Heatsink providing equivalent cooling for multiple in-line modules

The present disclosure relates to the field of electronic circuits, and specifically to cooling devices used in electronic circuits. Still more specifically, the present invention relates to heatsinks used as cooling devices in electronic circuits

A heatsink structure cools a first electronic chip and a second electronic chip that are positioned serially and in-line to cooling fan(s), where cooling air from the cooling fan(s) passes above the first electronic chip before passing above the second electronic chip. The heatsink structure includes a first heatsink and a second heatsink. The first heatsink is positioned over the first electronic chip and has a specific geometric shape. The second heatsink is positioned over the second electronic chip and has the same geometric shape as the first heatsink. The second heatsink is rotated 180 degrees relative to the first heatsink and pairs together with the first heatsink to form the heatsink structure. The first heatsink and the second heatsink are oriented lateral to each other and orthogonal to the cooling air flow direction such that cooling air from the cooling fan(s) passes across the second heatsink without passing through some or all of the first heatsink.

In an embodiment of the present invention, a circuit board includes the heatsink structure described above.

In an embodiment of the present invention, a computing device includes a cooling fan(s) and a circuit board that includes the heatsink structure described above.

FIG. 1 depicts an exemplary computing device into which a novel heatsink is incorporated;

FIG. 2 illustrates an alternative embodiment of the novel heatsink depicted in FIG. 1 ;

FIG. 3 depicts another alternative embodiment of the novel heatsink depicted in FIG. 1 ;

FIG. 4 depicts exemplary details of the novel heatsink depicted in FIG. 3 into which heatpipes are incorporated; and

FIG. 5 depicts exemplary detail of the heatsinks depicted in FIG. 4 .

The present invention addresses a problem that arises when two electronic components (e.g., integrated circuit (IC) chips such as central processing units (CPUs), graphics processing units (GPUs), etc.) are aligned in series respective to a cooling fan(s). That is, assume that two chips are mounted on a printed circuit board (PCB), and that each of the chips has a finned heatsink mounted on top. Assume further that the two chips are oriented such that the second chip is downstream (downwind) of the first chip, such that the second finned heatsink above the second chip gets air that has been warmed by the first finned heatsink that is above the first chip. Since sufficient cooling must be provided to keep both chips at or below their thermal specification, the fan speed must be increased to keep the second chip within its thermal specification, and as a result provides significant overcooling to the first chip. This overcooling comes at the cost of increased fan speed, increased fan acoustics, increased fan power consumption, and decreased fan reliability.

The amount of preheating from an upwind heatsink is significant. For example, assume that air is receiving bulk heating from a 300 W heat source (i.e., a first module). If 25 CFM of air flows through a first (upwind) heatsink and it picks up 300 W of heat, the air will be ˜22° C. warmer coming out of the first heatsink. This causes a second (downwind) heatsink and its associated module to be 22° C. warmer than the first module. This temperature difference is regardless of the heatsink technology employed (i.e., material choice, heat spreader, heatpipes, etc.).

One solution in the prior art to this problem was to simply install different types of heatsinks on the first and second chips (modules) described above. For example, the first finned heatsink on top of the first chip described above may have smaller and/or fewer fins than the second finned heatsink. This lowers the overall flow impedance of the path, allowing the fan to operate at a lower pressure drop/higher flowrate point on a fan curve. By removing fins on the first heatsink, the temperature of the first heatsink increases, while the added airflow decreases the temperature to the second heatsink, such that the first and second chips can be cooled to roughly the same temperature. However, this requires different types of finned heatsinks to be installed on the PCB, which is more expensive than using a same type of finned heatsink on both chips. Furthermore, if the module and heatsink are stocked as a pre-assembled pair, then upstream and downstream pairs must be stocked worldwide, instead of just a single assembly if the same heatsink is used upstream and downstream.

Having two types of finned heatsinks also creates the risk of installing the two finned heatsinks in the reverse orientation, such that the first chip gets the more efficient finned heatsink and the second chip gets the less efficient finned heatsink. This results in the first chip running very cool, while the second (downwind) chip overheats.

Thus, the present invention provides heatsinks that are geometrically identical (and thus cannot be incorrectly installed on the PCB), and yet still provide the same amount of cooling to both in-line modules.

With reference now to FIG. 1 , an exemplary computing device into which a novel heatsink is incorporated is presented.

As shown in FIG. 1 , the (two-part) heatsink structure 101 is made up of a first heatsink 103 a and a second heatsink 103 b . As depicted in FIG. 1 , the first heatsink 103 a and the second heatsink 103 b are geometrically identical, but second heatsink 103 b has been rotated 180 degrees relative to the first heatsink 103 a , such that the first heatsink 103 a and the second heatsink 103 b nest together to form the complete heatsink structure 101 . In one embodiment, there is an air gap 105 between the first heatsink 103 a and the second heatsink 103 b when nested together, such that there is no direct contact (and thus no metal-to-metal heat transference) between the first heatsink 103 a and the second heatsink 103 b . In a preferred embodiment, air gap 105 is through a mid height of the depicted modules.

The first heatsink 103 a has a first heatsink base 107 a (e.g., a plate of metal, which may include heat transfer structures such as heatpipes, vapor chambers, etc.). Mounted atop the first heatsink base 107 a is a first set of heat radiating fins 109 a , which dissipates heat away from first heatsink base 107 a.

The second heatsink 103 b has a second heatsink base 107 b . Mounted atop the second heatsink base 107 b is a second set of heat radiating fins 109 b , which dissipates heat away from second heatsink base 107 b.

Fins may span the entire left-to-right expanse of the heatsink, or may start and stop somewhat before the left and right edges of the heatsink (as depicted in FIG. 1 ).

First and second heatsinks 103 a - 103 b , including first and second heatsink bases 107 a - 107 b and first and second sets of heat radiating fins 109 a - 109 b , are constructed of any heat conducting material.

The first heatsink 103 a is mounted atop a first electronic chip 111 a (e.g., a central processing unit—CPU, a graphics processing unit—GPU, a network interface card (NIC), etc.). More specifically, a set of securing hardware (not shown) holds the first heatsink 103 a securely against a Thermal Interface Material (TIM) that is atop the first electronic chip 111 a . Similarly, a second heatsink 103 b is mounted atop a second electronic chip 111 b , which may be a same type of chip or a different type of electronic chip than first electronic chip 111 a.

As shown in FIG. 1 , the heatsink structure 101 may be utilized on a circuit board 115 , such as a PCB. Furthermore, the circuit board 115 may be part of a computing device 117 (e.g., a server, a desktop computer, a laptop computer, a smart phone, etc.) that includes cooling fan(s) 113 or another air moving device, such as blowers.

As shown in FIG. 1 , the first heatsink 103 a and the second heatsink 103 b are oriented lateral to each other and orthogonal to the cooling fan flow direction 113 , such that cooling air from the cooling fan(s) 113 passes across the second heatsink 103 b without passing through some or all of the first heatsink 103 a.

With reference now to FIG. 2 , an alternative embodiment of the novel heatsink depicted in FIG. 1 is presented as heatsink structure 201 where electronic chips 111 a and 111 b are spaced farther apart. For purposes of clarity, the first and second sets of heat radiating fins 109 a - 109 b and the cooling fan(s) 113 depicted in FIG. 1 are not shown in FIG. 2 , but should be assumed to still exist in the embodiment depicted in FIG. 2 .

As shown in FIG. 2 , the first heatsink 203 a and the second heatsink 203 b (similar to the first and second heatsinks 103 a - 103 b described in FIG. 1 ) are identical in shape. However, the first heatsink 203 a is rotated 180 degrees relative to the second heatsink 203 b , such that the first and second heatsinks nest together to form heatsink structure 201 .

In FIG. 2 , the first heatsink 203 a is mounted atop the first electronic chip 111 a , and the second heatsink 203 b is mounted atop the second electronic chip 111 b.

Although not shown, other embodiments of the shape of the heatsink between the electronics chips 111 a and 111 b are possible, wherein the heatsinks 203 a and 203 b are of the same shape. For instance, one possibility is that the angled shape is eliminated. Here heatsink 203 a covers the left half of the region between the chips and heatsink 203 b covers the right half of the region between the chips. Alternatively, heatsink 203 a could cover the top half of the region between the chips, while heatsink 203 b could cover the bottom half of the region between the chips. Other configurations are also possible.

As depicted in FIG. 2 , the portion of the first heatsink 203 a within the range depicted as Y 1 completely receives fresh and identical cooling to that which passes over the second heatsink 203 b within the range depicted as Y 3 . However, the portion of the first heatsink 203 a within the range Y 2 receives fresh air, whereas the portion of the second heatsink 203 b within the range Y 2 receives preheated air. When Y 1 and Y 3 are large compared to Y 2 (e.g., Y 1 has an area that is at least twice that of Y 2 and Y 3 also has an area that is at least twice that of Y 2 ), the effect of preheating that occurs in range Y 2 is small.

As shown in FIG. 3 , heatsink structure 301 includes the first heatsink 303 a that cools the first electronic chip 111 a . Similarly, the second heatsink 303 b cools the second electronic chip 111 b . In a preferred embodiment, however, additional structure is used to thermally interface the depicted heatsinks with the electronic chips.

With reference now to FIG. 4 , further details of the embodiment of FIG. 3 is shown which depicts a heatsink structure 401 , which incorporates the use of a thermal enhanced structure such as a set of heatpipes. A first set of heat radiating fins 409 a is mounted atop a first heatsink base, which includes a first set of heatpipes 402 a on the first heatsink 403 a , and a second set of heat radiating fins 409 b are mounted atop a second heatsink base which includes a second set of heatpipes 402 b on the second heatsink 403 b . As depicted, first heatsink 403 a and second heatsink 403 b have the same physical structure/shape (e.g., are identical to one another) but are rotated 180 degrees relative to one another.

As shown in FIG. 4 , the first set of heatpipes 402 a is part of the first heatsink 403 a (which is devoted to cooling the first electronic chip 111 a ). Similarly, the second set of heatpipes 402 b is part of the second heatsink 403 b (which is devoted to cooling the second electronic chip 111 b ). These depicted heatpipes transfer heat from their respective electronic chips 111 a / 111 b . Due to the fact that cool air from the cooling fan(s) passes across the heatpipes and/or their associated heat radiating fins at the same rate and temperature in both the first heatsink 403 a and the second heatsink 403 b , the heat removal rate and/or efficacy of the first heatsink 403 a and the second heatsink 403 b is the same.

In one or more embodiments of the present invention, the first set of heatpipes 402 a extends beyond the perimeter of heatsink 403 a and the second set of heatpipes 402 b extends beyond the perimeter of heatsink 403 b.

With reference now to FIG. 5 , exemplary detail of first heatsink 403 a and second heatsink 403 b shown in FIG. 4 is presented in a front view. Depicted above the first electronic chip 111 a from FIG. 4 is the first set of heatpipes 402 a (also shown in FIG. 4 ), including the front-most heatpipe to be visible from the front view. The first set of heatpipes 402 a is an integral part of first heatsink 403 a (also shown in FIG. 4 ), thus allowing heat to be transferred from the first electronic chip 111 a to the first set of heat radiating fins 409 a in first heatsink 403 a via heatpipe 402 a.

Similarly, a second set of heatpipes 402 b (including the visible front-most heatpipe from the second set of heatpipes 402 b ), which is positioned above the second electronic chip 111 b , carries heat from the second electronic chip 111 b to the second set of heat radiating fins 409 b in second heatsink 403 b.

Note that FIG. 5 is intended as a schematic depiction of the heatpipes. As such, additional mechanical support (not shown) for the first heatsink 403 a and/or the second heatsink 403 b heat sink to the module (e.g., first electronic chip 111 a and/or second electronic chip 111 b ) or the printed circuit board (card) upon which such modules are mounted may be provided.

The first and second sets of heatpipes 402 a / 402 b utilize heatpipes which each include a casing, which is an integral part of heatsinks. Within each casing is a wick, which holds a fluid, in either liquid form or as a vapor.

Thus, as first electronic chip 111 a generates heat, one end of heatpipes from the first set of heatpipes 402 a are also heated, which causes fluid within the casing to transition from a liquid to a vapor. After the fluid turned into a vapor (and thus absorbed thermal energy), it is expelled into from the wick and transferred to the end of the heatpipes from the first set of heatpipes 402 a that are distal to the first electronic chip 111 a . The vapor then cools condenses back into a liquid at the low temperature end of the heatpipes in the first set of heatpipes 402 a . This liquid then migrates back along the wick until it seeps back into the cavity that is near the first electronic chip. The process of vaporizing the liquid in the heatpipes in the first set of heatpipes 402 a back into a vapor to remove heat from the first electronic chip 111 a then repeats in a continuous manner to remove heat from the first electronic chip 111 a . The second set of heatpipes 402 performs the same function for the second electronic chip 111 b.

In one embodiment of the present invention, the first set of heatpipes 402 a is brazed to the bottom of the first heatsink 403 a , and the second set of heatpipes 402 b is brazed to the bottom of the second heatsink 403 b.

In one embodiment of the present invention, the first set of heatpipes 402 a is embedded into the internal structure of the first heatsink 403 a and the second set of heatpipes 402 b is embedded into the internal structure of the second heatsink 403 b.

As described herein in various embodiments, a novel heatsink structure cools a first electronic chip and a second electronic chip that are positioned serially and in-line to a cooling fan(s) by utilizing a first heatsink and a second heatsink that provide equivalent, or nearly equivalent, levels of cooling capacity.

Having thus described embodiments of the present invention of the present application in detail and by reference to illustrative embodiments thereof, it will be apparent that modifications and variations are possible without departing from the scope of the present invention defined in the appended claims.