IDB 3047

STUDENT INDUSTRIAL PROJECT (SIP)

Title:

WIRE BONDING PROCESS OPTIMIZATION

REPORT

May 2021 – Aug 2021

PREPARED BY:

NAME: MUHAMMAD AIMAN BIN IBRAHIM

ID: 17005815

PROGRAMME: MECHANICAL ENGINEERING

HOST COMPANY: Carsem (M) Sdn. Bhd.

UTP SV: AP IR DR MASRI BIN BAHAROM

HC SV: NORHALIZA BINTI KASIM

I VERIFICATION

i
II ACKNOWLEDGEMENT

First and foremost, I would like to thank Carsem (M) Sdn. Bhd. For giving me an
opportunity to undergoes my internship at the company. Even during the crisis of COVID-19
and during the movement conditional order take commence, they able to accept me into their
company as an intern is big bless for me. I would also like to thank anyone who support me
throughout the whole journey. This thankfulness is for all my family members, colleagues,
friends, and anyone associate with it.

I would also like to thank Universiti Teknologi PETRONAS (UTP), my educational

institute for giving me opportunity to undergoes my internship. This chance in a way helps
students to develop themselves and improve their capabilities. This initiative also helps their
students to be prepared upon embarking their journey in their working life once graduated. All
the experience I able to gain throughout the internship period helps to make my studies more
enjoyable.

Not to forget my supervisor appointed to me in Carsem (M) Sdn. Bhd., Norhaliza Binti
Kasim. I really appreciate all the time and effort she gave to help me during my whole
internship. She taught me a lot of new things and help me understands the job scope and task
given to me. With her guidance, I was able to learn a lot of new information as I can.

This also marks the end of my Student Industrial Training (SIP), the second half process
of my whole internship. I believe all the knowledge I gained during SIP could be utilised during
my career as an engineer in the future. During the SIP, I finally able to adapt to the atmosphere
of the company. I also able to gain insight of how engineering industry is really like.

ii
III TABLE OF CONTENTS
I VERIFICATION .......................................................................................................................................... i

II ACKNOWLEDGEMENT ........................................................................................................................... ii

III TABLE OF CONTENTS ....................................................................................................................... iii

IV LIST OF FIGURES ................................................................................................................................. v

V LIST OF TABLES ...................................................................................................................................... vi

VI LIST OF EQUATIONS ........................................................................................................................ vii

1.0 ABSTRACT ............................................................................................................................................. 1

2.0 INTRODUCTION ................................................................................................................................... 2

Background Of Study ......................................................................................................................... 2

Problem Statement .............................................................................................................................. 5

3.0 OBJECTIVE ............................................................................................................................................ 6

Scope of Work ..................................................................................................................................... 7

4.0 LITERATURE REVIEW ....................................................................................................................... 8

Integrated Circuit Packaging Assembly Process (Front Of Line) .................................................. 8

Wire Bond Process (Ball-Wedge Bonding) ..................................................................................... 10

Gold Wire........................................................................................................................................... 12

Copper Wire ...................................................................................................................................... 14

5.0 METHODOLOGY ................................................................................................................................ 15

Methods and Tools ............................................................................................................................ 15

Project Activities ............................................................................................................................... 17

5.2.1 Measuring wire length ............................................................................................................. 17

5.2.2 Measuring UPH ........................................................................................................................ 21

5.2.3 Determining standard usage .................................................................................................... 23

5.2.4 Collecting output devices ......................................................................................................... 24

5.2.5 Machine planning simulation .................................................................................................. 26

Gantt Chart and Milestone............................................................................................................... 28

6.0 RESULTS AND DISCUSSION ............................................................................................................ 30

Revalidating Standard Usage Value ................................................................................................ 30

Optimising UPH ................................................................................................................................ 39

7.0 CONCLUSION AND RECOMMENDATION ................................................................................... 41

iii
 Conclusion.......................................................................................................................................... 41

Recommendation ............................................................................................................................... 43

8.0 REFERENCES ...................................................................................................................................... 44

iv
IV LIST OF FIGURES

Figure 1: Whole process of MLPN-FOL ............................................................................ 2

Figure 2: Singulated wafer on the membrane attached to wafer ring ................................. 3
Figure 3: Example of bonded wire ..................................................................................... 4
Figure 4: Wafer Dicing ....................................................................................................... 8
Figure 5: Illustration of the wire bonding process cycle .................................................. 11
Figure 6: Flowchart of the project .................................................................................... 15
Figure 7: Wire bonding machine ...................................................................................... 17
Figure 8: High-power measuring microscope .................................................................. 18
Figure 9: Anatomy of the bonded wire ............................................................................. 18
Figure 10: The lens is focused on the bond pad ............................................................... 19
Figure 11: Lens is focused onto the tip of the loop .......................................................... 19
Figure 12: Lens on the middle of the ball ......................................................................... 19
Figure 13: Lens on the centre of the stitch........................................................................ 19
Figure 14: Template for recording devices output............................................................ 24
Figure 15: Template for simulation (Excel) ..................................................................... 26
Figure 16: Old vs. new Std. Usage (Device A) ................................................................ 30
Figure 17: Device A Output Finance vs. Study (Full Spool)............................................ 31
Figure 18: Old vs. new Std. Usage (Device B) ................................................................. 32
Figure 19: Device B Output Finance vs. Study (Full Spool) ............................................ 33
Figure 20: Old vs. new Std. Usage (Device C) ................................................................. 34
Figure 21: Old vs. new Std. Usage (Device D) ................................................................ 35
Figure 22: Device D Output Finance vs. Study (Full Spool)............................................ 36
Figure 23: Old vs. new Std. Usage (Device E) ................................................................. 37
Figure 24: Device D Output Finance vs. Study (Full Spool)............................................ 38
Figure 25: UPH before vs. after (5 devices) ..................................................................... 39
Figure 26: Daily Output before vs. after (5 devices) ........................................................ 40

v
V LIST OF TABLES

Table 1: Standard Electrode Potentials of Metals (Retrieve from Gold Bulletin 2002) ... 13
Table 2: Tools and methods used during the project ........................................................ 16
Table 3: Gantt Chart ......................................................................................................... 28
Table 4: Summary for standard usage changes ................................................................ 41
Table 5: Summary for UPH .............................................................................................. 41

vi
VI LIST OF EQUATIONS

Equation 1: Calculation for percentage difference ........................................................... 20

Equation 2: Time each unit ............................................................................................... 21
Equation 3: UPH ............................................................................................................... 21
Equation 4: Std. Usage...................................................................................................... 23
Equation 5: Expected Output ............................................................................................ 24
Equation 6: Percentage Output ......................................................................................... 25
Equation 7: Output ............................................................................................................ 26

vii
1.0 ABSTRACT

In Carsem’s department of MLPN-FOL, one of the main process in the assembly is wire
bonding. Wire bonding use copper and gold wire in its operation thus an optimisation in
standard usage and UPH values is needed to ensure the sustainability in the future. The standard
usage must be determined and revalidate either there is increase or decrease in its usage. UPH
also must be optimised to ensure the output is maximised. By using these two approach, it is
hoped that any wire wastage can be reduced. The objective of this project is to reduce wire
wastage to ensure the sustainability of the process itself in the future. Another objective is to
revalidate and determine the new standard usage. By determining new standard usage on high
running device, output expectancy for one spool of wire can be predicted and reduce any excess
wire by the end of the operation. Lastly, the project also aims to maximise and optimise the
UPH values. By this optimisation, proper machine planning can be achieved thus reducing the
administration of wire spool which in the long run reduce wire wastage. To successfully carry
out this project, the wire length used on each device is measured so that new standard usage
can be determined. Also, output percentage must be collected whenever a wire spool has
finished its usage in order to validate the measured standard usage. Machine UPH must be
recorded frequently to optimise and detects any reduction in the UPH values. To carry out this
project 5 high running devices were selected to be focused on. As a result, from the project, 3
out of 5 devices standard usage have achieved the set criteria in order it to be changed. It is
also can be observed that the UPH for 4 devices were able to increase and 1 devices remain the
same. From this, the wire bonding process is successfully optimised.

1
2.0 INTRODUCTION

Background of Study

In Carsem, MLPN-FOL department consist of 3 main steps of processes which are: -

1. Die attach
2. Surface Mount Technology (SMT)
3. Wire Bond

Figure 1: Whole process of MLPN-FOL

As the wafer is received from the customer, the wafer will be stored in the kitting room for
few processes. It will be mounted onto the membrane held by a wafer ring. The wafer will then
be sent to the wafer cutting machine which will divide the wafer into dies. During the process

2
of singulating the wafer, the wafer may crack or contaminated. To ensure the quality of the
dies, it will be inspected to mark any defected dies.

Figure 2: Singulated wafer on the membrane attached to wafer ring

After the wafer cutting, the die will be attached to the lead frame. Two ways to attach the
die are using solder paste or epoxy. For this project, the devices mainly use epoxy for die attach.
The epoxy will serve as a glue for the die to stay intact to the lead frame. The devices will then
be sent to epoxy oven in order for the curing of epoxy to takes place. Commonly, the time
required to cure the epoxy is around 2 hours per magazine.

Once the epoxy is cured, the devices will be cleaned using plasma cleaner. This is to ensure
any organic and non-organic residue on the products is removed. The importance of this step
is to make sure no problems occur during the wire bonding process.

In wire bonding, Carsem uses two types of wire. Gold and Copper wire are commonly
used in Carsem as both of the materials property are great for electrical devices. However, in
the recent years the availability of the materials in raw are harder to find which causes the wire
price to hike. Because of the circumstance, it is important for the company to reduce the
wastage. The wire usage also needed to be optimised.

3
 Figure 3: Example of bonded wire

The recorded standard usage available in the company might needed to be revalidate. By
revalidating the standard usage of the wire, the usage of wire for each spool can be optimised
further. UPH value also is important in order to reduce wastage of wire. The wastage could be
reduced by reducing the assigned spool for the production.

4
 Problem Statement

Wire bonding is one of the main process in the assembly of MLPN packages. This process
involves the bonding between the wire itself with the bonding pad situated on the lead and die.
In Carsem, two types of wire are commonly used in the production namely Gold and Copper.
Since Gold can be considered as valuable raw materials, over usage of the materials in the
process can be considered as non-sustainable in the long run. This required a mitigation action
to be put in plan to reduce any unnecessary wire wastage.

The usage of the wire on each device already recorded by Carsem. However, over the
years, with some improvement on the technology and the machine parameters and accuracy,
the standard usage for the devices might change. The change might be increased or decreased
in values. There also might be no change in the values. It is up to the capability of the machine
and the technicians for the value to have changes. Thus, the new values of standard usage must
be determined in order to ensure no excess wire usage will happen.

The number of spools administered daily also can be one of the reasons of excess usage of
wire. This are true especially when higher number of output volumes are required in a day.
With higher output, more machines are required to be running in a day. The higher number of
spools administered causes the probability of wire to be waste will be increased. This is due to
not all of the spool placed are guaranteed to be used up. Even though the unused portion can
be reused in the future, setting up the spool or changing the wire spool on the machine’s
capillary tube will somehow consumed a portion of the wire which will leads to a wastage. To
avoid this, the number of machines in used must be reduced thus the UPH of the machine must
be determined and optimised.

5
3.0 OBJECTIVE

To overcome the problems stated as before, few objectives were produced in line with the
experiment which includes: -

1. To reduce wire wastage, occur in wire bonding process by optimising the wire usage
through the standard usage and UPH values to ensure the sustainability of the wire
bonding process economically and environmentally in the future.

2. To revalidate and determine the standard usage of the wire used by each device
compared to its currently available data by measuring the wire length used by each
device to ensure no wire excess at the end of operation.

3. To determine and optimise the UPH value of wire bonding machine in the production
site for each device through comparing it to existing data available by taking the
average value of UPH from each machine running with same devices in order for the
output of the machine can be optimized for daily production.

6
 Scope of Work

To achieve the objectives mentioned before, few activities were conducted. For this
project, 5 high running devices were put on focus. The value of UPH were recorded daily from
machine that were assigned for the same devices from 5 focused devices. The average value
will then be calculated. To determine the standard usage of the devices, each wire length
bonded on the devices needed to be measured to get total wire length used for each device.
Through optimizing the machine productivity and revalidate the devices’ standard usage, the
wire wastage could be reduced.

7
4.0 LITERATURE REVIEW

Integrated Circuit Packaging Assembly Process (Front of Line)

In assembly process for integrated circuit few generic processes are needed. Focusing in
Front of line process, the processes include the preparation of the wafer. Integrated circuit in
wafer are usually probed electrically in order to tests its functionality. Usually, wafer that are
fabricated usually have the thickness approximately around 650 µm and also can be thinned
out until 400 µm depending on the package (Lim, 2013). Lim (2013) also states that thinning
is necessary to achieve reduced thermal stress causes by the difference in thermal expansion
CTE of the silicon and the packaging material. Probed wafers are then to be mounted on a tape
which were priorly assembled to a frame. The frame will be mounted to sawing machine which
have diamond blade to dice the wafer into die.

Figure 4: Wafer Dicing

After the dicing of the wafer into die, die attach process will be in place. Die attach is to
permanently mount the die to the package. In Carsem IC package assembly, epoxy is used
acting as the adhesive for the die attach. As mentioned by Lim (2013), epoxy will be dispensed
onto the lead frame trough nozzle and as the epoxy gives better adhesion for silicon or silicon
dioxide, no metallization is required to the back side of the die. After attaching the die using
the epoxy, it will be cured at elevated temperature in order the die bond to complete. Curing
conditions usually range from 125°C to 175°C around 1 to 2 hours (Lim, 2013).

8
 After curing the epoxy, the next process is wire bonding. In Carsem, usually the choices
of wire are copper or gold. According to Lim (2013), gold usually used as they could bond
well to the bond pads of the idea as well as metallized part of the package forming metallurgical
diffused materials. Copper has lower tendency to form inter-metallic compound with
aluminium. This can weaken the bond of the wire which will leads to bond lifting or other
problems (Lim, 2013). At Carsem, after wire bond is complete it will undergo a process called
3rd optical for quality inspection before proceeding to the End of Line for packaging processes.

9
 Wire Bond Process (Ball-Wedge bonding)

The process of wire bonding is to create an electrical interconnection in the integrated

circuit (TWI Ltd., n.d.). The connection between the semiconductors and silicon chips are
established using fine wires usually made of gold and aluminium. Before the process could
take place, the die attachment process is required to be completed where the die will be
mounted on the substrates such as lead frame or multilayer substrate (Brand, 2008). The
process technology can be divided by the bonding method which are ball-wedge or wedge-
wedge bonding (Schuettler & Stieglitz, 2013). Schuettler and Stieglitz (2013) also mentioned
the process can be divided by its actual mechanism in creating the metal interconnection
between both the substrate and the wire itself. During the wire bonding cycle, generally for
both copper and gold wire both have the same general process which it will undergoes ball
bond, loop, and stitch sequence (Small Precision Tools, n.d.). Firstly, consistent EFO
(electronic flame-off) is fired in order to form FAB (Free Air Ball). Consistent tail length which
is obtained after second bonding is required for the FAB size to have its consistency. According
to Schuettler and Stieglitz (2013), the wire is threaded in the machine using capillary and Small
Precision Tools (n.d.) states that FAB is focused in the capillary chamfer area using wire
tensioner before proceeding to be lowered down to the bonding pad.

The first bonding will be achieved when the ball is pressed onto the bond pad situated on
the die (Brand, 2008). As mentioned by Small Precision Tools (n.d.), FAB will be deformed
using the impact force and with the help of ultrasonic, force, temperature and time will help
further deformation to the FAB accordingly to the shape of the capillary’s chamfer angle. This
first bonding will be called as ball bonding. After the ball bonding completed, loop will be
formed as the capillary raises (Small Precision Tools, n.d.) and travel along the trajectory
programmed on the machine (Brand, 2008). Small Precision Tools (n.d.) also states as the
capillary arrived at the desired position on the lead (substrate bonding area), stitch bonding will
be formed by using the same aspects applied during the ball bonding formation. As written by
Schuettler and Stieglitz (2013), the capillary’s rim will be used to form the wedge-like
impression on the bonding area. The wire will be broken off as it is being clamped while the
bonding tools will travel upwards. The amount of the tail bond left is important in order to form
the next FAB for the next cycle of the bonding process.

10
Figure 5: Illustration of the wire bonding process cycle

11
 Gold Wire

Goodman (2002) mentioned gold is mainly used in electronics such as wire bonding
process. This is caused by the gold having excellent corrosion resistance, high electrical
conductivity and also has good wear resistance (Goodman, 2002). Gold widely used in
thermosonic bonding as it is easily bonded by heat, pressure, and ultrasonic energy. While
comparing gold to copper, silver and even aluminium in terms of conductivity, it offers higher
level of conductivity compared to the other stated metals (Monroe Engineering, 2019). High
conductivity in gold means that electricity can flow through the gold smoothly without high
resistance. This property makes gold a superior material to be used as wire.

Another property of gold making it widely used in wire bonding is its ductility. The soft
and pliable properties of gold as opposed to the mass assumption of where gold is hard and
strong makes it is easy to work with (Monroe Engineering, 2019). This ductile property allows
fine thin wire to be drawn out such as an ounce of gold are able to be drawn up to 80 kilometres
in length with 5 microns thick (American Museum of Natural History, n.d.). Gold is considered
to be reliable for wire bonding in semiconductors ICs at high rate due to its ability to be ductile
(Goodman, 2002) and can be included in small electronics whereas harder and stronger metals
are not possible to be done (Monroe Engineering, 2019). However, as mentioned by Goodman
(2009) in his article, the downside of gold being ductile is that gold is not a suitable choice for
situations where it is used for contacts where wear will occur. Few developments towards gold
were made for wear application to make it hard such as electroplating it on to Palladium (Pd)
or Nickel (Ni).

Besides being ductile and having low resistivity for electrical flow, gold is also known to
have good resistance from tarnishing (Monroe Engineering, 2019). This is the result of gold
not able to mix well with oxygen avoiding it from oxidized. As listed in the Table 1 resistance
of gold compared to other metals can be compared in terms of their standards electrode
potential. The larger the positive value of the potential indicates the better resistance of the
metals from oxidizing (Goodman, 2002). The reason as why it is better to have low rate of
oxidization is because as the oxygen reacts with metal, it will leads the metal to be tarnished
(or other form of corrosion) (Monroe Engineering, 2019) which will form a compound oxides
which in most cases are insulators (Goodman, 2002). This property also helps the packages or
devices not to be sealed hermetically.

12
 Table 1: Standard Electrode Potentials of Metals (Retrieve from Gold Bulletin 2002)

Another reason wire is used in wire bonding is that when the process needs to be carried
out thermosonically (combining heat and ultrasonic energy) or ultrasonically (without
involvement of heat), the process can be done with higher reliability when using gold
(Goodman, 2002). Goodman (2002) reports that this is important where millions of devices
with over 100 leads are being manufactured. Gold wire also have the upper hand where ball
bonds can be form in the rate of 20 bonds per seconds as compared to Aluminium where also
used in wire bonding but cannot formed a reliable ball formation thus needing it to be bonded
in wedge bonding operation occupying more space and at lower rate of 8 bonds per seconds.

13
 Copper Wire

Copper is the main choices for electrical connectors. This may stem from few advantages
copper possess itself. Copper has high conductivity when comparing with other metal only
below silver (Monroe Engineering, 2018) thus making copper an excellent conductor as it able
to conduct electricity more efficiently (Weinhandl, 2019). Monroe Engineering (2018) also
states that this property allow copper to provide greater distance for electrical current travel.
As copper oxidised, it does not corrode like any other metal. Instead, as mentioned by
Weinhandl (2019), it will produce a greenlike patina on its surface known as copper oxide.
Unlike rust on steel, this actually avoid the metal from further oxidixation whilst not
compromising its conductivity.

Even though gold is considered an exemplary conductor, however the cost of using gold
to manufacture electrical wires can be considered waste of money when comparing to gold if
to be talked about cost (Monroe Engineering, 2018). It is also proven as stated by Monroe
Engineering (2018) that copper is more conductive while being affordable comparing to its
aforementioned counterpart.

Copper also have high ductility means that copper could be bend without breaking
(Monroe Engineering, 2018). The continuously shrinking size of electronics requires metal that
is easy to shape and can be processed in which copper excel at (Shanghai Metal Corporation,
2018). As a result, copper able to transmit energy without compromising its power strength
due to its shape and deformity.

Lastly, the property of copper which is also deemed important is that copper have high
thermal resistance. This often-overlooked property of electrical wires is actually giving a great
benefit (Monroe Engineering, 2018). Shanghai Metal Corporation (2018) noted that heat
caused by electrical components during its usage could cause the components to age faster than
its actual estimated lifetime. With copper having high thermal resistance, the problem
associated with heat can be solve as it will dissipate the heat also providing a faster cooling.

14
5.0 METHODOLOGY

Methods and Tools

Figure 6: Flowchart of the project

15
 Table 2: Tools and methods used during the project

NO. Tools Methods

1 High power microscope Measuring the wire length and wire loop used on
the unit.
2 Stopwatch Measure time taken for a unit/strip to be bond
3 Carsem Web System Accessing FMEA+ available on the wire bond (IC)
process
4 Excel Data storing and documentation
5 Device output template Data collection
6 Ruler Measuring stray wire and cut-off wire

16
 Project Activities

5.2.1 Measuring wire length

Measuring wire length used for each device is one of the activities conducted during the
project. The objective for the activity is to validate the current standard usage available in the
system. This activity will be done manually also will be compared with the measurement
obtained by the bonding machine itself. To measure the wire length used for the device, a high-
power microscope is needed to perform the activities since the wire is too fine to be measure
using other measuring instruments, i.e., ruler.

Figure 7: Wire bonding machine

To obtain the reading of the measurement for each wire from the machine, a technician
will be asked to access it since the machine needs to stop its operation. Another reason why
technician is needed is because in order to access to the parameters, a password key is needed
which is only possessed by certain technician. Once the reading is taken, the technician will
continue the machine operation for its production. To obtain wire usage for the whole device,
the parameters recorded will be added to obtain its total wire usage. The unit for the parameters
is in μm. It will later be converted to kft.

17
 Figure 8: High-power measuring microscope

To measure the wire manually, high-power measuring microscope will be needed. To

measure it, the wire will be divided into two parts anatomically. The parts mentioned is loop
height and length. The loop height is measured vertically in Z-axis. For the loop length
however, it could be measure in horizontal plane either in X-axis or Y-axis. This will be
according to the direction of the wire connected. This process will be repeated following to the
number of wires connected on the device. For instance, if the wire count is 80, thus the process
will be repeated for 80 times. The two value will be added up to get total one wire length used
for one bond. The values will be summed up to obtain the total wire used on one device.

Figure 9: Anatomy of the bonded wire

18
 To measure the loop height, the microscope lens is first needed to be focusing on the bond
pad. While it is focused on the bond pad, the measuring screen of the microscope is required
to be reset to zero value in the Z-Axis reading. This is to establish its reference point. Then, the
lens will be refocused onto the tip of the loop where the final reading of the Z-axis measurement
on the screen is the loop height reading. The reading shown on the screen is in mil.

Figure 10: The lens is focused on the bond pad Figure 11: Lens is focused onto the tip of the loop

Moving on to measuring the wire length, the microscope is required to be focused on the
middle of the wire ball which were form during the bonding process (created by EFO) situated
on the bond pad. The measuring screen of the microscope is reset to zero in either X-axis or Y-
axis (according to the direction of the wire bonded). The microscope is then following the wire
direction to its bonded site on the lead. It will focus on the middle of the crescent-like stitch.
The reading recorded on the measuring screen is taken (either in X-axis or Y-axis direction
accordingly).

Figure 12: Lens on the middle of the ball Figure 13: Lens on the centre of the stitch

19
 After measuring all bonded wire loop height and length, the value for both height and
length will be totalled together accordingly to the wire index. Then, to gain the wire usage for
the whole one device, the values will be summed up altogether. The unit is in mil which will
later converted to kft.

Conversion from mil to kft: 1 mil ≈ 8.333×10-8 kft (approximation)

Conversion from μm to kft: 1 μm ≈ 3.281×10-9 kft (approximation)

The machine and the manual value will be compared with each other to see the validity of
the measurement from the machine. The allowable percentage difference between the two
readings is below 15%. The calculation for the percentage difference can be acquired as below:

𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 − 𝑀𝑀𝑀𝑀𝑀𝑀ℎ𝑖𝑖𝑖𝑖𝑖𝑖 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟

× 100% ≤ 15%
𝑀𝑀𝑀𝑀𝑀𝑀ℎ𝑖𝑖𝑖𝑖𝑖𝑖 𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟
Equation 1: Calculation for percentage difference

20
5.2.2 Measuring UPH

UPH is the capabilities of the machine to produce a certain amount of output in an hour.
In the wire bonding process, the values could be obtained easily by the machine. However,
there are few machines which are in older models does not have this function. In the absence
of this function, the value can be calculated manually.

To calculate the UPH value, few steps are needed: -

1. The number of the units per strip is counted.

2. Stopwatch is start from when the first unit is bonded. However, if the indexer does not
ready the strip before the last strip ends, stopwatch must be start from when the strip is
out of the magazine.
3. The timer will be stopped once the last unit is bonded. If the indexer is programmed to
not prepared the strip continuously, the timer is required to be stopped when the strip
has entered the magazine.
4. The time taken is recorded.
5. The time taken is converted to seconds (1 min = 60 seconds).
6. The number of units and time taken is compute into Equation 2.
7. Value from Equation 2 is entered into Equation 3.

𝑡𝑡
𝑇𝑇 =
𝑢𝑢
Equation 2: Time each unit

Where:

T = Time taken per unit (s)

t = Time taken per strip (s)
u = number of unit

3600
𝑈𝑈𝑈𝑈𝑈𝑈 =
𝑇𝑇
Equation 3: UPH

21
 UPH value is essential in order to optimized full capabilities of the machine in performing
required operation. It also helps for schedule planning for which device and machine could be
used for optimum output.

22
5.2.3 Determining standard usage

Standard usage is the amount of length of the wire used to bond every 1000 units of device.
The unit is in kft/k. This value can be obtained by using the value of total wire length used for
a unit of each device. To calculate standard usage, the equation below can be used: -

𝑤𝑤𝑤𝑤𝑤𝑤𝑤𝑤 𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙𝑙ℎ 𝑓𝑓𝑓𝑓𝑓𝑓 𝑒𝑒𝑒𝑒𝑒𝑒ℎ 𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢 𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑𝑑 (𝑘𝑘𝑘𝑘𝑘𝑘)

𝑠𝑠𝑠𝑠𝑑𝑑. 𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢 =
1000
Equation 4: Std. Usage

The value obtain will be used to be compared with the value that is already provided by
the finance. If the difference between the values is greater than 0.17 kft/k. further action will
be needed in order to verify either the changes is true or not. If the difference is lesser than
aforementioned value, thus no further action will be taken as it is automatically verified that
there is no change in the wire length used.

To verify the new standard usage, a deeper investigation on the actual output produced vs
the expected output from the old standard usage and new standard usage will be compared.

23
5.2.4 Collecting output devices

In order to determine change of standard usage value is required for certain device,
estimated output is calculated using both standard usage provided by the finance and new
collected one. The expected output data can be calculated using the following formula: -

𝑠𝑠𝑠𝑠𝑠𝑠. 𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝
𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 =
𝑠𝑠𝑠𝑠𝑠𝑠. 𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢𝑢
Equation 5: Expected Output

The standard pack value is provided by the wire manufacturer. Usually, the value is written
on the spool packaging. However, the number of output obtained here is hypothetical. The real
value of device output needed to be collected manually from the data provided by the wire
bonding machine. The task of noting the value of the output was given to the operator. They
were responsible to fill in the number of device output onto the template provided.

Figure 14: Template for recording devices output

From the template, the wire part indicates the type and specification of wire used by the
machine at the time of the recording. The serial number indicates which wire administered by
the company to operate the machine. This serial number is important as to detect either any
same spool of wire as used for other machines or at other time which indicates that the spool
is a partial spool. Bond unit column otherwise is to record number of devices bonded by the

24
spool of wire meanwhile the left-most column is to record amount of times threading of wire
were needed during the operation.

From the output recorded, percentage of real output is calculated against both expected
output for new and old expected output obtained from standard usage. The calculation can be
done using below formula: -

𝑟𝑟𝑟𝑟𝑟𝑟𝑟𝑟 𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝𝑝 𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜 % = × 100%
𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒𝑒 𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜𝑜
Equation 6: Percentage Output

The percentage output between new and old will be compared using few data. The number
of data where new is closer to 85% threshold occur in 75% of times will be considered as
accepted, thus proving the new value of standard usage is true.

25
5.2.5 Machine planning simulation

In order to optimise the operation in wire bonding, a proper planning is required to achieve
maximum output. The importance of good plan for an operation could reduce the number of
wire spool needed to be administered. By reducing the number of spool administered, the
wastage of wire can be reduced.

The simulation can be done merely in excel. The spreadsheet needed to be formulated in
order for it to calculate the production output automatically by entering the devices name and
the number of machine needed to produce the devices

Figure 15: Template for simulation (Excel)

In Figure 15, the tables shown are the formulated table in excel to simulate the planning.
The tables columns of “CUSTOMER” and “UPH” (blue table) are formulated in a way that it
will fetch the name and the values automatically from another excel file which contains the
“Masterlist.xlsx” where the name of the device together with customer and its UPH values
were recorded. The device name will be fill in according to needed devices output in the
“DEVICE” column. Only then the “CUSTOMER” and “UPH” columns fetch the values from
the “Masterlist.xlsx” file using “XLOOKUP” function. “NO. OF MACHINE” columns will be
entered by determining the number of machine will operate the device. The “OUTPUT”
column will calculate the output for the devices by using formula:

OUTPUT = 𝑈𝑈𝑈𝑈𝑈𝑈 × 𝑁𝑁𝑁𝑁. 𝑂𝑂𝑂𝑂 𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀𝑀 × 22

Equation 7: Output

26
 From Equation 7, the “OUTPUT” is calculated by multiplying the “UPH” and “NO. OF
MACHINE” with 22. 22 is hours of operation. It can be changed to suit the timeframe of the
operation. Since the output calculated is for daily output, the value of 22 is decided to be used.
This value already factors out all the trivial amount of time such as operator rest time, shift
change and etc. In the “TOTAL” row, it will sum up the number of machine to be operate and
the total of output.

For the green table, “COSTUMER” column is pre-filled with all customers of Carsem. For
the “OUTPUT” column, by using SUMIFS function, the number of output by each customer
is calculated. This is to easily see which customer have the right number of output for the day.
It is also formulated to be highlighted with red if the customer’s output is not achieved.

27
 Gantt Chart and Milestone

Table 3: Gantt Chart

Activities/Week 1 2 3 4 5 6 7 8 9 10 11 12 13 14
Obtain project
Understand the wire
bond process
Collect data for wire
length
Collect data for UPH
Produce template for
output percentage
record
Brief operator (record
output)
Collect output
percentage
Reproduce new
template
Collect data for
occurrence of thread in
wire
Update data file
Data analysation
Data documentation

28
 Produce report
/Presentation slide
Presentation
Mark submission

29
6.0 RESULTS AND DISCUSSION

Revalidating Standard Usage Value

Device A

Device A STD USAGE

0.45 FINANCE
0.412
0.4

0.35

0.3

STUDY
0.25 0.238

-0.174
0.2

0.15

0.1

0.05

Figure 16: Old vs. new Std. Usage (Device A)

Figure 16 shows the difference between the Finance standard usage and Study standard
usage. The standard usage currently being implement for device A is 0.412 kft/1000. However,
during the study, the measured value for standard usage stands at 0.238 kft/1000. The
difference between the two values is at 0.174 kft/1000. If such latter value is true, for every
10000 devices, 1.74 kft of wire can be saved which could bring more benefit to the company
in terms of sustainability. Since the difference between the values is bigger than 0.12 kft, thus
further action is required to prove the new value can be used.

30
 Device A (FULL SPOOL)
350.0%
300.0%
250.0%
200.0%
150.0%
100.0%
50.0%
0.0%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

FINANCE STUDY MIN OUT

Figure 17: Device A Output Finance vs. Study (Full Spool)

From Figure 17, graph shows the percentage of real output with using both standard usage
from finance (recorded) and collected during the project. The minimum percentage output set
as the benchmark was at 85%. By using the standard usage given by finance, the full spool of
wire can produce 290.8% at maximum and 102.9% at minimum of output which means
production output can be more than expected output. When the values of percentage output
were readjust using the new standard usage, one full spool of wire can give an output of 168%
maximum and 59.4% minimum. However, these values already considered the factors of the
stray wires due to thread in wire causes by error during operation. With setting the tolerance of
percentage output from the benchmark percentage at -5%, 87.5% of times the minimum output
is achieved which can be concluded that the new standard usage is acceptable.

31
Device B

Device B STD. USAGE

0.7
FINANCE
0.636

0.6

0.5

STUDY
0.395
0.4
-0.241

0.3

0.2

0.1

Figure 18: Old vs. new Std. Usage (Device B)

Figure 18 shows the difference between the Finance standard usage and Study standard
usage for Device B. The standard usage currently being use for device B is 0.636 kft/1000.
However, while carrying out the project, the measured value for standard usage is 0.395
kft/1000. The new standard usage is 0.241 kft decreased as compared to the original value. If
such latter value is true, for every 10000 devices, 2.41 kft of wire can be saved which could
bring more benefit to the company in terms of sustainability. Since the difference between the
values is bigger than 0.12 kft, thus further action is required to prove the new value can be
used.

32
 Device B (FULL SPOOL)
300%
250%
200%
150%
100%
50%
0%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16

FINANCE STUDY MIN OUT

Figure 19: Device B Output Finance vs. Study (Full Spool)

From Figure 19, graph shows the percentage of real output with using both standard usage
from finance (recorded) and collected during the project. The minimum percentage output set
as the benchmark was at 85%. By using the standard usage given by finance, the full spool of
wire can produce 273% at maximum and 93% at minimum of output. The high percentage of
maximum amount of output shows that the standard usage might have decreased. When the
values of percentage output were readjust using the new standard usage, one full spool of wire
can give an output of 170% maximum and 58% minimum. However, these values already
considered the factors of the stray wires due to thread in wire causes by error during operation.
With setting the tolerance of percentage output from the benchmark percentage at -5%, 81.25%
of times the minimum output is achieved which can be concluded that the new standard usage
is acceptable.

33
Device C

Device C STD. USAGE

0.3
STUDY
0.268

FINANCE
0.25
0.233

0.035

0.2

0.15

0.1

0.05

Figure 20: Old vs. new Std. Usage (Device C)

Based on Figure 20, currently the standard usage implemented for Device C is 0.233
kft/1000. When the standard usage is measured for the project, the value is 0.268 kft/1000. It
can be seen that there is an increment of value between the two values. However, the difference
between the two values is only 0.035 kft. Since the difference is less than 0.12 kft, it is assumed
that there is no change in the standard usage value. The slight increase might result to slight
difference of distance between the die bond pad and lead bond pad of the sample device taken
for measuring and the original sample. The slight increase in distance will thus resulting in
slight increase of wire usage. This situation is normal as it is still within permissible tolerance.
It can be concluded, the standard usage will not be change.

34
Device D

Device D STD. USAGE

0.5 FINANCE
0.472

0.45

0.4

0.35
STUDY
0.3 0.295
-0.177
0.25

0.2

0.15

0.1

0.05

Figure 21: Old vs. new Std. Usage (Device D)

In Figure 21, it shows the difference between the Finance standard usage and Study
standard usage for Device D. The standard usage currently being use for device D is 0.472
kft/1000. However, while carrying out the project, the measured value for standard usage is
0.295 kft/1000. The new standard usage is 0.177 kft decreased as compared to the original
value. If such latter value is true, for every 10000 devices, 1.77 kft of wire can be saved which
could bring more benefit to the company in terms of sustainability. Since the difference
between the values is bigger than 0.12 kft, thus further action is required to prove the new value
can be used.

35
 Device D (FULL SPOOL)
300%
250%
200%
150%
100%
50%
0%
1 2 3 4 5 6 7 8 9 10 11 12 13

FINANCE STUDY MIN OUT

Figure 22: Device D Output Finance vs. Study (Full Spool)

By referring to Figure 22, the graph shows the percentage of real output while compared
to both standard usage from pre-existing recorded and collected during the project. 85% were
set as the minimum output benchmark. Using the old standard usage, it can be seen that the full
spool of wire can produce 275% at maximum and 120% at minimum of output. The high
percentage of maximum amount of output shows that the standard usage might have been
optimised to lower values to able producing more output. By readjusting the graph using the
new standard usage, one full spool of wire can give an output of 172.09% maximum and 75%
minimum. These new percentage already considered by factoring out the stray wires due to
thread in wire causes by error during operation. With setting the tolerance of percentage output
from the benchmark percentage at -5%, 84.61% of times the minimum output is achieved
which can be concluded that the new standard usage is acceptable.

36
Device E

ALL Device E STD. USAGE

0.5
FINANCE
0.442
0.45

0.4

0.35 STUDY
0.318

0.3 -0.124

0.25

0.2

0.15

0.1

0.05

Figure 23: Old vs. new Std. Usage (Device E)

Based on Figure 23, it is shown that the currently being implemented standard usage is
0.442 kft/1000. During the project study, the standard usage measured is at 0.318 kft/1000. The
two values difference is at 0.124 kft which is a decrease for the newly measured value as
compared to the pre-existing value. This indicates that if the measured is true, a spool of wire
could save up to 1.24 kft of wire for every 10000 devices produced. This would bring the
company a lot of benefits in the long run in terms of sustainability. As the difference between
the new values is bigger than 0.12 kft, further action will be needed to validate the new value.

37
 Device E (FULL SPOOL)
300.00%
250.00%
200.00%
150.00%
100.00%
50.00%
0.00%
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19

FINANCE STUDY MIN OUT

Figure 24: Device D Output Finance vs. Study (Full Spool)

Referring to Figure 24, it shows the output percentage of the device using both old and
new standard usage. When using the pre-existing standard usage, the maximum output
produced is at 268.33% and the minimum output is 92.38%. The high maximum output shows
that it might be able to produce more output than expected. When the standard usage s adjusted
using the newly studied standard usage, the maximum output percentage is 193.05% and its
minimum output is at 66.46%. These new percentage already considered by factoring out the
stray wires due to thread in wire causes by error during operation. By setting the acceptable
tolerance for minimum output at -5%, 63.15% of the times the minimum output is achieved
when using the new standard usage. However, since the minimum occurrence for the new
standard usage to be acceptable is 75% thus the new standard usage is rejected.

38
 Optimising UPH

UPH Values
400.0

350.0

300.0

250.0
UPH

200.0

150.0

100.0

50.0

0.0
A B C D E

Before After

Figure 25: UPH before vs. after (5 devices)

The Figure 25 shows the difference between initial UPH at the start of the project and after
improvement. For device A, the initial value is 282.1 and after optimisation, the value was
increased 293.1 which shows the increment of 3.90%. For device B the initial value was
recorded at 225.9 while after improved the value is 228.6 shows the increment of 1.20%.
Device C however shows no changes in UPH which the value stays at 364. This shows that the
device C already fully optimised in its operation. For device D, the UPH value able to be
maximised from 304.6 up to 313.8 which shows the increment of 3.02%. Finally for Device E,
the UPH value increase by 2.51% from 259.2 up until 265.7 after optimisation.

39
 Daily Output
9000
8000
7000
6000
Daily Output

5000
4000
3000
2000
1000
0
A B C D E

Before After

Figure 26: Daily Output before vs. after (5 devices)

Figure 26 shows the output production daily when adjusted to the UPH. For device A
before optimising the UPH, the output can only be achieved at 6206 per machine as compared
to 6448 output per machine after optimisation. For device B, before optimising the UPH value,
the daily output can be reach is 4970 in average as compared to 5029 output after maximising
the UPH of the machine. For device C since no change in the average UPH thus the daily output
before and after the project remains the same where the output stands at 8017. For device D
there is increment in the daily output from 6701 per machine to 6904 output per machine.
Finally, for device E the output for each machine initially starts at 5702 per machine. After the
optimisation in the operation process, the output can be increased up to 5845 output per
machine daily.

40
7.0 CONCLUSION AND RECOMMENDATION

Conclusion

Table 4: Summary for standard usage changes

Std. Usage (Old) Std. Usage (New)

Device Difference (kft) Accepted to change
(kft/1000) (kft/1000)
A 0.412 0.238 -0.174
B 0.636 0.395 -0.241
C 0.233 0.268 +0.035
D 0.472 0.295 -0.177
E 0.442 0.318 -0.124

To summarize, based on Table 3 3 out of 5 devices have fulfilled the criteria in order to
change the standard usage. The devices which able to be changed includes device A, device B
and device D. For device A, the standard usage was reduced by 0.174 kft from 0.412 kft/1000
to 0.238 kft/1000. The reduction could save up to 42.23% of wire for every 1000 devices.
Device B, the standard usage is reduced by 0.241 kft which is reduction from 0.636 kft/1000
to 0.395 kft/1000 means that the reduction could save up to 37.89% for every 1000 devices.
For device D, the reduction of the standard usage is by 0.177 kft from 0.472 kft/1000 to 0.295
kft/1000. The reduction could save up to 37.5% of wire for every 1000 devices. For deice C,
the proposed new standard usage is rejected as the reason was that the difference between the
old and new standard usage is below 0.12 kft. For device E, the new standard usage is rejected
even though the difference was more than 0.12 kft as 63.15% of the times the minimum output
is achieved when using the new standard usage which is below 75% times of the occurrence.
By knowing the new standard usage of the device, the usage of wire can be expected thus helps
to reduce the wastage of wires.

Table 5: Summary for UPH

Daily output Percentage

Device UPH (old) UPH (new)
(new) changed (%)
A 282.1 293.1 6448 3.90
B 225.9 228.6 5029 1.20
C 364.4 364.4 8017 0

41
 D 304.6 313.8 6904 3.02
E 259.2 265.7 5845 2.51

For UPH optimisation, Table 4 summarise the data for the project. For device A, the UPH
values were able to be increased by 3.9% from the old value of 282.1 to 293.1 which by average
a machine running the device for production will be able to produce 6448 output daily. For
device B, the UPH is maximised to 1.2% of new value at 228.6 compared the old value at 225.9
which by daily output, average per machine would be able to ramp up production up to 5029
output per machine. However, for device C there is no change in the UPH values. For device
D and E, the UPH able to be increase by 3.02% and 2.51% respectively. Device D will be able
to be produced around 6904 averagely per machine daily while device E will be able to be
produced around 5845 at average per machine daily. The difference UPH values between
devices is the result of different number of wire to be bonded by each type of devices. As
observed, the more the number of wire to be bonded, the lower the value of UPH as more time
is needed to establish the connection for the wires. Thus, a proper planning of which devices
to be produce using the right combination is needed to ensure the daily quota production is
achieved. By optimising the UPH values, a smaller number of machine needed to be in used
thus reducing the number of wire needed to administer which will help to reduce wire wastage.

42
 Recommendation

A recommendation can be proposed to ensure the project can be conducted more easily in
the future. For the standard usage validation, it can be suggested that for the real output per
wire spool to be always recorded in the system. This can help in the future if a new validation
or other devices need new standard usage validation. This also will help to reduce the time
taken to collect data and also would help the future person in charge to trace back data as to
when did the output percentage increased or decreased.

For the UPH optimisation, in order to increase the productivity, a greater number of
workforce as in the operators can be put in charge per shift. This will help to ensure that when
an error on the machine occurs it can be fixed more efficiently as there are more people can be
alerted. By reducing the time down error, more output could be produced as the machine
capable to be used without any stoppages. It is also can be suggested that for the technicians to
revise the recipes for focused devices for every machine. They can reduce any unnecessary
parameters that could compromised the time and apply it to every machine with the same
models. By this every machine can be run optimally and almost at the same rate.

43
8.0 REFERENCES

American Museum of Natural History. (n.d.). Gold Properties. Retrieved from American
Museum of Natural History: https://www.amnh.org/exhibitions/gold/incomparable-
gold/gold-
properties#:~:text=Gold%20wire%20Gold%20is%20ductile,out%20into%20the%20t
hinnest%20wire.&text=Gold%20conducts%20heat%20and%20electricity,conductive
%20for%20a%20longer%20time.

Brand, O. (2008). Packaging. In Comprehensive Microsytems (pp. 431-463). Elsevier.

doi:https://doi.org/10.1016/B978-044452190-3.00019-7.

Goodman, P. (2002). Current and Future Uses of Gold in Electronics. Gold Bulletin 2002, 21-
26.

Lim, S. K. (2013). Assembly Techniques and Packaging. Universiti Tunku Abdul Rahman.
Retrieved from
http://staff.utar.edu.my/limsk/Process%20Integration%20and%20IC%20Manufacturi
ng/03%20Assembly%20Techniques%20and%20Packaging.pdf

Monroe Engineering. (2018, September 25). Why Copper is Used to Make Electrical Wires.
Retrieved from Monroe Engineering: https://monroeengineering.com/blog/why-
copper-is-used-to-make-electrical-wires/

Monroe Engineering. (2019, October 18). why Gold Is Used in Electronics. Retrieved from
Monroe Engineering Web Site: https://monroeengineering.com/blog/why-gold-is-
used-in-electronics/

Schuettler, M., & Stieglitz, T. (2013). 4 - Microassembly and micropackaging of implantable

systems. In Implantable Sensor Systems for Medical Applications (pp. 108-149).
Woodhead Publishing. doi:https://doi.org/10.1533/9780857096289.1.108.

Shanghai Metal Corporation. (2018, August 7). The Use of Copper Increases in Electronics.
Retrieved from Shanghai Metal Corporation Web Site:
https://www.shanghaimetal.com/8674-8674.htm

Small Precision Tools. (n.d.). Wire Bonding Cycle. Retrieved from Small Precision Tools:
http://www.smallprecisiontools.com/products-and-solutions/chip-bonding-

44
 tools/bonding-capillaries/technical-guide/basics-of-ball-bonding-process/wire-
bonding-cycle-process/?oid=563&lang=en

TWI Ltd. (n.d.). What Is Wire Bonding? Retrieved from TWI: https://www.twi-
global.com/technical-knowledge/faqs/what-is-wire-bonding

Weinhandl, D. (2019, July 9). Why Is Copper The #1 Choice For Electrical Connectors.
Retrieved from Mead Metals, Inc.: https://www.meadmetals.com/blog/copper-
electrical-connectors-
conductivity#:~:text=Why%20is%20copper%20used%20for,to%20push%20the%20c
urrent%20through.&text=Copper%20has%20low%20resistivity%2C%20and,less%20
oxidative%20than%20other%20metals.

45