External tray hose with integrated pump

Abstract

The present disclosure relates to an external tray hose with an integrated pump. One aspect described in the present application provides an apparatus comprising an external tray hose with an integrated booster pump for cooling computing devices. The booster pump can increase the fluid flow and fluid pressure of a coolant fluid pumped toward computing devices on a server tray. The booster pump includes a stator mounted around a booster pump housing having a hole. The stator includes one or more C-shaped laminated stacks for supporting coil windings. The energized coil windings generate a changing magnetic field that rotates an impeller within the housing hole to pump fluid along the housing hole of the booster pump. The apparatus includes: a monitoring module that monitors a sensor group to obtain information about one or more of a fluid temperature, a fluid pressure, and a device temperature; and a control module that adjusts the performance of the booster pump based on the information obtained from the sensor group.

Description

[0001] CROSS-REFERENCE TO RELATED APPLICATIONS

[0002] This application is related to U.S. application No. 16 / 215,498, entitled “AXIAL FLOW PUMP WITH REDUCED HEIGHT DIMENSION,” filed on December 10, 2018 by inventors John P. Franz and Tahir Cader; and is further related to U.S. application No. 16 / 579,254, entitled “SWIVEL-CAPABLE LOW-PRESSURE-DROP HOSE BARB FITTINGS,” filed on September 23, 2019 by inventors John P. Franz and Tahir Cader, both of which are incorporated herein by reference in their entirety. Background Art

[0003] The present disclosure generally relates to the field of cooling systems for processing units in computer systems. During operation, the processing unit generates heat, and when the heat generated exceeds the normal temperature range specified by the manufacturer, the operation of the computing device (or processing unit) may be affected. Therefore, in order to maintain the temperature of the computing device within the normal temperature range, different types of cooling systems are provided to cool the computing device. Some computer systems include a coolant distribution unit (CDU) that delivers a coolant fluid to a liquid-cooled computing device (e.g., a central processing unit (CPU)) to cool the computing device, thereby facilitating the normal operation of the computing device. BRIEF DESCRIPTION OF THE DRAWINGS

[0004] Figure 1A An expanded view of an exemplary micro axial flow intensifier pump according to one aspect of the present application is shown.

[0005] Figure 1B Shown are approximate dimensions of an exemplary micro axial flow intensifier pump according to one aspect of the present application.

[0006] Figure 2 According to one aspect of the present application Figure 1A A perspective view of the laminated stack of micro axial flow pumps is shown.

[0007] Figure 3A An exemplary micro axial flow booster pump integrated into an external tray hose assembly and attached to a server tray according to one aspect of the present application is shown.

[0008] Figure 3B An exemplary micro axial flow booster pump integrated into an external tray hose assembly with a protective sleeve according to one aspect of the present application is shown.

[0009] Figure 3C An exemplary micro axial flow booster pump integrated into an external tray hose assembly and connected to a rack of a server according to one aspect of the present application is shown.

[0010] Figure 4 An exemplary micro axial flow booster pump mounted within a server tray according to one aspect of the present application is shown.

[0011] Figure 5A A graph showing the relationship between fluid pressure and fluid flow rate for a system with and without a micro axial flow intensifier pump according to one aspect of the present application is shown.

[0012] Figure 5B A bar graph showing how to deploy servers in a single chassis with mixed fluid pressure drops using a micro axial flow booster pump according to one aspect of the present application is shown.

[0013] Figure 6 A flow chart illustrating an exemplary process for controlling a fluid flow rate of a micro axial flow intensifier pump according to one aspect of the present application is shown.

[0014] Figure 7 An exemplary computer system for facilitating control of fluid flow rate of a micro axial flow intensifier pump according to one aspect of the present application is shown.

[0015] In the drawings, like reference numerals identify like drawing elements. DETAILED DESCRIPTION

[0016] The following description is presented to enable any person skilled in the art to make and use the examples, and is provided in the context of a specific application and its requirements. Various modifications to the disclosed examples will be apparent to those skilled in the art, and the general principles defined herein may be applied to other examples and applications without departing from the spirit and scope of the present disclosure. Therefore, the scope of the present disclosure is not limited to the examples shown, but is intended to be based on the maximum scope consistent with the principles and features disclosed herein.

[0017] With the rapid development of computer technology and the growth of computationally intensive applications (e.g., applications based on artificial intelligence (AI)), there has been an increase in the hybrid deployment of central processing units (CPUs) and graphics processing units (GPUs) in computer systems. GPUs facilitate computationally intensive workloads involving parallel computing or accelerated computing.

[0018] Further, the amount of power consumed by the CPU and GPU increases rapidly. For example, in some computing systems or servers, the power consumption of the CPU may exceed about 500W, while the power consumption of the GPU may exceed about 700W. In addition, since the high bandwidth memory (HBM) system is co-packaged with the CPU and GPU dies, the amount of heat generated may cause the temperature to approach 50 degrees Celsius, and may easily exceed the manufacturer-specified device chassis temperature (usually below 60 degrees Celsius). In order to prevent the computing device from being damaged at a temperature close to or exceeding the device chassis temperature, several cooling mechanisms are applied, such as one or more air cooling mechanisms and one or more liquid cooling mechanisms. Some equipment manufacturers often support liquid-cooled cold plate solutions for their equipment. One of the disadvantages of cold plate technology is that they are typically fin-based and have developed slowly compared to the growth of computer technology. Therefore, cooling solution providers improve the cooling effect caused on computing devices by forcing more liquid to pass through the cold plate (i.e., increasing the liquid flow rate through the cold plate) to meet the equipment specifications.

[0019] Coolant distribution units (CDUs) are used to deliver coolant fluid to liquid-cooled servers, and traditionally these coolant distribution units can support four cabinets; and the ratio of cabinets to CDUs is typically 4:1 (one cabinet can be equivalent to two server racks). Increasing the liquid flow rate of computing equipment may result in an increase in the total flow rate of servers (e.g., servers deployed for high-performance computing (HPC) systems), which may reduce the ratio of cabinets to CDUs to 3:1 or further to 2:1. A smaller ratio of cabinets to CDUs indicates that one CDU can support a smaller number of cabinets, and additional CDUs may need to be deployed, which may increase the cost of the liquid cooling solution. In addition, deploying additional CDUs may result in the use of more floor space, increased warranty costs for users, increased maintenance costs, etc.

[0020] Furthermore, with the transition to increased heterogeneous computing, where CPU and GPU servers may be mixed in a single chassis, fluid flow rates and fluid pressure drops may be different across different servers. Existing liquid cooling solutions may not be well suited for such heterogeneous computing systems. This is because the CDU may distribute fluid flow unevenly across different servers. For example, a blade server with high impedance may result in a higher pressure drop and fluid flow rate compared to other servers in the server chassis.

[0021] Some aspects of the present application solve the above problems by applying devices and mechanisms that increase the fluid flow rate without reducing the ratio of the cabinet to the CDU. A small micro-axial flow booster pump can be integrated into an external hose assembly that connects each tray to the rack fluid distribution system. The micro-axial flow booster pump can improve the flow rate and pressure on some trays that need performance improvement. In addition, the pump integrated with the hose can be easily maintained and replaced without performing a time-consuming maintenance event to remove the server from the cabinet. The micro-axial flow booster pump can be optionally installed on the server tray. Due to its small size, the micro-axial flow booster pump occupies less space and can be deployed in many space-constrained applications. Both embodiments (i.e., the booster pump integrated into the external tray hose and the booster pump installed in the server tray metal plate volume) can improve the fluid flow rate and fluid pressure on the specific tray that needs to enhance performance. The micro-axial flow booster pump can allow different trays with different pressure drops (or impedances) to be mixed in a given server chassis and on a rack scale. Therefore, the system can balance the fluid flow distribution between different servers in a given chassis. In other words, the system can ensure a particular fluid flow rate to a server such that the fluid flow rate matches the fluid flow rates associated with other servers in the server chassis.

[0022] The terms "liquid" and "fluid" are used interchangeably in this application.

[0023] The term "server" refers to a high-density server computer whose design is optimized to reduce physical size and power consumption.

[0024] The term "rack" or "server rack" refers to a cabinet-like structure that can house IT equipment.

[0025] The term "chassis" or "server chassis" may refer to a metal frame into which one or more servers may be assembled.

[0026] Micro axial flow booster pump

[0027] Figure 1A FIG. 2 shows an expanded view of an exemplary micro axial flow booster pump according to one aspect of the present application. It should be noted that the micro axial flow booster pump described below is for illustration purposes only. Different types of micro axial flow pumps may be used in the system of the present invention. Figure 1AThe micro axial flow boost pump 100 shown can be connected to the cooling circuit via a hose barb connector. The micro axial flow boost pump 100 can include an inlet barb connector 102, which can be attached to a hose, thereby allowing the coolant fluid to flow to the pump. The micro axial flow pump 100 can also include an outlet barb connector 104, which can be attached to another hose to allow the coolant fluid to flow out of the pump. The inlet barb connector 102 and the outlet barb connector 104 can provide an expanded coolant flow channel, allow the use of a hose with an increased diameter, reduce pressure drop and corrosion (using a plastic hose barb connector), and contribute to improving the efficiency and life of the liquid cooling system. The micro axial flow boost pump 100 can also include an outlet barb plug-in 106, a hole 128, an inlet housing structure 130 and an outlet housing structure 108. When connected together, the inlet housing structure 130 and the outlet housing structure 108 can form a fluid-tight housing with a hole 128. The micro axial flow booster pump 100 may also include a bias spring 110, a cap 112, a magnet (inside) 114, an impeller 116, a shaft 118, a coil frame 120, a laminate 122, a copper winding 124, a stator 132, and a printed circuit assembly (PCA) 126. The motor control circuit may be located on the PCA 126, which may be mounted on the inlet portion or the outlet portion of the pump 100. The PCA 126 may be connected to the overall control system in a variety of ways based on the usage or application.

[0028] The impeller 116 and the magnet 114 can be separate components assembled together so that they are fixed relative to each other in rotation around the longitudinal axis along the shaft 118. The impeller 116 can be mounted on the shaft 118 to achieve enhanced support and balance, higher speed performance with less vibration and longer life. The impeller 116 can be optionally made of plastic or metal material. The stator 132 can be installed around the pump housing and can include a coil device, i.e., a coil winding 124, which can be suitable for applying a changing magnetic field through the housing (e.g., the inlet housing structure 130 and the outlet housing structure 108) to drive the magnet 114 around the longitudinal axis. The stator 132 can include a laminate 122, which can have a C-shaped stack, such as two stacks, and each laminate stack can have a coil winding 124 in the coil frame 120. This arrangement can define two poles for each laminate stack, with one pole at each end of the C-shaped section.

[0029] The coil winding 124 can be wound around the middle section of the C-shaped laminated stack. The pole teeth of each laminated stack can be positioned at spaced or adjacent partitions of the hole and provide magnetic poles opposite to each other. Therefore, at least four magnetic poles can be arranged around the hole. However, the body of each laminated stack and its associated coil winding 124 can be positioned on either side of the hole so that the housing around the hole occupies most of the height dimension of the pump. This arrangement allows the flow cross section of the pump to be increased while the height dimension of the pump is increased.

[0030] In addition, the coil winding 124 can be elongated, and the axis of the coil winding 124 can be defined parallel to the longitudinal axis (see Figure 2 The elongated nature of the coil windings 124 can help increase the flux linkage and can increase the power of the pump without increasing the height of the pump. Figure 1B Schematic dimensions of an exemplary micro axial flow booster pump according to one aspect of the present application are shown. Figure 1B In the example shown, Figure 1B (a) presents a top view of the micro axial flow boost pump 140 compared to the size of two 9V batteries 144 placed side by side. Figure 1B (b) shows a perspective view of a micro-axial flow booster pump 140 with approximate dimensions. In this example, the micro-axial flow booster pump 140 may be approximately 15 mm high, 44 mm long (along the direction of fluid flow), and 52 mm wide. The compact nature of the micro-axial flow booster pump may facilitate its use in many different space-constrained applications, thereby increasing the number of use cases or applications of the micro-axial flow booster pump. It should be noted that Figure 1B The dimensions presented are for illustration purposes only. Micro axial flow pumps of other dimensions may also be used.

[0031] Figure 2 According to one aspect of the present application Figure 1A A three-dimensional view of the laminated stack of micro axial flow booster pumps shown in FIG. Figure 2 As shown, the micro axial flow booster pump 200 may include a stator 220. In one embodiment, the stator 220 may be outside the housing structure and include a coil arrangement to generate a variable magnetic field through the hole to interact with the magnet. In addition, the variable magnetic field can drive the magnet and the impeller to rotate around the longitudinal axis 224, which in turn can move along the housing hole (see Figure 1A The stator 220 may include two laminated stacks, 222a and 222b, which may be C-shaped and perpendicular to the longitudinal axis 224 of the micro axial flow booster pump 200. The longitudinal axis 224 may also represent a magnet sleeve impeller assembly (see Figure 1AThe lamination stack 222a and 222b can define a stator in the form of a general cube, the width of which can be greater than the height. Each lamination stack (222a and 222b) can further include two parts 226a-226b and 228a-228b. Each plate of each lamination stack part (226a-226b, 228a-228b) can be substantially identical, thereby achieving substantially identical lamination stack parts (226a-226b, 228a-228b) and substantially identical lamination stacks 222a-222b. The lamination stack parts (226a-226b, 228a-228b) can be assembled together with matching protrusions 238, and the coil winding 232 is placed around the matching protrusions. This arrangement can guide the magnetic flux generated by the coil to the pole tooth area 236.

[0032] When the laminated stack portions (226a-226b, 228a-228b) are as Figure 2 When assembled as shown, they can support a coil assembly 230, which can include one or more coil windings 232. Each coil winding 232 can be optionally mounted on a bobbin ( Figure 2 The coil windings on the laminated stack 222a are not in the Figure 2 234 in the laminate to illustrate the cutouts 234 in the laminate which may be used to receive and accommodate the windings in each laminate stack portion (226a-226b, 228a-228b). The coil windings 232 may be wound around the waist or middle section of the C-shaped laminate stack (222a and 222b). During operation, current may be energized to each coil winding, and the energized coil windings may generate a rotating magnetic field in the corresponding laminate stack portion (226a-226b, 228a-228b), which in turn may generate magnetic poles at the tooth region 236 of each laminate stack. Typically, the pole tooth region 236 may be configured in such a manner that the magnetic flux generated may be concentrated toward the cylindrical permanent magnets of the rotor of the pump. In Figure 2 In the example shown, the tooth region 236 may have a concave surface facing the rotor, thereby increasing the efficiency of the pump. A rotating magnetic field that may be generated by the alternating current in the coil may cause the cylindrical permanent magnet to rotate about the longitudinal axis 224. An impeller fixed to the permanent magnet may rotate and drive the fluid along the housing bore.

[0033] The coil area 240 of the laminate stack portion (226a-226b, 228a-228b) may be an area of ​​the laminate that is perpendicular to the longitudinal axis 224. The coil area 240 may be occupied by the coil windings 232 in their passage through and around the laminate.

[0034] In addition, the protrusion 238 of each laminated stack portion (226a-226b, 228a-228b) can penetrate each coil winding 232 and connect the magnetic flux generated by the coil 232 with the tooth area 236. It should be noted that Figure 2 The illustrated protrusion 238 has a cubic shape. Other shapes may also be used, such as a cylindrical shape or a shape having an elliptical cross-section. By alternating the direction of the current in each winding 232, the polarity (N / S) of each pole tooth 236 may be alternating. The magnetic field generated by the coil 232 in the laminate stack 220 may interact with the magnetic field of the permanent magnet sleeve to drive the magnet sleeve and the connected impeller sleeve. Thus, the impeller may rotate about the longitudinal axis 224 and thereby drive the fluid in the bore of the impeller sleeve from the fluid inlet 102 of the micro axial flow booster pump to the fluid outlet 104 (at Figure 1A ). The magnetic poles of one lamination stack 222a can be arranged opposite to the corresponding magnetic poles of lamination stack 222b. Each magnetic circuit can be driven by a corresponding coil 232 around the waist of the C-shaped lamination stack (222a and 222b) defined by the protrusion 238.

[0035] The AC current in the coil 232 can be controlled by the control board (PCA) 234 ( Figure 1A The PCA 126 shown in FIG. 126 is used to achieve this. The rotational speed can be adjusted by changing the voltage and alternating frequency applied to both ends of the coil 232. For example, the micro-axial flow booster pump can be configured to operate at a pressure difference (psid) of 2 pounds per square inch and a rotational speed in the range of 20rpm to 30000rpm. When pulse width modulation (PWM) is used to control the pump, the PWM frequency is usually higher than 50Khz to ensure that a sufficient number of AC cycles occur within one commutation cycle so that the current can be adjusted when the current does not reach a stable (DC) state. In addition, the DC voltage supply of the pump can have a range of 0V to 12V. Other voltage ranges and frequencies can also be used. The parameters of the drive control can be determined in different ways according to the application, such as the heat generated by the fluid temperature, the power consumed and / or the device (computing device) cooled by the coolant pumped from the micro-axial flow booster pump.

[0036] Some aspects of the present application provide one or more use cases of a micro-axial flow pump that solves the difference in pressure drop of servers or computing trays in a supercharger manner. Typically, a cabinet or rack may include one or more servers (or computing trays); and one or more cabinets may represent a cluster. In a deployed cluster, the CDU may need to meet the server with the highest pressure drop in a 4-cabinet cluster supported by the CDU (i.e., the ratio of cabinets to CDU is 4:1). When the CDU encounters an increase in pressure drop, its ability to provide more coolant flow is reduced toward the computing device being cooled, thereby reducing the number of cabinets that the CDU can support. Typically, the servers with the highest pressure drop account for a smaller percentage of the servers deployed in a given cabinet or cluster. When the number of cabinets that the CDU can support is small (for example, the ratio of cabinets to CDU is 2:1 instead of 4:1), more CDUs may need to be deployed to support a given number of cabinets, which may increase costs, and may also result in occupying more floor space, increasing maintenance, and increasing warranty costs for users, etc.

[0037] Current systems are designed to use treated water as the coolant. Numerous instances of biological growth in systems using treated water have forced a transition to a propylene glycol / water (PGW-based) chemistry to prevent biological growth in the system while providing antifreeze protection during transport. Changing to a PGW mixture has a twofold impact on the cooling system. First, the fluid has a higher viscosity, which reduces the flow output of the CDU. Second, the inferior thermophysical properties of PGW (relative to treated water) have a negative impact on thermal performance due to reduced heat transfer, which forces the tray to require higher flow rates to achieve the desired thermal performance. In one aspect of the present application, the improved performance provided by the micro axial flow booster pump can allow the system to be converted from treated water-based to PGW-based without loss of cooling performance.

[0038] With the progress towards more heterogeneous computing solutions, current systems encounter more challenges. For example, the latest developments in computer technology and the demand for fast computing have led to the deployment of servers with a combination of high-power CPUs, GPUs, application-specific integrated circuits (ASICs), and switch chips. Some of these servers may include high-density designs with high power consumption, which may result in high impedance and may require cooling fluids with higher flow rates. The pressure drop on such high-impedance servers may be about one and a half times (or 1.5 times) the pressure drop on most servers in the cluster. If the flow is not controlled at the individual server level, the CDU is forced to meet the worst-case server flow, which means that most servers in the cluster will be oversupplied. This uneven distribution of fluid flow may result in improper cooling of computing equipment.

[0039] Typically, the above problems are solved by deploying more or larger CDUs, greater flow rates, and / or lower coolant temperatures. Regarding the ratio of IT racks to CDUs, current systems can change the ratio of IT racks to each CDU by deploying fewer IT racks for each CDU, or adding additional CDUs for a given number of racks. In the case of inconsistent servers, a common solution is to increase the impedance of the available servers to match the worst-case server impedance in the chassis. For example, when a chassis includes a mix of inconsistent servers (i.e., CPU and GPU servers), some CPU servers can be removed and replaced with virtual servers to match the impedance of the high-impedance GPU servers. This can result in inefficient use of server slots and increased costs for IT equipment.

[0040] In other words, different trays in a server rack may have different pressure drop and flow targets, which may be difficult for a row-level CDU to accommodate without incurring significant waste (i.e., meeting the worst-case trays (even for a single tray in the chassis) and over-provisioning the trays that already meet performance standards). Adding a booster pump inside the server tray sheet metal volume (the space within the metal shell of the server tray) may be difficult and unattractive due to lack of space and size limitations of existing pumps. Furthermore, in the event of a failure, repair and / or replacement time may interfere with the service provided by the IT equipment.

[0041] Some aspects described in the present application solve the problems associated with pressure drop on high impedance servers by integrating a small pump into an external hose assembly that connects each server tray to the rack fluid distribution system. Therefore, deploying a micro-axial flow booster pump for each of the high pressure drop servers or server trays can increase the flow and pressure on the trays that require additional performance. In other words, by using such a micro-axial flow pump positioned along the coolant supply line leading to a single server or tray, the system can have more refined control over the cooling capacity on a server-by-server or tray-by-tray basis. In addition, to simplify the deployment of the booster pump, one aspect of the present application can incorporate the pump into a hot-swap tube having a self-sealing fluid quick disconnect device at each end; a connector with a valve; or a mechanism that can reduce or eliminate fluid loss during maintenance events. This configuration allows easy use of the pump and therefore reduces maintenance costs. In addition, placing the micro-axial flow booster pump on the external hose can save space in the server tray and can reduce the risk of pump leakage within the server tray.

[0042] In addition, the use of embedded boost pumps can facilitate the mixing of server trays with different pressure drops (or impedances) in a given chassis and at the rack scale. Therefore, the deployment of micro-axial flow boost pumps can meet the growing demand of users to transition to heterogeneous computing, which includes a combination of CPU and GPU servers in a server chassis.

[0043] In one aspect, the micro axial flow boost pump can serve as a boost pump for a CDU pump. In addition, due to its small size (eg, about the size of two 9V batteries placed side by side), the micro axial flow boost pump can have various space-constrained applications.

[0044] Figure 3A A micro axial flow booster pump integrated into an external tray hose assembly and attached to a server tray is shown according to one aspect of the present application. Figure 3A In the example shown, the hot-swappable micro axial flow boost pump 302 is coupled to a plug 308 that is attached to a server contained in a server tray 312. The boost pump 302 can be powered by a connector 310 that is plugged into the server, the server tray 312, or another location in the server rack (e.g., a rack manifold), or another power source designed to provide power and control to the boost pump 302. The external tray hose 318 can include a first fluid coupler 304 that can be plugged into the front of the server along with the boost pump cable connector. The other end of the external tray hose 318 can include a connector or a second fluid coupler 316 that can be attached to a fluid exhaust channel, such as a rack manifold.

[0045] Connector 310 can be coupled to the front of the server tray to receive power and control signals from the power and control source. Optionally, when the rack has a manifold integrated to control each boost pump, the cable connection of the boost pump to connector 310 can be at the rack manifold. A variety of conventional connections are supported, which can include power connections, ground connections, tachometers, pulse width modulation (PWM) modules, etc.

[0046] The outer tray hose 318, which may be made of flexible tubing, may have an outer diameter of about half an inch. Other sizes may also be used. The first fluid coupler 304 may be connected to the front of the server, and the second fluid coupler 316 may be connected to a fluid exhaust channel, such as a rack manifold. Figure 3B Shown is how the wiring 314 is in a protective sleeve 306 that can be engaged and secured to the hose to prevent damage. In one embodiment, the protective sleeve 306 can be a mesh tube or a heat shrink tube. Other materials can also be used.

[0047] Figure 3CA micro axial flow booster pump integrated into an external tray hose assembly and connected to a rack of servers according to one aspect of the present application is shown. In this example, the external tray hose assembly (collectively represented by 318, 304, 316) can be coupled to a rack of servers 334 via a hose connector. Typically, the servers in rack 334 can have two hose connectors, namely, a supply connector 332 and a return connector 330. Fluid connector 304 can be coupled to supply connector 332 to supply fluid into the fluid network to cool computing devices on server trays within server rack 334.

[0048] Figure 4 A micro axial flow booster pump installed in a server tray according to one aspect of the present application is shown. Figure 4 The example shown depicts another use case for the micro axial flow booster pump. The booster pump can be mounted within the server tray 402, which means that the booster pump is installed inside the server tray 402. The booster pump occupies a small footprint within the server tray 402 and can provide a desired flow rate to a high impedance server. For example, the micro axial flow booster pump can occupy approximately 2.09in 3 , and can provide about 0.697W / in 3 With a power density (i.e., pump power consumption divided by pump volume), the booster pump can support both open-loop liquid cooling solutions and / or closed-loop liquid cooling solutions. In addition, the compact nature of the booster pump can facilitate its use in space-constrained environments. In other words, the micro axial flow booster pump can be optionally mounted inside the server tray.

[0049] In one example, two or more micro axial flow booster pumps mounted within a server tray may be connected in series or in parallel to allow adjustment of pump pressure and / or flow specifications. For example, booster pumps may be arranged in parallel to increase flow rate without increasing pressure drop, or the pumps may be operated in series to increase both pressure drop and flow rate. For both series and parallel operation, the pumps may be arranged so that their respective impellers rotate in opposite directions to reduce vortex effects of the fluid upstream of the pumps.

[0050] In one aspect, a compound pump may be provided that may include two boost pumps in series, wherein the fluid outlet of a first pump is connected to the fluid inlet of a second pump. This series connection of two boost pumps may increase the pressure associated with a single pump. The impellers of each pump may be configured to rotate in opposite directions relative to each other. This configuration may enable the downstream impeller to mitigate the rotation of the coolant fluid caused by the upstream impeller, thereby increasing the efficiency of the compound pump. Figure 4In another arrangement of the pumps shown, the fluid inlets of each pump in the compound pump including two pumps 404 and 406 connected in parallel can be interconnected by a T-connector 403 and commonly connected to a fluid inlet 408 (arrow 410 indicates the direction of fluid flow). The fluid outlet of each pump can be connected to two fluid inlets of a server 416. The coolant fluid flowing out of the server 416 can be directed to the outlet 418. In this arrangement, pumps 404 and 406 are connected in parallel, and the fluid flow rate can be increased. In addition, this parallel arrangement can provide parallel fluid flow paths 412 and 414 through a split pipe configuration. The split pipe configuration can contribute to a balanced fluid flow path toward the server 416.

[0051] In one aspect, multiple pumps can be packaged in a small form factor (SFF) hard drive enclosure that is approximately 15 mm high. Such packaging can provide hot-swap capability for the pump package. Within the pump package, various connectors such as T-connectors can be used to connect the pumps to provide parallel or series flow paths, which may be desired for performance and resiliency.

[0052] Figure 5A A graph of fluid pressure versus fluid flow rate for a system with and without a micro-axial flow booster pump according to one aspect of the present application is shown. The performance of the micro-axial flow booster pump in terms of boosting capability can be measured with an exemplary setup that can include a CDU or auxiliary pump in series with the micro-axial flow booster pump. The role of the auxiliary pump is similar to that of a CDU in a cluster or server rack. Figure 5A In the example shown, graph 500 shows a graphical representation of pressure in pounds per square inch (psi) versus flow rate in gallons per minute (gpm). Curve 504 indicates the relationship between pressure and flow rate values ​​when using a CDU system without a micro-axial flow booster pump and including only a CDU system pump. Curve 506 indicates the relationship between pressure and flow rate values ​​when using a 10.3W micro-axial flow booster pump, and curve 502 indicates the relationship between pressure and flow rate values ​​when the CDU system includes a micro-axial flow pump as a booster pump. Since the auxiliary pump and the micro-axial flow booster pump operate in series, their respective heads are cumulative, as shown in FIG. Figure 5A As shown, that is, the corresponding values ​​of curves 504 and 506 are added to obtain curve 502. Figure 5A The shown example indicates that the micro axial flow booster pump can be effectively used as a booster pump in a CDU system.

[0053] In one exemplary use case, the micro axial flow booster pump may be used in a standalone rack that may include two chassis (e.g., eight servers per chassis), an in-rack CDU, a rack manifold (or fluid exhaust channel) for fluid distribution, rectifiers for power conversion, switches, etc. The rack may support conduits for supply and exhaust fluids that connect each server to the rack manifold, which may be connected to the in-rack CDU. This fluid distribution arrangement may be easily used as shown in FIG. Figure 3B The hot-swap pumps and tubes shown replace conventional fluid supply tubes. When hot-swap pumps are deployed on each available server in a separate rack or cabinet, the micro-axial flow booster pump can improve the performance of the CDU pump. When the micro-axial flow booster pump is not running, the flow rate may be lower, for example, about 16.7 gpm. With the micro-axial flow booster pump running at about 16,700 rpm, the flow rate can be increased, for example, to about 18 gpm, an increase of about 1.3 gpm for the CDU compared to a system without a booster pump. Therefore, the micro-axial flow booster pump can effectively increase the flow rate and reduce the system impedance perceived by the CDU pump, for example, the flow rate of the CDU pump can be increased by about 7.7%.

[0054] In another exemplary use case, the micro-axial flow booster pump can be used to deploy servers with mixed pressure drops (impedances) in a single chassis. Typically, each slot in the chassis can accommodate one server. For example, a chassis with eight servers may include eight vertical slots, each slot accommodating a single server. In this case, it can be assumed that a server with a higher pressure drop than a nominal server (referred to as a 100% server) (e.g., increased to 121%) can be placed in the leftmost slot 1 and the rightmost slot 8 of the chassis that can accommodate eight servers (or vertical blades). When the micro-axial flow booster pump is included in a high-impedance server (e.g., having a pressure drop of 121% compared to the nominal server), the booster pump can reduce the effective impedance of this server. Therefore, including a micro-axial flow booster pump can make a high-impedance server or a 121% server run like a nominal server or a 100% server by reducing the impedance of the 121% server.

[0055] In current systems, it is often desirable to deploy servers with higher impedance in a chassis, for example, a heterogeneous computing system may include CPU and GPU servers mixed in a single chassis. Existing liquid cooling solutions may not be well suited for such heterogeneous computing systems. This is because the CDU may distribute fluid flow unevenly across different servers, and in order to balance the fluid distribution across the servers, the current system increases the impedance of all other servers, thereby increasing the impact on the CDU.

[0056] Figure 5BA bar graph is shown that demonstrates how servers are deployed in a single chassis with mixed fluid pressure drops using micro-axial flow booster pumps according to one aspect of the present application. For example, assume that approximately 150% of the servers with a higher pressure drop than the nominal servers (referred to as 100% servers) are placed in the leftmost slot 1 and the rightmost slot 8 of a chassis that accommodates eight servers (or vertical blades). In Figure 520, the white columns indicate that the fluid flow is unevenly distributed among the eight servers in the chassis when the micro-axial flow booster pumps are not deployed on the servers in slots 1 and slots 8 in the chassis of eight servers. The patterned columns (i.e., columns with a diagonal pattern) indicate that by deploying micro-axial flow pumps on servers 1 and 8, the fluid flow distribution on all eight servers in the chassis can be approximately uniform or balanced. The patterned columns shown in 522 and 524 indicate how the deployment of micro-axial flow booster pumps increases flow on servers one and eight and reduces the burden on the CDU.

[0057] Adding a boost pump to an external hose assembly and attaching the external hose assembly to the server tray and rack manifold (or fluid discharge channel) via a fluid quick disconnect device (or a mechanism that can reduce or eliminate fluid loss during maintenance events) can increase flow and improve performance without taking up additional floor space in the data center. In addition, the micro axial flow boost pump can mix different trays with different pressure drops in a given chassis and on a rack scale. The increase in performance provided by the boost pump can allow the use of PGW instead of treated water-based fluids. The pump installed with a hose can be easy to repair and replace. Due to the compact nature of the boost pump, it can have a variety of applications in space-constrained environments. For example, the boost pump can be installed within the sheet metal volume of the server tray to increase fluid flow and pressure.

[0058] Figure 6 A flow chart showing an exemplary process of controlling the fluid flow rate of a micro axial flow booster pump according to one aspect of the present application is shown. Figure 6 In flowchart 600, during operation, i.e., when the server rack, CDU, and micro-axial flow boost pump are powered on (operation 602), the system can monitor multiple servers deployed in the server rack (operation 604). For example, the server rack may include a fluid pressure sensor, an equipment temperature sensor, a fluid temperature sensor, etc. The micro-axial flow boost pump can be connected to the server and provide a coolant fluid thereto. In some cases, the performance of the micro-axial flow boost pump can be controlled to adjust the pressure of the fluid flow driven from the rest of the fluid network. For example, in one aspect, a control module that can include dedicated hardware and / or software and is located in a server or server rack can control the behavior of the corresponding micro-axial flow boost pump. In order to control the behavior of the micro-axial flow boost pump, this control module can be configured to monitor multiple sensors.

[0059] In one embodiment, the rack-level control module may include a monitoring circuit that monitors a pressure flow sensor, and may determine whether the sensor value exceeds a sensor threshold (operation 606). Accordingly, the system may adjust the fluid pressure of the micro-axial flow booster pump (operation 608). For example, a control module within a rack manager may monitor the pressure flow sensor and analyze their values. Based on the analysis, the control module may speed up or slow down the micro-axial flow booster pump. For example, in a server chassis including eight servers, the server in the first slot may be a high-impedance server, and the micro-axial flow booster pump connected to this server may increase the fluid flow toward the computing device in the first slot. When the control module at the server level identifies that the temperature of the corresponding server has reached a normal range based on the temperature sensor, the control module may slow down the corresponding micro-axial flow booster pump.

[0060] In one aspect, the temperature sensor and the pressure sensor can be located at the server level instead of the rack level. Accordingly, the speed of the micro-axial flow booster pump can be controlled by the server level control module based on the device temperature sensor (or pressure sensor) located at the server level.

[0061] Computer Systems

[0062] Figure 7 An exemplary computer system that helps control the fluid flow rate of a micro-axial flow booster pump according to one aspect of the present application is shown. In this example, a computer system 700 may include a processor 702, a memory 704, and a storage device 706. The computer system 700 may be connected to a peripheral input / output (I / O) user device 718, such as a display device 710, a keyboard 714, and a pointing device 716. The storage device 706 may store instructions for an operating system 720. The computer system 700 may be connected to a network 730, a set of sensors 732, and one or more micro-axial flow booster pumps 734 via one or more network interfaces. The computer system 700 may be at a rack level or at a server level in a rack.

[0063] The computer system 700 may be equipped with an ASIC or circuit 708 that may include a controller logic module 722 that performs the methods and / or processes described in the present disclosure. The controller logic module 722 may implement Figure 6 The process shown in the figure may include a monitoring module 724, an analysis module 726, and a module 728. The controller logic module 722 may include one or more hardware components and / or software components (i.e., software components that include instructions that can be executed by a processor to perform the processes described in this application). The corresponding modules may also include dedicated circuits.

[0064] The monitoring module 722 can monitor information or telemetry data corresponding to one or more sensors 732, which can include a fluid pressure sensor, an equipment temperature sensor, a fluid temperature sensor, etc. The analysis module 726 can analyze the monitoring information about the sensor 732 and determine whether the sensor value exceeds the sensor threshold. Based on this analysis, the control module 728 can communicate with the micro-axial flow booster pump 734 to control or adjust the speed of the micro-axial flow booster pump 734.

[0065] One aspect described in the present application can provide an apparatus for cooling computing devices. The apparatus may include a booster pump integrated into an external tray hose that increases the fluid flow and fluid pressure of a coolant fluid pumped toward a computing device on a server tray. The booster pump includes a stator mounted around a booster pump housing having a hole. The stator includes one or more C-shaped laminated stacks for supporting coil windings. The energized coil windings generate a changing magnetic field that rotates an impeller within the housing hole to pump fluid along the housing hole. Monitoring logic monitors a sensor group to obtain information about one or more of a fluid temperature, a fluid pressure, and a device temperature; and control logic adjusts the performance of the booster pump based on the information obtained from the sensor group.

[0066] In a variation of this aspect, a first end of the external tray hose assembly is coupled to the server tray via a first fluid coupler. A second end of the external tray hose is coupled to the fluid drain channel via a second fluid coupler. The first fluid coupler is attached to the first end of the external tray hose. The second fluid coupler is attached to the second end of the external tray hose.

[0067] In variations of this aspect, the boost pump is electrically coupled to a power and control source that provides power and control to the boost pump.

[0068] In variations of this aspect, the external tray hose may include an detachable tube having a mechanism to reduce repair and / or replacement time in the event of a boost pump failure.

[0069] In other variations of this aspect, the apparatus may further include a coolant distribution unit including a coolant distribution unit pump that pumps coolant fluid toward computing devices on a server tray, and a booster pump that increases the fluid flow and fluid pressure of the coolant fluid pumped by the coolant distribution unit pump.

[0070] In other variations, a micro-axial flow supercharger may be attached to a server in a chassis to ensure that flow matches flow associated with other servers in the chassis, where the server in the chassis has a different fluid pressure drop than the other servers in the chassis.

[0071] In other variations, the device may also include: analysis logic that determines whether the fluid pressure of the boost pump needs to be adjusted based on information obtained from the sensor group; and control logic that adjusts the fluid pressure of the boost pump in response to determining that the fluid pressure of the boost pump needs to be adjusted.

[0072] In another aspect, the device may include a booster pump mounted in a server tray that increases the fluid flow and fluid pressure of a coolant fluid pumped toward a computing device on the server tray. The booster pump includes a stator mounted around a booster pump housing having a hole. The stator includes one or more C-shaped laminated stacks for supporting coil windings. The energized coil windings generate a changing magnetic field that rotates an impeller in the housing hole to pump fluid along the housing hole. The monitoring logic monitors a sensor group to obtain information about one or more of a fluid temperature, a fluid pressure, and a device temperature. The control logic adjusts the performance of the booster pump based on the information obtained from the sensor group.

[0073] In a variation of this aspect, a booster pump may be mounted within the server tray and coupled to another pump on the server tray to serially pump coolant fluid to the computing devices.

[0074] In a variation of this aspect, a booster pump mounted within the server tray is coupled to another pump on the server tray to pump coolant fluid to the computing devices in parallel.

[0075] One aspect described in the present application can provide a system and method for increasing fluid flow and adjusting the performance of a boost pump. During operation, the system can receive a coolant fluid from a coolant distribution unit at a boost pump integrated into an external tray hose, wherein the boost pump includes a stator mounted around a boost pump housing having a hole, wherein the stator includes one or more C-shaped laminated stacks for supporting coil windings, and wherein the energized coil windings generate a changing magnetic field that rotates an impeller within the housing hole to pump fluid along the housing hole. The system can increase the fluid flow and fluid pressure of the coolant fluid pumped toward a computing device to be cooled on a server tray; and adjust the performance of the boost pump based on information associated with a sensor group, wherein the information includes one or more of fluid temperature, fluid pressure, and device temperature.

[0076] The methods and processes described in the detailed description section may be implemented as code and / or data, which may be stored in a computer-readable storage medium as described above. When a computer system reads and executes the code and / or data stored on a computer-readable storage medium, the computer system executes the methods and processes implemented as data structures and codes and stored in the computer-readable storage medium.

[0077] In addition, the above methods and processes may be included in hardware modules or devices. Hardware modules or devices may include, but are not limited to, ASIC chips, field programmable gate arrays (FPGAs), dedicated or shared processors that execute specific software modules or code fragments at specific times, and other programmable logic devices now known or later developed. When the hardware modules or devices are enabled, these hardware modules or devices execute the methods and processes included therein.

[0078] The foregoing description of various aspects has been presented for the purpose of illustration and description. The description is not intended to be exhaustive or to limit the scope of the present disclosure to the disclosed forms. Accordingly, many modifications and variations will be apparent to those of ordinary skill in the art.

Claims

1. An apparatus for cooling a computing device, comprising: a boost pump integrated into an external tray hose, the boost pump for increasing fluid flow and fluid pressure of a coolant fluid pumped toward computing devices on a server tray, wherein the external tray hose includes a first hose portion and a second hose portion, the first hose portion including a first end of the external tray hose and the second hose portion including a second end of the external tray hose, wherein the boost pump is integrated into the external tray hose between the first hose portion and the second hose portion, wherein the first end is configured to be removably coupled to a supply coupling of a server tray and the second end is configured to be removably coupled to a fluid supply channel of a rack housing the server tray, wherein the boost pump includes a stator mounted on a boost pump housing having an aperture, wherein the stator includes one or more C-shaped laminate stacks for supporting coil windings, and wherein the energized coil windings generate a changing magnetic field that rotates an impeller within the housing aperture to pump fluid along the housing aperture; a monitoring module for monitoring the sensor group to obtain information about one or more of fluid temperature, fluid pressure, and computing device temperature; and A control module is configured to adjust performance of the boost pump based on the information obtained from the sensor group.

2. The device according to claim 1, wherein: The boost pump is electrically coupled to a power source and to the control module, the power source providing power to the boost pump.

3. The device according to claim 1, wherein: The external tray hose includes a pluggable tube having a self-sealing quick disconnect fluid coupling, or a combination of a valve and a fluid coupling, at each of the first end and the second end.

4. The apparatus of claim 1, further comprising: a coolant distribution unit, the coolant distribution unit comprising a coolant distribution unit pump for pumping the coolant fluid toward the computing devices on the server tray; The booster pump is used to increase the fluid flow rate and fluid pressure of the coolant fluid pumped by the coolant distribution unit pump.

5. The device according to claim 1, wherein: The boost pump is attached to a server in a chassis to ensure flow rates match flow rates associated with other servers in the chassis; and wherein the server in the chassis has a different fluid pressure drop relative to the other servers in the chassis.

6. The apparatus of claim 1, further comprising: an analysis module, the analysis module being used to determine whether the fluid pressure of the boost pump needs to be adjusted based on the information obtained from the sensor group; as well as The control module is configured to adjust the fluid pressure of the boost pump in response to determining that the fluid pressure of the boost pump needs to be adjusted.

7. A method for increasing fluid flow and adjusting the performance of a booster pump, comprising: receiving coolant fluid from a coolant distribution unit at a boost pump integrated into an external tray hose, wherein the external tray hose includes a first hose portion and a second hose portion, the first hose portion including a first end of the external tray hose and the second hose portion including a second end of the external tray hose, wherein the boost pump is integrated into the external tray hose between the first hose portion and the second hose portion, wherein the first end is configured to be removably coupled to a supply coupling of a server tray and the second end is configured to be removably coupled to a fluid supply channel of a rack housing the server tray, wherein the boost pump includes a stator mounted around a boost pump housing having an aperture, wherein the stator includes one or more C-shaped laminate stacks for supporting coil windings, and wherein the energized coil windings generate a changing magnetic field that rotates an impeller within the housing aperture to pump fluid along the housing aperture; increasing the fluid flow rate and fluid pressure of the coolant fluid pumped toward the computing devices to be cooled on the server tray by the boost pump; monitoring one or more of fluid temperature, fluid pressure, and computing device temperature via a sensor set; The performance of the boost pump is adjusted based on information output by the sensor set, wherein the information is indicative of one or more of monitored fluid temperature, fluid pressure, and device temperature.

8. The method of claim 7, wherein: The boost pump mounted on the server tray is serially coupled to another pump on the server tray to serially pump the coolant fluid to the computing devices.

9. The method of claim 7, wherein: The boost pump mounted on the server tray is coupled in parallel to another pump on the server tray to pump coolant fluid to the computing devices.

10. The method of claim 7, wherein: The external tray hose includes a pluggable tube having a self-sealing quick disconnect fluid coupling, or a combination of a valve and a fluid coupling, at each of the first end and the second end.

11. The method of claim 7, wherein: The boost pump is attached to a server in a server chassis to ensure flow rates match flow rates associated with other servers in the server chassis, wherein the server in the server chassis has a different fluid pressure drop relative to the other servers in the server chassis.

12. The method of claim 7, wherein: Adjusting the performance of the boost pump based on the information associated with the sensor group further includes: Reduce the boost pump speed.

13. The method of claim 7, wherein: Adjusting the performance of the boost pump based on the information associated with the sensor group further includes: Increase the speed of the boost pump.

14. The method of claim 7, further comprising: The boost pump increases the fluid flow rate and fluid pressure of the coolant fluid pumped by a coolant distribution unit pump, wherein the coolant distribution unit includes the coolant distribution unit pump to pump the coolant fluid to the computing devices on the server tray.

Images