AMOL KHARAT et al. Citation: 10.

2348/ijset07150940 ISSN (O): 2348-4098

Volume 3 Issue 4: 2015 Impact Factor- 3.25 ISSN (P): 2395-4752

HEAT TRANSFER ANALYSIS IN ELECTRONIC CIRCUIT BOARD

1AMOL KHARAT, 2SUNIL INGOLE, 3VISHAL MESHRAM

1,2,3Mechanical (Heat power) Department, Savitribai Phule Pune University, Indira College of Engineering and

Management, Pune

Email: [email protected], [email protected], [email protected]

ABSTRACT

In recent years, electrical equipment are drastically developing and miniaturizing. Higher integration of semiconductor
devices leads to these developments. On the other hand, higher integration of semiconductor devices on one chip and higher
integration of electrical devices in one package means increase of heat generation density. The operation of integrated
circuits (IC) at elevated temperature is a major cause of failures in electronic devices and a critical problem in developing
more advanced electronic packages. This is because the life expectancy of electronic components reduces exponentially as the
operating temperature rises. One such example of an electronic device is the electronic control unit (ECU) in automotives
whose function has increased and is expected to further rise in the foreseeable future.

Thermal design of vehicle controller is an important element in the total design process, because of the impact of temperature
on performance and reliability. A thermally well designed electrical component applied in a thermally poor designed
controller, will still result in a poor total design. And it is important that thermal design on controller should be included as
early as possible in the overall design process.

Keywords‐ Electronic Control Unit, Computational Fluid Dynamics, ANSYS Icepack, Temperature Distribution, Electronic
packages.

1. INTRODUCTION automobiles is increasing staidly as numerous

applications now use electronic control. Application from
Nowadays, electronic circuit devices can be found entertainment and comfort of power train and body
everywhere around us. Electronic Circuit Boards (ECB) electronics increasingly use electronic control circuits for
are an essential part of computers, audio and video better reliability and performance. Severe competition
devices, automobiles or airplanes, and also in less and stringer pollution and safety norms and fuel economy
common applications as diverse elements of spacecrafts are some of the factors influencing auto industry to look
and telecommunication satellites. The boards are made for new technologies using electronic controls. For many
from glass reinforced plastic with copper tracks in the electronic circuits printed circuit board forms the basic
place of wires. Components are fixed in position by foundation for interconnecting and packaging.
drilling holes through the board, locating the components Microcontroller can be described in simple words as
and then soldering them in place. The copper tracks link heart and brain of electronic circuits.
the components together forming a circuit. The resistance
to the flow of electrical current through the leads, poly‐ The thermal characteristic of a component due to own
silicon layers, and transistors comprising a power and upstream air heated by components were
semiconductor device, results in significant internal heat studied [1] study is concerned with forced convective
generation within an operating microelectronic heat transfer of the channel flow with line arrays of
component. As an individual electronic component heated electric components mounted on a printed circuit
contains no moving parts, it can often perform reliably for board. The experimental results were compared with
many years, especially when operating at or near room those of numerical solution for various conditions. The
temperature. In practice, integrated circuits operate at experimental results agree well with the numerical
substantially higher temperatures and, unfortunately, solution. The correlation between Nu (Nusselt number
most electronic components are prone to failure from and Re (Reynolds Number) was presented to make
prolonged exposure to elevated temperatures. This possible the direct calculation of the adiabatic heat
accelerated failure rate results from mechanical creep in transfer coefficient on a non uniform heated array. The
the bonding materials, parasitic chemical reactions, and IPM controls the rotational motion of the tube in a
do pant diffusion, to mention just a few possibilities. washing machine. A heat sink is installed in the IPM to
Thus, a rise in temperature from 75°C to 125°C can be prevent overheat during its operation [2]. In the design
expected to result in a five‐fold increase in failure rate. process of heat sink, the performance and heat
Under some conditions, a 10°C to 20°C increase in chip dissipation performance are simultaneously considered.
temperature can double the component failure rate. Three dimensional heat and fluid analysis of two plastic
lead chip carrier packages mounted in tandem
Electronic control unit (ECU) in automotives whose arrangement on a printed circuit board [3] exposed to the
function has increased and is expected to further rise in free stream velocity. The decrease in the junction
the foreseeable future The electronic content in today’s temperature of the packages with the increase in

International Journal of Science, Engineering and Technology- www.ijset.in 940

AMOL KHARAT et al. Citation: 10.2348/ijset07150940 ISSN (O): 2348-4098
Volume 3 Issue 4: 2015 Impact Factor- 3.25 ISSN (P): 2395-4752

approach air velocity is clearly observed. The results also Figure 2: actual experimental layout
show that the spacing between packages influences the
thermal resistances and averages Nusselt number for In this experiment ECU present in the drivers cabin As
both packages at a particular approach air velocity. shown in figure power supply provided to the ECU. AC
power supply provided to the temperature indicator.
Heat generated within the unit coupled with ambient Vehicle starts normally after three hours steady state
temperature makes the system reliability susceptible to condition occur and different readings taken on the
thermal degradation which ultimately may result in different packages.
failure [4] With studies suggesting that current TIMs
contribute about 60% interfacial thermal resistance. A 3. CFD SIMULATION
higher thermal conductivity material could improve heat
dissipation from ECU [5]. The optimal placement of The ANSYS Icepack software among various CFD tool
electronic component on a printed circuit board requires specifically tailored for use in the electronic industry to
satisfying conflicting design objectives as most of the perform the board and system level thermal analysis. In
component have different power dissipation, operating icepack a variety of thermal limitation source elements
temperature, types of material and dimension. The SOGA are available, including heat dissipation, fixed
(self organising genetic algorithm) gives a better optimal temperature sources or heat flux sources. Typically
solution as compared to the other method [6]. The new controllers in vehicle are housed in sealed case, and rely
placement configuration for the LED array could be used solely on free convection and radiation for cooling. The
lowers the individual LED temperature. [7] heat flow path in controller is a series of conduction,
convection and radiation path. Much of the generated in
2. EXPERIMENTAL SET UP an integrated circuit chip is first conducted into printed
circuit board. From here, heat is transported through the
Figure 1; figure 2 shows the experimental layout and case wall, and finally removed to the environment by free
actual experimental set up respectively. J type convection.
thermocouple mounted on the different packages. On the
packages thermocouple end attached with the help of 3.1 SIMULATION MODELLING
thermal grease. Figure 2 shows the different
thermocouple attached on the board packages. These Figure shows the actual ECU of vehicle. When the vehicle
thermocouples are connected to the temperature is driving under loaded condition the total power
indicator. Due to space limitation only three consumption is 10 to 15W. Figure1 and figure2 the
thermocouples attached at one time. This procedure controller consists of the printed circuit board with
repeated a three times and total 11 reading taken on power electronic components mounted on it which is
different packages. sealed in a metal case.

Figure 1: experimental layout
Figure 3: ECU of automotive vehicle

Figure 4: Electronic control board with enclosure

The following table indicate the detail specification of the

electronic control unit for normal and loaded condition.
International Journal of Science, Engineering and Technology- www.ijset.in 941
AMOL KHARAT et al. Citation: 10.2348/ijset07150940 ISSN (O): 2348-4098
Volume 3 Issue 4: 2015 Impact Factor- 3.25 ISSN (P): 2395-4752

TABLE I: SPECIFICATION OF THE ECU Figure 6: The whole solution domain

The model case was constructed as being entirely sealed

as a close approximation to the real case geometry. As
four edges of the PCB (printed circuit board) are close in
contact with the case, thermal contact resistance is also
considered in the model. The CFD software requires the
model placed within a solution domain which was created
by enclosure, heat and mass transfer through enclosure is
allowed.

The PCB element in the model is assumed to be

homogeneously populated and consist of top layer and
bottom layer copper were considered as 0.035mm thick.
Copper traces that cover 50% of the board area.

Figure 5 shows the whole solution domain with six open

faces. The finite volume method is adopted for converting
governing equations to algebraic equation that can be
A figure three 3 shows the detail model of ECU with solved normally. The simple algorithm is used to solve the
different component on the PCB. A number of simplifying pressure and convection‐velocity coupling term.
assumptions were made in the modelling. All the smaller Discretization method is first order upwind scheme. In
and lower power dissipation components on the circuit the solution domain the maximum size of element in
board (capacitor, resistor, mount IC) were neglected. The three mutually perpendicular directions is less than 1/20
main generating heat sources (MOSFET, FET PLCC, QFP, of every domain dimension. Figure 5 shows the meshing
Voltage regulator) were omitted. distribution of whole domain. The computational domain
is discretized into unstructured hexahedral meshes that
have a total of 178219 nodes.

Figure 5: Thermal model for ECU Figure 7: Meshing
Small outline packages MOSFETs (metal oxide The mesh elements are smaller near objects, to take
semiconductor field effect transistor) are simplified into accounts thermal and velocity gradient that are often
planar heat sources. QFP (quad flat packages) and PLCC ( present near the boundaries of an object. By contrast, the
plastic lead chip carrier) packages are directly mounted open spaces between objects are meshed with large
on the board. Other components which are represented elements, to minimize computational costs.
by 3D solid volumes. The power supply provided to the
all the components. The controller shell was modelled The graph shows the x, y, .z velocity, continuity, energy, k
using enclosure which is 4mm thick and the material is and epsilon. The 150 iterations are sufficient to calculate
aluminium. the above problem. Prior to solving the model, ANSYS
Icepack will determine whether the flow will be
dominated by forced or natural convection. For flows
dominate by natural convection means buoyancy driven
flow , ANSYS Icepack computes approximates value of the
Rayleigh number and the Prandtl number

International Journal of Science, Engineering and Technology- www.ijset.in 942

AMOL KHARAT et al. Citation: 10.2348/ijset07150940 ISSN (O): 2348-4098
Volume 3 Issue 4: 2015 Impact Factor- 3.25 ISSN (P): 2395-4752

CFD analysis temperature inside the chip is also detected.

Figure 8 shows also surface temperature of the board for
validation purpose.

Once the normal condition temperature is validating with

the CFD calculated temperature the loaded condition
apply on the board in different result plotted with the
help of ansys icepack.

Figure 8: temperature distribution at normal condition
Graph1: Different CFD values for 150 iterations
4.1 CFD ANALYSIS FOR LOADED CONDITION
If the Rayleigh number is greater than 5E + 7 and the
Prandtl number is around is 0.71, the turbulent model The following figure 9 shows the temperature field
with natural convection will be recommended. Based on distribution of the printed circuit board and components
the physical characteristic of the above model as defined , for loaded condition.
the calculated Rayleigh number and Prandtl number by
software are 4.6090E + 9 and 0.7085 respectively.

4. RESULT AND DISCUSSION

The following table shows the experimental and

computational values.

Packages Experimental CFD

QFP 51 52.56

QFP 46 46.15

PLCC 47 49.17 Figure 9: Temperature distribution of the board at loaded

condition
Voltage Regulator 45 48.07
After calculating 150 iterations the board temperature
SMD 51 52.88 varies from 31 to 82.52˚c. The simulation predicted
maximum temperature in QFP package and MOSFET with
SMD 46 46.23 a value of 82.47oc. The three MOSFET are very near that’s
why the temperature increases on the board. The QFP
IC 41 42.24 package and MOSFET are also close due to that average
temperature of the board increases.
MOSFET 48 48.48
4.2 EFFECT OF COPPER PERCENTAGE ON THE BOARD
MOSFET 50 50.23
The copper percentage increases the thermal
MOSFET 48 48.54 conductivity of the board. In this PCB the copper
percentage is 50% the average thermal conductivity of
MOSFET 47 48.08 the board in plane and normal direction is 9.45688
W/mk, 1.04551W/mk respectively. When the copper
percentage decreases from 50% to 20% average
temperature of the board increases to 4˚c.
The maximum error between experimental and
computational is 6.38. Figure 8 shows the temperature Above 50% the temperature of board slowly decreases
distribution on the board. The temperature range is 38.32 but at the same time cost increases of printed circuit
to 64.21. in the experimental measurement only surface board. Figure (7) indicating, when the copper percentage
temperature of different packages is measured and in increases up to 65% the average thermal conductivity of

International Journal of Science, Engineering and Technology- www.ijset.in 943

AMOL KHARAT et al. Citation: 10.2348/ijset07150940 ISSN (O): 2348-4098
Volume 3 Issue 4: 2015 Impact Factor- 3.25 ISSN (P): 2395-4752

the board in plane and normal direction is 11.9939 application such as home computer, vehicle electronics
W/mk, 1.0456 W/mk respectively and car entertainment products.

The following table gives the different thermal propriety

of FR4 material and composite epoxy material.

TABLE II: MATERIAL PROPERITY OF FR4 AND CEM

FR4 CEM

CONDUCTIVITY(W/mK) O.35 1.0

DENSITY(kg /m3) 1250 1000

SPECIFIC HEAT(J/kg.K) 1300 750

CONUCTIVITY TYPE ISOTROPIC ISOTROPIC
Figure 10: Effect of copper percentage on board

After increasing the copper percentage temperature
decreases from 82.52oc to 80.94oC. Due to manufacturing Figure (12) shows the predicted temperature field
consideration percentage of the copper cannot increases distribution of the board with substrate material of CEM‐
above 65 percentages but when copper percentage 3(composite epoxy material)[8]
increases cost of the board increases.

4.3 EFFECT OF COMPONENT PLACEMENT ON

TEMPERATURE FIELD

The component placement is very effective and chi pest

technique in electronic cooling. Figure (11) indicating the
well component placed electronic circuit board. After the
component placement temperature decrease 2 to 3oc in
this ECU.

Figure 12: Effect of material on electronic circuit board

The composite epoxy material has high thermal

conductivity compare to FR4 also specific heat is also low
that’s why it improve thermal management in the board
.The maximum temperature is reduced to 78.94, which
has lower compared to the79.33. The average
temperature decreases from o.5 to 1.5○c. Thus it is more
advantageous to reduce the board temperature to select
high conductivity substrate materials.
Figure 11: Effect of component placement on the board
5. CONCLUSION
and components
Thermal management and thermal management design in
The three MOSFET on the board are very close due to that
automotive ECUs has long been a challenging issue as its
concentration of temperature increases in that particular
compact on the devices reliability and endurances.
region. Component placing temperature of the board
decreases from 80.53 to 79.22. This is not final optimum A thermal analysis model for sealed vehicle controller is
placement of the board. Taking number of trials demonstrated in this paper, the main conclusion are
temperature will decrease for PCB. summarised as the following
4.4 EFFECT OF PCB MATERIAL ON TEMPERATURE More than 65% coverage percentage is recommended for
DISTRIBUTION a better thermal performance.
The printed circuit board itself becomes the heat sink. Selecting high conductivity substrate material is more
The FR‐4 material which has very low conductivity is due advantageous to reduce board temperature
to that temperature field concentration increases. Two
types of most commonly substrate material used board Component placement is also an important factor for the
substrate material are also examined. FR4 and composite board temperatures. Multiple trials for components
epoxy material are very similar. Composite epoxy location can be performed in order achieve an optimal
material is available at low cost reduced for FR4 in thermal design effect.
International Journal of Science, Engineering and Technology- www.ijset.in 944
AMOL KHARAT et al. Citation: 10.2348/ijset07150940 ISSN (O): 2348-4098
Volume 3 Issue 4: 2015 Impact Factor- 3.25 ISSN (P): 2395-4752

REFERENCES

[1] Kwang soo kim, Won Tae Kim, Ki Baik Lee, “cooling
characteristic on the forced convection of an array of flat‐
form electronic components in channel flow”: KSME
international journal, vol12, 1998.

[2] Chung‐Hyo Jung, Young‐Suk Chung, Hyung‐Woo Lee,

“Investigation on thermal characteristic of heat sinks for
power module using STM” :Journal of Mechanical Science
and technology , 686‐697, 2008.

[3] S.Yusoff, M.Mohmad, K.A. Ahmad, M.Z.Abdullah, “3D

conjugate heat transfer analysis of PLCC packages
mounted in‐line on a printed circuit board, International
communication in heat and mass transfer”: 813‐
819,2009.

[4] K.C. Otiba, N.N. Ekere, R.S.Bhatti,S.Mallik, “Thermal

interface materials for automotive electronic control
unit”:Microelectronic Reliability,2031‐2043,2011.

[5] Sabuj Mallik, Ndy Ekere, Chris Best, Raj bhatti,

“Investigation of thermal management materials for
automotive controls units”:Applied Thermal Engineering,
355‐362, 2011.

[6] Fatimah Sham Ismail, Rubiyah Yusof, Marzuki Khalid,

“Optimization of electronic component placement design
on PCB using self organizing genetic algorithm”: springer
science +business media, 2010.

[7] K.C.Yung, H. Liem, H.S.Choy, “Heat transfer analysis of

a high brightness LED array on PCB under different
placement configurations”: International Communication
in Heat and Mass Transfer, 79‐86, 2014.

[8] K Sauter, Evaluating PCB design, manufacturing

process, and laminate material impacts on CAF
resistance, in: IPC Printed Circuits Expo Technical
Conference Proceedings, March 2002.

International Journal of Science, Engineering and Technology- www.ijset.in 945