth does not teach the importance of flow resistance, or that gross changes in flow direction and velocity can have a very negative impact on flow resistance, nor does Wolgemuth disclose the use of shaped protrusions to efficiently cool hot spots.

Romero et al. U.S. Pat. No. 5,915,463 issued Jun. 29, 1999 instructs the use of an optimized fin array to cool discrete components and a method of manufacture. Romero asserts that the fin surfaces perpendicular to coolant flow do not significantly contribute to heat transfer, directly contradicting a large body of published literature.