Design and Performance of Novel Low-Profile Heat Sinks Created Through Additive Manufacturing

Traditional thermal management techniques such as air-cooled plate- and pin-fin heat sinks are today being pushed to their limits by the increasing power densities of computing hardware (power supplies, controllers, processors, and integrated circuits). In comparison, direct immersion cooling within an alternative cooling medium such as commercial dielectric fluids offers the ability to handle high power densities while also accommodating tighter printed circuit board spacing. Together, these attributes are critical to facilitating higher computing densities. However, this type of high density setup also requires that any heat sink present be low profile so as to not obstruct adjacent printed circuit boards. Such a stringent limit on heat sink height can make achieving cooling targets challenging with existing designs. In this work, the performance of several low profile (height less than 6 mm) heat sinks of varying design are evaluated within a carefully controlled direct immersion cooling environment. Commercial copper heat sinks fabricated through conventional manufacturing (CM) approaches serve as baselines for these performance tests. These same heat sink designs are also replicated via additive manufacturing (AM) utilizing a conductive, carbon-filled printable polylactic acid (PLA) composite material. The performance of these AM heat sinks are then compared to the CM heat sinks, with special emphasis on differences in thermal conductivity between the constituent materials. Finally, novel bio-inspired heat sink designs are developed which would be difficult or impossible to achieve using CM approaches. The most promising of these designs were then created using AM and their performance evaluated for comparison. The overall goal of this is to ascertain whether the design and fabrication flexibility offered by AM can facilitate low profile heat sink designs that can meet or exceed the performance of conventional heat sinks even with perceived deficiencies in material properties for AM parts. Experiments were carried out within Novec 7100 dielectric fluid for single-phase natural convection scenarios as well as two-phase subcooled boiling conditions at atmospheric pressure. A custom test rig was constructed consisting of mirror-polished stainless steel plates and polycarbonate viewing ports to allow visual access. A rotating sample stage allows for data to be obtained at varying heat sink orientation angles from 0° to 90°. For two-phase experiments, multi-angle video capture allows for analysis of the two-phase dynamics occurring at the heat sink samples to be visualized and temporally linked to the associated temperature and heat flux data.

Pool Boiling of HFE 7200–C4H4F6O Mixture on Hybrid Micro-Nanostructured Surface

Steadily increasing heat dissipation in electronic devices has generated renewed interest in direct immersion cooling. The ideal heat transfer fluid for direct immersion cooling applications should be chemically and thermally stable, and compatible with the electronic components. These constraints have led to the use of Novec fluids and fluroinerts as coolants. Although these fluids are chemically stable and have low dielectric constants, they are plagued by poor thermal properties like low thermal conductivity (about twice that of air) and low specific heat (same as that of air). These factors necessitate the development of new heat transfer fluids with improved heat transfer properties and applicability. C4H4F6O is a new heat transfer fluid which has been identified using computer-aided molecular design (CAMD) and knowledge-based approaches. A mixture of Novec fluid (HFE 7200) with C4H4F6O is evaluated in this study. Pool boiling experiments are performed at saturated condition on a 10 mm × 10 mm silicon test chip with CuO nanostructures on a microgrooved surface, to investigate the thermal performance of this new fluid mixture. The mixture increased the critical heat flux moderately by 8.4% over pure HFE 7200. Additional investigation is necessary before C4H4F6O can be considered for immersion cooling applications.

Impact of TXV and Compressor in the Stability of a High-End Computer Refrigeration System

The combination of increased power dissipation and increased packaging density has led to substantial increases in chip and module heat flux in high-end computers. The challenge is to limit the rise in chip temperature above the ambient. In the past, virtually all commercial computers were designed to operate at temperatures above the ambient. However, researchers have identified the advantages of operating electronics at low temperatures. Until recently, large-scale scientific computers used direct immersion cooling of single-chip modules. The current research focuses on mainframes (computer system), which uses a conventional refrigeration system to maintain chip temperatures below that of comparable air-cooled systems, but well above cryogenic temperatures. Multivariable control of compressor speed along with thermostatic expansion valve (TXV) opening can give better stability and performance. TXV is a mechanical controlling device used in the refrigeration system. The compressor is the only mechanical-working component in the refrigeration cycle that circulates refrigerant through the system continuously. Hence, controlling the compressor is an important aspect. The control objective is defined as improving the transient behavior of the vapor compression cycle for the refrigeration system operating around an evaporator set-point temperature. The system behavior is studied in two cases, TXV being the only control element in the first case, while TXV and a compressor both act as control elements in the other case.

Performance of a Masterflux Compressor in a Refrigeration Unit for High End Computer System Application

The combination of increased power dissipation and increased packaging density has led to substantial increases in chip and module heat flux in high-end computers. Challenge has been to limit the rise in chip temperature above the ambient. In the past, virtually all-commercial computers were designed to operate at temperatures above the ambient. However researchers have identified the advantages of operating electronics at low temperatures. Until recently large scale scientific computers used direct immersion cooling of SCM’s. The current research focuses on mainframe (computer system), which uses conventional system to maintain chip temperatures below that of comparable air-cooled systems, but well above cryogenic temperatures. Multivariable control of compressor speed along with Thermostatic expansion valve opening can give better stability performance. Compressor is the only mechanical-working component in the Refrigeration cycle that circulates refrigerant through the system in a continuous cycle. Hence, the role of a compressor in the unit is significant and should be carefully studied. The control objective is defined as improving the transient behavior of the vapor compression cycle for the Refrigeration System around operating an evaporator set point temperature. The control system can give the better performance stability in regulating Evaporator Temperature. Also the interface temperature (temperature between the evaporator and heat dissipation device) should be maintained within a tolerance range. In this paper, first, the characteristics of the thermostatic expansion valve with the Masterflux compressor, the effect of variation of evaporator outlet superheat on the flow through the TXV at varying evaporator temperature, and effect of sudden changes in evaporator heat load and condenser pressure variation on the temperature oscillations at the evaporator is examined to determine the appropriateness of the compressor for such an application. Once that has been determined, the effect of Multivariable controlled compressor speed working with TXV and the advantages it offers regarding the steadiness of the unit will be reported.