LEAN THINKING IN THE CLEAN
ROOM AT VDL ETG ALMELO
By
RICHARD VEENENDAAL
University of Groningen
ETG
LEAN THINKING IN THE CLEAN
ROOM AT VDL ETG ALMELO
Author
R. Veenendaal BSc
Company supervisor
drs. ing. R. Couwenbergh
Supervisor
dr. ir. I. ten Have MBA
Co-assessor
SUMMARY
This research addresses the introduction of lean thinking at VDL ETG Almelo, a contract manufacturer for mechatronic systems and supply. The economic downturn emphasized the focus on improvement of the organization. The wafer stage, one of the products manufactured in Almelo, which has the biggest impact on turnover, is the object of this research. By investigating this production process with value stream mapping and benchmarking similar companies, experienced in lean thinking, recommendations are proposed to improve this process and therefore the organization.
By analyzing the production process with value stream mapping, a central tool in lean thinking, value and waste in the production process are highlighted and a good overview of the process is created. The assembly process, which consists of seven sub processes, has the biggest waste in the bearings, y beams and frame workstations. Main types of waste seen are waiting and superfluous motion. Compared to the current situation of 326 clean room assembly hours, a reduction of 39 assembly hours can be realized. This is an improvement of almost twelve percent. At an expected customer demand of three products a week also three to four operators fewer are needed, who can then be employed elsewhere in the organization. Takt-based pull planning is an important aspect of the realization of this reduction in assembly hours.
Other elements of lean thinking which are advised to be incorporated, are multi moment recordings for realizing accurate assembly time measurements, a 2x2 matrix for an overview of improvement actions, an up-to-date skills matrix to identify which operators can perform which task, training to learn employees about lean thinking and to keep them motivated and zero mistakes for a focus on error-proof assembly.
Lean thinking should be based on a long-term focus and full management commitment. The fight for the reduction of waste, value stream mapping for insight in the processes and cross-functional teams with expertise from direct employees should be central points in this philosophy. The same holds for the NPI process.
As lean thinking is all about (motivated) people, change management should play a central role in the new way of working.
PREFACE
In front of you lies the report of my internship at Van der Leegte Enabling Technologies Group (VDL ETG) Almelo. This thesis is the final piece of my study Technology Management at the faculty of Management and Organization of the University of Groningen. The time span of the internship was August 2008 till December 2008.
VDL ETG is a worldwide supplier of advanced mechanical components, modules and complete systems. The research took place at the Systems department in Almelo and was focused on the clean room assembly of the wafer stage, a positioning module for a wafer stepper.
Lean thinking, as mentioned in the title, is the central philosophy in the thesis. It is a way of working to improve an organization in which basic thinking, waste elimination and value creation are central. Several concepts, mentioned in the thesis between quotation marks may sound unfamiliar. Therefore I would like to draw your attention to the glossary in this thesis.
This thesis could not have been realized without the help of a large group of people. Therefore I would like to thank Bas Westerbos of 4People, for introducing me at VDL ETG Almelo, Richard Couwenbergh, my supervisor at VDL ETG for his patience and being available all the time for questions, Renze Keuning, colleague at VDL ETG, for accompanying me during most of my company visits, and the other colleagues at the Systems department, for their time and attention.
Furthermore I would like to thank Ingrid ten Have and Jan Riezebos, the University supervisors for their time, useful tips and extensive feedback. Finally I would like to thank David Jennings and Merlijn Kooiman for critically reading my thesis and my parents for their unconditional support and encouragement.
Richard Veenendaal
TABLE OF CONTENTS
1 Introduction 11
1.1 Company 11
1.2 Product 12
1.3 What is lean thinking? 14
1.4 Relevance of research 14 1.5 Current context 15 1.6 Report structure 16 2 Research model 17 2.1 Problem statement 17 2.2 Goal 18 2.3 Conceptual model 19 2.4 Research question 19 2.5 Research objects 20 2.6 Definitions 20 2.7 Scope 21 2.8 Research model 21 2.9 Research strategy 22 2.10 Sources 24 2.11 Planning 25 3 Theoretical framework 27
3.1 Brief introduction of lean thinking 27
3.2 Methodology 28
3.3 Lean elements at VDL ETG Almelo 29
3.4 Change management 33
4 Analysis of the production process 35
4.1 The current production process examined 35
4.2 Waste and its location 38
4.3 Lead-time improvement by eliminating waste 40
4.4 Takt-based planning and its necessary input 42
4.5 Lessons learned for the NPI protocol 45
5 Lean thinking outside of VDL ETG Almelo 47
5.1 Company selection 47
5.2 Miedema 48
5.3 Nido 50
5.4 Eaton Holec 52
5.5 Nefit 53
5.6 Applicability at VDL ETG Almelo 55
6 Conclusion and recommendations 59
6.1 The research question answered 59
6.2 Limits of the research 61
6.3 Suggestions for further research 62
Appendices 67 Appendix A Production parts list of the wafer stage XT4i 68
Appendix B Map of the clean room 69
Appendix C Current state map 70
Appendix D Bottlenecks categorized according to waste type 71
Appendix E Matrix with improvement ideas 74
Appendix F Future state map 75
1 INTRODUCTION
1.1 Company
VDL Group Van der Leegte Enabling Technologies Group (‘VDL ETG’) is
part of the VDL Group, an internationally operating industrial organization, specialized in development, production and sales of semi-finished and final products. The organization started in 1953 with Metaalindustrie en Constructiewerkplaats P.
van der Leegte and has since then grown to 77 subsidiaries in 14 countries all over the
world employing around 7200 people. Through multiple acquisitions the group has become what it is today (Figure 1.1). From the nineties onwards VDL is also active in the bus and coach sector (VDL, 2009).
Figure 1.1 Organization chart VDL1
1 The Systems department of VDL ETG Almelo is the department, which is the subject of this research.
VDL Group
Supplies Buses & Coaches Finished products
Metalworking
Mechatronic systems & system supply
VDL ETG VDL ETG is part of the VDL Group since the end of 2006. VDL
ETG itself is much older than this as it was formerly known as Philips
Machinefabrieken. VDL ETG has four production locations: Eindhoven, Almelo,
Suzhou (China) and Singapore. The headquarters of VDL ETG is situated in Eindhoven (VDL Enabling Technologies Group, 2009). It employs around 1100 people of which about a fourth can be found in Almelo.
VDL ETG Almelo produces complex ‘mechatronic’ modules. The product portfolio consists of semiconductors, solar energy and medical appliances. VDL ETG acts as a contract manufacturer for well-known companies such as ASML, Thales, Oerlikon and PANalytical. The production can be categorized as low volume, high tech. Production volumes per product vary from a single product to several hundreds a year.
The organization in Almelo consists, next to the regular staff functions, of three departments. These are Parts, Projects and Systems. Parts produces components for the external market as well as for Projects and Systems. In the Projects department prototypes and small one-off series are made.
Systems department When products have reached the release for volume
milestone they are transferred to the Systems department. The main product is the ‘wafer stage’ (see section 1.2). Several generations of wafer stages already have been assembled at the department. Next to modules also spares and repairs of the wafer stage are assembled. In the past also the reticle stage (another part of the ‘wafer stepper’) was assembled at the Systems department. All this assembly takes place in a U.S. Federal Standard 209e class M5.5 or 10.000 clean room. This means that a maximum of 10.000 particles of 0,5 micron per cubic foot of air are allowable in this clean room (Institute of Environmental Sciences, 1992). This class is also known as ISO class 7 in the ISO-14644-1 classification of air cleanliness (Cleanroom international standard, 2009).
1.2 Product
position underneath a lens. Because the lens position is fixed only a portion of the wafer can be exposed to light. This is why the wafer must be able to move. Because there are two silicon disks the process can go on continuously. As one disk is being removed, the other disk is positioned underneath the lens. A single wafer stage measures about two and a half by two and a half meters with a height of about half a meter. Several models are produced in the Systems department. These are the XT3, XT3i, XT4 and XT4i. The “i” means immersion, which means that the air between lens and object is replaced by water. The XT4i, which is the most recent module, currently is the module with the biggest demand (more details on the wafer stage and its production process are found in Chapter 4).
Figure 1.2 Schematic view of XT4 wafer stage
Figure 1.3 Wafer stepper
X beam
Operating environment Because computer chips are very sensitive products,
distorting dust is unwanted in the production. Manufacturing takes place in a clean room to make sure that the wafer stages are contaminated as little as possible. The immersion technology makes that checking the system is even more important. A leak in the water supply system in combination with the advanced electronics can lead to disturbances and that is not an option in such expensive equipment. To realize the highest possible output it is important that the wafer steppers run with no downtime. A high-speed, nanometre precision, clean defect-free and continuous production process of chips is the goal for the wafer steppers.
1.3 What is lean thinking?
‘Lean thinking’ is the central philosophy in the research. Womack and Jones (1996) stated the principles in a five-step thought process to guide managers through a lean transformation. The five principles are: Specify ‘value’, identify the ‘value stream’, create ‘flow’, create ‘pull’ and ‘perfection’. The last principle means that the process is an iterative circle. It is a way of working to improve an organization in which basic thinking, ‘waste’ elimination and value creation are central. More details on lean thinking can be found in Chapter 3.
1.4 Relevance of research
Due to a stagnating economy cost reduction is on top of the agenda. The management of VDL ETG has indicated that a lot of waste is present in the production process of the wafer stage. The elimination of waste and a more efficient method of manufacturing should reduce costs and ‘lead-time’ considerably. The philosophy chosen or method for reaching this goal is lean thinking.
As ASML is one of the large accounts of VDL ETG Almelo it is reasonable to assume that their products are representative for the product portfolio of VDL ETG Almelo. This is why an ASML product, the wafer stage, is central in this research. ASML, the main customer of the Systems department, has asked VDL ETG to use the lean thinking philosophy for improving assembly, as ASML themselves has already adopted this philosophy.
implications for VDL ETG remain a bit unclear. More knowledge about lean thinking at VDL ETG is therefore welcome. Because of the complexity of a complete overnight reorganization, it has been decided that the Systems department will act as a pilot for the rest of the organization. The lessons learned on this journey are also planned to be used in other parts of the organization, and applied to other production processes.
A first look at the production process of the wafer stage has shown that room for improvement of the process seems possible. Semi-finished products were found waiting in the process line. A single wafer stage represents a large amount of stock, when its value is compared against annual turnover. Next to this visible indication, there also is a feeling that the process can be carried out in a more efficient and effective manner. Based on the current market situation, ASML can decide to make the wafer stages themselves, if they consider it is cheaper to manufacture this product in Veldhoven at their own factory. Placing the focus on the improvement of this process, by keeping it in-house in Almelo, is therefore important.
1.5 Current context
Due to a stagnating economy cost reduction is on top of the agenda. Because the semiconductor industry is very sensitive to fluctuations in the market, and VDL ETG is dependent on this market, a substantial part of the turnover in Almelo has disappeared over the last year. Measures must be taken to survive during the current downturn. The departure of temporary employees and those with a limited contract was unavoidable. A request for a reduction in working hours was put to the Dutch government in November. This request was granted in early December, which means that about two thirds of the organization in Almelo will be working part-time for the first couple of months. This situation can be extended to a period of up to six months.
1.6 Report structure
2 RESEARCH MODEL
In this chapter the problem statement, goal, conceptual model, research question and its sub questions, research objects, definitions used and scope will be presented. Furthermore the research strategy followed, necessary sources and research planning are addressed. The research model is based on the design of a research, as proposed by Verschuren and Doorewaard (2007).
2.1 Problem statement
VDL ETG Systems manufactures wafer stages for the ASML wafer steppers. Systems wants to assemble the wafer stages faster so the associated capital is locked shorter and demand can be followed more closely. Due to an economic downturn the focus on cost reduction has become an even higher priority. There is a possibility that ASML will attempt to reallocate the wafer stage assembly if they think another party can do it cheaper. Therefore it is important that VDL ETG assembles the product as cost-efficient as possible.
To improve the production process lean thinking is chosen as the method. This journey is called “Lean at VDL ETG, the next level” (Adriaans, 2008). Previous improvement programs such as Manufacturing Excellence and the ‘5S’ program
ALlemaal MEedoen LOont seem to be discontinued. Successes were achieved though.
This new initiative has been welcomed with mixed feelings. On the one hand people want to improve, but on the other hand they are afraid of yet another non-completed journey.
2.2 Goal
The aim of this research is to present recommendations to the management of VDL ETG, in order to improve the assembly process of the wafer stage.
Value stream mapping will be conducted to locate waste in the production process of the wafer stage. Input for realizing a ‘takt’ concept will be presented. And other companies will be visited to find out if VDL ETG is on the right track with lean thinking.
General applicability of the above-targeted recommendations for other and future products of VDL ETG Almelo is also in the interest of the management.
Is the goal useful and SMART2? The use of the recommendations in Chapter
6 will be that waste can be eliminated from the assembly process of the wafer stage. Less waste leads to less cost and less frustration. The waste identified can be quantified in terms of purchase value of the components and it is attainable to reduce waste, as it is available in considerable quantities. It is realistic to search for waste because it is present and needs to be removed. The sooner the waste can be eliminated, the better it will be as waste only acts as ballast.
A reduction of lead-time will be useful because this leads to more flexibility towards the customer and means that the associated production space and tools are more quickly available for other jobs. This is attainable with the current employees. Systems wants and needs to be ready for increased volume in the future. It is important that the operators become involved in this process, because they are the key to realizing lead-time reduction. It is necessary to pursue lead-time reduction because this has been requested by the customer. ASML management also pursues this themselves. Due to current low demand the pressure on the ‘direct employees’ to improve is low, so this factor has to be taken into account. The result is time-based as it is all about the time required. The urge seems less time-based due to the current low demand and the ability to keep up with this.
Input for realizing a takt concept is useful, as VDL ETG Almelo currently is not very familiar with takt-based production. Next to studying available theories, companies will be visited to find out what is essential in takt-based production. Inquiring at other organizations is attainable as contacts are present and the researcher
is welcome. The future execution of takt-based production is once again dependent on the direct employees. Clear communication and involvement are important elements. It is realistic to learn more about takt, as theory is widely available and clear planning is important for the realization of the aforementioned lead-time reduction.
In general the time-based component is present in the goal of the research, as the goal must be reached by the end of the research and therefore timely implementation is needed.
2.3 Conceptual model
The focus in the research is on the reduction of lead-time of the clean room production process of the wafer stage by implementing lean thinking. Eliminating waste will reduce lead-time. An increase in waste reduction therefore leads to a shorter lead-time. An increased use of takt-based planning leads to a reduction of lead-time too, with the exception that the lead-time must be longer than the time interval in which a product is demanded. When the lead-time, which follows from the current feasible lead-time in ‘Baan’, is shorter than the time interval in which a product is demanded, takt-based planning increases lead-time up to the time interval in which a product is demanded. Takt-based planning therefore increases lead-time when customer demand is very low. The research question and its sub questions, research objects and scope follow from this model (Figure 2.1).
Figure 2.1 Conceptual model
2.4 Research question
How can lean thinking help to improve lead-time of the clean room production process of the wafer stage at VDL ETG Almelo?
2. Where is waste located in the lead-time of the production process of the wafer stage?
3. Where are improvements by eliminating waste possible, which will improve lead-time?
4. Which information is needed to create a takt concept for the production of the wafer stage?
5. What can we learn from the production process of the wafer stage to improve the new product introduction protocol3?
6. Which elements of lean thinking are essential according to selected other organizations?
7. Which of the above mentioned lean thinking elements are of use at the Systems department of VDL ETG Almelo?
8. What is the general applicability of the proposed improvements for the rest of the organization?
2.5 Research objects
The research objects are waste in the clean room production process of the wafer stage, the lead-time associated with this process and takt-based planning of this production process.
2.6 Definitions
To avoid errors in communication, central objects in this research are defined and should be clear to everyone. ‘Waste’ in this research is based on the well-known seven wastes of Toyota (Hines & Rich, 1997), but the specific sorts of waste are slightly different. The definition used in this research is proposed by Adriaans (2008), waste consists of the following elements: ‘Defects’, ‘waiting’, ‘motion’, ‘rework’, ‘inventory’, ‘control’, ‘over-processing’ and ‘unused knowledge’ (Figure 3.1).
‘Lead-time’ generally means the time between the initiation and completion of a production process. Combined with the constraints, mentioned in the next section, this means that lead-time in this research points to the time interval from the moment the material for a wafer stage XT4i enters the clean room until a finished wafer stage XT4i leaves this same clean room at VDL ETG Almelo.
3 The NPI protocol refers to the release for volume (R4V) checklist, which is a required document for
‘Takt’ means the pace at which production should operate. It is the available production time divided by customer demand per time unit (for example day) (Marchwinski & Shook, 2006). Takt-based planning means that cycle times of the various assembly steps must fit within the takt time.
Background on the above-mentioned definitions is available in Chapter 3. Terms between quotation marks can be found in the glossary.
2.7 Scope
Only the XT4i wafer stage is taken into account. Other products made at VDL ETG Almelo are not part of the research.
Only the clean room assembly is taken into account. Storage of components before they enter the clean room is not taken into account.
Storage of finished products outside of the clean room, before shipment, is not taken into account.
2.8 Research model
Figure 2.2 Research model
Chapter 3 Chapter 4 Chapter 5 Chapter 6 Theory about lean concepts Suggestions for improvement: Insight in location and possibilities for the elimination of waste
The goal of this research is to make recommendations about waste reduction, lead-time improvement, takt-based planning, the ‘NPI’ protocol and general applicability of lean thinking at VDL ETG. Based on literature research on lean thinking (Chapter 3), the production process will be described (Chapter 4, sub question 1). Waste in the production process will be located with the help of a ‘current state map’ (Chapter 4, sub question 2), and possibilities for eliminating waste will be pointed out with the help of the comparison of current and ‘future state map’ (Chapter 4, sub question 3). Input for a takt-based planning concept will be proposed with the help of a future state map (Chapter 4, sub question 4) and lessons learned from the production process for the NPI protocol will be proposed (Chapter 4, sub question 5). Several organizations similar to VDL ETG Almelo, who already apply lean thinking, will be visited to learn their view on essential lean thinking elements (Chapter 5, sub question 6). Lean elements that can be of use to VDL ETG Almelo and the way lean was introduced there are highlighted (Chapter 5, sub question 7). The general applicability of the answers on sub questions one to seven will be made clear in the concluding chapter (Chapter 6, sub question 8). The research methodology based on the theory acts as the roadmap for Chapter 4. The benchmark from Chapter 5 is used as input for the future state map and therefore the companies visited already are referred to in chapters preceding Chapter 5. The answers on the sub questions together lead to the answer of the main question and thus on reaching the goal. The suggestions for improvement are stated in Chapter 6. A schematic representation of this structure can be found in Figure 2.2.
2.9 Research strategy
The research will take place in Almelo at the Systems department of VDL ETG. As the research will be done on-site, it will focus on the practice. Attention will be paid to the design of elements of lean thinking, which can be of use to VDL ETG Almelo. The production process will be analyzed and other organizations experienced in lean thinking will be visited as a benchmark.
recommendations can be characterized as more qualitative than quantitative and the research itself has an empirical nature. The focus on depth and qualitative nature of the recommendations makes the survey and experiment less logical as a strategy. The empirical approach makes the grounded theory approach and desk research less applicable.
Approach The small number of research objects makes it possible to go more
in-depth. To identify waste, lead-time improvements and input for takt-based planning all the employees working on the wafer stage in the clean room will be interviewed. The environment also will be checked on visual waste. Pre-calculated cycle times and lead-time are available in the Baan enterprise resource planning (ERP) system, which is accessible during the research. ‘Indirect employees’ such as the planner, quality engineers and factory engineers also are available to answer questions, which may arise during the research. Next to interview-based information gathered, observation of the production process, weekly staff feedback meetings and ‘lean steering committee’ meetings will provide input for the recommendations. To prevent an inward tunnel vision on the research objects at hand a benchmark will be executed. Selected other organizations will be visited to check how they fit lean thinking in their production process. These organizations are picked in a way that they resemble VDL ETG Almelo as much as possible in terms of production volume, product variety and product type. The focus of the research stays on VDL ETG Almelo, in the sense that the information gathered at the organizations visited will be used for comparison to VDL ETG Almelo. This approach means that several sources are tapped for information and that several techniques such as observation, interviewing and studying literature and operating datawill be used. Source and method triangulation is thus insured.
Advantages and disadvantages of case study research Case study research
narrower approach of case study research compared to for example surveys can lead to a smaller external validity, which can be seen as a disadvantage. The general applicability of the recommendations can be small, but that is not too big a problem as the goal is to do recommendations to the management of VDL ETG.
2.10 Sources
Verschuren and Doorewaard (2007: 215) state that research objects consist of persons, situations, objects and processes. Information sources consist of persons, media, reality, documents and literature. Persons can be interviewed and observed, media can be analyzed and observed, reality can be observed, measured and analyzed, documents can be analyzed and studied and literature can be analyzed and studied also. In Table 2.1 the sources needed for each sub question are stated together showing how they will be disclosed.
Sub question
Source Sort Disclosure
Direct employees (20) Face-to-face interview Persons
Indirect employees (12) Face-to-face interview Reality Production process XT4i Observation
1
Documents Technical product documentation Content analysis Direct employees (20) Face-to-face interview Persons
Indirect employees (12) Face-to-face interview Reality Production process XT4i Observation
Documents ‘TPD’4 XT4i Content analysis
Adriaans (2008) Content analysis Ballé and Ballé (2005) Content analysis Hines, Holweg & Rich (2004) Content analysis Hines and Rich (1997) Content analysis 2
Literature
Womack and Jones (1996) Content analysis Direct employees (20) Face-to-face interview Persons
Indirect employees (12) Face-to-face interview Reality Production process XT4i Analysis
TPD XT4i Content analysis 3
Documents
SCM quick scan XT4i5 Content analysis Lean steering committee (8) Observation Persons
Hosts at visited companies (4) Face-to-face interview Production process XT4i Observation
Reality
Production process other companies (4) Observation Demand forecast from ASML Content analysis TPD XT4i Content analysis SCM quick scan XT4i Content analysis Documents
Value stream map XT4i Content analysis Adriaans (2008) Content analysis Ballé and Ballé (2005) Content analysis Hines, Holweg & Rich (2004) Content analysis 4
Literature
Hines and Rich (1997) Content analysis
4 Technical product documentation
Rother and Shook (2003) Content analysis Womack and Jones (1996) Content analysis NPI coordinator Face-to-face interview Project leader NXT Face-to-face interview Persons
Lean steering committee (8) Face-to-face interview Reality Production process XT4i Observation
5
Documents NPI protocol Content analysis Persons Hosts at visited companies (4) Face-to-face interview Media Websites visited companies (4) Analysis
Reality Production process visited companies Observation Documents Lean principles other companies Content analysis 6
Literature Marchwinski and Shook (2006) Content analysis 7 Persons Lean steering committee (8) Face-to-face interview
Reality Production process XT4i Observation
8 Persons Manager Systems department Face-to-face interview Reality VDL ETG Almelo Observation
Documents Research report questions 1-7 Content analysis
Table 2.1 Sources by sub question
The relevant material for the research comes from a respectable number of sources. The use of the several sources and the time it takes to retrieve information from these sources must be traded off. In Table 2.1 the sources needed for the sub questions are accompanied with their type of disclosure. The source list may appear to be very long, but it must be noted that several sources will be used for multiple sub questions, so the number of unique sources is smaller than the list indicates and therefore acceptable. The face-to-face interview is mentioned several times, but a distinction can be made here in the degree of structure in which the interview takes place. Interviews vary from semi-structured with a list of subjects to informal conversations. The numbers in the sort-column present the number of people that are interviewed or the number of websites visited.
Next to the sub questions more information is needed. The websites from VDL and VDL ETG for example are used too, but are not mentioned in the list. The same holds for the theoretical background, several authors are pointed out in Table 1.1, but more will be referred to in the report.
2.11 Planning
Step Description Nominal duration (weeks)
Lead-time (weeks) 1 Getting familiar with the organization 2 16 2 Literature research 2 6 3 Describe production process 1 4 4 Read technical product documentation 1 4 5 Gather needed ERP data 1 8 6 Waste walk through clean room 1 1 7 Interviews at VDL ETG 2 16 8 Company visits (4) 2 12 9 Description future state production process 1 2 10 Create takt concept 1 4 11 Compare concepts 1 2 12 Describe general applicability of recommendations 1 4
Total duration (weeks) 16
Table 2.2 Action plan
3 THEORETICAL FRAMEWORK
In this chapter lean thinking and its origins are explained further. The five steps of the lean thinking philosophy, value, value stream, flow, pull and perfection, (Adriaans, 2008; Ballé & Ballé, 2005; Womack & Jones, 1996) are used as the methodology and roadmap for this research. To implement lean thinking good change management is a critical success factor (Bhasin & Burcher, 2006; Vavra, 2008; Worley & Doolen, 2006). Change management therefore also is addressed in this chapter.
3.1 Brief introduction of lean thinking
The origins of lean thinking can be found at Toyota. Sixty years ago they started with a program to improve production by basic thinking rather than automation. Concepts that were central in this program, better known as the Toyota Production System (TPS), are ‘kanban’, pull, just-in-time, respect for people and waste elimination (Hines, Holweg & Rich, 2004). First the efforts were aimed at manufacturing because concepts can be made visible relatively easy. As the theory was used for other parts of organizations too, lean thinking became the more general name for the method or philosophy. Depending on how you look at lean it is more a method or a philosophy. When seen as a toolbox it is a method, when looked at a higher level it is more a philosophy. At a higher level it is not only about things as kanban or waste, but more as Ballé and Ballé (2005) say: “It’s all about people.”. The change management component (section 3.4) and the eighth waste of unused knowledge (Figure 3.1) also point in this way.
3.2 Methodology
Womack and Jones (1996) described lean thinking in the following five steps: Specify value from the perspective of the customer, analyze the value stream, create flow, create pull and strive for perfection.
Value It all begins with value, as the purpose of an organization is to make
profit. To make profit in a value chain, it is necessary to add value with as little waste as possible. Value is what the customer, not the producer, defines. In lean thinking this value concept has to become a central concept in the organization.
Value stream mapping Value stream mapping is a tool to see where value is
added in the production process and to make clear where the biggest forms of waste occur (a description of waste can be found in section 3.3). It is a system to map material and information flows. The flow is followed from the customer to the supplier (Marchwinski & Shook, 2006), as the customer is the one who eventually pays for the product. The advantage of a value stream is that waste is seen in connection to its environment and therefore sub-optimization can be avoided. Value stream maps can be made of any subset of the value stream. Value stream mapping (‘VSM’) is oriented fundamentally to productivity rather than to quality, which does not mean that quality is not an issue. Quality just is not the focal point. The reason for this is that improved productivity leads to leaner operations, which help to expose further waste and quality problems in thesystem. Therefore the systematic attack on waste also is a systematic assault on the factors underlying poor quality and fundamental management problems (Hines & Rich, 1997).
Following the value stream is more straightforward than it sounds. The current state map describes the production process as it currently is. After the current state map has been made, an ideal state is thought off. This ideal state usually is unattainable, but it gives the direction in which improvements should go, just as with Homer’s Ithaka (Cavafy, 2009). On this route from the current state to an ideal state a short-term future state map is plotted. This future state map describes the production process, as it should look in a year from the moment the current state map was made. Between current state and ideal state more future states can be plotted. A distinction can be made between future states realized next year and future states in five or ten years from the current state on.
Flow and takt To make it possible that the products are made at a constant
created. 5S also is a base for flow as inhibitors to flow are removed (Wood, 2004). When flow is reached, inventories on the floor are reduced drastically and lead-time will therefore be shorter.
To link the various steps in production to each other planning can help. Takt-based planning links production to customer demand. Opposed to the traditional production rate stands takt time. Takt means the rhythm at which production should operate. It is the available production time divided by customer demand per time unit (for example day) (Marchwinski & Shook, 2006). The several assembly steps cycle times must fit within the takt time and therefore line balancing is an important issue when takt time is implemented. Takt-based planning theoretically can lead to a longer lead-time with low demand as takt follows current demand. If this is the case reducing work time can be an alternative.
Pull The fourth lean thinking principle. An upstream workstation can only
send his output to the next station when this downstream workstation is ready to receive a product. Takt-based planning is very useful for realizing pull production. To implement pull production the mindset, especially of the planners, must be changed first. As this is the fourth lean thinking principle it becomes more important later on in the lean journey.
Perfection The fifth lean thinking principle. Perfection in a process is reached
when it provides pure value, as defined by the customer, with no waste of any sort (Marchwinski & Shook, 2006). When this fifth principle is approached for the first time, much improvement has been realized already. Perfection is like reaching the ideal state, it is not attainable, but it should be the goal. This concept emphasizes the idea that lean thinking is about the journey, rather than the destination. Continuous improvement (see section 3.3) is the key effort, which will keep the attention towards lean thinking. After some big improvement steps at first, smaller steps are taken later on.
3.3 Lean elements at VDL ETG Almelo
Waste The activities can be categorized into three categories. These are
non-value adding (NVA), necessary but non-non-value adding (NNVA) and non-value adding (VA) activities (Hines & Rich, 1997). The first is waste that should be removed immediately and the second category usually is waste too, but removing this is more difficult. Examples are: long walking distance to get parts, packing and unpacking and transport to a different physical location. Waste is not the same for every organization. Lean thinking at a car manufacturer focuses on different sorts of waste than lean thinking in an office. Compare for example machines, which are not fit to their job at a car manufacturer, to the meeting culture (vergadercultuur) in an office.
In the Toyota Production System (TPS) there are 7 wastes (Figure 3.1). These are: Overproduction, waiting, transport, inappropriate processing, unnecessary inventory, unnecessary motion and defects (Hines & Rich, 1997). At VDL ETG, Adriaans (2008) introduced the following list: Defects, waiting, motion, rework, inventory, control, over-processing and unused knowledge. What is different here is that rework, control and unused knowledge appear here, while overproduction and transport seem to have disappeared. This is a difference based on definitions only, as transport is incorporated in motion and overproduction is seen as over-processing. An important notion is that Adriaans (2008) added an eighth waste: unused knowledge, which means that the available knowledge in a company or value chain is not put to optimal use. It seems that it is not present in the TPS as they took this for granted, but it is important to note because people power is the key asset (Vavra, 2008).
Figure 3.1 Waste according to Hines & Rich (1997) and Adriaans (2008)
Continuous improvement The key to get the most added value from a system
is to eliminate waste and therefore to improve the system. There are two ways to improve: kaikaku, improvement via breakthrough events and ‘kaizen’, continuous improvement in smaller steps (Hines, Holweg & Rich, 2004). The advantage of the last is that through its continuous character a culture is created and that the cost attached to it usually is smaller. Continuous improvement appears under several names: kaizen (Hines, Holweg & Rich, 2004) rapid improvement events (RIE) (Papadopoulos & Merali, 2008) and small group activities (‘SGA’) (Adriaans, 2008) but although the name is different, the meaning is the same. SGA consists of eight steps based on the plan-do-check-act circle of Deming (Figure 3.2). The difference is that more focus lies on the plan phase, which is split in five steps. The steps are: Choose subject (onderwerp kiezen) (P1), fix target (doel vaststellen) (P2), investigate problem (probleem onderzoeken) (P3), find solutions (oplossingen bedenken) (P4), make implementation plan (invoeringsplan maken) (P5), execute implementation plan (invoeringsplan uitvoeren) (D6), measure effects (effecten meten) (C7) and standardize method (werkwijze standaardiseren) (A8).
Figure 3.2 Small group activity circle (Adriaans, 2008)
and therefore focus on continuity, also is a critical success factor. Small group activities are organized to remove waste from the production process. By removing waste, such as inventory, from the production process new problems usually appear, which can be removed subsequently. At Toyota Taiichi Ohno used the analogy of manufacturing being like a river full of rocks (Figure 3.3). The water level in the river represents inventory and the rocks represent manufacturing waste. The aim of the Toyota Production System, says Ohno, is to continually drop the water level and expose the rocks, so they can be eliminated. The rocks of poor quality, long set up times, poor factory layout and so forth all must be eliminated to turn the factory into a smooth, rapidly flowing stream of production (Waddell, 2005).
Figure 3.3 River full of rocks analogy
5S A method to organize the workplace is 5S. 5S stands for the Japanese
words: Seiri, Seiton, Seiso, Seiketsu, Shitsuke (Marchwinski & Shook, 2006). The English translation is somewhat ambiguous as several authors translate it differently. Marchwinski and Shook (2006) respectively talk about sifting, sorting, sweeping clean, spic and span and sustaining, whereas Becker (2001) names sort, systemize, sweep, standardize and self-discipline and Slomp (2008) replaces systemize with straighten. Although the translations differ, the goal is the same. The work floor is organized so that waste will be minimized.
A remark on usability of lean elements Every company has to find its own
way to implement the lean method: there is no universal way that will apply to all (Taj, 2005). This is especially true for VDL ETG as the original principles were developed for the high volume industry. The big difference of the production volume is that changes in demand have a relatively bigger effect in the low volume sector in which VDL ETG is active. Wallace and Sackett (1996) and Bellgran and Aresu (2003) state that the low volume characteristic already must be incorporated in design. Design for assembly and component modularity therefore is useful. Furthermore Jina, Bhattacharya and Walton (1997) advise the use of a product change coordinator. The low volume character of the production at VDL ETG possibly makes common lean elements such as ‘supermarkets’, ‘milk runs’ and ‘pick to light’ systems less applicable.
3.4 Change management
Lean thinking is all about people as Ballé and Ballé (2005) already mentioned. Therefore it is important not just to use the lean tools such as value stream mapping, small group activities and takt-based planning but also to look at the people side of the organization. Teamwork and discipline are important, but so is communication. As lean thinking is about change, Vavra (2008) emphasizes the role of change management. Management must be fully committed to the lean thinking philosophy and be aware of the fact that their role changes too. ‘Cross-functional’ teams in small group activities for example focus on empowerment.
4 ANALYSIS OF THE PRODUCTION PROCESS
In this chapter the production process of the wafer stage will be analyzed by answering the first five sub questions. First the assembly process in the clean room will be described by means of value stream mapping. Then the current state value stream map of the production process will make clear where waste is located. The future state map is made to describe the production process with expected improvements and implemented takt planning. The current and future state map will be compared to each other to calculate the expected lead-time improvement. The next subject is takt-based planning and what is needed for its implementation. Finally lessons learned from the production process come to light to recommend improvements for the NPI protocol.
4.1 The current production process examined
The wafer stage (Figure 1.2 and 4.1) is a module for the wafer stepper that is built in various versions. Although the different models vary in detail, the production process is still roughly the same. The type investigated is the XT4i. The wafer stage is a module that measures about two and a half by two and a half metres with a height of about half a metre. A module consists of several sub assemblies. These are the pre assembly, frame, y beams, x beams, cable slabs or interconnection flexes and bearings. The sub assemblies are constructed at specific workstations. Some of the sub assemblies can be assembled concurrently (Figure 4.2).
Figure 4.2 Production process of the wafer stage
The frame is the base of the module. In the first workstation the frame is put on a trestle and after the base is cleaned wiring and tubing is mounted. Concurrently the supply cabinets are prepared for assembly on the frame. These supply cabinets are similar to drawers placed on both sides of the frame. The contents are connected to the tubing and wiring on the frame and these connections are tested. At the same time there are parallel workstations for the assembly of the interconnection flexes, x beams, y beams, pre assembly and bearings. The interconnection flexes connect the chucks, plateau on which the silicon plate is placed, to the x beams and the x beams to the y beams. Bearings are prepared at one workstation, but are needed at the x beam workstation, y beam workstation and build together workstation. The pre assembly is a workstation in which small sub assemblies are made for several other stations. These are made at one place because they do not fit in the available time for the other stations. The x beams consist of several elements, which need to be glued together. It is important to note that the glue has a drying time of five days. The y beams are fitted with three engines each. These are the N (not connected) engine, the C (connected) engine and the cable shuttle engine. Each engine is independently linked to the y beam with bearings. In the build together workstation the stone is mounted in the frame as a counterweight. The y beams are placed, the interconnection flexes are fitted and the x beams are mounted. Also some of the sub assemblies from the pre assembly workstation are fitted here. When the module is complete it i.s first adjusted before being measured and packaged.
The production process in Baan The XT4i6 is the module that is checked in detail. As mentioned earlier, this currently is the most ordered type. Globally speaking the assembly consists of, bearings, pre assembly, frame, y beams, x beams, interconnection flex and building together and packaging. These processes consist of subassemblies again. From the production parts list structure (Appendix A) the combined assembly hours per process can be deducted (Table 4.1).
Process Hours Bearings 25 Pre assembly 15 Frame 74 Y beams 50 X beams 52 Interconnection flex 28 Build together & packaging 82 Total 326
Table 4.1 Assembly hours per process
Date : 07-10-08 [11:48] ROUTING (CONDENSED) VDL ETG Almelo BV
--- Job Task Description Dpt Mach. Setup Piece
time time --- Make item : 4022 639 03861 Xt4 waferstage base 1 piece
Routing : 10 1 t/m 1 Standard routing: No
10/ 1 8996 Material issue. AG ASML 699 MAFG-6 288 0 15/ 1 5996 IND Waiting time WF Systems 996 240 0 30/ 1 8186 Making Products dustproof 600 STWA 0 96 40/ 1 9066 Clean room mech. assembly 610 S1MO 0 351 45/ 1 9246 Mechanical classifying 610 MCON-6 0 1 50/ 1 5996 IND Waiting time WF Systems 996 120 0 55/ 1 9066 Clean room mech. assembly 610 S1MO 0 157 60/ 1 5996 IND Waiting time WF Systems 996 60 0 70/ 1 5996 IND Waiting time WF Systems 996 60 0
Table 4.2 Example of routing in Baan for item 4022 639 03861
The hours needed for assembly can be obtained from the Baan ERP system. For each twelve-digit article number the routing can be printed. Each routing (see Table 4.2) consists of several tasks, which are numbered in ascending order. Only the
6 The XT3, XT3i, XT4 and XT4i are different modules, but on average the hours that are needed for
tasks performed in the clean room (department 610) are taken into account. Picking material, washing the parts and making the parts dust free are beyond the scope of this research. In this case the assembly hours for the wafer stage base are 351 + 1 + 157 minutes = 509 minutes or almost eight and a half hours. These assembly hours are about half of the hours needed for the pre assembly process (Table 4.1, Appendix A). The total number of clean room assembly hours needed for a wafer stage is 326 hours. Assembly takes place five days a week, seven and a half hours a day (eight hours minus two times fifteen minutes coffee break). Currently the planned lead-time is fifteen days or three weeks.
Other activities in the clean room The assembly of the XT modules is not the
only job to be done in the clean room (see Appendix B for a map of the clean room). A corner of the clean room is reserved for the NXT module. This also is a wafer stage module, but the main difference with the XT modules is that the NXT is based on magnetic levitation technique. The positioning of the silicon disks is not longer done with linear engines but with the help of magnetic levitation. Apart from modules, spares and repairs for wafer stages are also made in the clean room.
Conclusion The assembly process consists of seven workstations. These are
frame, pre assembly, y beams, x beams, interconnection flexes, bearings and building together. A large part of the activities can be done concurrently. For a wafer stage 326 hours of clean room assembly are pre-calculated in Baan.
4.2 Waste and its location
stream map made at the end of September 2008 and the map made with the lean steering committee at the end of October 2008.
The current state map (Appendix C) consists of the various processes, which take place in the clean room. For each process it is made clear how long operators are working on it, how long it takes them to complete their part of the job, the number of assembly hours (based on historical results) that are available for each process in Baan, what the throughput time is according to Baan and how many hours the operators are available each day. Time studies were not conducted. Next to the processes and their push lines (use of push system is confirmed by planners) from one to the other, in-process inventory is marked. The information flow from the customer to VDL ETG and from VDL ETG to their suppliers is digital. Each process gets a weekly production planning based on the main production schedule.
The various bottlenecks found in the current state map, are categorized according to the types of waste, which have been defined in Chapter 3. Some of the afore mentioned bottlenecks appear at more than one place. If this is the case, only the first instance is listed to prevent double counting. Table 4.3 (the full list can be found in Appendix D) clearly shows that waste mostly appears in the form of waiting and superfluous motion. This does not directly imply that waiting and superfluous motion should have the biggest impact on lead-time. It is only a numerical representation of the issues sorted to forms of waste. Having mentioned this, it also must be noted that waiting has the most direct link towards waste in lead-time of the production process. Unnecessary waiting therefore should be eliminated. The available production time also should be used as effectively as possible, which means that time currently used for producing defects or over-processing, should be used for required production in a way that waiting will be minimized. Waiting is also an issue that pops up in the routings in Baan. For the sake of planning, dummy lead-times are added in these routings (see waiting time in Table 4.2). Although these dummy lead-times come in handy for predicting an expected completion date, they blur the actual time needed for assembly as operators try to complete their jobs in the time allotted by Baan.
Process Waste (issues7) D ef ec ts W ai tin g M ot io n R ew or k I nv en to ry C on tr ol O ve r-pr oc es si ng U nu se d kn ow le dg e To ta l Bearings 4 7 5 5 2 1 1 0 25 Pre assembly 1 5 2 1 0 0 0 1 10 Frame 1 3 3 1 1 2 2 1 14 Y beams 2 4 15 2 0 1 0 0 24 X beams 1 1 4 4 0 1 0 1 12 Interconnection flex 2 1 1 3 1 0 1 2 11 Building together and packaging 1 11 2 0 3 2 1 1 21 Total 12 32 32 16 7 7 5 6 117
Table 4.3 Bottlenecks categorized according to waste type
Conclusion Waste appears to be concentrated in the bearings, y beams and
frame processes. Furthermore the frame and building together are the processes, which take the most time to complete. These processes therefore seem to need initial attention for improvement. The main types of waste seen are waiting and motion. Waiting is an activity, which is added in the routings in Baan for the sake of planning. For clarity this dummy lead-time could be removed when more reliable data are available.
4.3 Lead-time improvement by eliminating waste
The link between lead-time and waste reduction can be made clear with the help of the value stream maps. By eliminating unneeded activities and restructuring the work, the hours needed to get the job done will be reduced. The current state map (Appendix C) is the base for a list of bottlenecks (Appendix D). These bottlenecks are translated in actions in the matrix for improvement actions (Adriaans, 2008; Appendix E). This ‘2x2 matrix’ (Figure 4.3) divides actions in easy and hard to implement improvements on the one hand and big and small improvements on the other. The big, easy to implement, improvements (Figure 4.3, cell 1) are also known as the quick wins or low hanging fruit. The improvements, which need more effort but are rewarding too (Figure 4.3, cell 2), are issues that can be solved with the help of, long duration efforts such as 5S and small group activities or SGAs. The improvements
S m al l B ig
that are small, but also do not require much effort are also worth implementing and are stated in cell 3 (Figure 4.3). Finally, the improvements that are small and hard to implement are not worth implementing at all or at least not at the short term (Figure 4.3, cell 4). The elimination of these bottlenecks paves the way for the new situation as stated in the future state map (Appendix F).
Easy to implement Hard to implement
Figure 4.3 2x2 matrix with improvement actions
In Table 4.4 the hours needed for assembly from different sources are compared. The processes mentioned are the same as in Table 4.1. The first data column shows the hours needed as planned in the ERP system. The second column shows the days needed for these processes. These numbers come from interviews with the direct employees (operators). The third data column is based on the data in the previous column multiplied by the number of employees needed for the process. The data in the final column for the current state map is based on interviews with direct and indirect employees. The hours needed for assembly in the future state map are estimations made by the direct and indirect employees based on expected improvements to be made in October 2009 compared to the current state map of October 2008. Baan (12/08) CSM (days) (09/08) CSM (09/08) CSM (10/08) FSM (10/09) Bearings 25 Unknown Unknown 19 11 Pre assembly 15 Unknown Unknown 17,50 16 Frame 74 3,50 105 106 85,25 Y beams 50 3,50 52,50 49,25 43,50 X beams 52 7 52,50 50 49 Interconnection flex 28 1,50 33,75 38 38 Building together 82 3,50 105 21,50 44,50 Total 326 19 348,75 301,25 287,25
Table 4.4 Hours needed for assembly compared
1 2
The bearings and pre assembly do not have an estimation of hours needed for assembly in the current state map of September 2008 because this data was not available at the time. The building together process in the current state map of October 2008 has a much lower number of hours needed for assembly because a part of this process is missing in the current state map of October 2008.
The hours in the current state map of September and October vary considerably; therefore an average is taken of these two columns of hours. The hours needed as planned in Baan are an average of the two current state maps of September and October 2008. For a rough estimate of the expected improvement in hours this number therefore are taken as a base. 326 – 287,25 = 38,75. According to Table 4.4 the implementation of the actions in Appendix E therefore leads to an expected reduction of 38,75 hours of assembly in the clean room. It is worthwhile noting, that compared to Baan, even though four of the seven processes should be done faster in the future, the other three are allotted more time to complete their job. The data in Baan should be updated to keep this system up to date.
In the previous section it became clear that the biggest wins are to be found in the reduction of waiting time and motion. It can also be noted that, numerically seen, some processes have more waste than others (Table 4.3). Reducing waste in the bearings, y beams and frame processes therefore appears to lead to the biggest wins. A lot of improvement opportunities arise from the current state map, but a plan has to be made to implement these expected improvements. The 2x2 matrix (Appendix E) is the answer to this question.
Conclusion By comparing Baan and the value stream maps, the expected
reduction of hours needed for assembly will be 38,75 hours. To establish this, improvement actions have to be executed as stated in Appendix E. Not all of the processes will be done faster in the future than currently stated in Baan, but the bottom line shows an improvement of almost twelve percent. As the data in Baan differ from reality considerably, the ERP system should be kept up to date regularly.
4.4 Takt-based planning and its necessary input
and this translates into production rate. The idea is that production follows actual instead of forecast demand. Planning therefore will be pull-based instead of push-based. With a demand of for example three products a week and 37,5 hours available for production in a week, this means that every 12,5 hour a product should exit the clean room. A big advantage of takt-based pull planning over push production is that products flow through the process and sub assemblies do not pile up halfway through the production process.
Production speed thus is directly linked to demand. Because people need to have a certain amount of clarity about when they are needed at their stations some smoothing is indispensable. A long term forecast is still necessary. To determine the required number of direct employees it is also necessary to know how much time each process takes. Next to the required number of assembly hours, drying time needs to be known as well. Not every employee can perform every task so the expertise of the employees is also an important input. The various assembly steps cycle times must fit within the takt time, therefore line balancing is an important issue when takt time is implemented. To keep the flow going, the products have to move from process to process in a steady rhythm. To realize this, it is necessary that the cycle times of the processes are in balance with each other.
When demand is stable, so is takt, but when demand fluctuates, takt does this too. To follow demand several takt-based plans should be available to follow. Although the employees are flexible, the demand fluctuations can only be followed to a certain level. The work to be done also requires certain stability in the hours available for it, so a balance needs to be found in following demand and working with a stable schedule. Nefit (Chapter 5) for example can change their takt time every four weeks with a limited set of possible settings.
time. At Nido (Chapter 5) they call these station times doorschuiftijd or move-up-time.
Table 4.5 Takt planning based on CSM Table 4.6 Takt planning based on FSM
Takt-based planning, an example In the current situation the assembly of a
takt time, which depends on the weekly demand. For the current state (Table 4.5) and future state (Table 4.6) the necessary number of direct employees is calculated for move rates from a half wafer stage per week to six wafer stages per week.
The expectation is that in the future state a move rate of three products per week will be common. This means that every 12,5 hour a wafer stage should leave the clean room. Theoretically 23 direct employees are needed. The processes take more than 12,5 hours, which means that more than one employee is needed per process. In a situation where the employees are only trained to work in one process, the number of employees needed rises to 26. In the situation that multi-skilled employees are present, at least 4 of the 23 employees needed should be cross-trained.
When takt-based planning is a fact, it will be clear how many products per week need to be assembled. The current layout, which is set up for a move rate of six wafer stages per week, can then be changed towards a layout for the average required wafer stages per week. The number of trestles and tooling can be reduced as a result of this.
Conclusion To produce wafer stages according to takt-based planning it is
necessary to know how much time is needed for assembly. Time studies or multi moment recordings therefore are useful. Takt-based planning also means that push-based production should be left behind in favor of pull-push-based production. Demand needs, at least in the long term, to be known. The number of available direct employees and especially cross-trained employees is also important. An up-to-date skills matrix (Baudin, 1996) can be useful here. Drying time of subassemblies is also a factor. To keep the planning reasonably stable the principles of a fixed takt for a couple of weeks from Nefit and move-up-time from Nido (Chapter 5) are useful.
4.5 Lessons learned for the NPI protocol
The focus on value and what the customer wants is not literally present in the current NPI protocol. Traditionally focus lies on cost and cost reduction, but the customer should be central because this actor actually is the reason that a company stays in business. Value stream mapping is a tool that promotes this value thinking. With a value stream map insight in the production process and how elements are related to each other, is gained.
In the release for volume document (De Bonth, 2008) there is a checkpoint for filling the ERP system. It is important to know how much time each process will take, but these expected cycle times must not solely be based on estimations of indirect employees. To cover input from the factory floor, direct employees should be present in the project team too.
To create flow, waste must be reduced. Awareness of waste is a crucial concept in value stream mapping. Creating slack or waiting hours in Baan therefore should be avoided because it can blur underlying problems (Goldratt, 2006). To handle waste reduction and to work towards an optimal production process a roadmap can be set up in the NPI process by means of a 2x2 matrix, as mentioned earlier on in this chapter (Figure 4.3). Another element of flow assembly is takt-based planning. This planning itself is an element, which should be added to the NPI protocol. An up-to-date skills matrix of direct employees is an element that can be very useful in creating a takt-based planning.
Conclusion The NPI protocol can be enriched with several elements to make
5 LEAN THINKING OUTSIDE OF VDL ETG ALMELO
In this chapter the sixth and seventh sub question are answered. The companies visited are presented as benchmarks for VDL ETG Almelo. These companies have been chosen to link up as much as possible with VDL ETG Almelo. The companies under investigation are: Miedema, Nido, Eaton Holec and Nefit. The goal of these company visits was to find out which lean thinking elements are essential according to these organizations. To be of use these findings are then judged for applicability for VDL ETG Almelo.
5.1 Company selection
VDL ETG is not a standard company, such as automobile manufacturers, for a lean implementation because of its production type. The products, which are made at VDL ETG, are contract manufactured at very low volumes and have a great variety. This comes down to a maximum of several hundreds of products of the same specification a year, but currently the runners are made at a speed of maximum one product per week. Next to the low volumes the products are high tech and fairly large. With this in mind low volume high tech companies were sought. The companies are chosen to identify the why and how of lean thinking, not to describe an average effect; hence cases often are not aimed at being representative, but rather exemplary (Stuart, McCutcheon, Handfield, McLachlin, & Samson, 2002).
The relatively low production volumes also can be found at Miedema and Nido, which produce several machines a day. Eaton Holec produces several dozens of units a day and Nefit even produces several hundreds of machines a day (Table 5.1). The companies visited all have a high product variety. Except for Nefit the products are mainly for the industrial market. Another characteristic of the wafer stage of VDL ETG is that demand depends heavily on the economic situation, as the computer chip industry is a cyclic market. This demand fluctuation can be compared with the seasonal fluctuations at Miedema and Nido opposed to stable demand.
