Journal of Marine Science and Technology, Vol. 18, No. 3, pp.

419-423 (2010) 419

FABRICATING A HEAT PIPE STRUCTURE WITHIN

A RADIATING PLATE FOR ELECTRONICS
FAN-LESS COOLING

Chun-Sheng Yu*, Shung-Wen Kang*, and Hong-Bin Chan**

Key words: fan-less, radiating plate, heat pipe. were to maintain key temperatures in the Pentium processor
case temperature. For operation and reliability; various am-
bient temperatures and conditions included 60°C, 55°C, 45°C,
ABSTRACT
35°C, and 25°C ambient. When electronics system adopted
This study presents a method for fabricating a conductive natural convection be the only thermal solution requirements
metal with a heat pipe structure integrally formed within a and constrained chassis design at the same time, heat plate
radiating plate, and the heat transfer channels are arranged in application plays an important role to enhance fin thermal
an orthogonal matrix. The developed heat transfer device is performance.
monolithic and seamless, thus achieving low heat resistance Electronic components thermal management must solve
and high heat transfer efficiency. The high conductivity radi- problems connected with the limitations on the maximum chip
ating plate discussed in this study has significantly improved temperature with the requirements on the temperature uni-
structural strength and a complete hot spot contact for efficient formity level. One can use air and liquid coolers as well as
cooling. Heat generated from a heat source is transferring coolers constructed on the phase change heat transfer principle
without passing through different interfaces of materials or in closed space to cool electronic components; i.e., immersion,
any solder paste within the high conductivity plate. The pro- thermosyphon, and heat plate coolers. Each of these methods
posed heat transfer device is suitable for mounting in various has its merits and draw-backs. One must take into considera-
heat transfer systems or heat dissipating modules. Therefore, tion the thermal parameters and system design and stability,
electronic system cooling using a fan-less design will have durability, technology, price, application [10].
potential improved heat transfer efficiency of about 33% by Heat plates, possessed of high thermal conductivity, pro-
natural convection. vide an essential isothermal environment with very small
temperature gradients between the individual components.
I. INTRODUCTION The high heat transfer characteristics, ability to maintain con-
stant evaporator temperatures under different heat flux levels,
This study focuses on recent applications of heat plates evaporator and condenser diversity and sizes variability make
in electronics industry. The operating principles, modeling, the heat plate an effective device for electronic components
design, and testing of heat plates as applied to electronic ap- thermal control.
plications with a comprehensive review of the specific appli-
cations are described.
Using Computational Fluid Dynamics (CFD) analysis to II. FUNDAMENTAL THEOREM
calculate the thermal phenomena of electronic systems is in-
1. Heat Plate Operation
creasing [1]. The benefits are undisputed when it comes to
performing parametric studies in early design phases. The Heat plates are sealed by vacuum chambers that are par-
heat transfer behavior of electronic system is very complex. A tially filled with a working fluid, typically de-ionized water,
common question is the numerical simulations accuracy as the heat transfer media. The heat plate envelopes are
compared to well-designed experiments. Many studies dem- made of copper of different sizes, rectangular, or any other
onstrate excellent agreement when the design requirements enclosed geometry. The heat plate embedded with extrusion
fins is as shown on Fig. 1. The envelope wall is lined with a
wick structure, which provides surface area for the evapora-
Paper submitted 12/12/08; revised 01/21/09; accepted 07/03/09. Author for tion and condensation cycle and capillary capability. Since the
corresponding: Chun-Sheng Yu (e-mail: 892340067@s92.tku.edu,tw).
*Department of Mechanical and Electro-Mechanical Engineering, Tamkang
heat plate is evacuated and then charge with the working fluid
University, 151, Ying-Chuan Rd., Tamsui, 25137, Taipei, Taiwan. prior to being sealed, the internal pressure is set by the work-
**ChrangRong Technology Co., LTD. ing fluid vapor pressure [6].
420 Journal of Marine Science and Technology, Vol. 18, No. 3 (2010)

ductivity of a heat plate [8].

160 mm
Fourier’s Law is used to investigate the thermal conduc-
tivity and compare with test data to understand the heat
10 transfer efficiency of sintered wick structure in a heat plate.
The equation is as follows:

∆Tmeasured
Qin = K eff Aeff (2)
Radiating plate Leff
structure

Fig. 1. Mechanic dimension of Radiating Plate.

Heat transfer between a solid surface and air flow

∆Tmeasured 1
Rconvection = = (3)
The working fluid is vaporized as heat is applied to the Qin hc A eff
surface of the heat plate. The vapor at the evaporator section is
at a slightly higher temperature and pressure than other areas. Equation (3) where:
This creates a pressure gradient that forces the vapor to flow ∆T = The driving temperature difference between average
into the cooler regions of the heat plate. For a heat plate to fluid and the average surface temperatures [°C]
function properly, the net capillary pressure difference be- Qin = Power transferred [W]
tween the evaporator (heat sources) and condenser (heat sink) hc = Convective heat transfer coefficient [W/cm2-°C]
must be greater than the sum of all pressures losses occurring Aeff = Surface area exposed to flow [cm2]
throughout the liquid and vapor flow path equation (1). This
relationship, referred to as the capillary limitation can be ex- Heat transfer coefficient is a function of flow regime.
pressed mathematically as follows [14]: From (3), the vapor and condense section temperature are the
most important parameters for performance evaluation. The
∆Pmax > ∆P1 ＋ ∆Pv ＋ ∆Pg (1) temperature differences (∆Tmeasured) are obtained from ther-
mocouple measurements, and calculate the heat loss by (3)
∆Pmax is the maximum capillary pressure difference gener- from the interface to the heat plate wall that causes thermal
ated within the capillary wicking structure between the resistance equation (4).
evaporator and condenser, ∆P1 and ∆Pv are the viscous pres- Illustrated thermal resistance by network, then:
sure drops occurring in the liquid and vapor phases, respec-
tively, and ∆Pg represents the hydrostatic pressure drop. Rconvection Rfin
When the maximum capillary pressure is equal to or greater
than the sum of these pressure drops, the capillary structure Tsource Tbase Tfin
Q Q
can return an adequate amount of working fluid to prevent the
evaporator wicking structure from drying out. When the sum
Rconvection with fin efficiency = Rfin efficiency + Rconvection (4)
of all pressure drops exceeds the maximum capillary pumping
pressure, the working fluid is not supplied rapidly enough to
the evaporator to compensate for the liquid loss through 3. Heat Plate on Electronic Cooling Application
vaporization, and the wicking structure becomes starved of The most versatile feature of using heat plates is the wide
liquid and dries out to prevent the evaporator wicking struc- variety of geometries that can be constructed to take advan-
ture from drying out [15]. When the sum of all pressure drops tage of the available space around the electronics to be cooled.
exceeds the maximum capillary pumping pressure, the work- Proper heat source thermal isolated prevents heat loss during
ing fluid is not supplied rapidly enough to the evaporator to heating the heat plate vaporization area [7]. In many applica-
compensate for the liquid loss through vaporization. The tions, the available heat sink size above the electronics is
wicking structure then becomes insufficient of liquid and dries limited by the board-to-board spacing. In this situation, heat
out. This condition, referred to as capillary limitation, varies plates are used in a low profile system design that equalizes
according to the wicking structure, working fluid, evaporator the heat to a large fin stack.
heat flux, operating temperature, and body forces [16]. In general, the smaller component cooling approaches trans-
port 10 W to 30 W each. Currently industrial computers use a
2. Performance Evaluation heat plate in the thermal management design. Most of these
A heat plate works as an effective thermal conductivity (Keff) heat plate heat sinks use a 160 mm × 90 mm as a platform to
device that can transfer a huge heat flux with a small tem- eliminate hot spot [2]. This type of passive design has been a
perature difference between vapor and condense zones. This very effective thermal management technique for CPUs which
shows high or low heat transfer efficiency from thermal con- power ratings less then 30 W without cooling fan [9].
 C.-S. Yu et al.: Fabricating a Heat Pipe Structure within a Radiating Plate for Electronics Fan-less Cooling 421

T3 Pure Al heat sink test T4

T5 T6
T1 T2

T2 T1
T6 T5

Fig. 3. Heat pipe structure in a radiating plate.

T4 T3
Pure Al heat sink test + Heat Spreader
T3 T4 tains precision measurement at ±0.1°C. The T-type thermo-
couple is a precise temperature measurement sensor (test cri-
teria is 0°C~100°C in this study). The measured temperature
is calibrated by referring to the atmospheric boiling point at
T5 T2 T6 100°C.
T1
In order to maintain the heat plate surface flatness and mean
thickness of is the most important issue during testing, because
thermal performance depends on the conductivity at each
T2 interface. The test system setup requires heat source thermal
isolation to prevent data acquisition errors, which is always
T5 T1 T6 affected by ambient temperature.
Heater No. 1 Heater No. 2
2. Experimental Results
Fig. 2. Experiment measured point description. Figure 3 shows an extrusion (160 × 90 × 10 mm [L × W ×
H]). Thermal test is running by natural convection (ambient
temperature is from 25°C to 60°C) with 2 major heat sources.
Most thermal designs incorporate a heat sink or fins using a The sizes are 25 × 25 mm for each heater and total power
heat plate. This concept permits the computer system designer consumption is 30 W. These tests were performed with the
to locate the CPU independent of the heat sink. This novel heat source in the center of the extrusion.
methodology helps system designer to design the one of most Different fin configurations and measured temperatures are
effective heat exchangers and an optimal airflow path. The shown on Table 1. Standard fins are generally used for ap-
optimal design helps to reduce airflow requirements and plications where high heat dissipation and airflow rate are low.
acoustic noise. For applications where heat dissipation is needed, embedded
fins are employed to enhance the air convection. These fins
must be optimized when designing the whole system to ac-
III. EXPERIMENTAL INVESTIGATION
count for the pressure losses [13].
In situations which electronic is mounted on heat plate The temperature profiles at the surface of the extrusion with
units inside electronic system. The structure strength of the and without heat plates appear on Figs. 4, and 5, respectively
heat transfer device in this study is significantly improved and [4]. The shape of the temperature profiles on Fig. 4 demon-
thus complete contact between the electronic device to be strates that heat sink temperature. With heat plate assisted heat
cooled and the heat transfer device may be achieved. The sink, thermal design needs to place the electrical components
proposed device is embedded in the plate and transports the in convenient locations and is not limited by spreading resis-
heat to an air-cooled fin section. There are several different tance to the center of the heat sink.
sized units like this being used in the field [11]. As shown in Fig. 5, the heat plates also significantly re-
duced the maximum surface temperature of the extrusion. By
1. Experimental Setup simply adding heat plates to the base of an existing heat sink,
The HP-34970 data logger and thermocouple (T-type #36) the overall sink to ambient temperature rises is reduced by
are used to measure temperature at each point, as shown in 30°C. Figure 6 shows that a radiating heat plate presents
Fig. 2. The measured temperature is transmitted to a per- better performance than original aluminum and copper plate
sonal computer via RS-232 from data acquisition system. The design on electronics components cooling, and well propor-
software observes and records the temperature variation on the tioned temperature distribution on chassis skin. Figure 7 shows
heat plate. The HP-34970 data acquisition instrument main- the lowest thermal resistance of this fan-less thermal solution
422 Journal of Marine Science and Technology, Vol. 18, No. 3 (2010)

Table 1. Temperature differences on various fin direction ARK1000

and contact wall materials.
Mounting T1 T2
Fin direction T3 T4 T5 T6 Ta
Case place (NB) (CPU)
to gravity
Material °C
Normal to
1 Rubber 78.7 97.3 72.2 68.7 66.3 74.4 25.1
gravity
Upside side
2 Rubber 83.4 102.4 77.5 73.9 72 79.8 24.9
down
Temperature (°C)
Normal to
3 Air space 78.2 95.3 72.1 68.2 66.2 73.8 24.7 > 135.34
gravity
Upside side 112.76
4 Air space 78.2 96.4 71.7 68.3 66.1 74 26.3
down
Normal to 90.171
5 Wood 78 95.7 71.5 68.2 66.1 73.6 25.4
gravity
Upside side 67.586
6 Wood 85 102.8 78.5 75 73.5 80.5 24.6
down
Normal to < 45
7 Al plate 73.4 94.3 66.6 63.3 61.4 69.5 24.2
gravity
Upside side
8 Al plate 71.2 92.4 62.1 61.7 57.5 68.3 24.4 Fig. 5. Temperature distribution without hot spot (Heat plate).
down
Normal to
9 Spreader 77.1 84.5 74 73.1 75.1 74.2 24.3
gravity
120 T2
T1
100 T4
T6
Temperature (°C)

ARK1000
80 T5
T3
60 Ta

40

20

0
Temperature (°C)
Rubber Air space Wood Aluminum Heat plate
> 116.73 Materials on mounting place

98.801 Fig. 6. Temperature description of different fin performances.

g 80.867
4.00 Ta = 25°C
62.934 Ta = 35°C
3.80 Ta = 45°C
Thermal resistance (°C/W)

< 45
Ta = 55°C
3.60 Ta = 60°C
Fig. 4. Temperature distribution with hot spot (Aluminum fin).
3.40

3.20
is the best choice in natural convection in this study [12].
Faced with high spreading resistance, the thermal solution 3.00
design must explore the feasibility of increasing the base
thickness in an effort to reduce both the heat transfer by con- 2.80
1 2 3 4 5 6 7 8 9
duction and mass.
The system design must decrease thermal resistance be- Case study
tween any hot component and heat sink interface. In high Fig. 7. Thermal resistance comparison of various ambient temperatures.
power electronics system and server system, heat plates heat
sink assemblies have been shown to offer both cost and per-
formance advantages. With spreading acceptance, production tion boundaries. In the past few years, the use of heat pipes for
volumes are increasing and costs are falling. This trend is, in the thermal management of industrial computers is a proven
turn, resulting in an expanding market with broader applica- and widely applied technology. Furthermore, heat plate ap-
 C.-S. Yu et al.: Fabricating a Heat Pipe Structure within a Radiating Plate for Electronics Fan-less Cooling 423

Table 2. Application field of fan less thermal solution. sign cooling approaches for the growing heat dissipation lev-
Fan-less thermal solution els of power electronic devices. Heat plate assemblies (heat
(Natural convection) dissipation from 10 W to 30 W) allow increased heat sink
Item TDP (W) performance within the volume available with little potential
Al extrusion Radiating heat
heat sink plate impact on the existing system design.
AMD GX3 (1.1 W)* OK OK Radiating heat plate applications can improve fin efficiency
ULV-400 MHz (4 W)** OK OK
and eliminate hot spot for a fan-less system. This improve-
ment, either in terms of permitting higher packaging densities
Celeron-M 900 MHz (7 W) +
CPU NA OK or permitting higher power dissipation within a compact vol-
Intel 852 GM (3.2 W)**
ume, a radiating heat plate heat sink, in this situation, extends
Pentium-M 1.1 GHz (12 W) +
NA OK the thermal solution to acquire additional surface area for heat
Intel 855 GME (4.3 W)**
dissipation by natural convection, thus eliminating the need
*www.amd.com [3] **www.intel.com.tw [5]
for a fan.
Heat plate applications become the accepted standard ther-
mal design in industrial computers. Whereas volume con-
plications have helped enhance the computer industry to the
straints limit the use of a natural convection cooling solution, a
currently advanced level. Heat plate designs with optimal fin
heat plate to a miniature fan/sink can be more economical than
performance are easily handling the current requirements and
a large system fan solution.
allow additional industrial computer design flexibility. The
next generation of industrial computers is limited by the in-
creasing power that can be handled by optimized thermal REFERENCES
management systems [2]. 1. CFD Molding of a Therma-Base Heat Sink, Thermacore, Inc. (1998).
Heat plates are now being combined with other technology 2. HTTP://WWW.ADVANTECH.COM.TW
to help meeting the thermal requirements. In practice, these 3. HTTP://WWW.AMD.COM
cooling units might be the best solutions for keeping pace with 4. HTTP://WWW.FLOTREND.COM.TW
the increasing heat loads of high power semiconductors. After 5. HTTP://WWW.INTEL.COM
cooling the electronics which generate the most heat, there is 6. Khrustalev, D. and Faghri, A., “Thermal characteristics of conventional
the path to transfer heat from electronics components in the and flat miniature axially grooved heat pipes,” ASME Journal of Heat
Transfer, Vol. 117, No. 4, pp. 1048-1054 (1995).
chassis. A sealed air-to-air heat exchanger is best for dealing
7. Leong, K.-C., Liu, C.-Y., and Sun, K.-H., “Vapor pressure distribution of
with the residual heat inside the chassis. a flat plat heat pipe,” International Communications in Heat and Mass
Heat plate reliability and flexibility has proven a valuable Transfer, Vol. 23, pp. 789-797 (1996).
attribute that provides the system designer with increased 8. McCrery, Glenn, E., “Liquid flow and vapor formation phenomena in a
layout possibilities and typically improves thermal perform- flat heat pipe,” Journal of Heat Transfer Engineering, Vol. 15, No. 4, pp.
ance. Heat plate assisted heat sinks can manage non-uniform 33-41 (1994).
heating and higher heat fluxes better than conventional heat 9. Rajesh, V. G. and Ravindran, K. P., “Optimal heat pipe design: A non-
linear programming approach,” International Communication in Heat
sinks. Heat plates can be installed in the base of an extrusion
and Mass Transfer, Vol. 24, No. 3, pp. 371-380 (1997).
to reduce the conduction spreading resistance. Spreading 10. Riehl, R. and Dutra, T., “Development of an experimental loop heat pipe
resistance occurs when a heat source’s chip area is smaller in future space missions,” Applied Thermal Engineering, Vol. 25, pp. 101-
than that of the heat sink’s base. The electronic system can 112 (2005).
also be customized to fit the application as Table 2 listed. Heat 11. Said, S. A. and Akash, B. A., “Experimental performance of a heat pipe,”
plates offer an attractive approach to supplementing the con- International Communication of Heat and Mass Transfer, Vol. 26, No. 5,
ventional heat sink solution. The fin stack height, width, pp. 679-684 (1999).
length, and fin spacing can be calculated using CFD numerical 12. Shah, R. K. and Bhatti, M. S., “Laminar convective heat transfer in
ducts,” in Handbook of Single Phase Convective Heat Transfer, Wiley,
optimization, which allows the designer to extend the per-
New York (1987).
formance of more conventional heat sinks such as extrusions 13. Shah, R. K. and London, A. L., Laminar Flow Forced Convection in
or castings. The heat plate and optimal fin design assembly Ducts, Academic Press, New York (1978).
permits the designer to achieve otherwise unreachable thermal 14. Vafai, K. and Wang, W., “Analysis of flow and heat transfer characteris-
resistance level via the efficient maximization of the convec- tics of an asymmetrical flat heat pipe,” International Journal of Heat and
tive surface area available for dissipation to the ambient air. Mass Transfer, Vol. 35, pp. 2087-2099 (1992).
Thermal simulation technology application means rapid com- 15. Zhu, N. and Vafai, K., “Vapor and liquid flow in an asymmetrical flat
plate heat pipe: A three dimensional analytical and numerical investiga-
puter performance increases.
tion,” International Journal of Heat and Mass Transfer, Vol. 41, pp. 159-
174 (1998).
IV. CONCLUSION 16. Zhu, N., Vafai, K., and Wang, W., “Analysis of asymmetric disk-shaped
and flat plate heat pipes,” ASME Journal of Heat Transfer, Vol. 117, pp.
This study presented a progression of advance thermal de- 209-218 (1995).