 Case
Studies
Electronic Liquid Cooling at Higher Pressures Using
Positive Displacement Pump Technology
By Steven Owen
Introduction
Many of the liquid cooling solutions available today employ cooling system devices that require system pressure drops limited to <0.5 bar. Recent advances in microchannel cooling have brought about pressure drop requirements >1.0 bar [1]. Higher pressure requirements can push cooling systems out of the pressure ranges of non-positive displacement pumps because of the prohibitive size and impeller speeds required for a centrifugal pump. A different technology is needed for these higher pressure requirements, yet the technology must meet the need for a small package size and fulfill the demand for low power consumption and high reliability. An external gear pump of the proper size can provide reliable, pulseless flow at elevated pressures and make it possible to design a cooling system that is small and uses minimal power.
The junction to inlet thermal resistance (R j-inlet), shown in Figure 1, is defined as:
where, T j, max is the junction temperature (°C) at the hottest point of the die, T in is the inlet liquid temperature (°C) in the microchannel cold plate, and the TDP is the thermal design power in watts (W) [2]. It has been shown that reducing the thermal resistance of the microchannels results in large pressure drops [1].
The Pumping Challenges of Microchannel Cooling
Table 1 shows that, from a technical standpoint, microchannel cooling compares favorably to other cooling technologies [2]. The interest in microchannel cooling results from the growing need for miniature heat exchangers in the electronic cooling and automotive markets. Increases in heat generated by faster processor speeds are rapidly approaching the limit of fan-cooled systems.
The advantages of microchannel heat sinks over conventional types are that they occupy less space, have high aspect ratios and heat transfer coefficients, and have low thermal resistance. The disadvantage is high coolant pressure drops, which are difficult to overcome with current technology. Recent gains in microchannel manufacturing techniques have been a factor in the growing interest of microchannel cooling.
Gear pump technology is well suited for high pressure cooling systems. As previously mentioned, a high pressure drop resulting from the use of a microchannel element lends itself to positive displacement gear pumps because they are capable of creating high differential pressures within a small package size. One gear pump can suit many different cooling system operating points because of its broader operating range, which is important to the thermal system designer. As shown in the graph of Figure 2, by varying the gear pump speed, operating points on three potential cooling system curves (where the red lines intersect the blue lines) can be reached. The performance curve of a gear pump indicates a predictable relationship of pump flow to speed. When coupled with the motor and control logic, this becomes an important factor because it allows for closed-loop control. In this way, the flow (pump speed) can be adjusted proportionally to the temperature of the coolant load of the chip, shown as T out – T in.
In contrast, the pressure generated by a centrifugal pump covers only a small range of the system curve. Centrifugal pumps can be more difficult to optimize to a given cooling system due to their characteristic nonlinear flow curve. A centrifugal pump will have a larger drop in flow with increased differential pressure. Use of a gear pump allows the system designer to provide flow at pressure for several different microchannel designs as well as ease of optimization. Figure 3 shows a comparison of centrifugal and gear pump performance curves.
Integrated Gear Pump
The main advantages of the integrated gear pump are:
- compact size
- low power requirements
- low noise
- reliability
- stiff pump curve
One of the requirements in electronic cooling is a small package size. An integrated gear pump makes an ideal candidate in this respect. In an integrated gear pump, the driven magnet in the pump becomes the rotor of a brushless DC motor, thereby minimizing the space required.
In this assembly, shown in Figure 4, the motor stator fits over a containment shell and aligns with the driven magnet. The containment shell, which also forms a fluid pressure barrier and allows clearance for the magnet to rotate, must have a thin wall section to minimize the air gap between the stator inside diameter (ID) and magnet outside diameter (OD). Because of the small package size, rare earth magnet materials are used to gain maximum motor performance. Electronic cooling applications sometimes use fluids that are corrosive, therefore stainless steel or engineered plastics are used to manufacture all pump body parts and the containment shell.
The rotating electromagnetic field, which is induced in the fixed and stationary stator by a controller, couples with and turns the permanent magnet sealed inside the pump. The speed of the pump can be monitored and maintained with feedback from the controller, which is mounted remotely. External control signals (0-5/10 VDC, 4-20 mA) can be used to control the speed of the pump, and the controller logic can be designed such that the input control signal is proportional to T out – T in of the microchannel element.
The integrated gear pump does not use dynamic seals, which eliminates the inherent failure mode of shaft seal leakage. The pumped fluid is used to lubricate the hydrodynamic bearings through the radial clearance between the rotating journal and bearing.
Basic Operation
The gear pump provides flow by sealing off at the gear mesh. As the gears rotate and the gear teeth open up, they create an expanding volume and an associated low pressure area. The low pressure area induces the pumped fluid to enter the pump inlet where it is trapped between the gear teeth and cavity wall. The rotation of the gears (Figure 5) moves the fluid around the cavity wall into the outlet area and passages, where a high pressure area is created as the gear teeth mesh displacing the pumped fluid.
In a gear pump, volumetric efficiency is determined by comparing the theoretically displaced volume to the actual pumped volume. The difference between these two volumes is the amount of slip in the pump, and this fluid is lost back to the inlet through the internal leak paths. Consequently, the efficiency changes with fluid properties such as viscosity. Fluids with higher viscosities require more power to pump, but reduce the amount of slip in the pump and increase volumetric efficiency.
The Key to Long Life
Gear pumps can be designed to have a long and predictable life as required in electronic cooling applications. The gear radial loads are primarily due to the discharge pressure acting on the gear cross-section. Each gear assembly has two journal bearings to withstand the pressure loading. The shaft dimensions required for small gear pumps usually result in very small Sommerfeld numbers (S). This puts the journal bearings into an operating regime where they have little or no lubricating film, and material choices are therefore made with boundary lubrication in mind. The Sommerfeld number is defined as:
where S is the bearing-characteristic number (Sommerfeld number), r is the journal radius, c is the radial clearance, µ is the absolute viscosity, N is the rotational speed (RPS), and P is the load/unit bearing area.
Bushing materials are initially chosen based on their pressure-velocity (PV) value, then verified with testing. Results to date show that combinations of materials such as ceramic shafts and engineered polymer bushings give wear factors on the order of 1x10 -10 mm 3/Nm, providing years of service. The inherent low friction factors in these materials help to further reduce pump power consumption.
Summary
Electronic microchannel cooling systems require high pressure capabilities, reliability, and small package sizes with low power requirements. The operational and design characteristics of the integrated gear pump will allow the system designer the flexibility to minimize thermal resistance by providing metered high pressure coolant across a microchannel element.
References
- Prasher, R.S., Chang, J., Sauciuc, I., Narasimhan, S., Chau, D., Chrysler, G., Myers, A., Prstic, S., Hu, C., “Nano and Micro Technology-Based Next-Generation Package-Level Cooling Solutions,” Intel Technical Journal, Volume 9, Issue 4, 2005.
- Sauciuc, J., Prasher, R., Chang, J., Erturk, H., Chrysler, G., Chiu, C., Mahajan, R., “Thermal Performance and Key Challenges for Future CPU Cooling Technologies,” paper 73242, Proceeding of IPACK 2005, July 17-22, 2005, San Francisco, CA.
- Garimella, S.V. “Advanced Thermal Management Technologies for Next Generation Micro-Electronics Systems,” ITHERM06 Short Course, San Diego, CA, May 30, 2006.
Author biographies and contact information:
Steven E. Owen is a staff engineer in the Product Development Group at Micropump Inc., a unit of IDEX Corporation. He holds two degrees in mechanical engineering; he obtained his B.S. in 1992 at the University of Washington and M.S. in 1997 at Northeastern University. Steve is the project team leader for the Integral Series upgrade and is a registered professional engineer in Washington State. Steve Owen proudly serves in the U.S. Naval Reserve and is also a qualified Navy Engineering Duty Officer.
Steven E. Owen
Micropump, Inc.
1402 NE 136th Avenue
Vancouver , WA 98684-0818
Tel: (360) 253-2008
Fax: (360) 253-8294
Email: sowen@idexcorp.com
|
