FUNDAMENTALS OF BGA BALL ATTACH REFLOW PROCESS

Michael Ko
Research International
Ronkonkoma, New York, USA
ABSTRACT
The fundamentals of reflowing solder balls for the BGA
ball attach process is discussed. The content is divided
into three major sections: modes of heat transfer,
convection technologies used in reflow ovens and
mechanics of temperature profiles.
There are three basic modes of heat transfer: conduction,
convection and radiation. All three modes are utilized in
achieving desired temperature profiles. For most of the
reflow ovens in the industry, forced convection is the
dominant mode of heat transfer. Understanding the modes
of heat transfer assists process engineers in
troubleshooting reflow applications.
In the reflow oven industry, there are two basic methods
of achieving forced convection: fans and pressure sources.
Typically, fans are used to achieve convection and
recirculate process gas. Examples of pressure sources are
compressors and nitrogen tanks. Benefits of both
methods are discussed. In convection heat transfer,
temperature and speed of the gas provide quantitatively
different effect in heating the product. Adjusting the
temperature difference between the process gas and the
product is a more effective method of controlling heat
transfer. In general, higher gas speed or volumetric flow
rate improves the temperature uniformity across products.
There are four different elements in reflow profiling:
preheat, dryout, reflow and cooling. Profiling is no longer
a shoot in the dark task. There is a definite procedure in
performing temperature profiling. Understanding the
relationship among heater set point temperatures, desired
profile temperature, gas speed and conveyor speed, one
can achieve desired temperature profiles with improved
efficiency and accuracy.
Understanding and practicing the fundamentals of the
reflow process for the BGA packaging can enhance the
productivity and lower operational cost.
Key words: BGA, reflow oven, temperature profile, heat
transfer
INTRODUCTION
In both the SMT [Surface Mount Technology] and BGA
packaging industry, reflow ovens are often considered as
black boxes. Many process engineers use reflow ovens
for many different applications without fully
understanding the science behind the heat control
technology. Especially in the semiconductor industry,
many engineers are faced with using conveyorized ovens or
furnaces for the first time. Others have used conveyorized
ovens, but only to the extent of duplicating old processes or
steps outline by their predecessors.
The papers goals are to explain the principles behind the heat
transfer in convection ovens and to outline the science of
proper temperature profiling.
Many process and quality issues in BGA ball attach process
can be improved by following the basic steps of ball attach
reflow profiling.
MODES OF HEAT TRANSFER
There are three different modes of heat transfer in reflow
ovens. They are conduction, convection and infrared
radiation. In most of the convection ovens the predominant
heat transfer method is convection, but other modes contribute
in transferring heat to the product. For heat to be transferred
there must be a temperature difference. In general, the greater
the temperature difference, the higher the heat transfer rate.
The heat always transfers from a body with higher temperature
to one with lower temperature.
Conduction heat Transfer
Conduction heat transfer occurs when bodies of different
temperature are in physical contact. The rate of heat transfer
depends on thermal conductivity of the materials and
geometric factors such as the thickness and contact area. In
conduction, the most important factor in determining the heat
transfer efficiency is the quality of surface contact. The
surfaces must be in perfect contact for efficient conduction
heat transfer. In general, the conduction heat transfer equation
is
Q = K A (T1 T2)
X
where
Q = Conduction heat transfer [W]
A = Cross sectional area [cm
2
]
K = Thermal conductivity [W/cm-C]
T1 = Temperature at location 1 [C]
T2 = Temperature at location 2[C]
X= Thickness between locations 1 and 2.
Another item to consider is the path of conduction. An effect
called contriction reduces the quantity of the heat transferred
via conduction because of the geometric factor. This can be
seen easily on the FEA [Finite Element Analysis] result of the
cooling fin analysis. If the conduction heat transfer path is
very small the efficiency is decreased. As an example, to
conduct heat away from a surface using thin cooling fins
the constriction effect needs to be minimized by
optimizing the fin shape, dimension, spacing and etc.
Convection heat transfer
In most reflow ovens, convection is the major heat
transfer mode. In general, convection heat transfer occurs
when a fluid contacts a body of different temperature. The
important fact is that the fluid is in motion and makes
contact with the body of different temperature. In all
three modes of heat transfer, convection is the only mode
where motion is essential. The rate of convection heat
transfer depends on the convection coefficient,
temperature difference and area. The general equation for
convection heat transfer is
Q = h A (T
a
T
t
)
where
Q = Convection heat transfer [W]
h = Convection coefficient [W/(cm
2
C)]
T
a
= Fluid temperature [C]
T
t
= Target temperature [C].
In general, the higher the fluid speed, the higher the
convection coefficient. There are two kinds of the
convection: natural and forced. Natural convection
occurs as the buoyancy of the fluid causes vertical
movement because of the temperature difference. Forced
convection is caused by some mechanical means such as
fan blades or pressure forces which causes the fluid to
move.
Infrared radiation heat transfer
Infrared radiation heat transfer occurs when any object is
above 273 C. [absolute zero]. Again, for any heat
transfer to occur there must be a temperature difference.
While conduction and convection modes of heat transfer
are linear function of temperatures, the infrared radiation
mode is exponential to the fourth power. This is why
infrared [IR] heat transfer is so efficient in heat transfer.
The general equation for IR heat transfer is
Q = F
v

s

t
A [T
s
4
T
t
4
]
where
Q = Radiation heat transfer [W]
F
v
= View factor

s
= Emissivity of the source

t
= Absorptivity of the target
= Stefan-Boltzmann Constant
(5.67 x 10
-12
W/(cm
2
K
4
)]
A = Cross sectional area [cm
2
]
T
s
= Temperature at source [K]
T
t
= Temperature at target [K].
Because infrared radiation depends on the line of sight for heat
transfer, the geometry of the objects plays a major role. In
BGA ball attach process, the IR view factor is very high
because of the direct line of sight between the heater panels
and products.
In most convection reflow ovens used for BGA ball attach
process, heat transfer via conduction is minimal compared to
convection and IR. In all convection ovens, IR heat transfer
contributes a high percentage to the overall heat transfer. As a
rough estimate, the ratio between convection and IR is about
70 to 30. This fact can be explained fairly simply. If the set
point temperature of the oven for a particular heating zone is
at 250C., the gas temperature and the heater panel with the
array of holes are at around 250C. In some convection oven
designs, the heater panel can be at a higher temperature than
the gas by design. The heater panel radiates IR at 250C, but
the IR energy is exponential to the fourth power of 250C. In
BGA ball attach process, the strips are flat and in direct line of
sight with the heater panels contributing to very efficient IR
heat transfer.
CONVECTION TECHNOLOGIES USED IN REFLOW
OVENS
There are many different methods in achieving convection.
For the scope of this paper, there are basically two different
means: mechanical or utilizing pressure difference. The
mechanical method is similar to using a regular household fan.
The fan blades exert a mechanical force causing the gas to
move or travel; it can be thought as hitting the gas to move.
The pressure difference causes the gas to move because the
gas wants to travel to a lower pressure area. The speed of the
gas depends on the magnitude of the pressure difference. In
convection ovens, the pressure difference is between the
heater cavity or the heater box and the chamber. The chamber
is defined as the volume where the products are heated during
transport on a conveyor system.
Higher pressure in heater cavity is created by one of two
methods. One method utilizes a fan to blow gas into the
cavity thus creating pressure. The other method utilizes a
pressurized source such as nitrogen tanks or compressors.
Because the speed of gas depends on the force caused by
pressure difference, increasing the pressure inside a heater
cavity increases the gas speed as it comes out from the holes
in the heater cavity. The pressure inside a heater cavity can be
controlled by either modulating the RPM of a fan or a flow
meter.
The benefits of using fans are lower cost of operation and the
ability to move high volume of gas. The benefits of
compressed source are easier control of the gas speed and
maintenance of the atmospheric purity inside the oven. On
ovens using fans to achieve convection, the pressure and
volume of gas movement depends on the fan curves and the
system curve. Both curves are not linear: the effect of gas
speed is not linear to the RPM of the fan. On ovens using a
compressed source, the pressure in heater cavities is controlled
via flow meters. The relationship between the gas speed
and flow meter setting is linear and thus provides better
control of gas speed.
The two control variables in convection heat transfer are
convection coefficient and fluid temperature. In
convection ovens, the convection coefficient can be
modified by either changing the RPM of fans or flow
meters. In general, the convection coefficient is
proportional to the gas speed. The fluid temperature can
be modified by changing the set point temperature of a
heater zone. Hence, in convection ovens the two control
variables of convection heat transfer are settings of fans
or flow meters and set point temperatures.
In most reflow ovens, changing the RPM of fans or the
flow meter settings have negligible effect to convection
heat transfer compared to the effect of changing heater set
point temperature. For example, the effect of increasing
the convection coefficient from 4 to 8 by increasing the
speed of the gas is small compared to increasing the set
point temperature from 220 to 260C. In optimizing the
ball attach process or programming the oven for a new
product, it is simpler and more efficient if only the set
point temperatures are modified. This is based on the
assumption that the fans or the flow meters of the oven
are already optimized during machine installation.
Temperature uniformity is defined as the temperature
difference among products processed in multiple lanes. It
describes the capability of an oven to heat products
uniformly across the process width of the conveyor belt.
Load factor in general is the percentage of process lanes
occupied by product. For example, if 50% of the process
lanes are occupied by product, the load factor is 50%.
This is illustrated in Figure 1. Temperature repeatability
is defined as the temperature difference among profiles
performed at different time and load factors while the
same oven program and product are used. The
repeatability of the oven is a design feature. It is a
thermal response characteristic of an oven at different
load factors and time. Repeatability is the capability of the
oven to provide the same temperature profile through out
its use as long as the oven program and the product are
not changed.
Figure 1. Example of 50% load factor. The length of
space is equal to the length of the product.
The importance of temperature repeatability of reflow ovens
cannot be stressed enough. It could be the difference in
profiling once a week versus several times a day while
manufacturing the same lot of product. Even though,
repeatability is a design feature of an oven, process engineer
must understand what conditions can change the temperature
profiles at no fault of the oven. If any of the three variables in
the reflow process is changed the profile will change. The
steps outlined in this paper can assist process engineers to
reduce unintended mistakes. Once the flow meters and fan
settings are optimized for the ball attach process, it should not
be changed unless there is a dramatic change in the product
package type. This leaves the conveyor speed and oven set
point temperatures as the only two variables. These two
variables make a unique oven program, sometimes called a
recipe.
MECHANICS OF TEMPERATURE PROFILES
Typically, a process engineer programs an oven to achieve a
particular temperature profile recommended by the flux
manufacturer. The basic components of a profile are preheat,
dryout, reflow and cooling as illustrated in Figure 2.
Figure 2. Temperature Profile
By understanding the different components of the profile, one
can program a reflow oven in less time with greater accuracy.
There are three major steps in programming an oven:
1. Calculating the conveyor speed and oven set point
temperatures.
2. Performing a temperature profile or measuring the actual
temperature of the product
3. Making minor adjusts to the oven program and re-
profiling.
Calculating the oven conveyor speed
The conveyor speed is calculated using the formula below:
Conveyor Speed = Total Heated Length
Profile Time
The Total Heated Length is provided by the oven
manufacturer and the Profile Time is recommended by the
Space
Space
Space
0.0
50.0
100.0
150.0
200.0
250.0
300.0
0.0 0.8 1.5 2.3 3.0 3.8 4.5
Time [minutes]
T
e
m
p
.


[
d
e
g
.

C
]
Preheat
Dryout
Reflow
Cooling
flux provider. The profile time is defined as the total time
the product takes to reach the peak temperature. It most
cases this is between 2 and 3 minutes for BGA ball attach
process. If the ovens total heated length is 48 inches and
the profile time is 3 minutes:
Conveyor Speed = 48 inches
3 min.
= 16 IPM [0.41 m/min]
Calculating Oven Set Point Temperatures
The next step in programming the oven is to set the
individual heater zone temperatures. The layout and
lengths of the heater zones in the oven are used to identify
which heater zone is used for what purpose. For the
sample profile in Figure 3, the first two vertical zones can
be used as preheat. The third zone and the fourth zones
can be used for dryout and reflow, respectively.
The resident time of the product in each zone is calculated
as
Resident time = Length of Zone
Conveyor Speed
This equation calculates how long the product is resident
in a particular zone. For example, if a zone is 12 inches
[30cm] long and the conveyor speed is 16 IPM [0.41
m/min], the resident time is 0.75 minute or 45 seconds.
Z 1 Z 2 Z 3 Z 4 Cool
205 210 165 265
Figure 3. Sample temperature profile
Using this information, different components of the
temperature profile can be matched to the heater zones of
an oven. After finding out which zone controls which
component of the profile, the next step is to calculate the
temperatures.
One important item to keep in mind is that because the product
travels through the oven, the temperature it reaches at each
heater zone is not equal to the zone set point temperature. For
example, when the oven set point for reflow zone is 275C,
the peak temperature might be only about 210C.
Normally, engineers try an oven program from their
experience and then profile to compare it to the desired profile
shape. This is a guessing game. By analyzing the science of
temperature profiling, the task of profiling can be fine-tuned.
The Delta T between the oven set point temperature and
measured temperature can be calculated from either previous
profile data or can be obtained by profiling 3 times on a new
oven. From these data, one can calculate a simple equation for
each heater zone that can predict within a reasonable accuracy
the set point temperature and final temperature of the product
as it leaves each zone.
As an example, lets examine an oven with 4 vertical zones,
total of 8 heating zones, with the following set point
temperatures. For this example, each oven zone is 12 inches
[30cm] long.
In Figure 3, both the zone set point and recorded temperatures
are shown. The Delta Ts between the zone set point and
measured temperatures can be calculated. The Delta T for the
reflow zone is calculated below:
Delta T = 265 215 = 50C
By using another profile with a different conveyor speed, a
linear equation as a function of conveyor speed can be
formulated to predict the profile temperature of each zone.
For each zone, the Delta Ts between the zone set and
measured temperatures for two different profiles are
Profile #1: TD1 = TS1 TM1
Profile #2: TD2 = TS2 TM2
Where
TS = Set Temperature
TM = Measured Temperature from profile
TD = Temperature difference
Using a simple linear equation, Y = MX + B, where M is the
slope and B is the Y-intercept, the data from two profiles are
used to come up with an equation:
M =
CS2 CS1
TD2 TD1

Where
CS = Conveyor Speed.
Hence the equation is
DeltaT = M [Conveyor Speed] + B.
25.0
92.5
160.0
170.0
215.0
100.0
0.0
50.0
100.0
150.0
200.0
250.0
300.0
0.0 0.8 1.5 2.3 3.0 3.8 4.5
T i me [ mi n u t es]
T
e
m
p
.


[
d
e
g
.

C
]
205
210
265
165
Using this equation, we can calculate the approximate
Delta T between the set and measured temperatures.
From this value, one can predict the profile temperature
from the oven set point temperature and vice versa:
Calculated Temperature = Oven set point T DeltaT.
This is a very simple, yet effective way of calculating
both the profile temperatures and oven set point
temperatures. The accuracy can vary from less than 5C
to about 10C depending on unique characteristics of
ovens. With additional profile data points, the equation
can be fine-tuned further for greater accuracy.
Sample calculation
Let us review the process using a set of sample data for
the reflow zone.
Set Point Peak T. Conveyor Speed
Profile #1 255C 210C 14 IPM
[0.35m/min]
Profile #2 270C 214C 16 IPM
[0.41m/min]
Profile #1: TD1 = 255 210 = 45C
Profile #2: TD2 = 270 214 = 56C
M =
14 16
45 56

= 5.5
Y = 5.5 X + B
Setting Y = 45 and X = 14, B is calculated as 32.
Hence the equation is
DeltaT = 5.5 [Conveyor Speed] - 32.
AT 16 IPM,
DeltaT = 5.5 (16) - 32 = 56 C.
If the set point temperature is 265C, then
Calculated profile peak temperature = 265 56 = 209C.
As illustrated the only variables are the conveyor speed
and the heater zone set point temperature. Because the
conveyor speed is dictated by the oven length and the
profile time requirement of the flux, the only control
variable is the oven set point temperature. Following a
method such as this can decrease the time it takes to
profile and improve the quality of BGA ball attach
process.
Performing a temperature profile
In order to obtain accurate temperature profiles, there are
two items that process engineers need to practice caution.
The first item is attaching thermocouples to the sample. The
goal is to measure the correct temperature for attaching the
solder balls to the pads. The thermocouple must be securely
fixed to the desired point on the substrate with a minimum
amount of compound. If the volume of the fixing compound
is too large, the temperature reading will be artificially lower.
The reason is that a preformed solder ball is very small
compared to the compound.
The most crucial point is that the thermocouple must be in
contact with the pad on the substrate. On a conveyorized
convection oven, the gas temperature is higher than the sample
temperature during profiling. If the thermocouple is not
contacting the pad on the substrate, the measured temperature
tends to be higher. The same is true if the thermocouple is not
securely fixed. The reason is that the thermocouple might be
reading the temperature of the gas, fixing compound or
combination of both.
The second caution is for the length of thermocouple wire
between the sample and the measuring device. Many
commercially available measuring devices or profilers are
available for process engineers. They are transported through
the oven along with the thermocoupled sample. Compared to
a typical BGA strip, a profiler is many times the volume and
hundreds times the weight. If the sample BGA strip and the
profiler are in the same heating zone as they are transported
inside the oven, the profiled temperature will be artificially
higher. Because of its weight, a profiler takes away
substantially more heat from the oven zones than BGA strips.
As the heat is taken away from the zone, the oven
compensates the heat loss by increasing the heater output
following a PID algorithm. Almost all advanced convection
ovens use a PID [Proportional Integral Derivative] method for
controlling temperatures inside the oven. The effect is that the
measured temperature from the sample could be artificially
higher. A good rule of thumb is that the length of the
thermocouple wire between the sample and the profiler must
be longer than the longest heater zone in the oven. This
method allows the profiler to be at least one zone downstream
from the sample in the direction of travel and minimizes the
interference effect.
After profiling, it is a good practice to critically analyze the
profile data. Many profilers provide pre-analyzed data such as
slopes of temperature rise and fall, dwell time at designated
temperatures, and etc. If the profile is for a product already in
production, it should match the previous profile(s) within the
ovens repeatability specification.
SUMMARY AND CONCLUSION
In BGA ball attach reflow process, there are three process
variables: gas speed, conveyor speed and oven set point
temperatures. By eliminating two variables, the task of
programming the oven and temperature profiling can be
simplified. The gas speeds in heater zones are controlled
by either the fan settings or flow meters. Optimizing the
fans or flow meters during installation is recommended.
The act of changing the gas speed to achieve a certain
temperature profile is shown to be negligible compared to
modifying the zone set point temperatures. The equation
for calculating the conveyor speed is explained.
The guess work of determining the zone set point
temperatures can be eliminated by utilizing simple
equations that can provide reasonable predictions. The
method of formulating prediction equations as a function
of conveyor speed is explained.
Items that affect the temperature profile data are
explained. The location of the thermocouple junction on
the sample, the amount of compound used in attaching the
thermocouple, and the length between the sample and
profiler can affect the temperature reading.
By understanding the principles of heat transfer inside
convection ovens and following the science behind
temperature profiling, process engineers can increase the
efficiency of profiling and BGA ball attach reflow
quality.
ACKNOWLEDGEMENT
I would like to thank my colleagues at Research
Incorporated for help in writing this paper.
REFERENCES
[1] Incorpera, Frank P. and DeWitt, David P.,
Introduction to Heat Transfer, Second Edition, John
Wiley & Sons, 1990, pp 311-349.
[2] Cox, Norman R., Reflow Technology Handbook, pp.
14-26.
[3] American Society of Heating, Refrigeration and Air
Conditioning Engineers, Inc., 1989 ASHRAE
Handbook Fundamentals, pp3.15-3.19
[4] Sparrow, Ephram, Lecture notes to Introduction to
Heat Transfer, Packet No. 136, 1990.