29-30 April 2010

29 April 2010

I started off the day continuing my journey into a deeper cognizance of die bonding process. As we had seen yesterday, die bonding process involves die select, die pick-up, die transfer and die attach. The first two sub processes occur in the pick-up area, while the die attach sub process happens in the attach area. Die transfer takes place between the two areas. Naturally, these two areas need to be aligned to a common level. However, there is always a tolerance. The acceptable height difference between the areas is designated at 100µm. If the pick up area is higher than the attach area, machine speed will plummet due to prolonged travel time. On the other hand, if the attach area is higher than the pick up area, the propensity of the transfer arm to prang into the heater block cover will increase drastically.

As we all know, there are copious die sizes. Hence, different collets with the suitable size corresponding to the type of die handled are required. With that information in mind, I helped Miss Manggala to complete a subservient task, which was to check for any die sizes of package SOT363/SC88, from the intranet, which were not listed in a table of die dimensions that NXP Semiconductor is currently having. She will then need to make the purchase of those collets.

After completing the task, I proceeded to gain a deeper understanding of wire bonding process. First off, I stumbled upon what PHICOM really means. It is actually the acronym for Philips Computer. PHICOM machine is meant for wire bonding process. Generally, six steps are involved. I will provide the steps with the aid of the following pictures.



1) The capillary is moving from the spark to ball bond position. When the Free Air Ball (FAB) touches the die, which is also known as touchdown, the ball bonding process starts. Note that FAB is bonded on the die at elevated temperature, pressure and ultrasonic energy.
2) After bonding the ball, capillary makes the wire loop.
3) The capillary lands on the lead and wedge bonding process starts.
4) The capillary moves up. Next, wire-clamp is closed. The capillary moves further up to break the wire and thus from the tail (wedge).
5) As the wire reaches spark height, a tail is hanging under the capillary.
6) A spark forms a ball at the end of the wire. The wire-clamp opens and the air jet pulls the ball inside the capillary.

If you were heedful when reading the steps aforementioned, you would wonder, what is the purpose of ultrasonic energy? Actually, it serves as an avenue to wipe away oxides or undesirable contaminations on the contact area. Also, it helps to mature the bonding in combination with thermal energy. In other words, this insinuates that the ball bonding process can be carried out in a higher temperature. In addition, the presence of ultrasonic energy can soften the metal of Free Air Ball, thus rendering easier ball bonding process.

Apart from that, I had learnt that the difference between normal, reserve and trapezoid loop is the wire loop length. Besides, I got myself a fun fact too! Do you know that the melting point of gold is 1064°C? Oh! By the way, gold wire is the only material used for wire bonding.

I realized that the insipidity of this industrial training had returned when I was again ordered to update bonding diagrams. Now for SC88 package. Again, due to my fast execution, I was able to finish it by today.

30 April 2010

Mundane, sedentary nature of work is fast becoming an idiosyncratic aspect of my industrial training. Sporadic clandestine self permitted time off from facing the computer proved to be the only humdinger of my industrial training life. I started off the day doing again petty task, which is to tabulate the values I obtained from previous measurements into Microsoft Excel. As an overview, I tabulated ball diameter, ball height, wire pull, break point, ball shear as well as free air ball.

Next, try guessing what I was assigned to do. It was none other than the exasperating bonding diagrams!! I was given new package to update, namely SC70 and SC75. At the end of the day, I was able to complete them. I even had the luxury of calculating the bonding diagrams that I had done so far.

26-28 April 2010

26 April 2010

I finished re-updating the bonding diagrams of SOT753P package. So, till today, I had re-updated SOT353P, SOT363P and SOT753P package, with one package having approximately 110 devices. An inexorable sense of achievement gushed up upon me. Next, I was given a new package, SC74 to update the bonding diagrams. I duly obliged and finished the task within considerably short period of time. I finished off the day with the ever-intriguing ball shear measurement. This time around, I had to perform it alone without the supervision of seniors or operators. It was challenging and fun. I put my flippancy aside and began to acquaint myself with the machine. Owing to my inexperience and slipshod nature, some dies were wrecked; some ball bonds were flattened and so forth. It would be preposterous to suggest that I was not traumatized. Fortuitously, the “debacle” that I had instigated was indiscernible.

27 April 2010

I was given the fifth package to update the bonding diagrams, SOT457P and was able to finish it within today. Subsequently, I was brought to the line where I gained knowledge on several things. Firstly, cross sectional view tests are performed on molded products or finished products to check for any copper interconnect and improper die attachment. Probable causes of imperfect die attachment are excessive bond force and exceedingly high temperature. One example of flawed die attachment is tilted die. Tilted die, along with lifted die or similar defects will culminate in the product be labeled as Rth units. These units will have high resistance, due to the voids between the die and the lead frame. On the other hand, copper interconnect can be reckoned as imperfect diffusion taken place between the die back metal and the lead frame substrate. Lead frames are made from copper. Under copper interconnect conditions; the back metal has copper particles diffused into it.

Next, Miss Manggala provided me with some SEM images which were done using devices far more superior than those powerful microscopes that I had used before. The SEM images were extremely clear. Besides that, high magnification can be attained.

I learned a new term today, by the name, collet. Collet is the protruding edge of a die bonding machine used to pick up individual die from a wafer. If the collet is filthy, the dies may be picked up in undesirable orientation or position, which may result in tilted dies. Hence, it must be cleaned regularly using propanol.

There are myriads of aspects that can engender defects, such as imperfections to the die, chipped die, cracked die; imperfections to the wire, offbond, exposed lead and so forth. Therefore, we need delayering analysis, where one can inspect and determine the root cause of failed products.

Lastly for today, some dies are large, which may pose inextricable difficulties in adhering them well to their respective bondpads. Hence, WBC or Wafer Back Coat is used to assuage this predicament. This is one type of glue bonding.

28 April 2010

I started off the day performing and finishing both ball shear measurement and wafer inspection. As I had mentioned in previous blog entries, a wafer may consists of few hundred thousands of dies. There is a possibility some dies are chipped, cracked or physically imperfect. That is the purpose behind me performing the wafer inspection.

Next, I was asked by Miss Manggala to have a roam around the assembly line. A casual stroll in the assembly line provided me with opportunities to gain some valuable knowledge. First off is the sporadic green light shining on lead frames in ADAT die bonding machines. By the way, ADAT is the acronym for Advanced Die Attach Technology. The occasional green light is meant as a visual or lighting aid for PRS, which is Pattern Recognition System. This system is designated for the ADAT machine to recognize good as well as bad dies. The picture below illustrates the PRS concept.


I was also introduced to Pepperpot by Mr Leong, an industrious engineer working along the line. Pepperpot is actually the vacuum head of the ADAT machine, which is used to ensure that the mylar does not stick along with the die when the die is picked up by the collet. It does so by adhering the mylar to the vacuum head.

Besides that, I gained a deeper understanding why BSOB is momentous in double dies wire bonding. Again, BSOB means Ball Stitched On Ball. Usually, in wire bonding, a ball is bonded on the die (ball bonding), while a wedge is bonded on the leadframe (wedge bonding). When double dies are involved, BSOB is implemented. One of the dies will be bonded first with a ball, then with a wedge on top of that ball. This combination is called a bump and acts as a cushion that protects the die from the high force entailed during wedge bonding.


Subsequently, Mr Leong explained the purpose of having two rectangular boxes just after several ADAT bonding machine. These boxes serve as ovens to dry or solidify the glue used for die bonding. Before I go into details of the boxes, let me deviate to the available types of die bonding. Die bonding process can be categorized into eutectic process, epoxy process and soft solder process. In NXP Semiconductor, soft solder process is not used. The discrepancies between eutectic and epoxy process are that eutectic process entails melting and solidifying two components (alloy), forming tri-alloy, such as AuSiAg (Gold Silicon and Silver) and AuGeAg (Gold, Germanium and Silver).
Epoxy process or glue bonding, however, involves the use of adhesive resin. Furthermore, epoxy or glue process is relatively cheaper than eutectic process. In addition, Epoxy process is conducted at a comparatively lower temperature than eutectic process. Now, back to the ovens; So, with the die bonding information in mind, it is equitable that only epoxy process will be needing ovens. If the glue is not dried, the die may budge or move around.

A pictorial representation of the types of die bonding process:


Besides the boxes, the discerning eyes of me caught sight of the changing colours of lead frames in different part of the process. This is due to the oxidation process. Apart from that, the relatively slow speed of some ADAT machines caught my attention too. After some cogent explanation from Mr Leong, I was able to know that these machines are producing power products which use dies with larger size. The low speed is indispensable to ensure that sufficient amount of time is present for the die attachment to take place properly.

I had also gained further understanding of die bonding process. There are 5 prominent issues with die bonding:
1) Must not transmit destructive stress to fragile chip
2) Must make intimate contact between chip and substrate with no voids and adhere well to both
3) Able to withstand high temperature
4) Must have good thermal conductivity: to remove heat generated within chip
5) Must be a good conductor or insulator, depending on application

I was also told by Miss Manggala that there are 3 parameters that are significant in die bonding process, which are temperature, bond force and process delay. Generally, die bonding process can be subdivided into 4 modules, die select, die pick up, die transfer and die attach.

20-23 April 2010

20 April 2010

I finished adding colours to the wires for SOT353 bonding diagrams. Basically, I need to choose from three colours to be used, namely black, blue and red. The criteria that I need to consider in selecting the correct colour is already there in the bonding diagram. If normal looping is written, the wires should be of black colour, blue colour for reverse loop and red colour for trapezoid loop.

Besides adding colours, I was told to check on erroneous lead frame diagrams implemented by previous document control personnel. The scope of search was limited to SOT353 and SOT753 package. I was able to find and remedy 9 incorrect SOT353 bonding diagrams as well as 7 incorrect SOT753 bonding diagrams.

Lastly, I was asked to obtain the amount of wires used for each NXP Semiconductor product by referring to the intranet. For your information, there are 1956 products.

21 April 2010

I continued the work of re-updating bonding diagrams after finishing SOT353 package. The current package that I am working on is SOT363. At the mean time, I am also implicating myself in ball height and diameter measurement as well as acquiring number of wires used in bonding diagrams via intranet. Besides all those easy but repetition-oriented tasks, I learned the term BSOB, which means Ball Stitched On Ball. The following picture provides the illustration.



My prosaic industrial training experience took an unexpected and invigorating twist today. I was asked to teach a lab operator on how to acquire ball bond and die images using microscopes. I guess all those repetition of work had made me a veritable measurement, data updating and acquisition specialist. Although tutoring was a brisk experience, I am yearning for engineering tasks which I can apply some engineering knowledge to challenge my brain.

22 April 2010

I had finished recording the wire amount for 1956 products in NXP Semiconductor. I continued my measurement of ball height and diameter. Just to refresh your memory. I have 720 measurements to be made. Although my work is saturated with utmost boredom, I kept indolence at bay and proceeded my work with swift execution, hoping that I can quickly escape this welter of simple jobs and probably get some challenging tasks.

Hours after lunch, Miss Manggala brought me and demonstrated wafer inspection under microscope. I was taught edge dies are usually chipped or flawed. This is acceptable. However, other than the edge dies, any chipped or flawed dies will be deemed as unacceptable. I will provide a hand drawn illustration of the wafer as well as what it looks like under microscope. Again, I could not take pictures of the wafer as Miss Manggala said it would probably infringe company rights.



Apart from that, I also inspected decapped products. As i had mentioned before, the finished dies will be molded using resin compound to provide some sort of protection or cover. By decapping, we can inspect and check for deficiencies in the dies that lie within the mold. I managed to discover mixed devices where wrong die type is used for that particular product.

I finished off the day performing free air ball measurement. Similar to ball height and ball diameter measurement, free air ball refers to the ball formed by the PHICOM wire bonding machine just before it is bonded on the die.

23 April 2010

I finished free air ball, ball height and diameter measurement as well as the tabulation of those data in Microsoft Excel. I had also finished re-updating SOT363 bonding diagrams and duly continued with SOT753 package. Besides all those antagonizing works, I managed to learn that there are generally two types of lead frames, namely bi lead frame and matrix lead frame. By referring to the picture shown, the upper lead frame is of bi type while the bottom one represents matrix lead frame.



Furthermore, I had also gone through verification flow chart for chipped package as well as scratched package. A picture speaks a thousand words, here goes:

19 April 2010

I completed the acquisition of ball images. Swiftly after that, I finished updating the additional ECO assigned. To be frank, the work that I have been doing does not seem to be of any significance in engineering point of view. I felt like and still feel like I am merely an office boy. As a matter of fact, one of the document control personnel told me that I did a colossal favour in helping them clearing their avalanche of unfinished tasks, like updating bonding diagrams. Not to mention helping the engineering department to run all the measurements, input acquisitions and so forth. These works, to be blunt, irk and antagonize me. I feel anguished. I feel like my engineering knowledge was put to no use at all. Those are the jobs that an ordinary person can accomplish. I feel like this industrial training is a huge waste of time. I could utilize this time to embark on something more beneficial. To add insult to injury, my eyes have been in excruciating pain as all I was doing was just staring at the merciless computer which just radiates incessant stream of baneful rays straight at my face. Worse still, no one seems to care even thought I had requested for a monitor ray filter. I was just rebuffed. I feel like I am a cheap labour. A very cheap labour. I had completed so much task in such a short time, and I only get myself a scant 500 per month. No subsidies, no leaves.....This is an unequivocal living purgatory.

I am not trying to be pedantic, but my internship just keeps getting obnoxious. I was required to add colour to wires in the bonding diagrams that I had just updated. It sound simplistic, but in fact, it entails a tremendously tedious and troublesome process using Microsoft Visio, Microsoft Powerpoint and Acrobat Reader. In other words, I need to rework the 300 odd bonding diagrams that I had completed. Mind you, it takes 6-7 minutes just to get one of these done. The reason behind this addition is to enhance operators' understanding of the bonding diagrams. Come on, there are already ample information on the bonding diagram. Why put something gratuitous? At least notify me when I started out updating the bonding diagrams early on. Again, are these what engineers should do?

14-16 April 2010

14 April 2010

By today, I can reckon myself as being highly efficient and dexterous in handling the microscope meant for ball diameter as well as height measurement. I finished off the measurement of the remaining samples and proceeded with the tabulation of the data in Microsoft Word and Excel. Also, I was required to calculate the three average values for each sample. As I had mentioned before, each sample will be evaluated a total of 15 times for ball height as well as ball diameter. These 15 evaluations were categorized into 3 groups for each sample.

Apart from that, I was taught of the reason behind using gold wires for wire bonding processes. According to Miss Manggal as well as the line operators that I had encountered. Gold wires offer stupendous conductivity and impervious to oxidation. These advantages overshadow gold wire exorbitant pricing. On the other hand, copper wire may be relatively cheap, but it is susceptible to oxidation. Hence, using copper in wire bonds will instigate rusting of wire bonds, thereby undermining the bond desired properties.

15 April 2010

I was introduced to ECO, which is the acronym for Engineering Change Order. I was taught on how to access NXP Semiconductor intranet where only certified employees can gain access to its cross nation database. Upon opening up the file containing the sale items, I was left in absolute awe as NXP Semiconductor deals with a plethora of product. Then, I was taught on how to update the ECO of some products here in NXP Assembly Plant Malaysia (APM).

Owing to my fantastic hand-eye coordination and diligence, I was able to complete updating ECO in the wink of an eye. Of course, I do not mean it literally. Next, to kill of time before Miss Manggala assign me tasks, I read through the verification flow for chipped package and scratched package. Then, Miss Dahlia asked me to stamp and sort those printed updated bonding diagrams that I had done.

After that, I was brought into the assembly line where I got to witness and perform ball shear measurement. Previously, I had dealt with the measurement of ball height and diameter. Now, the gist of this measurement is to test for the ball type wire bonding on each individual die. This task was really a much sought after break away from all the needs of facing computer all day long. In fact, it was extremely mind-boggling and exciting! My fun buds definitely got aroused. Not to mention hilarious as Miss Manggala and Miss Sally were providing animated and amusing tutorials.

This task really entails splendid hand-eye coordination as while one is looking at the die through microscope, one needs to utilize the joystick-like controls to position a special protruding section of the device just beside the ball bond on the die. Next, with a press of the ball shear button, a force will be applied as to displace the ball. Then, the required force to attain that will be displayed and to be recorded. Owing to my inexperience, my movements were completely sluggish and tentative. Worse-still, I "managed" to obliterate or flatten a ball because of incorrect positioning of the ball shear device which incur nothing but laughter from Miss Manggala and Miss Sally. After completing the measurement, I returned and recorded the values in Microsoft Excel under the same file as the one I recorded the ball diameter and height values. Average values were calculated too.

Ya. Another test that eluded me is wire pull. Probably due to its complexity, I was not taught on how to perform it. This test is to determine the strength of the wire bonding.

16 April 2010

I finished marking and sorting SOT353, SOT 363 and SOT 753 printed bonding diagrams. This consists of around 500 sheets of paper encompassing previous flawed bonding diagrams. Upon completion of the task, I was summoned to the New Product Introduction Lab where I was granted my wish of viewing my fingers under microscope. It was irrefutably an awesome experience. I got to see the sweat droplets, finger prints, hair and so forth.

Well, I spent the rest of the day acquiring ball images using microscope. If you had notice, I used the term microscope for numerous times. Actually, they are not the same device. I regard them generally as microscopes as I have no idea of what their specific names are. Again, coming back to the 16 samples that I had worked on for ball diameter, height and ball shear, I need to acquire 3 images for each sample. I will provide one of the images that I had captured.

12-13 April 2010

This is the start of a brand new week. I started off with yet another educational visit around the assembly line, led by Miss Manggala. This time around, I was brought to a few compartments designated for research and experiment purposes. I was introduced to several experienced personnel, who greeted me with some typical questions like how long will I be having my internship, where do I live, where am I studying and so forth. All of them were undeniably jovial and gregarious as uproarious mirth filled the air. However, after I left those rooms, I could sense their taciturnity as they cut a forlorn and morose figure beneath those cheerful semblances as they returned to their respective operating position. Maybe this is what we known as Monday Blues.

Later on, Miss Manggala told me to roam around, be acquainted with the environment, and to observe the actions carried out by the assembly line staffs, be it front end, mid end or back end staffs. First off, I stumbled upon the terms batch number, index number and shift number. Basically, they refer to the output quantity. However, there are few distinctions. Batch number refers to the current die type output quantity. Index number refers to the number of dies produced, regardless of the die type, as the machine senses the number of index hole on the lead frame instead of the individual dies. Customarily, batch number is the same as index number. Next is shift number which refers to the output quantity that the current shift has produced. For your information, there are three working shifts. Apart from that, I also came across terms such as processed die, remaining die and units per minute. As I had mentioned in previous posts, each wafer may contain hundred thousand of dies. Processed die connotes the idea of how many dies that had been extracted by the die bond machine for the bonding processes; and as the name itself implies, remaining die tells us the number of available die on the wafer. Note that the processed die value is inadvertently greater than that of batch number as some dies may be substandard and duly discarded. Moving on, units per minute suggest the speed at which the machine operates. For typical die bond processes using ADAT 2 Machine, the speed is approximately 302 units per minute!

Besides that, I had also observed how the technicians carry out wafer substitution to replace exhausted wafer. Normally after replacing the wafer, they will use Propane to clean up the tip of the machine which is used to extract dies from the wafer. Moreover, these technicians will rely on the alarm triggered or visual inspection of operation light on each machine to realize erroneous processes had occurred. As usual, green light brings about the idea of normal, error-free production process, whereas red light impinges on the notion of error occurrence. Yellow light, on the other hand, indicates that the machine is in idle mode. Furthermore, I had also witnessed how welding processes were done to combines separate lead frames. Lead frames are stored in circular reels. These reels of lead frames serve as the input to almost all machines. As these reels run out of lead frames, production process at that machine will be stopped, pristine reel of lead frames will be attached to the machine, replacing the exhausted reel. The lead frames of the previous reel will now be welded together with that of the new reel, using a welding machine.

As I saunter more, I discovered more. I perceived how the indefatigable operators conduct wafer mounting processes using an expansion device. I also got to see how the dies, wire bonds look like under microscopes. The operators were obliging and helpful as they effortlessly explained coherently what variables that could culminate in imperfect dies or bonds. For instance, excessive bond force exerted by the PHICOM wire bonding machine will engender flawed bond shape.

After those horizon broadening encounters, Miss Manggala came back for me and taught me how to measure wire bond height as well as bond diameter. The microscope meant for this is not just ordinary microscope. It is a highly sophisticated microscope that can enable us to see minute object measuring few micrometers clearly. I spent few hours measuring 16 samples of lead frames. Each sample will have 15 entries of bond height as well as 15 entries of bond diameter. Indolence and nonchalance started to kick in due to the monotonicity but my forbearance prevailed. In fact, I got a little bit exhilarated as I was able to see how the dies, the bonds, the lead frames look like under the microscope.

I could not finish the measurement today as there were a lot of operators who had queued up the utilize that machine. I felt like I was an eminent singer who has long queue of ardent fans jostling for my autograph. Finally, they asked me if I could return tomorrow for the measurement and I duly obliged. I returned to my desk where I was given the third package, SOT363 to update the bonding diagrams.

By the end of 13 April 2010, I completed updating the SOT673 package bonding diagrams, which takes the amount of bonding diagrams that I had encountered so far, to approximately 300. However, the same cannot be said of the measurement of bond height and bond diameter as the device was so in demand I could only use it for short periods of time.

9 April 2010

This entry is going to be relatively compendious. I continued what I had left off yesterday, which was the updating of bonding diagrams. I felt like I was a clerk instead of an engineer. I felt like I was a machine instead of a human. I felt like I was languishing in a sea of PDF files, Powerpoint slides...

There a always two sides to a coin. The only positive impact I had felt was that my speed in dealing with Powerpoint slides and PDF files had exponentially increased. I could literally shut my eyes and performed all those tasks impeccably.

By the end of the day, I had successfully completed updating the bonding diagrams of SOT753 and SOT 353 packages, as well as their respective flat location, if any. This would amount to a total of approximately 200 bonding diagrams. Fuiyoh! I am really fast! Oh ya! Flat location means the position of the wafer on the ADAT die bonding machine. There are a total of 8 possible orientations.

7-8 April 2010

It is unimpeachable that honesty is a virtue; and I intend to uphold that virtue in me. That is why I am going to say that the third and the forth day of my industrial training had been absolutely humdrum! Not to mention physically and mentally tormenting!!!

Upon arriving at the New Product Introduction Unit, I was introduced to my direct supervisor, Miss Manggala by my other supervisor, Mr. Melvin. She is an affable person with a warm smile emblazoning her face.

Then, both of them briefed me about documentation of bonding diagrams as well as the ratification of the flat location. What are those? Bonding diagrams are actually meant to show how the individual die or crystal containing certain circuitry is connected via tenuous gold wire to the bondpad on lead frames. There could be five connections or four connections, depending on the specifications. At the other end of the spectrum, flat location is the position of the wafer as it is used by the machine for die extraction. According to them, this is a momentous task as the assembly line hinges on the information that was about to be recorded. Any mistake that may be indiscernible may culminate in an execrable effect to the assembly line. Not to be rattled, I am a meticulous person. So I performed the task given after detailed explanation and demonstration from the Document Room's colleagues, Miss Dahlia and Miss Shree.

When I say it was mundane, I am not trying to pontificate. I really mean it as what i did for these two days was to face the computer whole day, copy, paste and edit approximately 200 diagrams. I would be preposterous if I had mentioned that I did not feel like I was a machine because what I did was just repetition of steps. My eyes were sore due to the prolonged exposure to the computer which is devoid of any means of harmful rays filtering. The only cheer that had brightened up my day was that I received compliments about the phenomenal rate of my work. To reward that, I was given more work of the exact type. To round it off, the work of these two days had been so monotonous that I had decided to create this new blog post which encapsulates these two days instead of individual post for each day.

6 April 2010

Today, I decided to give my watch a day off. To me, this was an astute decision as I was oblivious to how slow the clock ticks. Time flew by today. Apart from that, I underwent several mind-challenging and intriguing activities.

To start off the day, I was instructed to complete a partial report containing just what I had learned during the first day. Some of the details that I had written are Assembly Plant Malaysia (APM) processes, introduction to semiconductor and organizational chart.



Upon completion of the first task, my supervisor, Mr. Melvin presented me with a mathematical problem involving wire bonding process. I was asked to find the percentage increase in gold wire cost if each individual die size is reduced by 4% from the two edges. Note that the die height is still the same and relevant parameter values were given according to their engineering specifications. So, by using the famous Pythagoras theorem, I was able to calculate the percentage increase in cost. Then, I was also required to devise a general equation for evaluating the percentage cost increase which consists of four variables, namely wire bonding horizontal length, percentage increase wire bonding horizontal length, die height and line height.

After the brief lunch break, I was introduced to DOE which stands for Design of Experiment as well as SPC (Statistical Process Control). In DOE, I had gained some basic knowledge on how to implement 2k full factorial design. It is unequivocal that for each process, there may be several input variables. Through DOE, and via Minitab software, I was able to produce information regarding the set of input variables as well as the output variables. There are four models of interest. Full model, final reduced model, graphical analysis and mathematical model. I was able to determine which input variables are statistically significant or otherwise. Those insignificant factors will be neglected for further analysis of the operation. At last, one would be able to determine the equation that governs the relationship between the sole output and a set of input variables. Besides that, the determination of process window, which can be esteemed as the range of allowable process variables, is of utmost importance. Process operation should be based on the values that are at the center of this particular process window.

In SPC, I was exposed to the use of process capability index, or capability index. This index relates the voice of the customer (specification limits) to the voice of the process. In other words, it is the ratio of the distance from the process center to the nearest specification limit divided by a measure of the process variability. This parameter is useful in determining whether the process if stable or otherwise as it inadvertently reduces the complex information about the process to a single number. If the result were to be greater than 1.67, then the process is reckoned to be stable. Customarily, we deal with two limits, one being the upper sided limit, while the other one, naturally, being the lower sided limit. So, this inextricably means that there will be two calculations of the capability index. The one which yields the lowest value is the one we are going to compare it with the standard value of 1.67. As an extra information, process capability index is influenced by the way the process data are collected, that is, the process view. Process view can be defined in terms of the time frame, sampling method (sampling frequency, size...) used to obtain process data.

Industrial Training First Day (5-4-2010)

Today was a mentally exhausting beginning to my industrial training. Time seems not to budge. Luckily I was sufficiently pertinacious in keeping myself awake from 8.30am till 5.30pm.

Customarily, the first day would be designated for introductions. I was first introduced to the organizational chart by my supervisor, Mr. Melvin. He is an amicable and helpful person. His concise, animated and lucid explanations not only alleviated my anxiety, but also had imbued me with fundamental information about the operation of NXP Semiconductor.

Basically, this company is a breakaway department from Philips which specializes in semiconductor manufacturing. Some of the terms mentioned were totally unfamiliar as those were not covered in any of the lectures in UNITEN before. For instance, wafer fab, die bonding, wire bonding and so forth.

As an brief overview, this company industrial processes can be divided into three categories, namely front end, middle end and back end. In the front end process, silicon ingots of exceedingly low impurities will undergo sawing process, culminating in disc-like wafer fabs. Each piece of this wafer usually impregnates thousands of individual units, deemed as dies. Each die will then undergo die bonding process, to be followed by wire bonding process.

Then, in the mid end section, molding and plating of the dies on lead frames will be conducted. While in the back end section, processes involved are trimming, testing, marking, forming, taping, to be followed by prepacking, quality assuring and finally shipping or delivery of end products.

Apart from the aforementioned introduction narrated by Mr. Melvin, he also took me for several horizon-broadening visits to their assembly lines in which I was inundated with awe as the sight of sophisticated, impeccable machines appeared in front of me. Every employee that enters this assembly line, which composed of front end, mid end and back end, is required to wear special attire. This is to mitigate the baneful effects of our body's electrostatic charges on the semiconductors.

Besides that, I was asked to look up the internet as well as the company's intranet to gain knowledge on semiconductor manufacturing involving wafer fabs. I was also introduced to miscellaneous probable defects of the semiconductor. For example, exposed wires, contaminated dies, cracked wafers, pimples, scratched wafers and so forth.

That's all from me. I hope there could be more breaks in between the working hours as it was really mentally tormenting for me to stare at the computer whole day long. For your information, the lunch break is the only break we are having, and it is only just a paltry 45 minutes.