METHOD AND APPARATUS FOR COOLING ELECTRONIC OROTHER

DEVICES

FIELD AND BACKGROUND OF THE INVENTION

The present invention relates to methods and apparatus for cooling various types of devices which may be adversely affected by a high temperature produced during the operation of the device. The invention is particularly useful for cooling electronic devices, such as CPU's, laser diodes, semiconductor chips, and the like, and is therefore described below with respect to this application.

Most advanced electronic devices (CPU's, laser diodes, etc.) generate substantial heat during operation. Such heat must be quickly and efficiently dissipated in order to prevent undue temperature rise which could effect the operation of the device or even destroy it. The problem of heat dissipation becomes more acute with the reduction in size of the electronic devices.

Many cooling techniques have been devised for quickly and efficiently cooling electronic devices. One technique involves direct cooling, wherein the electronic device is immersed in a liquid coolant. Another technique involves indirect cooling, wherein the electronic device is brought into contact with one face of a cold plate having high thermal conductivity, the heat transferred from the electronic device to the cold plate being dissipated from one or more other faces of the cold plate. In some cases, where the amount of heat to be dissipated is not particularly large, forming the outer surface of the cold plate with fins may be sufficient to produce the required heat dissipation. In other cases, involving the need to dissipate the heat at a faster rate, a liquid coolant is used for removing the heat from the cold plate.

Examples of various cooling techniques that have been devised are described in the US Patents 5,316,075, 6,650,542, 6,675,875, 6,708,501, 6,741,469 and

6,866,067. Nevertheless, there is a continual need, which is made particularly acute with the continual miniaturization of electronics, for providing more efficient cooling devices of small, compact construction, and low power consumption.

The method and apparatus for cooling electronic or other devices in accordance with the present invention are similar to those described in the above- cited US Patent 5,316,075. That patent describes a method of cooling by bringing the device to be cooled into contact with the outer face of a cold plate having high

thermal conductivity, and applying a plurality of liquid jets of a liquid coolant to the inner face of the cold plate. The liquid jet is produced by a nozzle plate having one side communicating with a source of the pressurized liquid coolant, and an opposite side along which the liquid coolant drains by gravity. As described therein, the high- velocity liquid jets improve cooling capacity by minimizing laminar liquid layers. One embodiment is described wherein the cooling surface of the cold plate is provided with a plurality of pins to increase the area exposed to the liquid coolant.

OBJECTS AND BRIEF SUMMARY OF THE PRESENT INVENTION

An object of the present invention is to provide a method and apparatus for cooling electronic or other devices capable of producing more efficient heat-transfer from a cold plate, of being implementable in a more compact construction, and/or of requiring less power than, for example, the method and apparatus described in the above-cited US Patent 5,316,075.

The invention thus involves a method and apparatus for cooling electronic or other devices by: bringing the device to be cooled into contact with one face of a cold plate having high thermal conductivity; and applying a plurality of liquid jets of a liquid coolant to the opposite face of the cold plate at spaced locations thereon.

According to one aspect of the present invention, the plurality of liquid jets are applied in pulses as pulsatile jets. According to another aspect of the present invention, the liquid coolant is also circulated as a planar-flow liquid in contact with and parallel to the opposite face of the cold plate, and the plurality of pulsatile jets of the liquid coolant are applied perpendicularly to the opposite face of the cold plate such as to flow as jet streams through the planar— flow liquid coolant circulated in contact with and parallel to the opposite face of the cold plate.

Preferably, the planar-flow liquid coolant circulated in contact with and parallel to said opposite face of the cold plate is also applied in pulses.

As will be described more particularly below, the feature of using pulsatile jets, rather than continuous jets as in the above-cited US Patent 5,316,075, substantially reduces the quantity of liquid coolant required, and thereby the power consumption for pumping such liquid coolant. The features of also circulating the liquid coolant as a planar-flow liquid in contact with and parallel to the opposite face of the cold plate, particularly when such planar flow is effected in pulses, enhances

the heat-transfer efficiency of the cold plate even when small quantities of liquid coolant are used.

Further features and advantages of the invention will be apparent from the description below.

BRIEF DESCRIPTION OF THE DRAWINGS

The invention is herein described, by way of example only, with reference to the accompanying drawings, wherein:

Fig. 1 is a block diagram illustrating one form of apparatus constructed in accordance with the present invention utilizing a single pulsating pump; Fig. 2 is a block diagram illustrating another form of apparatus constructed in accordance with the present invention utilizing two pulsating pumps;

Fig. 3 is a pictorial view illustrating one form of cold plate for use in the apparatus of either Figs. 1 or 2;

Fig. 4 is an exploded perspective view of the components of the cold plate of Fig. 3;

Fig. 5 is an enlarged transverse sectional view of the cold plate of Fig. 3; Fig. 6 is a further enlarged fragmentary view of a part of Fig. 5;

Fig. 7 is a sectional view along line VII VII of Fig. 5;

Figs. 8a-8c illustrate three examples of nozzle orifices in the orifice plate included in the cold plate of Fig. 3 ;

Fig. 9 illustrates the high— pressure and low-pressure cycles in the operation of the apparatus of Fig. 1 utilizing a single pulsating pump;

Fig. 10 illustrates the high-pressure and low-pressure cycles in the operation of the apparatus illustrated in Fig. 2 utilizing two pulsating pumps; Fig. 11 is a graph illustrating the flow rate through a nozzle during a high- pressure cycle;

Figs. 12a-12f illustrate the jet streams produced through a nozzle during different time segments of a high-pressure cycle as illustrated in Fig. 11;

Fig. 13 is a graph illustrating the heat dissipation per jet during the different time segments illustrated in Figs. 11 and 12a-12f;

Fig. 14 is a block diagram illustrating another form of apparatus constructed in accordance with the present invention utilizing a continuous-flow pump and a valve distributor;

Fig. 15 illustrates the apparatus of Fig. 14 used for cooling a plurality of electronic or other devices; and

Fig. 16 illustrates the high— pressure and low— pressure cycles in the operation of the apparatus of Figs. 14 and 15. It is to be understood that the foregoing drawings, and the description below, are provided primarily for purposes of facilitating understanding the conceptual aspects of the invention and possible embodiments thereof, including what is presently considered to be a preferred embodiment. In the interest of clarity and brevity, no attempt is made to provide more details than necessary to enable one skilled in the art, using routine skill and design, to understand and practice the described invention. It is to be further understood that the embodiments described are for purposes of example only, and that the invention is capable of being embodied in other forms and applications than described herein.

DESCRIPTION OF PREFERRED EMBODIMENTS

The Embodiments of Figs. 1 and 2

The apparatus illustrated in Fig. 1 includes a reservoir 10 for the liquid coolant, which is preferably filtered water, but may include various additives, such as anti-foaming, surface-tension and/or biocide agents. The liquid coolant may be any other material commonly used for this purpose, provided it has adequate thermal- hydraulic characteristics meting the flow requirements. Reservoir 10 includes a pressure regulator 11 effective to accommodate volume changes due to temperature changes. Reservoir 10 is connected to the opposite sides of a reciprocatory pump 20.

Thus, the reservoir is connected via suction line 12 and check-valve 12a to the high- pressure side 21 of pump 20, and via suction line 13 and check valve 13a to the low- pressure side 22 of the pump.

Pump 20 includes a reciprocatory operator 23 coupled at the high-pressure end 21 to a relatively small diaphragm 24 movable within a chamber 25 of relatively small volume so as to produce an output pulse of a relatively high pressure and low volume. The opposite end of operator 23 is coupled to a diaphragm 26 of relatively large diameter and movable within a chamber 27 of relatively large volume so as to output a low-pressure, high volume pressure pulse. The high-pressure pulses and low-pressure pulses from pump 20 are applied to a cold device, generally designated 30, used for cooling the electronic or other device to be cooled, as will be described more particularly below. Thus, the high- pressure, low volume pressure pulses from pump 20 are applied, via line 14, check valves 14a and 14b, and filter 15, to a high-pressure inlet 31 of cold device 30; whereas the low-pressure, high-volume pulses from pump 20 are applied via line 16 and check valve 16a, 16b to a low-pressure inlet 32 of cold device 30.

The construction of cold device 30 is described below with respect to Figs. 3— 7. Briefly, it includes a cold surface 33 brought into contact with the electronic device to be cooled, generally designated CD in Fig. 1, to effect an efficient heat- transfer from device CD to the liquid coolant before exiting, via exit 34. The so- heated liquid coolant is then circulated via line 35 through a radiator 36 to dissipate the heat, before being returned via return line 37 to the reservoir 10.

Fig. 2 illustrates an alternative construction wherein, instead of using a single reciprocatory pump 20, the apparatus includes two reciprocatory pumps, generally designated 20a and 20b, respectively. Pump 20a is used for producing the high- pressure pulses, and is connected to the high-pressure inlet 31 of cold device 30; whereas pump 20b is used for producing the low-pressure pulses and is connected to the low-pressure inlet 32 of the cold device.

Thus, pump 20a includes a reciprocatory operator 23a coupled at one end to a small-diameter diaphragm 24a movable within a small-volume chamber 25a, and at its opposite end to a similar small-diameter diaphragm 24b movable within a similar small-diameter chamber 25b. Similarly, pump 20b includes a reciprocatory operator 23b coupled at one end to a relatively large diameter diaphragm 26a movable within a relatively large volume chamber 27a, and at the opposite end to another relatively large diameter diaphragm 26b movable within a similar large-volume chamber 27b. The liquid coolant is fed from reservoir 10 to both high-pressure chambers 25a, 25b of pump 20a via suction line 12 and check— valves 12a, 12b, which produce the high- pressure pulses applied to the high-pressure inlet 31 of cold device 30 via line 14, check-valves 14a-14c and filter 15. The low-pressure chambers 27a, 27b of pump 20b are similarly fed with liquid coolant from reservoir 10 via line 13 and check- valves 13a, 13b and apply the low— pressure pulses to the low-pressure inlet 32 of cold device 30 via line 16 and check-valves 16a, 16b.

While using two pumps as illustrated in Fig. 2 adds to the cost and size of the system, it provides a number of significant advantages, in permitting more precise control, as will be described more particularly below. The Cold Device 30 Figs. 3-7 illustrate the construction of cold device 30. As shown particularly in Fig. 3, and as briefly described earlier, it includes a high-pressure inlet 31, a low- pressure inlet 32, and a drain outlet 33 for the liquid coolant used to dissipate the heat from the electronic device CD brought into contact with the outer surface of cold device 30. As shown particularly in Figs. 3 and 4, cold device 30 includes a body member, generally designated 40, of a material having high thermal conductivity defining an outer fiat, cold plate 34 brought into contact with the device CD to be cooled. As seen in Fig. 4, the inner face of body member 40 is formed with a rectangular cavity 41 for receiving a thin nozzle plate 42 formed with a plurality of

nozzle orifices 43. Nozzle plate 42 is spaced from the bottom surface of cavity 41 by a plurality of short spacer posts 44 engageable with the under surface of the nozzle plate.

Cavity 41 is closed by a cover 45 secured to body member 40 by a plurality of fasteners 46. The cavity is sealed by a sealing ring 47 interposed between the cover and the body member.

As further seen in Fig. 4, the high-pressure inlet 31 includes a nipple 31a applied to cover 45; the low-pressure inlet 32 includes a nipple 32a applied to body member 40 at one side of cavity 41, and the drain outlet 33 includes a further nipple 33a applied to body member 40 at the opposite side of the cavity.

As shown particularly in Fig. 5, cover 45 is also formed with a cavity 48 which defines, with one side of orifice plate 42, a high-pressure chamber HPC communicating with the high-pressure inlet 31. The opposite side of orifice plate 42 defines with the bottom surface of cavity 41 a low-pressure chamber LPC communicating with the low-pressure inlet 32 on one side, and with the drain outlet 33 on the opposite side. The nozzle orifices 43 formed in nozzle plate 42 establish communication between the high-pressure chamber HPC and the low-pressure chamber LPC as best seen in Fig. 6.

The configuration of the low-pressure chamber LPC is defined by cavity 41 formed in body member 40 as best seen in Fig. 7. Its configuration would depend to a great extent on the configuration of the device CD to be cooled. As shown in Fig. 7, cavity 41 is of substantially square configuration. It is straddled on one side by an inlet manifold 51 communicating with the low-pressure inlet 32 of the cold device, and on the opposite side by an outlet manifold 52 communicating with the drain outlet 33 of the cold device. Body member 40 is further formed with a plurality of feed lines 53 from the inlet manifold 51 to cavity 41 defining the low-pressure chamber LPC, and with another plurality of feed lines 54 leading from cavity 41 to the outlet manifold 52. Feed lines 53 and 54 may be conveniently produced by drilling bores through body member 40 and plugging the ends of the bores, as shown by end plugs 55.

Fig. 7 also illustrates the flow paths of the low-pressure liquid coolant from the inlet manifold 51 through the feed lines 53 to cavity 41 defining the low-pressure chamber LPC, and from cavity 41 via feed lines 54 and outlet manifold 52 connected to the outlet 33 of the cold plate. It will be appreciated that this flow of the coolant

δ

fluid cools the bottom surface of cavity 41, and thereby the inner surface of cold plate 34 in contact with the device CD to be cooled.

Fig. 6 illustrates various significant parameters in the construction of the cold device 30. Thus, as shown in Fig. 6, "A" is the thickness of the portion of body member formed with the cavity 41 and defining the cold plate 34 contacted by the device CD to be cooled: preferably, this thickness should be from 0.4-0.6 mm.

The height of the low-pressure chamber LPC is defined as "B". It depends upon the size of the device CD to be cooled, the power required to be dissipated, the type of system in use (single pump as in Fig. 1, or dual pump as in Fig. 2), as well as other parameters, as will be more particularly discussed below.

The thickness of nozzle plate 42 is defined as "C". As briefly noted above, and as will be described more particularly below, it serves to separate the high- pressure chamber HPC from the low-pressure pressure LPC, and also to discharge the liquid coolant from the high-pressure chamber to the low-pressure chamber in the form of a plurality of liquid jets. Because of the pressure difference between the two chambers, nozzle plate 42 should be thick enough to maintain its stiffness; for example, it should preferably have a nominal thickness of 0.5 mm.

The speed of the jets emerging from the nozzle orifices is dependent upon the diameter of the nozzle and its height, as both control the nozzle pressure drop (measured between the nozzle upstream and downstream sides). Figs. 8a-8c illustrate examples of different nozzle orifices which may be provided. In these figures, the thickness of the nozzle plate 42 is indicated as "C"; the diameter of the nozzle orifice is indicated as "d"; and the height of the nozzle orifice is indicated as "H".

Fig. 8a illustrates a nozzle construction having minimum entrance losses, but expensive to produce. Fig. 8b illustrates a novel construction having higher entrance losses, but less expensive to produce. If the construction of either Fig. 8a or 8b is used, preferably the ratio of H/d is about 0.2. Fig. 8c illustrates a tapered construction having an inlet diameter "D", which is the easiest shape to produce. However, such a nozzle has a "hydraulic diameter" which is dependent upon the speed (and the inlet pressure) of the exiting jet, and also has lower stability than either of the above two constructions. Examples of Operation

The operation of the apparatus will first be described with respect to the single-pump embodiment illustrated in Fig. 1.

At the start of the operation, the two chambers, namely the high-pressure chamber HPC and the low-pressure chamber LPC, in cold device 30 are first filled with the liquid coolant in any convenient manner. The high-pressure end of pump 20 is then connected to the high-pressure inlet 41 of cold device 30, and the low— pressure end of the pump is connected to the low— pressure inlet 32 of the cold device. Pump 20 is then operated to produce the high-pressure pulses as shown in Fig. 9 applied to the high-pressure inlet 31 of cold device 30 alternating with the low- pressure pulses applied to inlet 32 of the cold device.

Fig. 11 illustrates a high-pressure-pulse applied to the high-pressure inlet 31 of cold device 30; and Figs. 12a-12f and 13 illustrate what occurs at different time segments along the applied high-pressure pulse as the liquid coolant in the high- pressure chamber HPC passes through nozzles 43 of the nozzle plate 42 into the low- pressure chamber LPC.

Thus, as shown in Fig. 12a, at the beginning of the high-pressure pulse, a jet 43 a of the liquid coolant first appears at the exit of nozzle 43 in nozzle plate 42 and grows in length as the pressure increases (43 b, Fig. 12b) until at the maximum pressure (Fig. 12c) at which time it impinges the bottom surface of cavity 41, perpendicularly thereto, and then begins to spread laterally along that surface, as shown at 43c (Fig. 12c) and 43d (Fig. 12d). The spreading continues even as the pressure decreases as shown in Fig. 12d, to the end of the high-pressure cycle, whereupon the jet breaks-off, as shown at 43 e in Fig. 12e.

This high-pressure cycle is followed by a low-pressure cycle. During this cycle the liquid coolant is applied at low pressure to the low-pressure inlet 32. This produces a full volume replacement of the liquid coolant in the low-pressure chamber LPC to ensure the complete energy dissipation. The liquid coolant is forced from the low-pressure chamber LPC through the outlet port 33 for circulation via line 35 to radiator 36 (Fig. 1) before returning via line 37 to the reservoir 10.

It will thus be seen that the liquid coolant entering the high-pressure chamber HPC via inlet 31 is converted by nozzles 43 of nozzle plate 42 into a plurality of liquid jets travelling perpendicularly to the cooling surface of cold device 30 through the low-pressure chamber LPC; and that the liquid coolant applied to the low- pressure chamber LPC via inlet 32 produces a planar flow of the liquid in contact with and parallel to the cooled surface of the cold plate. As a result, that the pulsatile jets flow as jet streams through the planar flow of liquid coolant circulated through the

low-pressure chamber. Such an arrangement produces a highly turbulent film on the inner surface of cold plate 34, thereby providing a high degree of heat-transfer with respect to the device CD to be cooled in contact with the outer face of the cold plate. Fig. 9 illustrates the pressure vs time relationship of high-pressure chamber HPC and low-pressure chamber LPC in the case of a single pulsating pump as illustrated in Fig. 1. The time period Ti and T ₂ are dependent on the pressure heads in the high-pressure cycles and low-pressure cycles, respectively; while periods T ₁ ' and T ₂ ¹ indicate control delays which can be selectively effected according to the heat dissipation requirements for any particular application. Thus, in the case where T ₁ =T ₂ , the reciprocatory operator 23 of the pump reciprocates at a constant frequency.

Fig. 10 illustrates the pressure vs time relationship with respect to the high- pressure cycles and low-pressure cycles when a two-pump apparatus is used as illustrated in Fig. 2. Thus, whereas the single-pump sequence illustrated in Fig. 9 produces high-pressure pulses alternating with low-pressure pulses, it is possible when using the two-pump arrangement illustrated in Fig. 2, to produce a plurality

(e.g., three) of high-pressure pressure pulses followed by a low-pressure pulse. Since a different pump is used for each of the two pressures, the synchronization between the pumps can be controlled as desired. Accordingly, the two-pump system is inherently more flexible and adaptable for a specific heat dissipation scenario, which is one of the peformance advantages of the novel system.

Fig. 13, when taken together with Figs. 11 and Figs. 12a-12f, illustrates the heat dissipation cycle. Thus, as shown in Fig. 13, the maximum heat dissipation is effective when the maximum film area has been reached, as shown in Fig. 12d, somewhat past the maximum pressure produced in the high-pressure cycle. Theoretically, the best dissipation performance will be proportional to the speed of the jets. For demonstration purposes, let us assume:

(a) Low viscosity coolant (like: water, lcps), fully Newtonian, and having low surface tension (35 dyne/cm);

(b) Given maxim value for the high-pressure pulse (1.1 bar at the peak);

(c) Given flow capacity of the high-pressure pulse (0.4cc per pulse); and

(d) Jet speed of 12m/sec;

The jet speed can be achieved using nozzles having a diameter larger than 60 microns, while H/d=0.2. The volume of the coolant injected by an array of 10 nozzles for Tl pulse length, is presented in Table 1:

TABLE 1

It is clear that for the given parameters, if the period of Tl = 800 msec is needed, and if 10 nozzles are required as well, the scenario will not be able to be supported by a single pump configuration (Fig. 1).

The maximal effective area per jet for different low-pressure chamber heights is presented in Table 2, which sets forth the maximal effective area as a function of nozzle diameter and low— pressure chamber height (B) for an array of 10 nozzles.

TABLE 2

It is clear that there is an optimal value for B. As the number of jets is increased, the overlap between jets is increased, and the capability to dissipate higher power densities will accordingly be increased. On the other hand, if the power density value is not high, it is not necessary to implement an "overkill" system.

The above examples demonstrate the flexibility of the present invention to adapt the system configuration according to the requirements of the chip designer, and to operate the system at frequencies related to the on-going heat generation.

The Embodiment of Figs. 14-16

Figs. 14-16 illustrate yet another embodiment of the invention which further demonstrates the flexibility of the system to be adapted according to the requirements of any particular design, and also according to the number of devices, such as chips, to be cooled.

Fig. 14 illustrates an apparatus utilizing, instead of a reciprocatory pump 20 in Fig. 1, a continuous-flow pump 120 for pressurizing the liquid coolant, and a valve distributor 122 connected to the outlet of the pump for alternatingly directing the pressurized liquid coolant at the pump outlet to the high-pressure chamber and the low— pressure chamber of the cold device 30. Thus, as shown in Fig. 14, one side of valve distributor 122 is connected to the high-pressure inlet 31 of cold device 30; whereas the other side of the valve distributor is connected to the low-pressure inlet 32 of the cold device. Thus, valve distributor 122 feeds both chambers of the cold device 30 alternatingly, according to the control signals, as shown in Fig. 16. In all other respects, the apparatus illustrated in Fig. 14 is substantially the same, and operates substantially in the same manner, as described above with respect to Fig. 1, and therefore the same reference numerals have been used to identify corresponding parts to facilitate understanding.

Fig. 15 illustrates the apparatus of basically the same construction as in Fig. 14, except that the continuous-flow pump, therein designated 130, supplies the pressurizes liquid coolant to a plurality of valve distributors 132a, 132b, etc. each supplying the liquid coolant, alternatingly at high-pressure and at low— pressure to a cooling device, 30a, 30b, etc., used for cooling an electronic chip or other device. Thus, basically the same apparatus, but using a continuous-flow pump of the required capacity, can be used for cooling a large number of chips or other devices. The valve distributors are fed in parallel via pump feed line 133, and the liquid coolant is returned to the reservoir 10 via return line 134, check-valve 135, and pressure- reducer 136.

As shown by the pressure vs time graph of Fig. 16, the embodiment of the invention illustrated in Figs. 14 and 15 reduces the flexibility of the control, but on the other hand, provides the opportunity to separate the pump from the control loop (its flow is the only control parameter), thereby making the apparatus more convenient for adoption to broad ranges of pumps or to multi-device cooling, as shown in Fig. 15.

While the invention has been described with respect to two preferred embodiments, it will be appreciated that these have been set forth merely for purposes of example, and that many variations, modifications and other applications of the invention made be made.