Radiator for immersion cooling

Abstract

A heat sink for immersion cooling is provided. A heat sink for immersion cooling a processor of an information handling device includes a base, heat dissipation protrusions, and a manifold. The base has a first face to thermally couple to the processor. The heat dissipation protrusions, such as fins, are coupled to and protrude from a second face of the base. The manifold is directly attached to the second face of the base, adjacent to the heat dissipation protrusions. The manifold includes an inlet and a plurality of outlets. When the information handling device is immersed in an immersion cooling tank, the manifold receives an input flow of liquid coolant via the inlet and discharges an output flow of the liquid coolant toward the heat dissipation protrusions via the outlets.

Description

Technical Field

[0001] This disclosure relates to radiators for immersion cooling. Background Technology

[0002] Information processing systems (such as servers and networking devices) are typically equipped with cooling systems to remove heat and maintain the system's components within the desired operating temperature range. Immersion cooling is a cooling method that is gaining popularity as an alternative to traditional air cooling systems and direct liquid cooling (DLC) systems. Immersion cooling involves immersing the information processing system in a non-conductive (e.g., dielectric) liquid coolant, which is typically contained in a large tank capable of housing multiple immersion systems simultaneously. Unlike DLC, where the liquid is contained within a closed loop and is not directly exposed to the system's electrical components, in immersion cooling, the liquid coolant diffuses throughout the system and directly contacts all its exposed components. Therefore, the liquid coolant can absorb heat from virtually all components of the system. The absorbed heat is then removed from the individual system via convection of the liquid through the tank. Coolers (such as heat exchangers) may be positioned at the tank level to remove heat from the liquid. Summary of the Invention

[0003] According to one aspect of this disclosure, a heat sink is provided for immersion cooling of a processor of an information processing device, the heat sink comprising: a base including a first surface configured to be thermally coupled to the processor when the heat sink is mounted in the information processing device; a heat dissipation protrusion coupled to and protruding from a second surface of the base; and a manifold directly attached to the second surface of the base and adjacent to the heat dissipation protrusion, the manifold including an inlet and a plurality of outlets, and the manifold being configured to receive an input flow of liquid coolant via the inlet and discharge an output flow of liquid coolant toward the heat dissipation protrusion via the outlets when the heat sink mounted in the information processing device is immersed in a pool of liquid coolant in an immersion cooling tank.

[0004] According to another aspect of this disclosure, an information processing apparatus is provided, comprising: a chassis; a system board supported by the chassis and including a processor mounted to the system board; an immersion-cooled heat sink mounted to the system board and including: a base including a first side thermally coupled to the processor; a heat dissipation protrusion coupled to and protruding from a second side of the base; and a manifold directly attached to the second side of the base and adjacent to the heat dissipation protrusion, the manifold including an inlet and a plurality of outlets; and a pump coupled to the chassis and fluidly coupled to the inlet of the manifold, wherein, while the information processing apparatus is immersed in a pool of liquid coolant in an immersion tank, the pump is configured to draw in a flow of liquid coolant from the pool and supply the flow as an input flow to the manifold via the inlet, and the outlet of the manifold is configured to discharge an output flow of liquid coolant toward the heat dissipation protrusion.

[0005] According to another aspect of this disclosure, an immersion cooling system is provided, comprising: an immersion cooling tank including a pool of liquid coolant, the immersion cooling tank being configured to receive a plurality of information processing devices immersed in the pool; information processing devices immersed in the pool, the information processing devices including: a chassis; a system board supported by the chassis and including a processor mounted to the system board; and an immersion cooling heat sink mounted to the system board and including: a base, the base including a first surface thermally coupled to the processor. A heat dissipation protrusion coupled to and protruding from a second surface of the base; a manifold directly attached to the second surface of the base and adjacent to the heat dissipation protrusion, the manifold including an inlet and a plurality of outlets; and a pump fluidly coupled to the inlet of the manifold and configured to draw in a flow of the liquid coolant from the pool and supply the flow as an input flow to the manifold via the inlet, the outlet of the manifold being configured to discharge an output flow of the liquid coolant toward the heat dissipation protrusion. Attached Figure Description

[0006] From the following detailed description, it can be seen either alone or in conjunction with appendices. Figure 1 To understand this disclosure, the accompanying drawings are included to provide a further understanding of the disclosure, and these drawings are incorporated in and form part of this specification. The drawings illustrate one or more examples of the teachings and, together with the description, explain certain principles and operations. In the drawings:

[0007] Figure 1 This is a schematic diagram illustrating an immersion cooling system, an information processing device, and an immersion radiator.

[0008] Figure 2 This is a perspective view of an example immersion radiator.

[0009] Figure 3 yes Figure 2 An exploded perspective view of an immersion radiator.

[0010] Figure 4 yes Figure 2 Another exploded perspective view of the submersible radiator.

[0011] Figure 5 yes Figure 2 A top view of an immersion radiator.

[0012] Figure 6 yes Figure 2 Side view of an immersion radiator.

[0013] Figure 7 yes Figure 2 Front view of an immersion radiator.

[0014] Figure 8 This is an exploded perspective view of another example of an immersion radiator.

[0015] Figure 9 yes Figure 8 A perspective view of an immersion radiator.

[0016] Figure 10 yes Figure 8 Another perspective view of the submersible radiator. Detailed Implementation

[0017] Immersion cooling is gaining popularity as an alternative to air cooling and DLC for several reasons. For example, the increasing heat generated by processors (e.g., central processing units (CPUs), graphics processing units (GPUs), etc.) and other components makes it increasingly difficult to efficiently cool systems via airflow. In some high-power systems, fans may not be able to adequately dissipate the generated heat. In other systems, fans may be able to dissipate heat, but doing so may require the fans to operate at high speeds, increasing their power consumption and noise generation, thus making cooling inefficient. In contrast, immersion cooling tanks typically consist of far fewer actively driven (electric) cooling elements, and therefore they can provide cooling with substantially less power consumption and less noise generation.

[0018] For example, the electrically powered cooling components of some immersion cooling tanks may be limited to one or a few pumps, which, for example, drive liquid through the tank and through some bulk circulation of the tank's cooler. A single tank can accommodate multiple servers, in some cases as many as can be accommodated by a conventional rack (e.g., in some cases, approximately twenty 2U servers). Therefore, in some cases, only a few electrically powered cooling components are needed to cool many servers. In contrast, an air-cooled rack with a comparable number of servers as the tank may require dozens of fans—for example, assuming a rack has approximately twenty 2U servers, there could be over one hundred fans. Therefore, immersion cooling tanks will draw significantly less power and generate significantly less noise compared to comparable air-cooled racks.

[0019] Furthermore, DLC typically requires providing a complex closed cooling loop to circulate liquid coolant throughout the information processing system. This cooling loop includes pipes, valves, pumps, cold plates, leak detection devices, and many other components of the liquid cooling infrastructure. In the case of a rack with multiple servers, the entire liquid cooling loop for the rack can comprise multiple sub-loops (one for each individual system within the rack). Therefore, DLC solutions tend to be very expensive and complex to implement, and thus DLC is generally reserved for high-power and high-cost systems, such as high-performance computing (HPC) systems (aka supercomputers), where cost can be reasonable. Often, DLC is used to cool the most powerful components, such as the CPU. This in itself can be expensive. Additionally, in some systems, DLC is used to cool numerous other components distributed throughout the system. This significantly increases the cost and complexity of the cooling solution, as each component may need to be thermally coupled to the liquid cooling loop, and there are many components with irregular shapes, disparate locations, and varying cooling requirements. To reduce some of the cost and complexity, some systems will use DLC to cool only high-power electrical components (e.g., the CPU) and will include air cooling to cool other auxiliary components. While this hybrid approach can be effective, DLC components will still be expensive and complex (albeit not as much as in DLC-only systems), and air-cooled components suffer from the same drawbacks as air-cooled systems (high power consumption and noise generation).

[0020] In contrast, immersion cooling typically eliminates the need for the large and complex custom liquid cooling circuits and associated infrastructure often required for DLCs, making it less costly and complex to implement in the long run. Furthermore, because immersion cooling allows coolant to diffuse throughout the system and reach all exposed components, it eliminates the need for special effort or cost to ensure that non-CPU components are cooled along with the CPU.

[0021] However, while immersion cooling can effectively cool many systems, in some cases, immersion tanks can provide less cooling than might be expected for certain high-power components with concentrated heat generation, such as some very hot processors. In particular, some processors can generate heat at a rate exceeding that of the surrounding liquid, which can absorb and remove heat via natural convection present in the immersion tank. This causes the processor temperature to rise and ultimately suppresses the processor's performance to avoid overheating.

[0022] To address these issues, this paper discloses a heatsink (“immersion heatsink”) specifically designed for immersion cooling of information processing devices, incorporating a forced convection manifold. This immersion heatsink can be configured to cool the processor of an information processing device, such as a CPU, GPU, system-on-a-chip (SoC), application-specific integrated circuit (ASIC), or other processor. The immersion heatsink includes a thermally conductive base having a first side thermally coupled to the processor to absorb heat therefrom. A set of thermally conductive heat dissipation protrusions (e.g., fins) are thermally coupled to the base, wherein these protrusions extend away from the base to a second side. When the information processing device is immersed in a pool of liquid coolant, the second side and the set of protrusions are exposed to the liquid coolant and transfer the heat absorbed by the base to the liquid coolant, thus cooling the processor. The second side and the protrusions together have an exposed surface area in contact with the surrounding liquid, which is significantly larger than the surface area of ​​the processor alone, and this allows the immersion heatsink to conduct heat into the liquid much faster than could be achieved with a single processor. Additionally, immersion heatsinks include a manifold integrated therein. The manifold is attached to a base that receives (by a pump) an inflow of liquid coolant drawn from a surrounding pool of liquid and discharges multiple outflows of liquid coolant from a series of outlets arranged to direct the outflows toward protrusions (e.g., fins). These coolant flows create forced convection through the immersion heatsink, which reduces thermal resistance and increases the rate at which heat is transferred from the protrusions and second surface into the liquid. Therefore, immersion heatsinks allow for a significant reduction in processor temperature compared to passive cooling in an immersion fluid. This can allow for more efficient operation of a given processor, or allow for the utilization of a more powerful processor while maintaining a given efficiency.

[0023] The immersion radiators disclosed herein offer superior performance compared to those achievable using radiators designed for air cooling in immersion cooling systems (“air radiators”). While air radiators may operate in immersion cooling environments, they may not be well optimized for immersion cooling. For example, many air radiators have very fine fin pitches (i.e., small spacing between each fin), which work well in air cooling but may lead to suboptimal results in immersion cooling. For instance, the liquid coolant may have a higher viscosity than air and therefore may not flow very freely through an air radiator. In contrast, the immersion radiators disclosed herein may have heat dissipation protrusions designed to work well in immersion cooling, and in some examples, these protrusions may include fins with a larger fin pitch than some air radiators, allowing better liquid flow through the fins and thus allowing for better cooling performance. In addition, the manifolds integrated into the immersion radiator provide forced convection of liquid through the protrusions, which can greatly improve the thermal performance of the radiator compared to air radiators that would lack such convection.

[0024] In some examples, a system originally designed for air cooling can be retrofitted or converted for immersion cooling, in particular, by removing the air radiator and replacing it with one of the immersion radiators disclosed herein. In some examples, the base of the immersion radiator may include mounting features compatible with the radiator mounting features of the system board, which were originally designed for mounting air radiators. Therefore, additional mounting features do not need to be added to an existing system or designed into a new system to accommodate an immersion radiator.

[0025] Furthermore, because the manifold is integrated into the immersion radiator (i.e., attached to the base), it is automatically installed in the system as part of the immersion radiator mounting process, residing within the space reserved for the radiator. In contrast, using separate radiators and manifolds could require multiple installation processes (one for the radiator, one for the manifold), complicating the installation process. Additionally, the manifold may be located outside the space reserved for the radiator, thus requiring additional space for it, which may be impractical in some space-constrained systems. A separate mounting feature may also be needed to secure the manifold to the chassis, which may be lacking in some existing systems and could increase the cost of designing new systems.

[0026] Furthermore, with the manifold integrated into the immersion radiator, the position of the manifold relative to the protrusion can be precisely controlled. This ensures that the output flow is guided precisely across the protrusion as the radiator designer intended. In contrast, precisely controlling the manifold position is more challenging when using a separate radiator and manifold—installation errors, tolerances, or various other factors can cause some variability in the manifold's positioning relative to the radiator. This can result in a suboptimal output flow relative to the protrusion, potentially reducing cooling performance.

[0027] Turning now to the accompanying drawings, various apparatuses, systems, and methods according to non-limiting aspects of this disclosure will be described.

[0028] Figure 1 This is a schematic diagram of a conceptual illustration system 100, which includes an immersion cooling tank 101 and an information processing device 180 having an immersion radiator 130. Figure 1 It is essentially illustrative and conceptual, and unless otherwise stated, is not intended to illustrate spatial relationships, shapes, dimensions, or other structural details. Furthermore, Figure 1 It is not intended to describe in detail all aspects of system 100, information processing device 180 or immersion radiator 130. Figure 1 The elements illustrated may be included in a different quantity and / or arrangement than those illustrated; some examples may include... Figure 1 Components not shown in the diagram, and / or some examples may be omitted. Figure 1 One or more of the elements shown in the figure. Figure 1 The complete system 100 is illustrated to aid understanding, but some examples disclosed herein may only include an information processing device 180 with a radiator 130 installed (but without an immersion tank 101), while other examples may only include the radiator 130 itself.

[0029] exist Figure 1 In the diagram, physical attachment / coupling between components is illustrated by double solid lines extending between the boxes depicting the components. Thermal coupling (e.g., contact) between components is illustrated by a single solid line extending between the boxes depicting the components, with arrows indicating the direction of heat flow during normal use. Dashed arrow lines extending between the boxes depicting the components are used to depict liquid flow paths and fluid connections between two components, with arrows indicating the direction of flow.

[0030] like Figure 1As shown, the system includes an immersion tank 101 (“Tank 101”). The immersion tank 101 includes a barrel or container or similar structure configured to provide immersion cooling to an information processing device (such as device 180). The tank 101 includes a pool 102 of a non-conductive (e.g., dielectric) liquid coolant 103 (“Liquid 103”). The tank 101 is configured to receive one or more information processing devices (e.g., servers), including information processing device 180 (“Device 180”) immersed in the pool 102. In this context, “immersion” means that at least a system board of the information processing device 180 is immersed in the pool 102. In some examples, the entirety of each information processing device 180 may be completely immersed. In other examples, a small portion of the device 180 (such as a support bracket of a support system) may protrude from the pool 102. The immersion of the device 180 in the pool 102 is... Figure 1 The middle is indicated schematically by a white arrow.

[0031] The immersion tank 101 may also include a cooler 104 configured to cool liquid 103. For example, the cooler 104 may include a heat exchanger that cools liquid 103 by exchanging heat with facility cooling fluid of the facility cooling system. The heat exchanger of the cooler 104 may be, for example, a liquid-to-liquid heat exchanger, in which case the facility cooling fluid may be liquid that runs through the facility in a facility liquid loop to cool one or more immersion tanks (including tank 101). In some examples, the cooler 104 is located outside the pool 102, in which case the cooler may receive a flow of liquid 103 from the pool 102, cool the flow of liquid 103 by exchanging heat with facility cooling fluid, and then return the now-cooled flow of liquid 103 to the pool 102. A tank pump may be provided in the tank 101 to facilitate the flow of liquid 103, which may also boost (or amplify) the overall movement of liquid through the tank 101. In other examples, the cooler 104 may be immersed in the pool 102 to directly cool the surrounding liquid coolant 103.

[0032] Information processing device 180 may include a server, a networking device, or other device configured to process digital information. Device 180 includes a system board 120 (e.g., a motherboard) on which a processor 121 is mounted. The processor 121 may include (or be a subset of) a CPU, GPU, SoC, ASIC, or any other processor. System board 120 may also include other electronic components (not shown), including one or more additional processors, memory modules, etc. System board 120 also includes mounting features 122 arranged around the processor 121 and configured to mount a heat sink (such as an immersion heat sink 130), as will be described in more detail below.

[0033] The device 180 may also include a chassis 110 that supports the system panel 120. In some examples, the chassis 110 houses (encloses) the system panel 120 and other components of the device 180, while in other examples, the chassis 110 merely supports the system panel 120 without housing it (e.g., the chassis 110 may have a tray configuration). In other examples, the chassis 110 may be omitted. As previously mentioned, Figure 1 It is essentially schematic and therefore does not show structural details of the chassis 110 or system board 120 or their positional relationship to each other or other components.

[0034] The immersion radiator 130 (“radiator 130”) is configured to cool the processor 121 when installed in the information processing device 180. Figure 1 A radiator 130 is shown in its installed state as part of device 180. In other examples, the radiator 130 is not installed and is not yet part of device 180. The submersible radiator 130 includes a base 140, a manifold 160, and a mounting feature 142.

[0035] The base 140 includes a heat-conducting body having a generally plate-like shape, meaning that the body has two wide, generally planar faces arranged opposite and parallel to each other, and a number of narrow edge surfaces extending between and connecting the two wide planar faces. Furthermore, both the width and length dimensions of the body parallel to its wide faces are significantly greater than (e.g., 5 times or more) the thickness dimension of the body perpendicular to its wide faces. In some examples, the base 140 is formed of a plate of a heat-conducting metal (such as aluminum, copper, steel, etc.). In some examples, the base 140 includes a vapor chamber comprising a sealed internal volume within the base, in which a working fluid is contained and configured to transfer heat between the wide faces of the base 140 through a phase change circulation of the working fluid.

[0036] The plate-like body of the base 140 has two wide surfaces, including a first surface 143 and a second surface 144. In the installed state, the first surface 143 is thermally coupled to the processor 121 (e.g., positioned in contact with it), such that heat generated by the processor 121 is transferred to the base 140 via conduction. Figure 1 The solid arrow indicates this. Heat is then transferred from the first surface 143 through the thickness of the base 140 to the second surface 144. In an example where the base 140 comprises a solid metal plate, heat is transferred through the thickness of the base 140 via conduction. In an example where the base 140 comprises a heat spreader, heat is transferred through the thickness of the base 140 via conduction and vapor change heat transfer cycles.

[0037] The heat sink 130 further includes a set of heat dissipation protrusions 150 (“protrusions 150”). Each of these protrusions 150 includes a body of thermally conductive metal (such as aluminum, copper, etc.) physically attached (directly or indirectly) to the base 140, wherein the protrusion 150 protrudes from the second surface 144 in a direction perpendicular to the second surface 144 and away from the first surface 143 (or in other words, along the height dimension of the heat sink 130). The protrusions 150 are also thermally coupled to the base 140, and thus heat absorbed by the base 140 from the processor 121 is transferred from the base 140 to the protrusions 150 by conduction.

[0038] The protrusions 150 are arranged in an array, wherein the array includes protrusion positions arranged in rows and columns, wherein the rows extend parallel to the width dimension of the heat sink 130 and the columns extend parallel to the length dimension of the heat sink 130. In some examples, the array of protrusions 150 is a one-dimensional array with only one row of protrusion positions. In other examples, the array of protrusions 150 may be a two-dimensional array with multiple rows of protrusion positions. In some examples, where the protrusions 150 are in a two-dimensional array, some rows are offset relative to other rows, meaning that each position within a row is not necessarily aligned with a corresponding position in another row (e.g., in the same column), and / or some columns are offset relative to other columns, meaning that each position within a column is not necessarily aligned with a corresponding position in another column (e.g., in the same row). Note that the references to one-dimensional and two-dimensional arrays herein refer to the location of the protrusions 150 and not to the structure of the protrusions 150 themselves, which are three-dimensional.

[0039] The protrusions 150 define channels extending along the length of the heatsink 130. These channels extend between adjacent rows of protrusions 150. The channels form liquid flow paths through which liquid 103 can flow, thereby allowing liquid 103 to contact the protrusions 150 and absorb heat therefrom. Thus, heat absorbed from the processor 121 is dissipated through the protrusions 150 into the surrounding liquid 103.

[0040] In some examples, the heat dissipation protrusion 150 includes fins, which are relatively thin and planar sheets or plates of thermally conductive metal (such as aluminum, copper, etc.). In such examples, the fins are physically attached to the base 140, wherein the fins are oriented perpendicular to the base 140 and parallel to each other. (As used herein, the orientation of the planar body (e.g., the fins and base 140) corresponds to the orientation of the wide face of the planar body). These fins protrude from the second face 144 along the height dimension of the radiator 130 and extend longitudinally along the length dimension of the radiator 130, wherein each pair of adjacent fins defines one of the aforementioned channels therebetween. The fins may be arranged in a one-dimensional array, meaning that the fins are arranged at locations spanning a width dimension distribution of the radiator 130. In other words, there may be a row of fins, where each fin location corresponds to a different column of the one-dimensional array. Although the fin locations span a width dimension distribution, the fins themselves may extend longitudinally along the length dimension. The fins may be formed together with a base (e.g., base 140 or another base) by extrusion, scraping, casting, or other manufacturing processes. For example, in some examples, the base and the fins are part of the same extruded metal body. Alternatively, the fins may be joined to the base, such as by welding, brazing, etc.

[0041] In some examples, the heat dissipation protrusion 150 includes pins or pin-shaped fins. Pins include columnar protrusions, such as cylinders, square pillars, or any other type of pillar. These pins can be arranged in a two-dimensional array, for example, with multiple rows and columns of pins. Pin-shaped fins include protrusions that are essentially generally planar (similar to fins) but relatively shorter along their length dimension compared to standard fins. For example, pin-shaped fins may have a length less than their height, while regular fins typically have a length exceeding their height (referring to the direction of projection from the base). Furthermore, pins or pin-shaped fins can be arranged in a two-dimensional array (e.g., multiple rows and columns of pins or pin-shaped fins), while regular fins are typically arranged relative to each other along one dimension (e.g., multiple parallel fins in a row). In some examples, pins or pin-shaped fins can be arranged with a thicker pitch along one or more dimensions than the pins or pin-shaped fins of some air radiators (e.g., larger spacing between rows and / or columns of pins or pin-shaped fins), thereby allowing better liquid flow and thus better cooling performance.

[0042] In some examples, the protrusion 150 is directly attached to the second face 144. For example, in some examples, the base 140 and the protrusion 150 are all integrally formed as part of the same monolithic body. Such a monolithic body can be formed, for example, by extrusion or machining from the same monolithic block of material, by casting or 3D printing together as a single part, or by other manufacturing techniques. In other examples, the protrusion 150 and / or the base 140 may comprise one or more portions that are initially formed separately and then later joined together (e.g., by brazing, welding, or other fastening / joining techniques).

[0043] In some examples, the base 140 comprises two (or more) portions that are initially separate and later joined together to form the base 140. In some of these examples, a protrusion 150 is attached to one of these portions. For example, in some examples, an insert assembly comprising an insert base and protrusions 150 may be formed, wherein the insert base is a first portion of the base 140, and wherein protrusions 150 are attached to the insert base. The insert base may be a plate-like structure. Protrusions 150 may be formed integrally with the insert base, or they may be formed separately from the insert base and then later joined to the insert base. The insert base (the first portion of the base 140) may then be joined to a second portion of the base 140 by brazing, welding, adhesives, mechanical fasteners, friction fits, or other fastening / joining techniques, thus completing the formation of the base 140. The second portion of base 140 may, for example, define the outer periphery of base 140 and may include mounting features 142 and may have a manifold 160 (described below) attached thereto, while the first portion of base 140 may occupy a more central region of base 140 and carry protrusion 150. For example, the second portion of base 140 may have a central recess recessed from a second face 144 into which the first portion of base 140 (i.e., insert base) is inserted. In some examples, the central recess may extend partially through the thickness of the second portion of base 140, while in other examples, the central recess may extend all the way through the thickness of the second portion of base 140. In some examples, the first face 143 may be formed entirely by the second portion of base 140, while in other examples, the first face 143 may be formed partially by the first portion of base 140 (e.g., by the bottom surface of the insert base of the insert assembly). In some examples, the first portion (insert base) and the second portion of the base 140 may be formed of the same material, while in other examples they may be formed of different materials.

[0044] The presence of multiple separately formed portions of the base 140 facilitates the formation of the heat sink 130 from a variety of different materials, allowing for a desired balance between strength, thermal performance, cost, and other factors. For example, copper can be used in the insert assembly to take advantage of its high thermal conductivity, while aluminum can be used in the second portion of the base 140 to take advantage of its lower cost and high strength. Of course, this is just one example, and any desired combination of materials can be chosen, each offering its own advantages and / or disadvantages in various situations. Furthermore, even when using the same materials, the insert assembly can be formed more easily, faster, or otherwise more conveniently separate from the second portion of the base 140, for example, due to the use of different manufacturing processes to form these portions.

[0045] Manifold 160 includes a hollow manifold body having an internal chamber (e.g., including orifices or empty volumes within the body), an inlet 161 fluidly coupling the internal chamber to an external environment, and a plurality of outlets 162 fluidly coupling the internal chamber to the external environment. Thus, manifold 160 is configured to allow an input flow 103a of liquid 103 to flow into the internal chamber via the inlet 161 and then output the liquid 103 back to the external environment as a plurality of output flows 103b from the internal chamber via the outlets 162. The manifold body extends along the width dimension of radiator 130, adjacent to one end of the set of protrusions 150, and the outlets 162 are distributed across this width dimension such that each outlet 162 faces a different subset of the protrusions 150 and / or a subset of the channels defined between the protrusions 150. The number and location of the outlets 162 may vary from one embodiment to another. In some examples, some protrusions 150 and / or channels do not have an opposite outlet 162.

[0046] Inlet 132 may include an opening in the manifold body that fluidly connects inlet 132 to an internal chamber. Inlet 132 may further include a liquid coupling or fitting configured to couple with a liquid conduit 141 (e.g., a hose, pipe, tube, or any other liquid conduit). Thus, inlet 132 may include coupling features to facilitate this coupling. Examples of such liquid couplings / fittings include hose barbs, threaded fittings, compression fittings, quick-disconnect (QD) fittings, and other liquid couplings / fittings. Examples of coupling features for these fittings may include barbs, threads, washers, quick-disconnect features (springs, lift valves, etc.). Outlet 162 may include openings through the manifold body that fluidly connect outlet 162 to an internal chamber. Outlet 162 may further include nozzles configured to focus the output flow 103b to increase its velocity and / or reduce its dispersion (i.e., make it more columnar).

[0047] Outlet 162 is positioned such that outlet flow 103b is directed toward protrusion 150. More specifically, in some examples, outlet 162 directs outlet flow 103b into and through channels between adjacent columns of protrusion 150 (e.g., through channels between adjacent fins). In some examples, some channels do not have outlet flow 103b directed through them, but at least some of the channels do receive such outlet flow 103b. The forced convection of liquid 103 around protrusion 150 (generated by these outlet flows 103b through the channels) reduces thermal resistance and increases the rate at which heat is transferred into liquid 103.

[0048] Some of the heat absorbed by the base 140 is transferred to the protrusions 150, from which it dissipates into the liquid 103. However, some of the heat absorbed by the base 140 can be transferred from the base 140 into the liquid 103 without flowing into the protrusions 150. In some examples, the outlet flow 103b also helps to increase the rate at which such heat is transferred directly from the base 140 into the liquid 103. In some examples, some (or all) of the outlet flow is slightly angled toward the base 140, such that the outlet flow 103b flows both longitudinally (through the channel) and toward the base 140. Thus, a portion of the outlet flow 103b may contact or flow adjacent to the base 140, which can help to facilitate a faster transfer of heat directly from the base 140 into the liquid 103.

[0049] Manifold 160 is attached to the second surface 144 of base 140. In some examples, at least a portion of manifold 160 is integrally connected to base 140, meaning they are formed as part of the same monolithic (integral) body. For example, a portion of manifold 160 and base 140 may be cast together in the same casting, formed together in an additive manufacturing process (e.g., 3D printing), etc. In other examples, manifold 160 may be formed as a separate part, which is attached to base 140, for example, via fasteners, thermal riveting, adhesives, welding, brazing, or other fastening mechanisms. Manifold 160 is attached to base 140 such that outlet 162 is aligned relative to base 140, so that output flow 103b is guided as described above.

[0050] In some examples, manifold 160 is formed as two (or more) parts coupled together. This allows for easier manufacturing. For example, if manifold 160 is manufactured as a single integral part, the internal chambers and other internal features within the manifold body may be difficult to manufacture, but if manifold 160 is manufactured as two separate parts coupled together, they can be manufactured very easily. In some examples, these parts may include a bottom portion attached to a second face 144 of base 140 and a top portion coupled to the bottom portion, wherein the internal chamber of the manifold is partially defined by the walls of the bottom portion and partially by the walls of the top portion. In other words, in these examples, the seam or connecting surface where the top and bottom portions are joined divides the internal chamber in two. This makes the manufacturing of the internal features of the chamber easier, as these features become external when formed in one of the parts and only become internal to the manifold once the parts are coupled together. In some examples, manifold 160 may be metal, and its parts may be cast, 3D printed, or manufactured using some other suitable technique. In some examples, the manifold may be plastic, and its parts may be injection molded, 3D printed, or manufactured using some other suitable technique. In some examples, the parts of manifold 160 are connected together by fasteners (such as screws). In some examples, the same fasteners that couple the parts of manifold 160 together may also couple manifold 160 to base 140.

[0051] The base 140 may further include mounting features 142 configured to mount the base 140 to the system board 120 and / or the chassis 110. In some examples, these mounting features 142 may include holes arranged to receive fasteners (such as screws). As another example, the mounting features 142 may include latching mechanisms formed into or attached to the base, such as half-turn or quarter-turn latches or any other desired latches. As another example, the mounting features 142 may include features configured to receive or engage with retainers (e.g., spring clips) attached to or movable to engage with the base 140. In some examples, the mounting features 142 are configured to engage with mounting features 122 of the system board 120. For example, if the mounting features 142 include holes for receiving fasteners, then the mounting features 122 may include holes in a corresponding position and configured to receive the same fasteners. As another example, if mounting feature 142 includes a latch, then mounting feature 122 may include a latch receptacle or a complementary latch element configured to engage with the latch.

[0052] In some examples, mounting feature 122 may be the same mounting feature used for mounting an air-cooled heatsink to processor 121. In such examples, mounting feature 142 of immersion heatsink 130 may be configured to mount to these same mounting features 122 instead of an air-cooled heatsink. In other words, mounting feature 142 of immersion heatsink 130 may have a location and attachment mechanism that mimics that of an air-cooled heatsink. This is useful in situations where it is desirable to convert a system originally designed for air cooling into one that can be used for immersion cooling. For example, to convert a system, fans, air-cooled heatsinks, airflow baffles, and other air-cooling fixtures that may already be part of the system may be removed and replaced with an immersion heatsink 130.

[0053] The device 180 may also include a pump 170 and liquid conduits (not shown) configured to supply an inlet flow 103a of liquid coolant 103 to a manifold 160. The pump 170 is configured to draw some of the liquid coolant 103 from the pool 102 and force the liquid 103 into and through the manifold 160, thereby generating an outlet flow 103b that is ejected from the outlet 162 toward the protrusion 150, as described above. Figure 1 As indicated by the dashed arrow in the diagram. Specifically, pump 170 may be coupled directly to inlet 161 or via one or more liquid conduits. Although in Figure 1 Not shown, but the output stream 103b ejected from outlet 162 returns to pool 102. Pump 170 can be powered by the same power source (not shown) (such as a power supply unit) that powers system board 120.

[0054] In other examples, pump 170 may be omitted from device 180, and instead, pump may be provided separately from device 180 to circulate liquid. For example, an immersion cooling tank may have one or more tank pumps to circulate liquid through the tank, and one of these tank pumps may be fluidly coupled to manifold 160 to supply it with input flow 103a.

[0055] Turn now Figure 2-7An example immersion radiator 230 will be described. Immersion radiator 230 is an example embodiment of the immersion radiator 130 described above. Therefore, some components of immersion radiator 230 are similar to corresponding components already described above (e.g., example embodiments or configurations), and thus the description of components of immersion radiator 130 above applies to similar components of immersion radiator 230, and therefore repetitive descriptions of certain aspects of immersion radiator 230 can be omitted. Corresponding components may be referred to using reference numerals with the same last two digits (e.g., 150 and 250). It should be understood that immersion radiator 230 is merely one possible embodiment of immersion radiator 130, and immersion radiator 130 is not limited to immersion radiator 230. Similarly, the various components of immersion radiator 230 are examples of corresponding components of immersion radiator 130, but the various components of immersion radiator 130 are not limited to the corresponding components of immersion radiator 230.

[0056] The order in which the components of the submersible radiator 230 are described below is chosen to aid in understanding the structural and functional relationships between them. However, the logical order that best facilitates this understanding is not necessarily the same as the order in which the components appear in the accompanying drawings. Therefore, the components will not be described in strict order below. Figure 2-7 Instead, when describing various aspects of an immersion radiator, certain figures deemed most helpful for understanding the particular aspect under discussion will be invoked, even if this requires moving back and forth between the figures in an unordered manner.

[0057] like Figure 2-7 As shown, the submersible radiator 230 includes a base 240, a manifold 260 coupled to the base 240, and fins 250 coupled to the base 240.

[0058] In this example, the base 240 includes a metal (e.g., aluminum) plate having a first surface 243 and a second surface 244 opposite to the first surface 243. The first surface 243 is configured to be thermally coupled to a processor of the information processing device, while the second surface 244 is configured to be exposed to the liquid coolant when the information processing device is immersed in a pool of liquid coolant in an immersion cooling tank. In some examples, a thermal interface material (TIM) may be disposed on the first surface 243. The TIM may be, for example, a graphite pad / mesh, a stretchable metal foil (e.g., indium foil), a carbon nanotube mesh, or any other TIM that is compatible with the liquid coolant (e.g., not dissolved by or reacting with the liquid coolant). When the base 240 is mounted on the processor, the TIM (if present) sits between and in contact with the first surface 243 and the processor, such that the TIM thermally couples the first surface 243 to the processor. In some examples, the TIM may be omitted, in which case the first surface 243 may directly contact the processor.

[0059] The base 240 includes mounting features 242. In this example, there are four sets of mounting features 242 (one set at each of the four corners of the base 240). In this example, each set of these mounting features 242 includes three mounting features 242 in the form of orifices through the base 240. These mounting features 242 can be configured to receive fasteners to secure the base 240 to the system board. In some examples, the fasteners include screws, studs, bolts, etc. In other examples, the fasteners include latching devices that are mounted to the base 240. For example, each set of three mounting features 242 may receive one such latching device, wherein two of the mounting features in that set secure the latching device to the base 240 and the third mounting feature 242 in that set allows the latching device to engage with a mounting feature of the system board. The mounting features 242 can be configured to allow engagement with an air-cooled radiator mounting feature of the system board, meaning that an air-cooled radiator will be mountable to a mounting feature on the system board.

[0060] A group of multiple fins 250 are coupled to the base 240, such as Figure 2 As shown (only a few of the illustrated fins 250 are labeled). Fin 250 is an example embodiment of protrusion 150. Fin 250 is a relatively thin and flat sheet or plate of thermally conductive metal (such as aluminum, copper, etc.). Fins 250 are oriented perpendicular to the base 240 and parallel to each other. Figure 2 As shown, fins 250 protrude from the second surface 244 along the height dimension of the heat sink 230 and extend along the length dimension of the heat sink 230. Each pair of adjacent fins 250 defines a channel 251 therebetween, as shown in the figure. Figure 5 and Figure 7 As shown (only one of channels 251 is labeled). Fins 250 can be arranged in a one-dimensional array, with the positions of these fins spanning a width dimension distribution of the heat sink 230. Although the fin positions span the width dimension distribution, the fins 250 themselves extend longitudinally along the length dimension, as shown. In this example, the fins 250 are integrally connected to the base 240, thereby forming a monolithic body. For example, the base 240 and the fins 250 can be integrally formed together by extrusion, scraping, casting, or other manufacturing processes. In some examples, the base 240 and the fins 250 are part of the same monolithic extruded metal body. In some examples, the fins 250 have a fin pitch that is coarser (i.e., wider fin spacing) than the air-cooled fins of an air-cooled processor heat sink.

[0061] like Figure 2-4 As shown, manifold 260 includes manifold body 266 (“body 266”), support member 267 coupled to body 266, inlet 261 coupled to body 266, and four outlets 262 coupled to body 266. Figure 3-4 As best seen in this example, the manifold body 266 is formed of two distinct parts—a bottom body portion 266a and a top body portion 266b. These two body portions 266a and 266b are joined together at a joint 269, as shown... Figure 2 As shown in the diagram. The body 266 is hollow, with the walls of the body 266 defining an internal cavity 263 within the body 266. The bottom body portion 266a defines the lower half 263a of the cavity 263 (see Figure 266). Figure 3 The top body portion 266b defines the upper half 263b of the chamber 263 (see...). Figure 4 This configuration allows the lower half 263a and the upper half 263b to join together and form a chamber 263 when the two parts 266a and 266b are joined. This two-part structure of the body 266 makes it easier to form the internal chamber 263 because each half 263a and 263b can be formed as an external feature on the corresponding body part 266a or 266b, rather than as an internal feature (in some cases, internal features may be more difficult to manufacture).

[0062] Support member 267 is coupled to the bottom body portion 266a of manifold body 266. In the illustrated example, support member 267 is integrally formed with bottom body portion 266a, i.e., they are part of the same monolithic body. In other examples, support member 267 may be different from bottom body portion 266a and coupled to it.

[0063] like Figure 3 and Figure 4 As shown, the top body portion 266b includes holes 268b at its opposite ends to receive fasteners (e.g., screws). The bottom body portion 266a also includes holes 268a, which are aligned with holes 268b and are also configured to receive the same fasteners. Holes 268a also extend through two of the supports 267 (e.g., Figure 4 As shown in the figure), and aligned with the corresponding hole 248 in the base 240 (as shown in the figure). Figure 3 As shown in the diagram, these corresponding holes are configured to receive the same fasteners. Therefore, there are two sets of aligned holes (hole 268a aligned with hole 248, and hole 268b aligned with hole 248), each set arranged to receive a fastener. When the fastener is inserted through these respective sets of aligned holes, it mechanically fastens the top body portion 266b and the bottom body portion 266a together, thereby forming the manifold body 266, and simultaneously mechanically fastening the manifold 260 to the base 240. In some examples, the fastener may include a screw, and the hole 248 may include threading to engage with the screw's threads.

[0064] like Figure 3 and Figure 4 As shown, inlet 261 and outlet 262 are fluidly coupled to the internal chamber 263 of manifold 260. Specifically, inlet 261 and outlet 262 each include a first opening through the wall of manifold body 266 to fluidly connect with the internal chamber 263. Furthermore, inlet 261 includes a fluid coupling element coupled to the manifold body through its first opening. Outlet 262 further includes nozzles coupled to the manifold body 266 through their respective first openings. The fluid coupling element of inlet 261 and the nozzle of outlet 262 each define an internal passage fluidly coupled to the first opening of the corresponding inlet 261 or outlet 262. Thus, internal chamber 263 fluidly connects inlet 261 to each of outlets 262. Additionally, the internal passages of inlet 261 and outlet 262 also include a second opening fluidly connecting to the external environment. Thus, inlet 261 and outlet 262 fluidly couple internal chamber 263 to the external environment. Therefore, the liquid input flow 203a from outside the manifold 260 (e.g., from the pump) can enter the manifold 260 by flowing into the internal channel of the inlet 261, flowing from the inlet 261 into the chamber 263, flowing through the chamber 263, flowing from the chamber 263 into the internal channel of the outlet 262, and then being ejected from the outlet 262 as a plurality of output flows 203b.

[0065] In the illustrated example, inlet 261 and outlet 262 are integrally formed with body 266, such as via additive manufacturing, casting / molding, etc. More specifically, inlet 261 is formed of two connected parts, namely a bottom inlet portion 261a and a top inlet portion 261b. The bottom inlet portion 261a is integrally formed with the bottom body portion 266a, and the top inlet portion 261b is integrally formed with the top body portion 266b, as shown. Figure 3 and Figure 4 As shown in the example, the fluid coupling portion of inlet 261 takes the form of a hose clamp fitting configured to couple with a hose, which in turn couples to a pump that pumps liquid coolant from the submerged tank into manifold 260. In other examples, the fluid coupling portion of inlet 261 can be replaced by any other desired fluid coupling, such as a threaded coupling, a quick-disconnect fluid coupling, or others.

[0066] The outlet 262 is also formed by two freely connected parts, namely the bottom outlet part 262a and the top outlet part 262b. The bottom outlet part 262a is integrally formed with the bottom body part 266a, and the top outlet part 262b is integrally formed with the top body part 266b, as shown. Figure 3 and Figure 4As shown in the diagram. Each outlet 262 includes an opening through the wall of the manifold body 266 and an internal conduit fluidly coupled to the opening, such that the internal chamber 263 of the manifold body 266 is fluidly coupled to the external environment via the outlet 262.

[0067] The outlet 262 is arranged adjacent to one side of the fin 250 and spans the width of the radiator 230, as shown in the figure. Figure 5 As shown in the example. In this example, outlet 262 includes a nozzle that is angled toward fin 250 and base 240. More specifically, as shown in the example. Figure 5 and Figure 6 As shown, the central axis 262' of each outlet 262 extends at an acute angle θ parallel to the fins 250 and relative to a horizontal plane 99 parallel to the second surface 244 of the base 240. Therefore, as the flow 203b flows through the channel 251 between the fins 250, a portion of each output flow 203b may impinge on the surface 244 of the base 240, which redirects the flow so that it subsequently flows approximately parallel to the surface 244. Thus, the output flow 203b can contact both the fins 250 and the surface 244, thereby absorbing heat directly from both. Additionally, the output flow 203b may cause other surrounding liquids to circulate, for example, by entraining liquid to flow with the flow 203b. Other liquids may also flow independently of the output flow 203b around the fins 250 and through the channel 251, for example, due to natural convection and / or the overall movement of liquid through the system (due to a tank or system-level pump). However, the forced convection provided by the output flow 203b can provide much greater cooling than natural convection alone or the overall movement of the liquid. In some instances, θ is approximately 25°. In some instances, θ is between 0° and 30°. In some instances, θ is between 30° and 60°.

[0068] In the illustrated example, each outlet 262 is centered around the corresponding one of the fins 250, as shown below. Figure 5 and Figure 7 As shown in the diagram. Therefore, the output flow 203b from the given outlet 262 is divided into two flows 203b by the corresponding fin 250. These two flows flow along two channels 251 adjacent to the corresponding fin 250, as shown in the diagram. Figure 5 As shown in the illustration. In the example shown, not every channel 251 receives output stream 203b from outlet 262.

[0069] The quantity and type of outlets 262 can vary from one example to another depending on the specific system in which the radiator 230 is designed. In some examples, different types of manifolds 260 with different quantities and / or types of outlets 262 can be produced, and each can be mounted to the base 240 using the mounting arrangement described above. This allows system designers to customize the radiator 230 as needed by selecting the manifold 260 that best suits the system.

[0070] In the illustrated example, there is an equidistant spacing between the outlets 262 along the width dimension; however, in other examples, the spacing between the outlets 262 may not be equidistant. In the illustrated example, there is a row of outlets 262 extending along the width dimension, where all outlets 262 have similar heights above the surface 244 (along the height dimension); however, in other examples, one or more outlets 262 may be positioned at a different height above the surface 244 than one or more other outlets 262. In the illustrated example, all outlets 262 discharge liquid at the same angle θ relative to the horizontal plane 99; however, in other examples, one or more outlets 262 may discharge liquid at a different angle than one or more other outlets 262. For example, in one embodiment, a first set of outlets 262 may discharge liquid at a first angle θ1, while a second set of outlets 262 may discharge liquid at a second angle θ2, where θ2 < θ1, such that the central axis of the liquid flow ejected from the first set of outlets 262 intersects the closer portion of the surface 224, while the central axis of the liquid flow ejected from the second set of outlets 262 intersects the farther portion of the surface 224. In the illustrated example, all outlets 262 have the same shape and configuration as each other, including output openings of the same size and shape, and therefore all liquid jets have substantially similar characteristics (e.g., the same flow width, the same flow velocity, etc.). In other examples, one or more outlets 262 may have a different shape or configuration than one or more other outlets 262, and therefore their respective jets may differ from each other in one or more characteristics. For example, the first set of outlets 262 may have a larger output opening than the second set of outlets 262, resulting in a larger diameter and slower flow ejected from the first set of outlets 262, and a smaller diameter and faster flow ejected from the second set of outlets 262. In the illustrated example, the cross-section of the opening of the outlet 262 is circular, so each outlet 262 generates a single flow with a substantially circular cross-section. In other examples, the opening may have other cross-sectional shapes, such as elliptical, rectangular (groove), etc., which can produce output flows of different shapes. Furthermore, in some examples, each outlet 262 may be provided with more than one opening, which allows multiple smaller output flows to be generated from a single outlet 262.

[0071] Turn Figure 8-10Another example of an immersion radiator 330 will be described below. Immersion radiator 330 is an example implementation of the immersion radiator 130 described above. Therefore, some components of immersion radiator 330 are similar to corresponding components already described above (e.g., example implementations or configurations), and thus the description of components of immersion radiator 130 above applies to similar components of immersion radiator 330, and therefore repetitive descriptions of certain aspects of immersion radiator 330 can be omitted. Corresponding components may be referred to using reference numerals with the same last two digits (e.g., 150 and 350). It should be understood that immersion radiator 330 is merely one possible implementation of immersion radiator 130, and immersion radiator 130 is not limited to immersion radiator 330. Similarly, the various components of immersion radiator 330 are examples of corresponding components of immersion radiator 130, but the various components of immersion radiator 130 are not limited to the corresponding components of immersion radiator 330.

[0072] The submersible radiator 330 includes a base 340, fins 350, and a manifold 360.

[0073] The radiator 330 differs from the radiator 230 described above at least in that, in the radiator 330, the base 340 is formed of two distinct portions. Specifically, the base 340 is formed of a base periphery 349 and an insert base 353, which is part of the fin insert assembly 355. The base periphery 349 has a generally similar shape to the base 240 described above, except that a socket 346 is formed therein, wherein the socket 346 includes an aperture extending through the thickness of the base periphery 349. The socket 346 is dimensioned and shaped to receive the insert base 353. Note that in the exploded view, the fin insert assembly 355 is shown above the base periphery 349, but in some examples, the fin insert assembly 355 may be inserted into the socket 346 from below rather than from above. Once inserted into the socket 346, the insert base 353 can be coupled to the base periphery 349, for example, by welding, brazing, friction fitting, adhesive, or other fastening / connection techniques. Once coupled, the base periphery 349 and the insert base 353 together form the base 340. Therefore, the bottom surface 343 of the base 340 includes two sections, one of which is formed by the bottom surface 343a of the base periphery 349, and the other of which is formed by the bottom surface 343b of the insert base 353.

[0074] As previously described, the heat sink 330 includes a fin insert assembly 355, which includes an insert base 353. Additionally, the fin insert assembly 355 includes a plurality of fins 350 coupled to the insert base 353, such as... Figure 8As shown (only one of the illustrated fins 350 is labeled). Fin 350 is an example embodiment of protrusion 150. Fin 350 is a relatively thin and flat sheet or plate of thermally conductive metal (such as aluminum, copper, etc.). Fins 350 are oriented perpendicular to the insert base 353 and parallel to each other. Figure 3 As shown, fins 350 protrude from the top surface (not marked) of the insert base 353, which faces the bottom surface 343b. Fins 350 protrude from the top surface along the height dimension of the radiator 330 and also extend along the length dimension of the radiator 330. Each pair of adjacent fins 350 defines a channel therebetween. The fins 350 can be arranged in a one-dimensional array, with their positions spanning a width dimension distribution of the radiator 330. Although the fin positions span the width dimension distribution, the fins 350 themselves extend longitudinally along the length dimension, as shown. In this example, the fins 350 are integrally connected to the insert base 353 to form a single integral body, namely the fin insert assembly 355. For example, the insert base 353 and the fins 350 can be integrally formed together by extrusion, scraping, casting, or other manufacturing processes. In some examples, the insert base 353 and the fins 350 are part of the same integral extruded metal body.

[0075] In this example, fin 350 has a finer pitch (narrower inter-fin spacing) than fin 250. However, this is just one example, and in other examples, fin 350 may have the same pitch as fin 250, a thicker pitch (wider spacing) than fin 250, or a larger pitch than fin 250. Figure 8-10 The pitch shown is a finer pitch. In fact, one advantage of using a fin insert assembly 355 separate from the base periphery 349 is that the fins 350 can be easily customized to suit a specific application in which the heat sink 330 is deployed. For example, multiple different fin insert assemblies 355 with different configurations of the fins 350 can be produced, and the fin insert assembly 355 best suited for a specific application can be selected and coupled to the base periphery 349 to form a heat sink 330 customized for that application. Another advantage is that, in some examples, the fin insert assembly 355 can be formed of a different material than the base periphery 349, which allows for a balance between thermal performance and cost (e.g., materials with higher conductivity (such as copper) can be used for the insert assembly 355, while less costly and less conductive materials (such as aluminum, steel, zinc, etc.) can be used for the base periphery).

[0076] The base periphery 349 includes a mounting feature (not labeled) similar to the mounting feature 242 described above. Additionally, the manifold 360 is coupled to the base periphery 349 in the same manner as the manifold 260 is coupled to the base 240. The manifold 360 may be similar to the manifold 260 described above, and therefore its repeated description is omitted.

[0077] It will be understood that both the general description and the detailed description provide examples that are illustrative in nature and intended to provide an understanding of the disclosure without limiting its scope. Various mechanical, compositional, structural, electronic, and operational changes may be made without departing from the scope of the description and claims. In some cases, well-known circuits, structures, and techniques have not been shown or described in detail to avoid obscuring the examples. Similar numerals in two or more figures represent the same or similar elements.

[0078] Additionally, unless the context otherwise indicates, the singular forms “a,” “an,” and “the” are intended to also include the plural forms. Furthermore, the terms “comprises,” “comprising,” “including,” etc., specify the presence of the stated features, steps, operations, elements, and / or components, but do not preclude the presence or addition of one or more other features, steps, operations, elements, components, and / or groups. Unless explicitly noted otherwise, components described as coupled may be directly coupled electronically or mechanically, or they may be indirectly coupled via one or more intermediate components. Unless the context otherwise indicates, mathematical and geometric terms are not necessarily intended to be used according to their strict definitions, as those skilled in the art will understand that, for example, substantially similar elements operating in substantially similar ways may readily fall within the scope of descriptive terms, even if those terms also have strict definitions.

[0079] And / or: Occasionally, the phrase “and / or” is used in conjunction with a list of items in this article. The phrase means that any combination of items in the list may be included—from a single item to all items and any permutation in between. Thus, for example, “A, B and / or C” means “one of {A}, {B}, {C}, {A, B}, {A, C}, {C, B} and {A, C, B}”.

[0080] Where practicable, an element and its associated aspects described in detail with reference to an example may be included in other examples in which they are not specifically shown or described. For example, if an element is described in detail with reference to an example but not with reference to a second example, the element may still be claimed as being included in the second example.

[0081] Unless otherwise noted herein or implied by the context, the use of approximate terms such as “substantially,” “approximately,” “about,” “roughly,” “approximately,” etc., should be understood to mean that mathematical precision is not required, and instead refers to a range of variation that includes, but is not strictly limited to, the stated value, property, or relationship. In particular, in addition to any range explicitly stated herein (if any), the range of variation implied by the use of such approximate terms includes at least any insignificant variations, as well as those typical in the relevant field for the type of article under discussion due to manufacturing or other tolerances. In any case, unless otherwise indicated, the range of variation may include values ​​at least within ±1% of the stated value, property, or relationship.

[0082] Given the disclosure herein, further modifications and alternative examples will be apparent to those skilled in the art. For example, apparatus and methods may include additional components or steps omitted from the figures and description for clarity of operation. Therefore, this description will be interpreted merely as illustrative and for the purpose of teaching those skilled in the art the general manner of implementing this teaching. It will be understood that the various examples shown and described herein will be considered exemplary. Elements and materials, and arrangements of those elements and materials, may replace those illustrated and described herein, and arrangements of those elements and materials may be reversed, and certain features of this teaching may be utilized independently, all of which will be apparent to those skilled in the art upon benefiting from the description herein. Changes may be made to the elements described herein without departing from the scope of this teaching and the following claims.

[0083] It will be understood that the specific examples illustrated herein are non-limiting and that modifications can be made to the structure, dimensions, materials, and methods without departing from the scope of this teaching.

[0084] Other examples based on this disclosure will be apparent to those skilled in the art upon consideration of the specification and practice of the invention disclosed herein. The specification and examples are intended to be considered exemplary only, and the following claims, including equivalents, shall have their full scope claimed in accordance with applicable law.

Claims

1. A heat sink for immersion cooling of a processor in an information processing device, the heat sink comprising: Includes a base on a first surface, which is configured to be thermally coupled to the processor while the heat sink is mounted in the information processing device; A heat dissipation protrusion, said heat dissipation protrusion being coupled to and protruding from the second surface of the base; as well as A manifold, which is directly attached to the second surface of the base and adjacent to the heat dissipation protrusion, wherein the manifold includes an inlet and a plurality of outlets, and the manifold is configured to receive an input flow of liquid coolant via the inlet and discharge an output flow of liquid coolant toward the heat dissipation protrusion via the outlets while the radiator installed in the information processing device is immersed in a pool of liquid coolant in an immersion cooling tank.

2. The radiator according to claim 1, in, The heat dissipation protrusions include fins.

3. The radiator according to claim 2, in, The base and the fins are integrally connected.

4. The radiator according to claim 3, in, The base and the fins are part of the extruded metal body.

5. The radiator according to claim 2, comprising: The base periphery includes a socket; and An insert base is inserted into the socket and coupled to the periphery of the base, the insert base and the periphery of the base together forming the base; The fins are coupled to the base of the insert.

6. The radiator according to claim 5, in, The fins are integrally connected to the base of the insert.

7. The radiator according to claim 5, in, The base of the insert and the periphery of the base are made of different materials.

8. The radiator according to claim 1, in, The base includes mounting features configured to mount the heat sink to the system board of the information processing device.

9. The radiator according to claim 8, in, The system board includes a radiator mounting feature configured to mount an air-cooled radiator to the system board. and The mounting feature is configured to engage with the radiator mounting feature to mount the immersion-cooled radiator to the system board instead of the air-cooled radiator.

10. The radiator according to claim 9, in, The heat dissipation protrusion includes fins, and the fins have a thicker pitch than the air-cooled fins of the air-cooled radiator.

11. The radiator according to claim 1, in, The manifold can be removably attached to the base by fasteners.

12. The radiator according to claim 1, in, The manifold includes a bottom portion coupled to the base and a top portion connected to the bottom portion, the bottom portion and the top portion defining an internal chamber of the manifold between the bottom portion and the top portion.

13. The radiator according to claim 12, in, The bottom portion and the top portion are coupled together by fasteners extending through the top portion and the bottom portion and coupled to the base.

14. The radiator according to claim 1, in, The outlet includes a nozzle.

15. The radiator according to claim 1, in, The outlet is configured to guide the output flow through a channel extending between the heat dissipation protrusions.

16. An information processing apparatus, comprising: Chassis; A system board, the system board being supported by the chassis and including a processor mounted on the system board; An immersion-cooled radiator, said immersion-cooled radiator being mounted to the system board and comprising: Base, the base including a first surface thermally coupled to the processor; A heat dissipation protrusion, said heat dissipation protrusion being coupled to and protruding from the second surface of the base; and A manifold, directly attached to the second surface of the base and adjacent to the heat dissipation protrusion, wherein the manifold includes an inlet and a plurality of outlets; and A pump, coupled to the chassis and fluidly coupled to the inlet of the manifold, wherein, while the information processing device is immersed in a pool of liquid coolant in an immersion tank, the pump is configured to draw in a flow of liquid coolant from the pool and supply the flow as an input flow to the manifold via the inlet, and the outlet of the manifold is configured to discharge an output flow of liquid coolant toward the heat dissipation protrusion.

17. The information processing apparatus according to claim 16, in, The heat dissipation protrusions include fins.

18. The information processing apparatus according to claim 17, in, The fins are integrally connected to the base.

19. The information processing apparatus according to claim 16, in, The system board includes a radiator mounting feature configured to mount an air-cooled radiator to the system board. and The base includes a mounting feature that engages with the radiator mounting feature to mount the immersion-cooled radiator to the system board in place of the air-cooled radiator.

20. An immersion cooling system, comprising: An immersion cooling tank comprising a liquid coolant pool, the immersion cooling tank being configured to receive a plurality of information processing devices immersed in the pool; An information processing device immersed in the pool, the information processing device comprising: Chassis; System board, the system board being supported by the chassis and including a processor mounted to the system board; and An immersion-cooled radiator, said immersion-cooled radiator being mounted to the system board and comprising: Base, the base including a first surface thermally coupled to the processor; A heat dissipation protrusion, said heat dissipation protrusion being coupled to and protruding from the second surface of the base; and A manifold, directly attached to the second surface of the base and adjacent to the heat dissipation protrusion, wherein the manifold includes an inlet and a plurality of outlets; and A pump, fluidly coupled to the inlet of the manifold and configured to draw in a flow of the liquid coolant from the pool and supply the flow as an input flow to the manifold via the inlet, wherein the outlet of the manifold is configured to discharge an output flow of the liquid coolant toward the heat dissipation protrusion.

Images