Next Patent: Heat dissipation device incorporating fan duct
Next Patent: Heat dissipation device incorporating fan duct
Plaque It!
Sponsored by: Flash of Genius |
This invention relates to electronics assemblies, for example to computer systems, and to thermal management thereof.
Certain systems may be employed as servers, for example in local area networks (LANs), or in wide area networks (WANS), telecommunications systems or other operations such as database management or as internet servers. Such servers may be used in so-called “horizontally scaled” applications in which tens or hundreds of corresponding servers are employed as part of a distributed system.
A typical computer employed for such purposes will comprise a pair of microprocessors mounted on a circuit board, together with other components such as power supply units (PSUs), hard disc drives (HDDs), fans, digital video disc (DVD) players, memory modules, ethernet ports etc. The servers may be designed in modular form so that they can be stacked together with other corresponding servers in large racks. For example, some servers may be designed with a height in the order of 10 cm and a width in the order of 5 cm.
Whatever the dimensions of the assembly, it can often be a problem to ensure that such systems are cooled adequately in view of the relatively large number of electronics components in an enclosed environment. The assembly enclosure may be provided with one or more enclosure fans in one of the assembly walls which provide a flow of air through the assembly, e.g. from front to back.
However, as microprocessors have become more powerful, they therefore generate larger quantities of heat, 20 to 30 watts of heat output or higher not being uncommon. It has accordingly been difficult to ensure adequate cooling of the microprocessors. In addition, the servers may not always be employed in data centres with air conditioned environments, and ambient temperatures may therefore be relatively high. For example, the NEBS standard requires that such a system will operate at ambient temperatures of up to about 55° C. without the junction temperature of the components exceeding their specification, for example about 85° C.
Furthermore, because of relatively restricted lateral dimensions of the enclosures, it is often the case that one of the microprocessors is placed in front of the other, with the result that the rearmost processor receives cooling air that has already been heated by the front processor.
According to one aspect of the present invention, there is provided an electronics assembly which comprises:
In any such assembly, there will typically be a number of electronics components that will generate heat to a greater or lesser extent. However, most components, for example memory circuits, passive components and the like, will generate relatively small quantities of heat, while a small number of components, and in particular microprocessors, will generate significantly larger quantities of heat. The term “heat-generating component” as used herein is intended to refer to components such as microprocessors that generate significantly more heat than the average component in the enclosure.
This form of assembly has the advantage that the efficiency of thermal management within the enclosure can be significantly improved. By virtue of this arrangement, it is possible to ensure that each heat-generating component has a heat-sink that is cooled by air coming directly from outside the enclosure without having previously been heated by any other such heat-generating component.
Each heat-sink may be mounted on one of the heat-generating components, and may have a cantilevered portion that extends over the other heat-generating component or one of the other heat generating components but is not in thermal contact therewith. Thus, each heat-sink may extend over both the components in plan view but be in thermal contact only with a single component, thereby enabling the same “footprint” of the components and heat-sinks to be maintained. The cantilevered portion of the heat-sink may be in physical contact with the other component or the other heat-sink by means of a thermal insulator, or they may be separated by an air gap.
In principle, the heat-sinks may be employed with any number of heat-generating components, for example with three or more, but often the assembly will contain only two such components. In such a case, each heat-sink may have a generally flat base that is mounted on its associated heat-generating component, and a cantilevered portion that is provided with cooling fins and which extends over up to one half of the transverse dimension of its associated heat-generating component in the direction of air flow. The cantilevered portion of each heat-sink may extend over the base of the other heat-sink, and have a lower surface that is sufficiently higher than the upper surface of the base to allow clearance between the cantilevered portion and the base of the other heat-sink.
According to another aspect of the present invention, there is provided a method of cooling a plurality of heat-generating components in an electronics assembly, which comprises:
The heat-sink described herein is novel per se and so, according to yet another aspect of the invention, there is provided a heat-sink for allowing cooling of a heat-generating electronics component, which comprises:
Embodiments of the present invention will now be described in detail by way of example with reference to the accompanying drawings, in which corresponding parts are given like reference numbers. In the drawings:
FIG. 1 is a schematic perspective view of a conventional electronics assembly forming a network server;
FIG. 2 is a schematic perspective view of an electronics assembly according to the present invention;
FIG. 3 is an isometric view of a heat-sink according to the present invention; and
FIG. 4 is a view of the heat-sink as shown in FIG. 3 together with a further heat-sink for an associated heat-generating component.
Referring now to the drawings, in which like reference numerals are used to designate corresponding elements, FIG. 1 is a schematic perspective view of a conventional narrow form factor computer that is intended to provide a rack mounted server for use with the internet or as part of a local area network (LAN) or for other telecommunications purposes. The computer is designed to fit into a rack along with a number of other such computers, for operation in a data centre. The assembly may be designed to be a so-called “high RAS” system, that is to say, to have high reliability, availability and serviceability, so that it can be operated with the minimum amount of down time, although for other systems, different criteria such as cost may be more important. The computer may, for some applications, be intended to function continuously with the down time in the order of minutes per year.
The computer comprises an enclosure 1 shown in ghost lines, and contains a printed circuit board (PCB) 2 on which all, or the majority of, the components are mounted. The computer contains a pair of microprocessors 4 and 6 located on the PCB 2 , to the side of a number of dynamic random access memory (DRAM) cards 8 that may have a variety of functions, for example to provide memory, input/output (I/O) functions or the like. The enclosure has a front wall 16 which may have a number of keys, switches and/or status output devices, and may include devices such as disc drives, CD ROM drives, digital video disc (DVD) drives, tape drives etc. To the rear of the enclosure, a mezzanine PCB 14 is provided which may support a number of additional modules or other components 15 . One or more connectors 12 may be located on a rear wall 10 of the enclosure to form a connection as the sever is pushed into the rack. In addition one or more fans may be located on the front wall 16 , or alternatively on the rear wall 10 , in order to generate a flow of air through the enclosure for cooling the components.
It will be appreciated that the enclosure may contain a large number of additional components, but these have been omitted for the sake of clarity. In addition, the assembly may have a number of different configurations or combinations of components.
In order to improve performance of the system two microprocessors 4 and 6 may be employed. Depending on the operating system (OS) and the capability of the software, this improvement may be achieved by symmetric multiple processing (SMP) and/or software applications that can split into separate threads and run concurrently, one on each processor. Because of the geometry of the assembly, especially its relatively small lateral dimensions, and the existence of the cards 8 that extend from the front to back along one side of the enclosure, the microprocessors 4 and 6 are placed one behind the other.
These microprocessors, by their nature, tend to generate more heat than other electronics components of the assembly, for example in the region of 20 to 30 W each or even more in some cases. For this reason, the microprocessors are each provided with a heat-sink 20 , 22 in order to maintain their temperatures within acceptable operating limits. In addition, (dependent on geometry and performance), the thermal conductivity across the flat heatsink base may need to be enhanced. This may be achieved via higher thermal conductivity materials, e.g. copper or graphite (aligned in the correct orientation) or an embedded phase-change fluid solution such as a vapour-chamber or heat-pipe(s).
The heat-sinks 20 and 22 are typically in the form of a block of metal, e.g. aluminium or other high thermal conductivity material having a number of fins 24 oriented so that the air flow will be parallel to the plane of the fins. As can be seen from FIG. 1, in order to provide an adequate degree of cooling, the heat-sinks 20 , 22 need to be quite sizable, and may take up a significant fraction of the lateral dimensions of the enclosure. For example, in the assembly shown in FIG. 1, the heat sinks 20 and 22 occupy slightly more than half the internal lateral dimensions of the enclosure 1 . However, because the two microprocessors are positioned one behind the other in the direction of the airflow, the heat-sink 22 of microprocessor 6 that is located closest to the front of the enclosure will receive cooling air at ambient temperature, while the microprocessor 4 and its associated heat-sink 20 located toward the rear of the enclosure 1 will receive cooling air only after it has been warmed up by the front microprocessor 6 and heat-sink 22 . This can cause a problem in trying to balance the cooling of the two microprocessors and in trying to maintain the rearmost microprocessor 4 within acceptable temperature limits. While one could modify the assembly in order to position the two microprocessors side-by-side, this will inevitably require a considerable modification of the design of the assembly, circuit boards etc., and may well not be possible without altering the size and shape of the enclosure. Furthermore, in order to optimise the length of the electrical connections to the DRAM cards 8 it is preferred to space the microprocessors 4 and 6 equally alongside the cards 8 rather than at one end.
FIG. 2 shows an assembly of the same general form as that shown in FIG. 1 which has been modified according to one aspect of the present invention by providing heat-sinks 30 and 31 that are each in thermal contact with one of the microprocessors 4 and 6 , but are essentially positioned side-by-side. FIG. 3 shows one heat-sink 30 for a heat-generating component such as a microprocessor that is employed in the assembly as shown in FIG. 2 in greater detail.
The heat-sink 30 has a generally flat base 32 for mounting on the component. The base is generally square as shown in FIG. 3, but will normally have the same shape as that of the component to be cooled, and may be bonded to the component by means of an adhesive, for example one having a relatively high thermal conductivity or alternatively may be attached by mechanical attachments e.g. screws or spring clips together with a high conductivity compliant gasket (thermal interface material). On the base 32 is located a an assembly of fins 34 , five as shown for the sake of clarity, but other numbers of fins may be employed, for example from 3 to 20 fins. In this case, the fins are longer than the base 32 so that they extend beyond the edge of the base in a cantilevered manner. The fins 34 may extend significantly beyond the base, and may, for example have a length that is at least twice the dimension of the base (in the direction of the fins). In this embodiment the fins generally have a length that is equal to twice the dimension of the base of the heat-sink plus the distance separating the microprocessors. In addition, the fins 34 do not extend over the entire dimension of the base 32 in a direction perpendicular to their length (i.e. perpendicular to the flow of air in use, referred to hereafter as the width of the base) but instead extends over somewhat less than one half of the width of the base 32 , and are located on one side of the base so that at least one half of the base 32 has no fins. In addition, the under edge or surface 36 of that part of the fins 34 that extends beyond the base 32 is raised slightly above the level of the base, for example by a distance of from 0.5 to 4 mm.
FIG. 4, shows how two such heat-sinks may be placed together after rotating one heat-sink through 180°, one heat-sink being shown in outline only. In this arrangement, the parts 38 of the fins of one heat-sink that extend beyond the edge of the base 32 are located over the base 32 of the other heat-sink but are not in contact with the other heat sink. For example they can be separated from the base and from the fins of the other heat-sink by a distance of about 1 to 3 mm. In this way two heat-sinks are provided that can be substantially the same size as the conventional heat-sinks shown in FIG. 1 with substantially the same plan view or footprint and with substantially the same height and transverse dimensions. The two heat-sinks will both receive ambient cooling air directly from outside the enclosure without being warmed by other components to a significant extent. However, because each heat-sink is in direct physical contact with only one of the microprocessors and does not contact the other heat-sink, each heat-sink will provide for cooling of its associated microprocessor only. The microprocessors are, therefore, cooled by the airflow in parallel with each other even though they are positioned such that they would normally receive cooling air sequentially.
As shown in FIGS. 3 and 4, the base of each heat-sink extends over substantially the entire heat-generating component so that the cantilevered part of the fins of the other heat-sink are located over the remaining part of the base 32 . While such a design has the advantage that thermal conduction between the component and its associated heat-sink is maximised, it is possible in some circumstances for the heat-sink to be located over only a part of a heat-generating component.
Other designs of heat-sink may alternatively be employed depending on the design and location of the heat-generating components. For example, three heat-generating components could be positioned in line with the flow of air with each heat-sink designed so that a portion thereof, e.g. the fins, extend over the other two components and so that the lateral dimension of the fins is only substantially one third the lateral dimension of the base of the heat-sink or of the component. Even more components positioned in the line of the airflow could be cooled by this arrangement.
The heat-sinks and the assembly have been described above with reference to cooling of microprocessors. However, other components that generate significant quantities of heat, for example power supplies and the like may also be cooled by such an arrangement.
The scope of the present disclosure includes any novel feature or combination of features disclosed therein either explicitly or implicitly or any generalisation thereof irrespective of whether or not it relates to the claimed invention or mitigates any or all of the problems addressed by the present invention. The applicant hereby gives notice that new claims can be formulated to such features during prosecution of this application or of any such further application derived therefrom. In particular, with reference to the appended claims, features from dependent claims can be combined with those of the independent claims and features from respective independent claims can be combined in any appropriate manner and not merely in the specific combinations enumerated in the claims.