ArticlePDF Available

Abstract and Figures

Purpose – The purpose of this paper is to develop thermal modelling to investigate the thermal response of sample boards (at board level) during the preheating stage of the reflow process and to validate with experimental measurements. Design/methodology/approach – A thermal-coupling method that adopted the Multi-physics Code Coupling Interface (MpCCI) was utilized. A forced-convection reflow oven was modelled using computational fluid dynamic software (FLUENT 6.3.26), whereas structural heating at the board level was conducted using finite-element method software (ABAQUS 6.9). Findings – The simulation showed a complex flow pattern having characteristics of a free-jet region, stagnation-flow region, wall jet-region, recirculation region and vortices. A sharp maximum heat-transfer coefficient was detected in the stagnation region of the jet, resulting in a spatial variation of local heat transfer on a thermal profile board (TPB). This coefficient affected the temperature distribution in the TPB with different specific heat capacitances and thermal conductivity of the structure. The simulation results were in good agreement with the experimental data and analytical model. The cold region and temperature uniformity (?T) increased with increasing complexity of the TPB. The cold region can occur in two possible locations in the TPB. Both occurrences can be related to the flow field of the reflow oven. ?T of the TPB decreased when the conveyor speed (v) was reduced. A suitable conveyor speed (1.0?cm/s) was determined to maintain ?T below 10°C, which prevented the thermally critical package from overheating. Practical implications – The paper provies a methodology for designing a thermal profile for reflow soldering production. Originality/value – The findings provide fundamental guidelines to the thermal-coupling method at the board and package levels, very useful for accurate control of ?T at the board and package levels, one of the major requirements in achieving a high degree of reliability for electronic assemblies.
Content may be subject to copyright.
Soldering & Surface Mount Technology
Emerald Article: Three-dimensional thermal investigations at board level
in a reflow oven using thermal-coupling method
Chun-Sean Lau, M.Z. Abdullah, F. Che Ani
Article information:
To cite this document:
Chun-Sean Lau, M.Z. Abdullah, F. Che Ani, (2012),"Three-dimensional thermal investigations at board level in a reflow oven using
thermal-coupling method", Soldering & Surface Mount Technology, Vol. 24 Iss: 3 pp. 167 - 182
Permanent link to this document:
http://dx.doi.org/10.1108/09540911211240038
Downloaded on: 28-06-2012
References: This document contains references to 22 other documents
To copy this document: permissions@emeraldinsight.com
Access to this document was granted through an Emerald subscription provided by Emerald Author Access
For Authors:
If you would like to write for this, or any other Emerald publication, then please use our Emerald for Authors service.
Information about how to choose which publication to write for and submission guidelines are available for all. Please visit
www.emeraldinsight.com/authors for more information.
About Emerald www.emeraldinsight.com
With over forty years' experience, Emerald Group Publishing is a leading independent publisher of global research with impact in
business, society, public policy and education. In total, Emerald publishes over 275 journals and more than 130 book series, as
well as an extensive range of online products and services. Emerald is both COUNTER 3 and TRANSFER compliant. The organization is
a partner of the Committee on Publication Ethics (COPE) and also works with Portico and the LOCKSS initiative for digital archive
preservation. *Related content and download information correct at time of download.
Three-dimensional thermal investigations
at board level in a reflow oven using
thermal-coupling method
Chun-Sean Lau and M.Z. Abdullah
School of Mechanical Engineering, Universiti Sains Malaysia, Nibong Tebal, Malaysia, and
F. Che Ani
Celestica Malaysia Sdn. Bhd, Kulim, Malaysia
Abstract
Purpose The purpose of this paper is to develop thermal modelling to investigate the thermal response of sample boards (at board level) during the
preheating stage of the reflow process and to validate with experimental measurements.
Design/methodology/approach A thermal-coupling method that adopted the Multi-physics Code Coupling Interface (MpCCI) was utilized.
A forced-convection reflow oven was modelled using computational fluid dynamic software (
FLUENT 6.3.26
), whereas structural heating at the board
level was conducted using finite-element method software (
ABAQUS 6.9
).
Findings The simulation showed a complex flow pattern having characteristics of a free-jet region, stagnation-flow region, wall jet-region,
recirculation region and vortices. A sharp maximum heat-transfer coefficient was detected in the stagnation region of the jet, resulting in a spatial
variation of local heat transfer on a thermal profile board (TPB). This coefficient affected the temperature distribution in the TPB with different specific
heat capacitances and thermal conductivity of the structure. The simulation results were in good agreement with the experimental data and analytical
model. The cold region and temperature uniformity (DT) increased with increasing complexity of the TPB. The cold region can occur in two possible
locations in the TPB. Both occurrences can be related to the flow field of the reflow oven. DT of the TPB decreased when the conveyor speed (v) was
reduced. A suitable conveyor speed (1.0 cm/s) was determined to maintain DT below 108C, which prevented the thermally critical package from
overheating.
Practical implications The paper provies a methodology for designing a thermal profile for reflow soldering production.
Originality/value The findings provide fundamental guidelines to the thermal-coupling method at the board and package levels, very useful for
accurate control of DT at the board and package levels, one of the major requirements in achieving a high degree of reliability for electronic assemblies.
Keywords Assembly, Flow, Heat transfer, Computer software, Thermal-coupling method, Forced-convection reflows, Computational fluid dynamics,
Finite-element method, Heat-transfer coefficient
Paper type Research paper
Nomenclature
A¼area of top surface of PCB (m
2
)
A
heat
¼surface area for heat transfer by convection
(m
2
)
c
p
¼specific heat capacity of structure (J/Kg.K)
D¼diameter of the orifices (cm)
H¼distances between orifice and the assembled
PCB (cm)
h
avg
¼average convective heat-transfer coefficient
(W/m
2
.K)
h¼convection heat-transfer coefficient (W/
m
2
.K)
k¼thermal conductivity of the gas (W/m.K)
P¼total area of top surface of package (m)
q¼convectional heat-transfer energy (W)
_
q¼convectional heat-transfer energy per unit
volume (W/m
3
)
r¼orifice pitch (cm)
T
g
¼temperatures of the environment (K)
T
o
¼initial temperature (K)
T
s
¼actual temperature of surface (K)
T¼temperature (K or 8C)
D
T¼temperature uniformity (K or 8C)
t¼time (s)
U
x
,U
y
,U
z
¼velocity component in x,y,zdirection (mm/
s)
v¼conveyor speed (cm/s)
V¼total volume of structure (m
3
)
x,y,z¼Cartesian coordinates (–)
Greek letters
r
¼density of structure (kg/m
3
)
L
¼thermal conductivity of PCB or package (W/
m.K)
t
¼time constant (s)
The current issue and full text archive of this journal is available at
www.emeraldinsight.com/0954-0911.htm
Soldering & Surface Mount Technology
24/3 (2012) 167– 182
qEmerald Group Publishing Limited [ISSN 0954-0911]
[DOI 10.1108/09540911211240038]
The authors gratefully acknowledge the financial support of the Ministr y
of Higher Education of Malaysia for the FRGS grant scheme.
167
1. Introduction
Reflow thermal profiling is affected by the miniaturization of
electronic packages and the increasing complexity of printed
circuit board (PCB) assemblies. The increasing complexity of
PCBs also leads to more complicated thermal responses when
a PCB assembly passes through a reflow oven (Florian et al.,
2009). The increasing trend toward lead-free soldering due to
environmental concerns is another challenge for reflow
thermal profiling. A lead-free solder typically requires a
narrower range of flow temperature and a workable melt
compared with a lead-based solder. An inadequate reflow
profile may not only result in a high level of thermal stress in
the package but may also result in a variety of soldering
defects (Tsai, 2009). Thus, soldering defects can result in
significant reliability issues in the electronics industry.
An offline tool for the reflow process will be greatly useful to
the electronics manufacturing industry. Therefore, modelling
the soldering process to predict the required parameter setting
is necessary. A simplified analytical model based on lumped
thermal capacity at the package level, as shown in equation
(1), has been described in a number of studies (Van
Steenberge et al., 2006; Gao et al., 2008; Ille
´s, 2010a, b).
Equations (2) and (3) show the solution obtained from
equation (1) for constant environment temperature. The time
constant is an important parameter, which describes the time
required for the package to reach a set temperature. The
convection heat-transfer coefficient (h) was assumed during
the simulation:
cp
r
VdT
dt ¼hAheat ðTg2TsÞ;ð1Þ
T¼ToþðTair 2ToÞð12e2t=
t
Þð2Þ
t
¼cp
r
V
hAheat ð3Þ
where c
p
is the specific heat capacity of the structure (J/kg ·K),
r
is the density of the structure (kg/m
3
), Vis the total volume
of the structure (m
3
), T
g
is the temperature of the
environment (K), T
s
is the actual temperature of the surface
(K), his the convection heat-transfer coefficient (W/m
2
· K),
A
heat
is the surface area of heat transfer by convection (m
2
), T
o
is the initial temperature (K) and
t
is the time constant (s)
given in equation (3).
The most popular thermal response analysis at the package
level is based on the finite-element method (FEM). Shen et al.
(2005) and Inoue and Koyanagawa (2005) built a FEM model
to obtain the temperature distribution of a ball grid array
(BGA) package for the reflow process. The average heat-
transfer coefficient (h
avg
) was calculated using experimental
equations for the multiple impinging jets shown in equations
(4)-(6) (Inoue and Koyanagawa, 2005). However, this method
is incapable of dealing with the changes in some parameters in a
reflow oven (e.g. the flow velocity and density of the gas). Ille
´s
(2010a, b) discovered an experimental method to determine
the distribution of the hvalues under the orifice-matrix of a
convection reflow oven. In those experiments, the temperature
changes were measured, and hwas calculated using the heat
equation of the investigated reflow oven. The results showed
that hof the heater and the gas changes were inconsistent in the
reflow oven:
havg ¼k
2DGg;H
D

·Re2=8·1þH=D
0:6=ffiffi
g
p

6
"#
20:05
·Pr
0:42;
ð4Þ
G¼2ffiffi
g
p122:2ffiffi
g
p
1þ0:2ððH=DÞ26Þffiffi
g
p;ð5Þ
g¼
p
4
D
r

2
;ð6Þ
where kis the thermal conductivity of the gas (W/m · K), Dis
the diameter of the orifices, His the distance between the orifice
and the assembled PCB, ris the orifice pitch, Re is the Reynolds
number and Pr is the Prandtl number.
Performing a thermal investigation at the board level is
necessary to investigate the thermal response within the solder
joints of a BGA package. Such investigation can be
accomplished by considering the conditions within the reflow
oven and the PCB assembly layout. Very few publications have
focused on the thermal response in a reflow oven at the board
level. Son and Shin (2005) used a 2D numerical model to
simulate the multi-mode heat transfer within a reflow oven and
the electronic assembly. The solder paste pad was modelled
using a multi-mode heat-transfer numerical formulation based
on finite difference method (FDM) (Tavarez and Gonzalez,
2003). However, the 2D numerical and FDM models were
incapable of capturing the heat transfer in all directions of the
complex shape, making their results unreliable.
The high level of complexity of a PCB assembly results in a
high risk of uneven heating across the board and packages
(Belov et al., 2007). Thus, maintaining temperature
uniformity (DT) across the board is important so that the
solder joints at different positions in the package can be
simultaneously reflowed (Wang et al., 2010).
With the growing trend of computer power and software
ability, code coupling of two commercial software,
e.g. ABAQUS, ANSYS, STARCD, FLOWMASTER and
FLUENT has been able to provide multidisciplinary
simulation solutions in various research. Previous studies
used the Multi-physics Code Coupling Interface (MpCCI)
software in biomedical, nuclear and aerospace engineering
(Molony et al., 2009; Brandt et al., 2009; Nikbay et al., 2009).
The commercial software codes permitted to be coupled by
MpCCI were used to facilitate high-quality simulation. To the
best of our knowledge, no study on the thermal response in
reflow soldering that uses MpCCI as a code coupling of
commercial software has been published. The multi-physics
activities of the characteristics of the flow field and
temperature in a reflow oven that affect the structure model
at the board level can be coupled by the MpCCI software.
This thermal-coupling method can be used to investigate the
effects of the complexity of a PCB assembly.
In the present study, a thermal-coupling method was
developed to investigate the thermal response at the board
level during the pre-heating stage of the reflow process. The
developed method comprised the internal flow of a reflow
oven modelled by the computational fluid dynamic (CFD)
software (FLUENT 6.3.26) and a structural heating board
level created using the FEM software (ABAQUS 6.9). The
MpCCI was used to couple both computer programs. The
flow field of the convection reflow oven was simulated
during the pre-heating reflow process stage. The convection
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
168
heat-transfer coefficient and temperature distribution within
the board level were also obtained. The model was validated
using experimental data and an analytical model within
acceptable errors. The DTs for various complexities of PCB
assemblies were analyzed. Finally, the effects of the conveyor
speed (v)onDTduring the reflow process were discussed.
2. Experimental procedures
This section describes the experiment conducted to validate
the thermal-coupling numerical model. The experiment was
conducted at a nearby semiconductor plant. The reflow oven
in the manufacturing line was used in the experiment. The
four indispensable elements in conducting thermal profiling
in a reflow soldering process are described in the following
sections.
2.1 The convection reflow oven used
A BTU Paragon 150 convection reflow oven was used to
reflow an assembled board and to form a good solder joint
between the PCB and its components. A schematic diagram
of the surface mount technology (SMT) reflow oven is shown
in Figure 1. The reflow oven consists of an entrance section
and ten heating zones and two cooling zones at the exit. The
reflow oven was designed to use dual conveyor lines, as shown
in Figure 2(a). The top and bottom layers of each heating
zone were equipped with two convection heaters in staggered
configuration. A partial dimension of the heater zone plenum
is shown in Figure 2(b). The dwell time of the assembled
board within each zone can be adjusted using the conveyor
speed (v) as it passes through the oven. In the current
experiment, vwas set to 1.25 cm/s.
A typical heated zone plenum for this oven consists of a
blower wheel located at the centre of the heating panel.
The heated zone plenum induces a gas flow that passes through
a series of heaters, as shown in Figure 3. As the gas passes
through the heaters, energy absorption increases the
temperature and pressure. The blower wheel distributes
the high-pressure gas, used to force the gas to pass through
the orifice array of the heating panels. The absorbed energy is
then transferred to the PCB. Finally, the gas passes through the
recirculation area to complete the cycle.
2.2 Thermal profile board used
The thermal profile board (TPB) was assembled using solder
paste printing, device placement and forced-convection reflow
processes. The TPB was subjected to controlled heat, which
melted the solder paste and solder balls and then permanently
formed the joint in the forced-convection reflow oven. In the
current experiment, a sample size of 80 TPBs was used, an
example of which is shown in Figure 4. The thermal profiler
was insulated before being placed on the conveyor of the
reflow oven. This precautionary step protects the thermal
profiler from the hot environment.
2.3 Target reflow profile
The zones were appropriately programmed into the following
temperature sections: pre-heating, soaking, reflow, peak and
cooling, to obtain a profile, as shown in Figure 5 (Tsai, 2009).
Each oven zone was specified in the experiment by setting the
temperature via computer for all the heating zones based on
Table I.
2.4 Instruments used for measuring the expected
temperature profile
Temperature measurements on substrates were performed
during the reflow process using a SlimKIC thermal profiler.
The profiler moved with the TPB through the reflow oven.
A hole at location A (Figure 4) was drilled from the bottom of
the PCB. Kapton adhesive tape was used to secure the
Figure 1 Schematic diagram of the SMT reflow oven
Lip vent
exhaust
Entrance
Conveyor Belt PCB
Heating zones 10 top/10 bottom Cooling zones 2 top/2 bottom
Exit
Lip vent
exhaust
12 345
67 8910
Figure 2 Parts of the SMT reflow oven
(a) Dual Conveyyor lines (b) Part of dimension of plenum
Length = 61.5 cm
Width = 29.0 cm
Diameter of the hole = 0.4 cm
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
169
thermocouple wire. A number of air velocity measurements
were conducted using hot wire anemometers. Room
temperature was maintained at 208C throughout the
experimental verification using an air conditioner. A data-
logging system was built into the reflow oven computer and
directly interfaced with the oven operating software. The system
provided related data, such as the pre-heating slope, based on
the temperature profile. These data were used to validate the
simulation results, as explained in Section 4.2
3. Numerical models
An offline tool, such as numerical modelling for the reflow
process, can enhance the prediction of the setting requirements
on the thermal profiling. In the current study, the numerical
thermal-coupling method was used to model the flow field
and thermal response at the board level during the pre-heating
stage (Zone 1). Three important components, namely, the
structural model, CFD convection model and coupling of CFD
and structural models, are described in the following sections.
3.1 Structural model
An actual TPB is very complex and cannot be modelled in
detail, so a simplified structural model that only considers
large thermal mass packages was developed (Sinkovics and
Krammer, 2009). To understand the variation of temperature
distribution in a TPB, the thermal diffusion is represented by
equation (7):
r
cp
T
t¼
l
2T
x2þ
2T
y2þ
2T
z2

þ_
q;ð7Þ
where
l
is the thermal conductivity of a PCB or package, _
q
represents the convection heat-transfer energy per volume into
the system, _
q¼q=V, and qwill be explained in Section 3.2.
In multi-material structures, the temperature distribution in
the TPB is simulated using FEM based on the structural
solver ABAQUS 6.9. In the current study, the heat-transfer
step was used for thermal analysis. The induced residual
stress in the previous process was initially disregarded (Yi and
Sze, 1998). The thermal properties used in the model are
listed in Table II. The mesh-size test for the TPB was
optimized for better accuracy and computational time.
3.1.1 Boundary conditions of the str uctural model
For the non-symmetrical problem, TPB comprised an entire
PCB and a selected package, as shown in Figure 6. Meshing
was performed using DC3D8 hexahedral element with
13,355 nodes and 8,808 elements. The node at the centre
of the PCB was constrained with boundary conditions U
x
¼
U
y
¼U
z
¼0 to prevent free-body translation. A uniform
temperature of 56.738C was set as the initial temperature in
ABAQUS. After the initial step, the applied convection
Figure 4 Example of a TPB
Thermal Profiler
Thermal Profile Board
Location A
Figure 3 Heated zone plenum
Heaters
PCB
High Pressure Blower Wheel
Low Pressure (vacuum)
Nitrogen Inlet 3 phase Blower Motor 1/8
HP 2840-3400 RPM
Figure 5 Typical reflow thermal profile
240
220
200
180
160
140
120
100
80
60
40
20
0 30 60 90 120 150
Time (sec.)
Temperature (°C)
180 210 240 300270
Preheating
Soaking temperature
Soaking
Soaking time Reflow time
Reflow Cooling
Peak temperature
Ramp-up
slope
Preheating
slope
T1 T2 T3 T4
Source: Adapted from Tsai (2009)
Table I Temperature setting in heating zones
Zone 1 2 3 4 5 6 7 8 9 10
Top and bottom (8C) 105 140 165 175 175 175 180 205 240 235
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
170
heat-transfer coefficient and film temperature were provided
using a CFD code, which will be explained in Section 3.2.
3.1.2 Complexity of the TPB
The complexity of the TPB can be described by non-
dimensional parameters known as the complexity ratio (CR),
defined in equation (8). Five CRs were generated to study
their effects on DTacross the board, as shown in Figure 7:
Complexity ratio ðCRÞ¼P
A;ð8Þ
where Pis the total area at the top surface of a package and A
is the area at the top surface of the PCB.
3.2 CFD convection model
The convectional heat-transfer energy, q, from the hot gas to the
PCB and the package can be expressed using equation (9). qwas
calculated using FLUENT 6.3.26 based on the boundary
conditions applied at the corresponding surface. Generally, the
convection heat-transfer coefficient (h) depends on flow
properties, such as velocity, viscosity and Re of air, which
determine whether the flow is laminar or turbulent:
q¼hAheat ðTg2TsÞð9Þ
3.2.1 Boundary conditions of the reflow oven
For the symmetrical problem (dual conveyor lines), a 3D half
geometry of the reflow oven without a TPB was created, as
shown in Figure 8. The surface and volume meshes were
generated using GAMBIT 2.3.16 and exported to FLUENT
6.3.26 for analysis. The mesh size was dense near the TPB, and
a hexahedral grid was used to create the volume mesh. The total
number of faces was 3,974,255 with 1,245,567 cells. In the
current study, the mesh and time-step sizes were optimized and
finalized for better accuracy and computational time.
A dynamic mesh layering was applied at the top and bottom
layers of the reflow oven. A suitable user defined function profile
was developed using the C program (through MS VISUAL
STUDIO.NET) based on vof 1.25 cm/s. The temperatures at
the top and bottom inlets (Zones 1 and 2) were set based on
Table I. The flow rate in the forced-convection reflow oven
was set to 9.81 m/s (approximately 70 per cent heating power).
A pressure boundary was assigned as the outlet. The wall
boundary was assigned to the TPB, and a symmetrical
boundary condition was also assigned to reduce the
computation time, as shown in Figure 8. Nitrogen was set as
the flow material in FLUENT (Holman, 2002).
The K-epsilon turbulence model that enables thermal
effects was used to model the behaviour of the turbulent flow
Figure 6 Computational meshes in the structural model of a TPB
Location A
Flow direction
Figure 7 Various complexities of a TPB
Case 4 (CR = 15.22) Case 5 (CR = 22.33)
Case 3 (CR = 10.22)Case 2 (CR = 6.33)Case 1 (CR = 2.78)
Table II Thermal properties used in the TPB
FR-4 PCB Packages
Density (kg/m
3
)1,700 1,820
Thermal conductivity (W/m.K) 0.2 0.6
Specific heat capacitance J/(kg.K) 920 236
Source: Shen
et al.
(2005)
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
171
field of the impinging jets over the surface (Yu and Kivilathti,
2002). The SIMPLE algorithm was used for pressure-velocity
coupling and for a second-order upwind scheme to discretize
each control volume for better accuracy.
3.2.2 Variations of conveyor speed
The dwell time of the TPB within each zone was adjusted
using the conveyor speed (v) as it passed through the reflow
oven. The relationship between the dwell time and complexity
of the TPB is important for the determination of DTof TPB
during the reflow process. The standard conveyor speed was
1.25 cm/s, changed to 1.0 and 0.8 cm/s for the sensitivity
study of DT.
3.3 Coupling of CFD and structural models
The external coupling software MpCCI 3.1 (Fraunhofer
SCAI, Germany) was applied for bi-directional coupling of
the CFD and the structural solvers. In this approach, the fluid
solver (FLUENT 6.3.26) and structural analysis codes
(ABAQUS 6.9) were simultaneously run. Coupling
information was exchanged during simulation. Interpolation
was used to transfer quantities between the fluid and structure
meshes. The coupling surface included the entire surface of
the structure. The coupling quantity data were the time-step
size, film temperature, wall heat-transfer coefficient and wall
temperature of the structure. The codes exchanged
information at the beginning of each time step. This
software allowed for non-matching meshes.
Accordingly, different time steps were tested and a time step
of 0.01 s was proven to be optimum for all cases. The data
exchange also occurred every 0.01 s, where FLUENT sent the
film temperature and the convective heat-transfer coefficient
to ABAQUS, and ABAQUS sent the current wall temperature
to FLUENT. The simulations were conducted on a PC with
an Intel dual-core CPU 3.33 GHz processor in a Windows
environment. A simulation for the time period from t¼0to
t¼24 s took approximately 36 h. Additional information is
available in the MpCCI.3.1, Documentation (2009).
4. Result and discussion
4.1 Grid independence test
A grid independence test was conducted on the fluid and solid
meshes. Three parameters were selected from the simulation
results to consider the influence of mesh density: maximum
heat fluxes from TPB at 24 s, temperature in location A at
24 s, and maximum hvalues of the TPB. A summary of the
grid independence test for fluid and solid meshes is presented
in Table III, where Sensitivity Case 1 is the coarsest and
Sensitivity Case 5 is the finest in terms of grid size.
All the results were normalized based on the values
predicted by Sensitivity Case 5, as shown in Figure 9. The
solid and fluid meshes were declared independent when the
critical parameter did not change by more than ^10 per cent
between successive meshes. Sensitivity Cases 4 and 5
performed better with a low percentage deviation between
them. Thus, Sensitivity Case 4 was chosen as the optimum in
terms of accuracy and computational costs.
4.2 Experimental validation
Figure 10 shows the comparison of the simulation profile with
the experimental and analytical results in Zone 1. The
temperature setting in the simulation was defined based on
the setting temperature of Zone 1 (Figure 10). The inlet
temperature of the convective gas and provided inlet flow
velocity, the input of the numerical model, were similar to
that of the experimental reflow oven. In the validation
experiment, the profiles were measured using a thermocouple
at location A, and the simulated profiles were derived from
the same location.
Different parameters, such as the pre-heating slope of Zone
1 and maximum temperature difference, were extracted from
the data-logging system based on the temperature profile.
These parameters were used as comparison values for the
simulation results, as shown in Table IV. The maximum
percentage error was approximately 7 per cent, indicating that
the numerical simulations were in good agreement with the
experimental data.
In addition, a simplified analytical model based on lumped
thermal capacity was calculated using equations (1)-(3). The
convection heat-transfer coefficient (h) at location A was
calculated using equations (4)-(6). The analytical model was
only valid if the structural model was subdivided into small
element cubes in which heat transfer was assumed to occur along
the depth of the cubes. Conduction in other directions was
disregarded. The analytical results shown in Figure 10 correlated
well withthe experimental and simulation data. The capability to
obtain a good fit with small percentage error between the results
for Zone 1 justified the study of a flow field at the board level.
Thus, thermal profile simulation is feasible for a forced air-flow
convection type reflow oven (BTU Paragon 150).
4.3 Flow characteristic of a convection reflow oven
The gas flow in Zone 1 shown in Figure 11 exhibits a change in
the position of the TPB at different levels of flow time (0, 5, 10,
15, 20 and 24 s). Figure 11 shows that the TPB moves with time
toward the flow field in Zone 1. The velocity of the TPB surface
also increases with flow time. The velocity at the inlet of
the heater is higher (9.81 m/s) than that of the TPB surface
(0.49-1.47 m/s), resulting in the transfer of energy into the TPB.
Figure 8 3D half geometry of the reflow oven without a TPB
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
172
To understand further the flow and heat-transfer
characteristics of multiple impinging jets on a plenum, a
detailed discussion is provided. The complex simulation flow
patterns within an array of impinging jets of a reflow oven
were reviewed by Weigand and Spring (2011). An illustration
of the flow pattern is shown in Figure 12. The individual jets
are possibly affected by three types of interactions
(Sarkar et al., 2004). The first type of interaction (indicated
as 1 in Figure 12) is the jet-to-jet interaction among the
neighbours of adjacent jets after their impingement onto the
target plate. The second type of interaction (indicated as 2 in
Figure 12) is the cross-flow interaction that occurs when the
jets are asymmetrical, and an exhaust port does not exist
between orifices. The reverse air flow near the plenum
interacts with the path of the impinging jets when flowing
toward the exhaust. The third type of interaction is the
interaction among surrounding jets in a mixing domain before
impingement, which depends on the orifice-orifice spacing.
These disturbances generally decrease the energy of a jet and
affect the heat transfer on the target plate.
Most studies on flow fields have mainly been based on a
static flat target plate (Sarkar et al., 2004). However, in
practical applications, the target plate is not flat and is always
in motion due to the conveyor movement. These two factors
affect the flow pattern of the reflow oven, as shown in
Figure 13. The complex flow pattern shown in Figure 13
shows characteristics similar to that in Figure 12. The pattern
consists of a free-jet region, stagnation-flow region, wall-jet
region, recirculation region and vortices. In the stagnation
region, the impingement and deflection of the jets rapidly
reduce the axial velocity. Nevertheless, the vortices generated
in Figure 13 are observed in the non-uniform velocity vector,
which may be due to the movement of the TPB. In addition,
Figure 13 also shows the jet-to-jet interaction and cross-flow
disturbances.
To examine the effect of spatial variations of heat transfer
on the TPB, the temperature within the reflow oven
environment must be determined. Temperature contours
were obtained from simulation results over a line passing
through the centre of the TPB, as shown in Figure 14. The
temperature in Zone 1 was slightly distributed to the standby
region at the initial state (0 s). The hot gas began to disperse
with time throughout the TPB. The TPB entered a zone with
a uniform environment temperature after approximately 15 s.
Then, the hot gas from the next zone (Zone 2) started to
affect the temperature of the TPB. Understanding the flow
characteristics of a reflow oven is important for the design,
evaluation, and improvement of the reflow soldering process.
4.4 Convection heat-transfer coefficient within the TPB
The convection heat-transfer coefficient (h) depends on air-
flow properties, such as flow velocity, viscosity of gas and Re.
However, the experimental equation based on equation (4)-
(6) specifies only an average heat-transfer coefficient. Thus, a
precise method is necessary to examine the effect of spatial
variations in local heat transfer on a TPB. Figure 15 shows the
hvalues and temperature contours of the TPB at different
flow times (0, 5, 15 and 24 s).
Figure 9 Percentage deviation versus various mesh sizes of fluid and
solid meshes
0
123
Sensitivity Cases 45
10
20
30
40 Max. Heat Flux
Temperature of Target
Max. h
Percentage Deviation (%)
Figure 10 Comparison of the simulation profile against the
experimental and analytical models
110
100
90
80
Temperature (°c)
70
60
50 0510
Time (s) 15 20
Experiment
Simulation
Analytical
Oven Setting Temp
25
Table III Summary of grid independence test for fluid and solid meshes
Sensitivity Case 1 2 3 4 5
Fluid elements 270,948 492,798 492,798 1,245,567 1,245,567
Solid elements 2,202 2,202 8,808 8,808 35,232
Max. heat flux at 24 s (W/m
2
)1,798 1,995 2,362 2,449 2,666
Temperature of location A at 24s (8C) 83.41 81.15 80.86 80.98 80.97
Max.
h
values (W/m
2
.K) 135.0 81.4 81.5 96.9 96.8
Table IV Comparison between temperature profile parameters
Parameter Experiment Simulation
Percentage
error (%)
Zone 1 (pre-heating
slope) 1.195 1.109 7.2
Max. temperature
different (8C) 84.45 80.98 4.1
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
173
During the initial state, the contours of the hvalues indicate heat
transfer on the TPB, attributed to the fact that hot gases from
the Zone 1 entrance already influenced the TPB temperature.
The changes in the hvalues become visible when the TPB starts
to move into Zone 1, as shown in Figure 15.
At flow times of 15 and 24 s, the maximum hvalues in the
stagnation region of the jet sharply increased because of a
significant deceleration in the stagnation region resulting from a
sharp temperature gradient in the fluid-structure interface.
This sharp temperature gradient increases the numerator
of equation (10), resulting in a high level of heat transfer
(Sarkar et al., 2004). Figure 15 also shows the low hvalues in the
wall-jet region attributable to the increasing thickness of the
boundary layer of the wall jet, as shown in Figure 13, resulting
in a softer temperature gradient at the fluid-structure interface:
h¼kð
T=
yÞ
Tg2Tsð10Þ
where kis the thermal conductivity of gas and yis the vertical
position.
The hvalues of the TPB were also affected by the orifice-to-
assembled PCB distance (H/Dratio). In the current study, the
reflow oven was designed with H
top
larger than H
bottom
.
The effect of the H/D ratio can be investigated using the
simulation results (Figures 15 and 16). H
bottom
or H/D was
smaller, so the bottom plate was closer to the orifice. Smaller
H/D ¼5 resulted in the failure of the jet to develop fully, as
shown in Figure 13. Hence, the flow in the stagnation region
may be laminar. The energy of the jet slightly decayed for the
laminar flow, resulting in high hvalues at the bottom surface
of the TPB, as shown in Figure 16. On the other hand, for H/
D¼9.5, the high level of energy of the jet dissipated because
of the turbulence in the free-jet region (Sarkar et al., 2004).
Thus, the hvalues at the top surface of the TPB were lower
compared with those at the bottom.
Figure 11 Simulation of the flow field in Zone 1
Flow
direction
Time = 0 s
Time = 10 s
Time = 20 s Time = 24 s
Time = 15 s
Time = 5 s
Velocity (m/s)
9.81e+00
9.32e+00
8.83e+00
8.34e+00
7.85e+00
7.36e+00
6.87e+00
6.38e+00
5.89e+00
5.40e+00
4.91e+00
4.42e+00
3.92e+00
3.43e+00
2.94e+00
2.45e+00
1.96e+00
1.47e+00
9.81e–01
4.91e–01
5.52e–05
Figure 12 Complex flow pattern within an array of impinging jets
Plenum
Vortices
Stagnation region Stagnation point
Recirculation
Target Plate
Wall Jet
Vortices Vortices
Free Jet
Orifice
2
1
Source: Adapted from Weigand and Spring (2011)
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
174
The effects of spatial variations in heat transfer on the
temperature distribution in multi-material structures within
the TPB were simulated, as shown in Figure 15. This
temperature distribution within multi-material TPB
structures was affected by the specific heat capacitance and
thermal conductivity of the structure. Less time was needed for
the low specific heat capacitance of a package to heat up
compared with that of the PCB, as shown in Figure 15 (flow
time ¼5 s). Moreover, the temperature was more rapidly
distributed throughout the smaller package. The spatial
variations of heat transfer were also affected by the complexity
of the TPB, resulting in a non-uniform temperature
distribution.
4.5 Effect of complexity on the TPB
The varying temperatures of the PCB assembly influence
soldering defects and reliability. If the temperature on the
PCB assembly differs significantly, some packages on the
board cannot receive an exact amount of heat or may receive
an uneven amount of heat, resulting in low-quality soldering.
Furthermore, the effect of complexity of the PCB assembly
worsens the problems. Figure 17 shows the temperature
distribution of the TPB under various cases. The increase in
CR affects the temperature distribution of the TPB under
similar conditions in a reflow oven. The cold regions (coloured
blue in Figure 17) increase and cover the entire area of the PCB
when more packages are included in the TPB. This finding can
also be attributed to the slow and gradual heating of grouped
packages.
Cold regions can occur at two possible locations in the TPB,
which can be related to the flow field in a reflow oven. The first
location is shown in Figure 17(a) where the cold region appears
around the package sides. This phenomenon can be explained
using the flow shown in Figure 18. Hot gas impinges on the
package surface, dividing the flow field into three regions,
namely, the free-jet, stagnation and wall-jet regions. At the
stagnation region, the direction of the gas stream was changed
to be parallel to the horizontal lines. Therefore, the gas stream
Figure 13 Velocity vector around the TPB at 24 s
Vortices
Recirculation Stagnation region
Flow direction
Free Jet Wall Jet
Velocity (m/s)
9.81e+00
9.32e+00
8.83e+00
8.34e+00
7.85e+00
7.36e+00
6.87e+00
6.38e+00
5.89e+00
5.40e+00
4.91e+00
4.42e+00
3.92e+00
3.43e+00
2.94e+00
2.45e+00
1.96e+00
1.47e+00
9.81e–01
4.91e–01
5.52e–05
Figure 14 Temperature contours in Zone 1 at various flow time
0
Time (s) Temperature Contour
5
10
Flow direction
Temperature (°C)
15
20
24
140
136
132
128
123
119
115
111
107
103
98
94
90
86
82
78
74
69
65
61
57
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
175
in the wall-jet region is in a straight streamline. The sudden
decrease in height between the package and PCB, resulting in a
gap between the package and the PCB, causes the existence of
fewer velocity streamlines, as shown in Figure 18 (red circle).
Thus, a cold region occurs because there were smaller hvalues
or heat transfers in this area.
The second location for the cold region is between the
package-to-package locations, as shown in Figure 17(b)-(e).
This occurrence can be explained using the flow pattern
shown in Figure 19. The velocity streamlines in the wall-jet
region are blocked by the outer package, as shown in
Figure 19(b), compared with that of Figure 19(a), creating a
space between the inner and outer packages with fewer
velocity streamlines, as shown in Figure 19 (red circle). In
addition, increasing the CR further enhances this effect
(Figure 19(c)).
DTis defined as the temperature difference between
the hottest package and the coldest solder joint of the
Figure 15
h
values and temperature contours at the top surface of the TPB at various flow times (
H/D
¼9.5)
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
176
largest package. Figure 20 shows the increase in DTon the TPB
for various cases or for increasing CR, which resulted because
the location of the hottest package changed when CR was
increased. The packages near the edges or corners of the front
TPB tended to heat up first, consequently becoming the hottest
package and causing DTto increase. The worst DTat 10.48C was
found in Case 5. DTshould not exceed 10-208C (Belov et al.,
2007). Exceeding this limit will result in board warpage.
In addition, the temperature profile at location A and the
location of the coldest solder joint of the largest package were
plotted for various cases, including the standby period, as
shown in Figure 21. Figure 21 shows that the temperatures at
these locations are not affected by the complexity of the TPB.
Moreover, the time delay parameters for the reference points
shown in Figure 21 indicate that the reference points do not
instantaneously heat up when the TPB enters Zone 1 in the
reflow oven. The time delay can be calculated based on the
conveyor speed and the distance between two reference
points. Time delay was not considered in thermal profiling
but was considered in the comparison of the experimental
data with the simulation results at different reference points.
4.6 Effect of conveyor speed on
D
Tof the TPB
Changing the conveyor speed (v) is a common practice in
manufacturing applications resulting from changes in product
size and complexity. Thus, the effect of von DTof the PCB
was investigated to mitigate problems with DT. Figure 22
shows the temperature distributions of the TPB attributable
to changes in v.
The area of a cold region between the gap decreased with
decreasing v, as shown in Figure 22. When the conveyor
speed decreased, the time consumed for hot gas impingement
on the package surface increased. A greater amount of heat
energy is delivered through the hot gas to mitigate the cold
gap temperature, resulting in the decrease in DTof the TPB.
DTwas reduced by 2.58C when vwas decreased from 1.25 to
0.80 cm/s, as shown in Figure 23. DTshould be maintained
below 108C to minimize board warpage and soldering defects.
Conveyor speeds of 0.80 and 1.00 cm/s were found to
maintain DTbelow 108C.
The overall TPB temperature will rise at low conveyor
speeds because of the increased time required to deliver heat
energy into the TPB. Figure 24 shows the plot for location A
at different conveyor speeds, resulting in an increase in
temperature when vdecreases. Excessive heat delivered to the
TPB due to an extremely slow conveyor speed can result in
early flux consumption (Tavarez and Gonzalez, 2003) and can
result in the overheating of thermally critical packages
(Florian et al., 2009). Thus, a conveyor speed of 1.00 cm/s
was selected to maintain DTbelow 108C and prevent the
overheating of thermally critical packages. Therefore, a
thermal-coupling simulation method is required to
determine the suitable conveyor speed precisely and
maintain DTwithin the desired range.
5. Conclusions
In the current study, a thermal-coupling model was developed
to investigate the thermal response at the board level during
the pre-heating stage of the reflow process. A thermal-
coupling method using the code coupling software MpCCI
was utilized. The reflow oven was modelled using CFD
software (FLUENT 6.3.26), whereas the structural heating at
the board level was conducted using the FEM software
(ABAQUS 6.9).
The complex flow pattern in the simulation was observed to
have a free-jet region, stagnation-flow region, wall-jet region,
recirculation region and vortices. The impingement and
deflection of the jet resulted in the rapid decrease of axial
velocity. Moreover, the movement of the TPB also generated
non-uniform vortices. Sharp maximum hvalues were found in
the stagnation region of the jet, resulting in a spatial variation
of local heat transfer in the TPB and affecting the temperature
distribution within the TPB with different specific heat
capacitances and thermal conductivities of the structure. The
simulation results were in good agreement with the
experimental data and analytical model.
ThecoldregionandDTincreased with increasing
complexity of the TPB. The cold region can occur at two
possible locations within the TPB. The first location is around
the package sides, and the second location is between the
package-to-package locations. Both occurrences can be
associated with the flow field of a reflow oven. The
packages near the edges or corners of the front TPB tend to
heat up first, consequently becoming the hottest package and
resulting in an increase in DT. Reducing the conveyor speed
Figure 16
h
values and temperature contours at the bottom surface of the TPB at flow time of 24 s (
H/D
¼5)
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
177
Figure 17 Temperature distribution of the TPB at various cases
(a) (b)
(c) (d)
(e)
Notes: (a) Case 1; (b) Case 2; (c) Case 3; (d) Case 4; (e) Case 5
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
178
Figure 18 Velocity vector around the TPB at 22 s
Note: CR = 2.78
Figure 19 Velocity vector around the TPB at 24 s for (a) CR ¼2.78, (b) CR ¼6.33 and (c) CR ¼22.33
(a) (b) (c)
Figure 20 Comparison of
D
T
s in the TPB for various cases
1
0
4
6
8
10
12
Temperature Uniformity (°C)
2
2
Case
345
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
179
Figure 21 Temperature profile of different location for various cases
80 Case 1 (Location A)
Case 3 (Location A)
Case 1 (Package)
Case 3 (Package)
Case 5 (Location A)
Case 5(Package)
Case 4 (Location A)
Case 4 (Package)
Case 2 (Location A)
Case 2 (Package)
75
70
65
60
55 05
Standby
period Time delay
10 15
Time (s)
20 25 30
Temperature (°c)
Figure 22 Effect of variations in conveyor speeds
Flow direction
(a)
(b)
Temperature (°C)
100.9
98.3
95.7
93.0
90.4
87.8
85.2
82.6
77.4
74.8
72.2
69.6
80.0
(c)
Notes: (a) 1.25 cm/s; (b) 1.00 cm/s; (c) 0.80 cm/s
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
180
reduces DTof the TPB. However, an extremely slow conveyor
speed can result in the overheating of thermally critical
packages. Thus, a suitable conveyor speed (v¼1.0 cm/s) was
determined to maintain DTbelow 108Candprevent
overheating of the thermally critical packages.
In future studies, other stages, such as the reflow and
cooling, will be included. Studies on the effect of
interconnection layer layouts (copper material) of the PCB
with a simplified chess-like pattern are important and will also
be considered in future studies.
References
Belov, I., Lindgren, M., Leisner, P., Bergner, F. and
Bornoff, R. (2007), “CFD aided reflow oven
profiling for PCB preheating in a soldering process”,
Thermal, Mechanical and Multi-Physics Simulation
Experiments in Microelectronics and Micro-Systems, 2007.
EuroSime 2007. International Conference on, 16-18 April,
pp. 1-8.
Brandt, T., Lestinen, V., Toppila, T., Ka¨hko¨nen, J., Timperi,
A., Pa¨ttikangas, T. and Karppinen, I. (2009), “Fluid-
structure interaction analysis of large-break loss of coolant
accident”, Nuclear Engineering and Design,Vol.240,
pp. 2365-74.
Florian, S., Denis, K. and Jo¨rg, F. (2009), “Influences on the
reflow soldering process by components with
specific thermal properties”, Circuit World, Vol. 35 No. 3,
pp. 35-42.
Gao, J.G., Wu, Y.P., Ding, H. and Wan, N.H. (2008),
“Thermal profiling: a reflow process based on the heating
factor”, Soldering & Surface Mount Technology, Vol. 20,
pp. 20-7.
Holman, J.P. (2002), Heat Transfer, McGraw-Hill Higher
Education, New York, NY.
Ille
´s, B. (2010a), “Distribution of the heat transfer coefficient
in convection reflow oven”, Applied Thermal Engineering,
Vol. 30, pp. 1523-30.
Ille
´s, B. (2010b), “Measuring heat transfer coefficient in
convection reflow ovens”, Measurement,Vol.43,
pp. 1134-41.
Inoue, M. and Koyanagawa, T. (2005), “Thermal simulation
for predicting substrate temperature during reflow
soldering process”, Electronic Components and Technology
Conference, 2005. Proceedings. 55th, 31 May-3 June, Vol. 1,
pp. 1021-6.
Molony, D.S., Callanan, A., Kavanagh, E.G., Walsh, M.T.
and McGloughlin, T.M. (2009), “Fluid-structure
interaction of a patient-specific abdominal aortic
aneurysm treated with an endovascular stent-graft”,
BioMedical Engineering Online, Vol. 8, p. 24.
MpCCI.3.1, Doccumentation (2009), User Manual,
Fraunhofer Institute for Algorithms and Scientific
Computing SCAI, Sankt Augustin.
Nikbay, M., O
¨ncu
¨, L. and Aysan, A. (2009),
“Multidisciplinary code coupling for analysis and
optimization of aeroelastic systems”, Journal of Aircraft,
Vol. 46 No. 6, pp. 1938-44.
Sarkar, A., Nitin, N., Karwe, M.V. and Singh, R.P. (2004),
“Fluid flow and heat transfer in air jet impingement in
food processing”, Journal of Food Science, Vol. 69,
pp. 113-22.
Shen, L., Wang, M., He, Y., Lam Tim, F. and Jiang Yu, Q.
(2005), “Reflow profile simulation by finite element
method for a BGA package”, Electronic Packaging
Technology, 2005 6th International Conference on,
30 August-2 September, pp. 419-22.
Sinkovics, B. and Krammer, O. (2009), “Board
level investigation of BGA solder joint deformation
strength”, Microelectronics Reliability,Vol.49No.6,
pp. 573-8.
Son, Y.S. and Shin, J.Y. (2005), “Thermal response of
electronic assemblies during forced convection-infrared
reflow soldering in an oven with air injection”, JSME
International Journal Series B Fluids and Thermal
Engineering, Vol. 48, pp. 865-73.
Tavarez, A. and Gonzalez, J.E. (2003), “Modeling
the thermal behavior of solder paste inside
reflow ovens”, Journal of Electronic Packaging, Vol. 125,
pp. 335-46.
Tsai, T.N. (2009), “Modeling and optimization of reflow
thermal profiling operation: a comparative study”, Journal
of the Chinese Institute of Industrial Engineers, Vol. 26,
pp. 480-92.
Van Steenberge, N., Limaye, P., Willems, G.,
Vandevelde, B. and Schildermans, I. (2006), “Analytical
and finite element models of the thermal behavior for
lead-free soldering processes in electronic assembly”,
Microelectronics Reliability, Vol. 47, pp. 215-22.
Figure 23 Comparison of
D
T
on the TPB for different conveyor speeds
1.25 1.00
Conveyor speed (cm/s) 0.80
12
10
8
6
4
2
Temperature Uniformity (°c)
0
Figure 24 Temperature profile at location A for different conveyor
speeds
90
85
80
75
70
65
Temperature (°C)
60
55 0 5 10 15 20 25
Conveyor Speed 1.25 cm/s
Conveyor Speed 1.00 cm/s
Conveyor Speed 0.80 cm/s
30
Distance (cm)
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 –182
181
Wang, Y., Cui, Z. and Wang, B. (2010), “Test and assessment
of the heating system of reflow ovens”, Electronic
Packaging Technology & High Density Packaging
(ICEPT-HDP), 2010 11th Inter national Conference on,
16-19 August, pp. 798-804.
Weigand, B. and Spring, S. (2011), “Multiple jet
impingement – a review”, Heat Transfer Research, Vol. 42,
pp. 101-42.
Yi, S. and Sze, K.Y. (1998), “Finite element analysis of moisture
distribution and hygrothermal stresses in TSOP IC
packages”, Finite Elements in Analysis and Design, Vol. 30,
pp. 65-79.
Yu, H. and Kivilathti, J. (2002), “CFD modelling of the flow
field inside a reflow oven”, Soldering & Surface Mount
Technology, Vol. 14, pp. 38-44.
About the authors
Chun-Sean Lau received his Bachelor’s degree
in Mechanical Engineering from the Universiti
Sains Malaysia in 2010 and currently is pursuing
a PhD degree in the School of Mechanical
Engineering, Universiti Sains Malaysia.
His research interests are in reliability
of microelectronics, computational fluid
dynamic analysis, finite element analysis, and heat transfer.
Chun-Sean Lau is the corresponding author and can be
contacted at: chun_sean@hotmail.com
Dr M.Z. Abdullah has been a Professor of
Mechanical Engineering at Universiti Sains
Malaysia since 2010. He obtained a Bachelor
degree in Mechanical Engineering from
University of Wales, Swansea, UK. His MSc
and PhD degrees are from University of
Strathclyde, UK. He has numerous
publications in international journals and conference
proceedings and his areas of research are heat transfer,
aerodynamics, CFD and electronic cooling. He is a member
of IEEE-CPMT, Malaysia chapter.
F. Che Ani is a Process Engineering Manager at
Celestica Malaysia (a multinational company
based in Toronto, Canada). He obtained a
Bachelor’s degree in Electrical and Electronic
Engineering from University of Strathclyde,
UK.Hehasextensiveknowledgeofand
experience (almost 15 years) in electronic
packaging, especially in the underfill (PoP, micro BGA, Flip
Chip), surface mount technology (screen printing, placement
and reflow technology (lead free/RoHS and leaded)), conformal
coating and wave soldering.
Three-dimensional thermal investigations at board level
Chun-Sean Lau, M.Z. Abdullah and F. Che Ani
Soldering & Surface Mount Technology
Volume 24 · Number 3 · 2012 · 167 182
182
To purchase reprints of this article please e-mail: reprints@emeraldinsight.com
Or visit our web site for further details: www.emeraldinsight.com/reprints
... Several studies were conducted to examine the effect of thermal management and heating transfer capability examination of reflow ovens. These are including the application of simple temperature distribution measurements [24] [25], simulation process [26] [26], thermal models of the reflow process [22] [27], heat gas circulation and heat transfer coefficient [22] [28] investigations. In practice, the defect on the PCB as the result of an inhomogeneous heat transfer requires additional repairs and reworking that can cause a loss of productivity of roughly 30-50% of the total manufacturing costs [29], and therefore they should be avoided. ...
... Usually, a significance level of 0.05 (5%) is reliable. If the P-value is larger than the significance level, the decision is to fail to reject the null hypothesis because there is not enough evidence to 27 ...
... There are a set of variables, which are demonstrating the significance on the quality of soldering process. A set 27  The first overpressure build-up is decelerating the new gas formations in the solder joint, and by the pressure release, the existing voids or the trapped gases are exhausted out of the solder joint leading to reduce voids. The significance analysis demonstrates that pressure has the highest significance in the magnitude of the effect on reduction of void ratio. ...
Thesis
Full-text available
Surface Mount Technology (SMT) is the primary method of permanently mounting electronic components on a substrate. A decisive improvement in the quality of the connection technologies is achieved by increasing the reliability of the solder joints and reducing heat dissipation. The soldering oven technology in combination with an overpressure module enables void-free solder joints. However, the configuration of the process-related mechanism has not yet been fully investigated. In addition, the method offered in the market is energy-intensive, and a deep understanding of the resource flow into the process is mandatory if minimization of consumption is to be achieved. This study deals with a multi-criteria/overall optimization solution based on a trade-off model for quality, energy, cost and time criteria. The approach to find the optimal solution for each criterion is possible. However, a holistically optimized multi-criteria solution is not possible. Using significance analysis a criteria approach for the overpressure soldering process is defined to maximize or minimize each goal/criterion using Artificial Neural Network (ANN) modeling approach. A compromise model has been developed, which enables the decision maker to achieve the setup goal while considering the conflicting requirements in the best possible approach.
... The consumption growth of electronic products has increased the demand for printed circuit boards (PCBs) [1]. In addition, the expansion of the electronic products market has led to an increase in the complexity of PCBs, which causes a more complex thermal response when a PCB passes through a reflow oven [2]. Inside this equipment occurs a process known as reflow soldering, which is achieved by forced convection using multiple air jet impingement technologies [3]. ...
... The average Nusselt number was determined as a function of the average convective heat transfer coefficient, ℎ , the nozzle diameter, D, and thermal conductivity of the air, k, according to Equation (2). While the nozzle's diameter was a constant geometrical parameter and k was temperature-dependent and obtained directly from the literature [33], ℎ had to be calculated by applying Equation (3). ...
Article
Full-text available
Multiple jet impingement is a widely implemented convective process for enhancing heat transfer over target surfaces. Depending on the engineering application, the impinging plate can have different configurations. However, the increased complexity of the surface induces complicated thermal behaviors that must be analyzed. In that sense, this study consisted of the experimental and numerical analysis of multiple jets impinging on a step surface. A particle image velocimetry technique was applied to measure velocity fields, while a heat flux sensor was mounted on the surface to determine the heat transfer. Numerical simulations, for both flat and non‐flat plates, were conducted in ANSYS FLUENT applying the SST k‐ω model, and experimental results were used to validate the model. Three surface configurations were analyzed, flat, 1 D, and 2 D steps, and the results show an increase in the average Nusselt number compared with the flat plate, 9% and 20%, respectively. This increase was mainly due to the intensification of the flow turbulence induced by the step. Numerical results were in good agreement with the experiments, but the heat transfer was slightly underpredicted for the 2 D step case due to the difficulty of predicting with accuracy the velocity field near the step.
... The expansion of the electronic products market has led to an increase in the complexity of Printed Circuit Boards (PCB), which causes a more complex thermal response when a PCB passes through a reflow oven [1]. Inside this equipment occurs a process known as reflow soldering process which is achieved by forced convection using the multiple air jet impingement technology [2]. ...
... The expansion of the electronic products market has led to an increase in the complexity of Printed Circuit Boards (PCB), which causes a more complex thermal response when a PCB passes through a reflow oven [1]. Inside this equipment occurs a process known as reflow soldering process which is achieved by forced convection using the multiple air jet impingement technology [2]. ...
Conference Paper
Jet impingement is a complex heat transfer technique which involves several process variables, such as nozzle-to-plate distance, jet diameter, Reynolds number, jet temperature, among others. To understand the effect of each variable in the process, it is important to study them separately and to identify the different interactions between them. In industrial applications that use forced convection by air jet impingement, the correct analysis of the flow structure and accurate definition of the variables values that affect the heat transfer over the target surface can lead to an increase of the process performance and to a decrease of the manufacturing costs. Considering the reflow soldering process, the uniformization of the flow structure over the Printed Circuit Board can decrease the number of product defects, minimizing manufacturing costs. To reduce costs and time related to the study of the process variables, the introduction of numerical methods has been fundamental. Using a Computational Fluid Dynamic software, the number of experimental tests necessary are highly reduced, being possible to focus on the phenomena that are more relevant for the purpose of the study. In this work, the nozzle-to-plate distance variable is analyzed. This is considered one of the most important parameters since it influences the entire structure of the jet flow as well as the heat transfer coefficient over the target surface. The results present a comparison between six different cases at Reynolds number equal to 2,000 and demonstrate the complexity of the jet impingement flow that lies in the transition region.
Article
This paper presents a finite element method and boundary element method (FEM-BEM) coupling scheme to study the steady-state 2D heat transfer problem for structures with multi-scale features. Based on the geometric information or material properties, the considered model is divided into FE sub-domain and BE sub-domain. The closed BE sub-domain in the mixed FEM-BEM model is defined as a finite super-element for which the effective stiffness matrix and effective fluxes are obtained by the self written BE code and hence can be assembled to the FE systems by the user subroutine (UEL) provided in ABAQUS. This method offers a number of key improvements compared with current analysis methods available for multi-scale structures. These improvements are: (i) the flexibility and convenience by using the powerful pre/post-processing in ABAQUS; (ii) the higher accuracy of the solution over the classical FEM; and (iii) the computational cost and time can be highly reduced by using the coupling method. Numerical results are compared with reference solutions and demonstrate the accuracy and effectiveness of the proposed approach.
Article
Purpose The purpose of this paper is to present the experimental and simulation studies on the influence of copper pillar bump structure on flip chip packaging during reflow soldering. Design/methodology/approach In this work, solidification/melting modelling and volume of fluid modelling were used. Reflow soldering process of Cu pillar type FC was modelled using computational fluid dynamic software (FLUENT). The experimental results have been validated with the simulation results to prove the accuracy of the numerical method. Findings The findings of this study reveal that solder volume is the most important element influencing reflow soldering. The solder cap volume reduces as the Cu pillar bump diameter lowers, making the reflow process more difficult to establish a good solder union, as less solder is allowed to flow. Last but not least, the solder cap height for the reflow process must be optimized to enable proper solder joint formation. Practical implications This study provides a basis and insights into the impact of copper pillar bump structure on flip chip packaging during reflow soldering that will be advancing the future design of 3D stack package. This study also provides a superior visualization and knowledge of the melting and solidification phenomenon during the reflow soldering process. Originality/value The computational fluid dynamics analysis of copper pillar bump structure on flip chip packaging during reflow soldering is scant. To the authors’ best knowledge, no research has been concentrated on copper pillar bump size configurations in a thorough manner. Without the in-depth study, copper pillar bump size might have the impact of copper pillar bump structure on flip chip packaging during reflow soldering. Five design of parameter of flip chip IC package model was proposed for the investigation of copper pillar bump structure on flip chip packaging during reflow soldering.
Article
Purpose In this paper, analytical modelling of heat distribution along the thickness of different printed circuit board (PCB) substrates is presented according to the 1 D heat transient conduction problem. This paper aims to reveal differences between the substrates and the geometry configurations and elaborate on further application of explicit modelling. Design/methodology/approach Different substrates were considered: classic FR4 and polyimide, ceramics (BeO, Al 2 O 3 ) and novel biodegradables (polylactic-acid [PLA] and cellulose acetate [CA]). The board thicknesses were given in 0.25 mm steps. Results are calculated for heat transfer coefficients of convection and vapour phase (condensation) soldering. Even heat transfer is assumed on both PCB sides. Findings It was found that temperature distributions along PCB thicknesses are mostly negligible from solder joint formation aspects, and most of the materials can be used in explicit reflow profile modelling. However PLA shows significant temperature differences, pointing to possible modelling imprecisions. It was also shown, that while the difference between midplane and surface temperatures mainly depend on thermal diffusivity, the time to reach solder alloy melting point on the surface depends on volumetric heat capacity. Originality/value Results validate the applicability of explicit heat transfer modelling of PCBs during reflow for different heat transfer methods. The results can be incorporated into more complex simulations and profile predicting algorithms for industrial ovens controlled in the wake of Industry 4.0 directives for better temperature control and ultimately higher soldering quality.
Article
Full-text available
This paper presents a practical methodology for static aeroelastic analysis and aeroelastic optimization via coupling of high-fidelity commercial codes. A finite-volume-based flow solver FLUENT is used to solve three-dimensional Elder equations, Gambit is used to generate mesh in the fluid domain, and CATIA is used to model parametric solid geometry. Abaqus. a structural finite element method solver. is used to compute the structural response of the aeroelastic system. The mesh-based parallel-code coupling interface MpCCI is used to exchange the pressure land displacement information between FLUENT and Abaqus to perform a loosely coupled aeroelastic analysis by a staggered algorithm, and modeFRONTIER software is used as the optimization driver for scheduling a nondominated sorting genetic algorithm initiated with design of experiments. First. an AGARD 445.6 wing configuration is optimized with objectives of maximum lift/drag ratio and minimum weight. Optimization variables are chosen as sweep angle at the quarter-chord and the taper ratio of the wing. Second, a more realistic wing model, ARW-2, is optimized for thickness values of the inner ribs and spars. Aeroelastic analysis produce consistent results with experimental data.. and the applied optimization methodology results in Pareto-optimal solutions.
Article
Full-text available
The thermal response of electronic assemblies during forced convection-infrared reflow soldering is studied numerically. Soldering for attaching electronic components to printed circuit boards is performed in a process oven that is equipped with porous panel heaters, through which air is injected in order to dampen temperature fluctuations in the oven. Forced convection-infrared reflow soldering process with air injection is simulated using a two-dimensional numerical model. The multimode heat transfer within the reflow oven as well as within the electronic assembly is simulated. Parametric study is also performed to study the effects of various conditions such as conveyor speed, air injection velocity, and electronic assembly emissivity on the thermal response of electronic assemblies. The results of this study can be used in the process oven design and selecting the oven operating conditions to ensure proper solder melting and solidification.
Article
The development of a mathematical model of the soldering process of actual pastes as used in surface mount technology (SMT) lines is described in this paper The coupled heat transfer processes between the solder Paste and the flux including changes in solder paste properties are considered in the model. Specifically, the loss of solvents in the vehicle system, melting, solidification and further single phase cooling of the solder paste are contemplated in the model. Experiments were conducted with the objective of validating the predictions of the solder paste temperature profile and of the loss of weight due to flux extraction. Results are shown in this paper for typical eutectic paste 63%Sn-37%Pb and experimental data is in good agreement with the numerical predictions. Simulations using the lead-free solder paste systems 965%Sn-3.5%Ag and 42%Sn-58%Bi are also reported in this paper. The proposed model is suitable for incorporation into existing three dimensional heat transfer models of PCBs for simulations in ovens with similar characteristics as those used in actual manufacturing applications.
Article
In this study, a comparative study of optimizing the reflow thermal profiling parameters using a hybrid artificial intelligence and the desirability function approaches without/with combining multiple performance characteristics into a single desirability is presented. Reflow soldering is the key determinant for the improvement of the first-pass yields of electronics assembly. A reflow thermal profile is a time-temperature contour with multiple performance characteristics utilized to monitor the heating effects on a printed circuit board (PCB) and surface mount components (SMCs) in the reflow oven. The use of an inadequate reflow thermal profile may not only produce a variety of soldering failures, but can also result in the needs for considerable reworking and waste. An L18 (2*3) Taguchi experiment design is conducted to collect the thermal profiling data. A quick propagation (QP) neural network is modeled based on experimental data to formulate the nonlinear relationship between the thermal profiling factors and responses, and a genetic algorithm (GA) is used in the optimization of thermal profiling factors with the fitness function based on the trained QP neural network model. Alternatively, the response columns for the experimental data can be transformed into a single measure of desirability which is then optimized by the desirability function approach with the response weightings derived from an analytic hierarchy process (AHP). The empirical evaluation results show that the desirability function approach with combining the multiple performance into a single desirability delivery superior soldering performance to that obtained by the hybrid artificial intelligence method without combining the multiple performance into a single desirability, as measured by the DPMO, yield rate, and process sigma.
Article
Nowadays reflow ovens apply forced convection heating with nozzle-matrix blower system which generates numerous heater gas streams. In these types of ovens the heat transfer coefficients of the gas streams determine mainly the efficiency of heating. In this paper a measurement method is discussed which can determine the heat transfer coefficient distribution under the nozzle-matrix of a convection reflow oven. During the experiments, temperature changes have been measured and the heat transfer coefficient has been calculated using the heat equation of the investigated reflow oven. In the first step, the heat transfer coefficient of the heater gas streams has been studied under the nozzle-matrix in function of height. In the second step, the heating efficiency of the nozzle-lines has been compared and the distribution of the heat transfer coefficient has been calculated. The results are confirmed by a theoretical model. It is also presented that how the contamination of the nozzles affects the heating efficiency of the reflow oven. This measuring method can be useful for monitoring the operation of the reflow oven, soldering failure prediction and measuring input parameters of thermal modeling.
Article
As the point interconnects between printed circuit board (PCB) and chips and owning a soft constitutive material property, solder joint interconnect is the crucial link for reliability of electronic systems. The behaviour and quality of solder joints are determined by solder material, geometry design and reflowing operation processed by reflow ovens. Reflow oven is thus the most important means in PCB soldering and its performance directly concerns the behaviour and reliability of electronic systems. Until now there was no unified standard of evaluation and acceptance for reflow oven in China domestic market. The existed are all laid down by different equipment manufacturers, which makes confused and blinded situation for customer acceptance. This situation has heavily affected the quality of solder joints. There is an urgent need to carry out relevant research for formulating uniform inspection subjects, inspection standards, inspection tools and inspection processes. This paper reports our work on testing, assessment and research of the heating system of reflow ovens. This is a systematic study of testing process, including building testing equipment, setting up testing curve, measuring temperature field and assessing reflow ovens. Using the method suggested in the current paper, various performance parameters of the heating system can be obtained. The performance of reflow oven and behaviour of soldering joints can be predicted. The result and method are relevant to enhancement of manufacturing quality and selection of appropriate ovens. This may also set a basis for setting up relevant standards for regularity of the manufacturing and assessment of reflow ovens. This paper is concerned with the first step of a whole systematic project. Other results will be reported in due course.
Article
Because of new requirements related to the employment of lead-free manufacturing and the diversity of components and metal finishes on high density printed circuit boards, better understanding and control of the reflow process is required in order to achieve acceptable yields and reliability of SMT assemblies. Accurate control of the temperature distributions within components and boards during the reflow process is one of the major requirements, especially in lead-free assembly. This paper outlines a scheme for reflow modelling and presents an oven-level model of the steady state flow-field inside a reflow oven, which will be needed in subsequent transient analysis and small-scale modelling. The model is constructed by utilising the advanced computational fluid dynamics (CFD) technology using commercial software. The computational results are discussed and compared with measured data.
Article
Purpose – This paper aims to offer a convenient method to develop an oven recipe for a specific soldering profile in a reflow process. The method is devised to quickly achieve proper profile shape and heating factor Q?, a measure of success for high reliability of the solder joints reflowed. Design/methodology/approach – An in-depth analysis of the heating mechanism and some experiments of the reflow soldering process are performed to research on how to realize a specific shape reflow profile were conducted. Findings – Heating mechanism analysis and experiments demonstrate that the combinatorial parameters based method is feasible to do thermal profiling. Research limitations/implications – The mapping function among a particular configured PCBA, an oven used, a target reflow profile and an optimal range of the heating factor should be further established for fast and reliable production of reflow soldering. Practical implications – Provides a methodology for designing an oven recipe for reflow soldering production. Originality/value – An oven recipe can be quickly attained with the approach established in this paper, facilitating the formation of solder joints with high reliability during the reflow soldering process.
Article
Purpose - The purpose of the paper is to focus on the research into components with specific thermal properties and their influences on the reflow soldering process. Design/methodology/approach - After a brief introduction, the paper gives an overview of the necessity of thermal management on printed circuit boards (PCBs) and the possible effects on the manufacturing of electronic devices. In the next sections, different test boards are presented for investigations into different thermal effects during soldering. The last section deals with the influences of molded interconnected devices (MIDs) on the reflow soldering process. Findings - The investigations show that components from the thermal management influence the reflow soldering process more or less. The highest impacts on the soldering process are from components with a thermal connection to the electrical component and its solder joint. All results from the investigations have in common that the thermal influence can only be compensated by increasing the temperature during soldering. However, this significantly increases the risk of overheating the electrical components or the PCB itself. Research limitations/implications - This paper shows only the influence of some of the effects caused by thermal management on the reflow soldering process. Furthermore, vapour phase soldering is not considered, but actual investigations are carried out on vapour phase soldering ovens as well. Originality/value - Thermal management becomes more and more important with the increasing functionality of electrical components and electronic devices. This topic has been the subject of a large number of articles. However, this paper deals with influences that thermal management has on the soldering process during the manufacturing of the electronic device.