THERMAL RESISTANCE OF A HEAT SINK
Norah binti Tuah
FACULTY OF INFORMATION TECHNOLOGY AND SCIENCE QUANTITATIVE
MARA UNIVERSITY OF TECHNOLOGY
Abstract
The main objective of this paper is to explain the behavior of thermal resistance in a heat sink. An assembly of a heat sink is built in a Solidworks program. The simulation of it is made by using a simulation package called Cosmos Flowork software. All the result from the simulation is compared with the available datasheet from Intel. Besides that, the result from the ‘wind tunnel’ experiment is also used. The experiment was done in a steady state condition. From the simulation there are good relationship data between these three types of works.
NOMENCLATURE
R = Thermal resistance in oC/W
= Temperature different between two locations in oC
LFM = Linear Feet per Minute
Q = Total power or rate of heat dissipation in Watt
Tj = Maximum junction temperature of the device in °C
Tc = Case temperature of the device in °C
Ts = Sink temperature in °C
Ta = Ambient air temperature in °C
Rjc = Thermal resistance between the junction and the case in OC/W
Rcs = Thermal resistance between the case and the sink in OC/W
Rsa = Thermal resistance between the sink and the ambient in OC/W
Rja = Thermal resistance between the junction and the ambient in OC/W
Rca = Thermal resistance between the case and the ambient in OC/W
Pmax = Maximum power dissipation in Watt
= Rate of energy inflow into the system, Kilo Joule (KJ)
= Rate of energy generated inside the system, KJ
=Rate of energy outflow from the system, KJ
=Rate of change of internal energy of the system, KJ
INTRODUCTION
Heat sink is an important device to control the heat removal from the Central Processing Unit (CPU). The rate of heat removal from the CPU must be monitored to ensure the computer operation is always good. Nowadays, the heat generated from the CPU can reach 100 oC (generated from Mobile Intel Pentium II processor). Besides that, increasing the density and decreasing the size of the CPU will make designing of a heat sink in the future mandatory.
In this research the CPU used is a Pentium processor with a clock speed of 60 MHz. It was designed by Intel in 1993. It is the second type of CPU that had installed with heat sink. The old chip is used here because this research is a basis to more advance research or component designing in the future.
In this research an experiment was done in a steady state condition. A system is said to be in a steady state when its temperature does not change with time. This means that the change of internal energy of the system is zero.
RESEARCH PROBLEM
A microprocessor is the heart of the microcomputer. If it is damaged all the system in the microcomputer will not function. Moore has made a research on the density and power dissipation in the microprocessor and made the conclusion through his famous Moore’s Law saying that the value of it will double roughly every 18 months. His insightful observation on device advancement has had a direct correlation with power dissipation. Figure 1 shows the power dissipation of some representative Intel microprocessor as a function of time. When the power dissipated by a microprocessor rises above a few watts, you might need to start thinking about adding a heat sink. An integrated circuit die can get only so hot before the transistors on the die are damaged. Thus, all microprocessors have a specified maximum temperature, as shown at Table 1 below. Typical maximum temperatures are in the range of 125oC to 150 oC.
This maximum temperature refers to the temperature of the actual circuit, or die, not the case. If the thermal resistance between die and the case is high, the case may not be too hot, yet the die might be very hot. If the die temperature ever exceeds the maximum temperature, the manufacturer will not give any guarantee on the defect of the device. Often, this is a short prelude to permanent degradation or destruction of the device. Maximum temperature refers to the temperature at which damage will occur. Performance often degrades well before the maximum die temperature is reached. The commercial temperature range is often specified as 0 to 70 oC. Beyond 70oC, then, you cannot expect that the device will perform within the guaranteed.
Figure 1. Time-line Plot of the Intel Corporation Microprocessor Power Dissipation.
Table 1: Processor Electrical Specification
CPU Max. Power Dissipation [ Watt ] Maximum case temperature [ Celsius ]
AMD 3.7 – 65 55 - 85
Cyrix/IBM 4.32 – 27.9 70 – 85
IDT 9.5 – 16 70
INTEL 1.1 – 47 45 – 100
SGS-Thomson 2.243 – 7.875 85
Rise 3.5 – 10.72 72
LITERATURE REVIEW
Lee (1995) made an analytical simulation model to predicate and optimize the thermal performance of hi-directional fin heat sinks in a partly confined configuration. The model is validated by comparing the results with existing experimental data and sample cases are presented with discussions on the parametric behavior and optimization of hi-directional heat sinks.
Forghan (2000) made an experimental and theoretical study to investigate the thermal performance of a variety of heat sinks. The experiments were done in a wind tunnel where free stream velocity ranged from natural convection to 5 m/s. Flowtherm, the software used for the development of the theoretical results, provides a detailed model of heat sink by solving the Navier – stokes and energy equations with the consideration of several turbulence generation schemes. From their observation, the theoretical model provides a good approximation to the actual experimental results.
Chapman (1994), made a comparison of the thermal performance of different fin geometries. Crosscut pin fin and straight or parallel plate fins were investigated and compared with elliptical pin fin heat sink. In their experiment, results were obtained with an aluminum heat sink made of extruded fin, crosscut rectangular pins and elliptical pins in laminar airflow. All three heat sinks had equal volume, and the total surface area was also calculated to be nearly identical. The heat sink and ambient temperature difference was used to calculate thermal resistance, they observed that the elliptical pin-fin heat sink was able to minimize the pressure loss across the heat sink by reducing vortex effects, and to enhance the thermal performance by maintaining a large exposed surface area available for heat transfer.
Thermal resistance is a quantitative measure of heat transfer efficiency across two locations of a thermal component, expressed as
(1.0)
From the Equation (1.0) above, if the value of R is small that means more heat transfers easily. This is the target in any designing of heat sink application devices. The basic concept of a thermal circuit is seen in Figure 1.1 below.
Figure 1.1: Thermal Circuit
It is specified by the device manufacturer (constant value) (1.1)
It is called Interface resistance (1.2)
It is called Heat sink thermal resistance (1.3)
It is called junction-to-ambient resistance. (1.4)
To meet the requirement for the satisfaction of the component, the first step is to determine the heat sink thermal resistance. By rearrangement of the previous equation, the heat sink resistance can perform as
(1.5)
in this expression, Tj,Q and Rjc are provided by the device manufacturer, and Ta and Rcs are the user defined parameters. Ta for cooling electronic equipment depends on the operating environment in which the component is expected to be used. It ranges from 35oC to 45oC (for external air is used) and from 50oC to 60oC (for component is enclosed or is placed in a weak state of another heat generating equipment). Rcs depends on the surface finish, flatness, applied mounting pressure, contact area, and of course the type of interface material and its thickness.
With all the parameters on the right side of the above expression identified, Rsa becomes the required maximum thermal resistance of a heat sink for the application. In other words, the thermal resistance value of a chosen heat sink for the application has to be equal or less than the above Rsa value for the junction temperature to be maintained at or below the specified Tj.
The smaller the value, the heat sink will more smoothly transfer the heat to the ambient air. If Tj is bigger than 150 oC, we need to add the fan/blower to support the heat sink performance.
STEADY STATE AND UNSTEADY STATE HEAT TRANSFER
Energy balance equation:
and have 3 types of heat transfer: conduction, convection and radiation.
A system is said to be in a steady state when its temperature does not change with time. This means that the change of internal energy of the system is zero. In an equation, this implies that . Consequently, the equation above becomes .
For an unsteady state system, the temperature of the system changes with time. Therefore, the internal energy of the system will vary from time to time. In the equation, this means . There are 2 types of unsteady state system:
1. Regular periodic unsteady state.
The temperature at the various space points, although changing with time, changes according to a definite pattern, namely a cyclic variation in time that is repeated. For an example, the temperature distribution within the cylinder wall of a reciprocating automobile engine operates at a constant speed over a period time.
2. Transient unsteady state.
The temperature variation at a space point, as time goes on, is not cyclic, for example, in jet engines and rocket engines during start up and shutdown.
EXPERIMENTAL REVIEW
The data from Intel was taken by its real working experiment on the Pentium processor. The data was also taken from the ‘wind tunnel’ experiment to make a comparison to the simulation. All the data was presented in Table 1.2.
A wind tunnel experiment is a measurement of thermal resistance using the wind tunnel device. By applying the temperature sensitive parameter (TSP) method, special design thermal test structures are used which are approximately the same size as the Pentium processor die. The test structure consists of resistor and diodes. Resistors are used to simulate the Pentium processor power dissipation and thereby heat up the package. Diodes located at the center of the thermal test die are used to measure the die temperature. The measurements are carried out in a wind tunnel environment. The airflow rate and the ambient temperature are measured 12 inches away from the package in the upstream air.
SIMULATION CONDITION
Applying the real condition from the real working experiment by Intel that did the simulation.
The heat sink size for this simulation is 0.5 x 2.1 square (High x Wide, inches), using a Pentium processor 60 MHz as the heat sources with the maximum power dissipation of 14.6 Watt. Considering only to Case – to – ambient thermal resistance, Rca. To remove the heat we use a passive heat sink (without fan assembly). The location of the fan is 15 millimeters from the heat sink assembly and thus produces airflow in the x-axis. It will help in reducing of thermal resistance in a heat sink assembly.
The environment of this heat sink assembly is in a steady state condition. The heat will dissipate from sink-finned side only. The heat sink is made from Aluminum with a rectangular pin fin model. The power distribution from the CPU is 8.7785 Watt. The ambient room is 23 oC with airflow distribution is around 175 to 200 LFM.
Figure 1.2 shows the heat sink assembly model and it meshing for the simulation. The part involved in this assembly is ZIF socket, Pentium processor P66, heat spreader, thermal grease and heat sink.
Figure 1.2: Heat sink assembly simulation
SIMULATION RESULT
Figure 1.3 and Figure 1.4 show the heat distribution result in 175 LFM and 198 LFM. It represents the front view of the model. The arrow signs show the velocity movement in the x-axis. From that picture we can see the change of temperature mood of the heat sink assembly. By adding the velocity value the heat around the heat sink assembly will decrease smoothly.
Figure 1.3: 175 LFM result Figure 1.4: 198 LFM result
Table 1.1 and Table 1.2 below show the calculation result for 175 LFM and 198 LFM. From that table average temperature of ambient air presented the average temperature around the heat sink assembly. Then the rest is the part that builds the heat sink assembly with their own temperature value result.
From table 1.1 the case temperature is 69.59oC and the ambient temperature is 36.79 oC. By using Equation (1.6) below the thermal resistance value for airflow with 198 LFM is 3.74 oC/W. From Table 1.2 the thermal resistance value for airflow with 198 LFM achieved is 3.19 oC/W. Thermal resistance value will decrease by increasing the velocity of the fan.
(1.6)
Table 1.1: The goal result for an experiment 3 assembly in 175 LFM
Goal Name / Unit / Value / Use In ConvergenceAverage Temperature of ambient air / [°C] / 36.785 / Yes
Average Temperature of Pentium processor / [°C] / 70.100 / Yes
Average Temperature of thermal grease / [°C] / 69.586 / Yes
Average Temperature of heat sink / [°C] / 69.348 / Yes
Average Temperature of heat spreader / [°C] / 69.586 / Yes
Average Temperature of ZIF socket / [°C] / 72.667 / Yes
Table 1.2: The goal result for an experiment 3 assembly in 198 LFM
Goal Name / Unit / Value / Use In ConvergenceAverage Temperature of ambient air / [°C] / 32.561 / Yes
Average Temperature of Pentium processor / [°C] / 60.999 / Yes
Average Temperature of thermal grease / [°C] / 60.578 / Yes
Average Temperature of heat sink / [°C] / 60.361 / Yes
Average Temperature of heat spreader / [°C] / 60.579 / Yes
Average Temperature of ZIF socket / [°C] / 63.316 / Yes
COMPARISON OF SIMULATION AND EXPERIMENTAL RESULTS