Teaching Chapter 13: Design for Manufacturing
Timing
Since design for manufacturing (DFM) concepts are useful even during the concept development phase, it would certainly be helpful to place this session near the beginning of a project course. However not everything can be placed up front, so we generally teach this session somewhere during the second half of the semester. This assures that students have at least struggled with DFM issues to some extent in their projects before the class discussion.
Objectives and Strategy
This session introduces the class to many concepts related to DFM. The intended message is that DFM is much more than cost reduction. It is a design strategy that requires the expertise of multiple team members and the use of basic design rules, guidelines, and cost models. DFM often results in significant cost reduction as well as improvement in product quality and the development of cross-functional expertise within the organization.
We have taught single sessions on DFM for many audiences, including in an MBA core course on operations management and in our more advanced graduate course on product development. We have also found this topic valuable for undergraduate courses and useful in executive education. The session outline below describes some of the material we have used in a variety of these contexts.
Session Outline
The session can follow this flow:
Introduction to DFM
The class session can begin by relating the hype about DFM over the last decade to significant gains in product quality, reductions in product cost, and improvement in the structure of development organizations. While DFM is not directly responsible for all of these effects, it can be closely linked to each of them as we will explore in this session.
Next, a student can be asked to summarize the main points of DFM, which usually brings up three or four of the items in the next section.
DFM Principles and Caveats
There are several points which should be brought out in discussion:
Cross-Functional Teams: This practice is particularly useful in DFM since a wide range of skills and perspectives is required for DFM success.
DFM Starts Early: Even during the concept development phase, manufacturing concerns can be used to guide the generation and selection of alternatives. There are many examples of DFM success where novel concepts were required to achieve significant improvements (e.g., Digital’s desktop mouse).
Cost Modeling: To guide the efforts and measure the benefits of DFM, it is useful to have cost models capable of accurately predicting the eventual production costs for any product. Ideally, a design team would be able to push a button on their CAD terminal and receive an accurate production cost for their proposed design. Unfortunately such CAD-to-cost systems do not exist – except for a few, very specialized, applications.
Cost Structure: It is useful to discuss alternative ways to break down costs, such as: 1) direct-indirect, 2) materials-labor-overhead, or 3) components-assembly-overhead. For DFM, we seek a cost structure which gives the most useful information for making the necessary decisions. It can be useful at this point to spend a few minutes discussing activity-based costing and the meaning of overhead costs.
Focusing on Assembly: Students will probably be able to list several reasons why it is useful to focus attention on the simplification of assembly operations. There are several "multiplier effects" to discuss, such as the multiple ways that reducing parts count reduces overhead costs. (Exercise 13.3 discusses some of these effects.)
Design Rules: The class can list a set of basic DFM guidelines (known as DFM rules of thumb) to follow, as many of these are described in the chapter. The discussion should include how to handle cases when these guidelines conflict with one another. (This trade-off is discussed in Thought Question 13.1.) Note that sets of DFM rules are common in many industries and are often specific to each manufacturing firm.
Boothroyd-Dewhurst DFA Software: Since this software is widely used today, we usually take a few minutes to explain and demonstrate to the class how the software works and how teams make use of this tool. We accomplish this by simply explaining an example of how to find the assembly times for components using tables obtained from Boothroyd-Dewhurst Inc. or those published in one of their books. We have not tried it, but a live software demonstration might be interesting.
Product Quality: It is important to discuss the relationship between DFM success and various dimensions of product quality. Can DFM improve design for disassembly, service, or recycling? Having been frustrated with products in the past, students should have quite a lot to add at this point.
Optional Video: IBM ProPrinter
In introductory classes, we have often used a videotape from IBM describing the design of their ProPrinter. The full tape is 30 minutes, so we usually just show the manual and automatic assembly processes (9 minutes). We then discuss the philosophy and principles that guided this project, as described in related articles. The "IBM ProPrinter Design" videotape that we have used is available for a nominal charge through Sunbelt Video (704-527-4152). Note that there are several suitable alternative DFM videotapes available from other sources.
DFM Exercise
Choose one of the in-class exercises described below. Each takes about 30 minutes of class time.
Discussion
At the end of the session, we like to discuss the broader strategic context of DFM. We open this discussion by contrasting two design approaches: the Sony Walkman and the Digital Equipment Corporation desktop mouse.
The Digital mouse design exemplifies the now popular DFA approach. It uses no discrete fasteners, and can be assembled in a matter of seconds. (Alternatives for this example include the IBM ProPrinter or the Polaroid Impulse or Spectrum camera.)
The Sony Walkman represents an alternative approach. One can easily see that the Walkman has numerous discrete fasteners; there are several visible even from the outside. Sony, of course, has a very effective product development organization. Yet they appear to take an approach quite orthogonal to DFM as described here.
To settle this question, it is important to note that a Walkman is very different than a camera or a mouse. The development process for Sony’s incremental products might be a matter of a few months and the model may only be in production for a few months.
As illustrated by the diagram below (which can be drawn on the board), the choice of a design strategy, such as DFA, depends upon the marketing strategy and the manufacturing strategy. As a result, design strategies and DFM implementations will differ across firms and across projects.
Since design for manufacturing (DFM) concepts are useful even during the concept development phase, it would certainly be helpful to place this session near the beginning of a project course. However not everything can be placed up front, so we generally teach this session somewhere during the second half of the semester. This assures that students have at least struggled with DFM issues to some extent in their projects before the class discussion.
Objectives and Strategy
This session introduces the class to many concepts related to DFM. The intended message is that DFM is much more than cost reduction. It is a design strategy that requires the expertise of multiple team members and the use of basic design rules, guidelines, and cost models. DFM often results in significant cost reduction as well as improvement in product quality and the development of cross-functional expertise within the organization.
We have taught single sessions on DFM for many audiences, including in an MBA core course on operations management and in our more advanced graduate course on product development. We have also found this topic valuable for undergraduate courses and useful in executive education. The session outline below describes some of the material we have used in a variety of these contexts.
Session Outline
The session can follow this flow:
- Introduction to or Summary of DFM
- DFM Principles and Caveats
- Optional Video: IBM ProPrinter
- DFM Exercise
- Discussion
Introduction to DFM
The class session can begin by relating the hype about DFM over the last decade to significant gains in product quality, reductions in product cost, and improvement in the structure of development organizations. While DFM is not directly responsible for all of these effects, it can be closely linked to each of them as we will explore in this session.
Next, a student can be asked to summarize the main points of DFM, which usually brings up three or four of the items in the next section.
DFM Principles and Caveats
There are several points which should be brought out in discussion:
Cross-Functional Teams: This practice is particularly useful in DFM since a wide range of skills and perspectives is required for DFM success.
DFM Starts Early: Even during the concept development phase, manufacturing concerns can be used to guide the generation and selection of alternatives. There are many examples of DFM success where novel concepts were required to achieve significant improvements (e.g., Digital’s desktop mouse).
Cost Modeling: To guide the efforts and measure the benefits of DFM, it is useful to have cost models capable of accurately predicting the eventual production costs for any product. Ideally, a design team would be able to push a button on their CAD terminal and receive an accurate production cost for their proposed design. Unfortunately such CAD-to-cost systems do not exist – except for a few, very specialized, applications.
Cost Structure: It is useful to discuss alternative ways to break down costs, such as: 1) direct-indirect, 2) materials-labor-overhead, or 3) components-assembly-overhead. For DFM, we seek a cost structure which gives the most useful information for making the necessary decisions. It can be useful at this point to spend a few minutes discussing activity-based costing and the meaning of overhead costs.
Focusing on Assembly: Students will probably be able to list several reasons why it is useful to focus attention on the simplification of assembly operations. There are several "multiplier effects" to discuss, such as the multiple ways that reducing parts count reduces overhead costs. (Exercise 13.3 discusses some of these effects.)
Design Rules: The class can list a set of basic DFM guidelines (known as DFM rules of thumb) to follow, as many of these are described in the chapter. The discussion should include how to handle cases when these guidelines conflict with one another. (This trade-off is discussed in Thought Question 13.1.) Note that sets of DFM rules are common in many industries and are often specific to each manufacturing firm.
Boothroyd-Dewhurst DFA Software: Since this software is widely used today, we usually take a few minutes to explain and demonstrate to the class how the software works and how teams make use of this tool. We accomplish this by simply explaining an example of how to find the assembly times for components using tables obtained from Boothroyd-Dewhurst Inc. or those published in one of their books. We have not tried it, but a live software demonstration might be interesting.
Product Quality: It is important to discuss the relationship between DFM success and various dimensions of product quality. Can DFM improve design for disassembly, service, or recycling? Having been frustrated with products in the past, students should have quite a lot to add at this point.
Optional Video: IBM ProPrinter
In introductory classes, we have often used a videotape from IBM describing the design of their ProPrinter. The full tape is 30 minutes, so we usually just show the manual and automatic assembly processes (9 minutes). We then discuss the philosophy and principles that guided this project, as described in related articles. The "IBM ProPrinter Design" videotape that we have used is available for a nominal charge through Sunbelt Video (704-527-4152). Note that there are several suitable alternative DFM videotapes available from other sources.
DFM Exercise
Choose one of the in-class exercises described below. Each takes about 30 minutes of class time.
Discussion
At the end of the session, we like to discuss the broader strategic context of DFM. We open this discussion by contrasting two design approaches: the Sony Walkman and the Digital Equipment Corporation desktop mouse.
The Digital mouse design exemplifies the now popular DFA approach. It uses no discrete fasteners, and can be assembled in a matter of seconds. (Alternatives for this example include the IBM ProPrinter or the Polaroid Impulse or Spectrum camera.)
The Sony Walkman represents an alternative approach. One can easily see that the Walkman has numerous discrete fasteners; there are several visible even from the outside. Sony, of course, has a very effective product development organization. Yet they appear to take an approach quite orthogonal to DFM as described here.
- How can this difference be explained? Students will offer several hypotheses; here are some of the stronger ones:
- Robustness–Perhaps Sony uses screws to improve the ruggedness of the product.
- Production Capability– Perhaps Sony has manufacturing equipment that installs screws easily.
- Appearance–Perhaps Sony believes that screws provide a look of quality.
- Flexibility–Perhaps the screws facilitate product line modularity.
- Size–Perhaps Sony uses screws because they take less space than snap fits.
- Agility–Perhaps Sony does not have the time to design complex moldings and to procure the mold tooling within the allotted development time. Complex molds can take many months to build.
To settle this question, it is important to note that a Walkman is very different than a camera or a mouse. The development process for Sony’s incremental products might be a matter of a few months and the model may only be in production for a few months.
As illustrated by the diagram below (which can be drawn on the board), the choice of a design strategy, such as DFA, depends upon the marketing strategy and the manufacturing strategy. As a result, design strategies and DFM implementations will differ across firms and across projects.
Sony’s marketing strategy for the Walkman line is one of rapid product changes. New models with new features are released every month. Sony’s automated production system allows screws and other small parts to be easily inserted. As a result, Sony’s design strategy, rather than following typical DFM/DFA guidelines, revolves around rapid development projects. Complex plastic parts with snap fits require many months to design and build the molds.
The conclusion is that design strategies differ across firms and products depending upon the marketing strategy and manufacturing capabilities. Accordingly, DFM must be implemented differently in these various situations.
Props
In-Class Exercises
We have developed three DFM-related exercises for classroom use at various levels.
1. Digital Mouse
This exercise takes about 30 minutes of class time. We have used it successfully in undergraduate, graduate, and executive classes.
Bring to class the following items:
This exercise involves simply disassembling and reassembling the Digital desktop mouse shown below (right). It is helpful to contrast this product with its predecessor, also shown below (left). The old mouse uses the traditional ball-cage design, whereas the new mouse uses a novel roller-foot design. The table below contrasts the characteristics of these two designs.
The conclusion is that design strategies differ across firms and products depending upon the marketing strategy and manufacturing capabilities. Accordingly, DFM must be implemented differently in these various situations.
Props
- Sony Walkman
- Digital Mouse or Polaroid Camera
- Materials for in-class exercise (see below)
- Some very complex parts resulting from component integration (e.g., parts from a Polaroid camera or the left side frame from an IBM ProPrinter)
In-Class Exercises
We have developed three DFM-related exercises for classroom use at various levels.
1. Digital Mouse
This exercise takes about 30 minutes of class time. We have used it successfully in undergraduate, graduate, and executive classes.
Bring to class the following items:
- 1 or more of the old-style Digital Desktop Mouse
- several of the new-style Digital Desktop Mouse
- 1 small (#0) Phillips screwdriver
This exercise involves simply disassembling and reassembling the Digital desktop mouse shown below (right). It is helpful to contrast this product with its predecessor, also shown below (left). The old mouse uses the traditional ball-cage design, whereas the new mouse uses a novel roller-foot design. The table below contrasts the characteristics of these two designs.
Old Mouse Design
|
New Mouse Design
|
caged ball
|
working principle
|
two inclined rollers
|
31
|
# of unique mechanical parts
|
16
|
10 screws
|
# of fasteners
|
none
|
30
|
# of unique electrical parts
|
28
|
11
|
# of adjustments
|
none
|
83
|
# of assembly operations
|
56
|
17 minutes
|
total assembly time
|
6 minutes
|
-
|
material cost savings
|
over 40%
|
Comparison of Old and New Digital Mice. (source: Plastice World, September 1990, pg. 29.)
After describing the old and new designs, ask the students to disassemble and study the new mouse design. We bring about one mouse for every two or three students in the class, so they can work in small groups. It can be tricky to remove the clamshell cover, so it helps to walk around with a pry tool to assist.
This exercise will challenge students to understand the working principles involved in the mouse’s functions. Students may be asked the following questions:
• What is the purpose of the slits in the code wheels? These are optical encoders used to track the motion of the mouse.
• How does the computer know which way the mouse moves? Most students will understand how the horizontal and vertical motion are decoupled. However very few will guess how the +/— directionality of each motion is sensed. There are two receivers for each emitter. Having two signal channels for each motion allows the mouse circuitry to decode directions. (When the two receivers show "dark/dark", the possible adjacent states are "dark/light" for one direction and "light/dark" indicating the other direction of motion.)
• What are the magnetic disks for? These pull the rollers down to create bias force against the desktop or mouse pad.
Students will surely notice how simple it is to assemble this product. They will only need about 30 seconds to put back together what they have disassembled. It must be emphasized that this redesign included an entirely new technical working principle in comparison to the old product. This should serve to illustrate the benefit of considering DFM as early as possible in a development project.
The discussion can then focus upon what it would take to achieve this kind of DFM success. Digital required about 10 person-years of effort to redesign the mouse, costing about $1M. Tooling costs added another $1M perhaps. At 150,000 units per year, with an estimated savings of $20/unit, this DFM effort paid back in less than one year.
2. Videocassette
We developed this exercise for our graduate course in which most of the students already have a basic familiarity with DFM concepts. This exercise also takes about 30 minutes of class time.
Bring to class:
This exercise involves each student disassembling a videocassette and considering the following questions:
The photos below show two very different videocassette designs, representing the extremes of the spectrum in terms of parts count.
This exercise will challenge students to understand the working principles involved in the mouse’s functions. Students may be asked the following questions:
• What is the purpose of the slits in the code wheels? These are optical encoders used to track the motion of the mouse.
• How does the computer know which way the mouse moves? Most students will understand how the horizontal and vertical motion are decoupled. However very few will guess how the +/— directionality of each motion is sensed. There are two receivers for each emitter. Having two signal channels for each motion allows the mouse circuitry to decode directions. (When the two receivers show "dark/dark", the possible adjacent states are "dark/light" for one direction and "light/dark" indicating the other direction of motion.)
• What are the magnetic disks for? These pull the rollers down to create bias force against the desktop or mouse pad.
Students will surely notice how simple it is to assemble this product. They will only need about 30 seconds to put back together what they have disassembled. It must be emphasized that this redesign included an entirely new technical working principle in comparison to the old product. This should serve to illustrate the benefit of considering DFM as early as possible in a development project.
The discussion can then focus upon what it would take to achieve this kind of DFM success. Digital required about 10 person-years of effort to redesign the mouse, costing about $1M. Tooling costs added another $1M perhaps. At 150,000 units per year, with an estimated savings of $20/unit, this DFM effort paid back in less than one year.
2. Videocassette
We developed this exercise for our graduate course in which most of the students already have a basic familiarity with DFM concepts. This exercise also takes about 30 minutes of class time.
Bring to class:
- several small (#0) Phillips screwdrivers
- several brands of VHS videocassettes (We ask each student to bring one to class.)
- one G-Zero videocassette, disassembled (by Global Zero, Westbrook, Maine)
This exercise involves each student disassembling a videocassette and considering the following questions:
- How difficult is it to assemble the product?
- How can the design be improved?
- How might the design evolve over time?
The photos below show two very different videocassette designs, representing the extremes of the spectrum in terms of parts count.
By simply asking the students to count the number of parts in their videocassettes, several observations can be made:
If none of the students brought a videocassette with only eight parts, it is nice to show the very impressive design by Global Zero. This product illustrates to the students how far DFM can be taken. In this case, the designers have developed a very low cost product which can be assembled in the United Stated by machine, uses only a single molded component for the entire shell, two springs, and two one-piece hubs. (The sixth part in the count is the clear leader tape.) The G-Zero videocassette shell is made from polypropylene and is therefore easily recycled. The G-Zero product itself can be made of recycled polypropylene as well. This videocassette shell seems ideally suited to the direct-mail promotion market and it has sold very well in that niche.
Here are some useful background facts about the industry and the production process for videocassette shells.
3. PaperMate Pen
This exercise takes at least 30 minutes of class time. We have used it successfully in undergraduate and graduate classes to introduce DFM and to emphasize two points:
Bring to class:
- The range of parts count for typical designs is 28-38. Even within this range there are striking differences that represent improvement and evolution of the design over several years.
- The "hub lock" mechanism can be implemented using five parts, or four, or three, or one.
- Some manufacturers color code the mirror-image parts to avoid confusion (as shown in Exhibit 13-14). This is useful even in automatic assembly situations because it avoids confusing the parts when loading the bowl feeders.
- Springs can be very difficult to install.
If none of the students brought a videocassette with only eight parts, it is nice to show the very impressive design by Global Zero. This product illustrates to the students how far DFM can be taken. In this case, the designers have developed a very low cost product which can be assembled in the United Stated by machine, uses only a single molded component for the entire shell, two springs, and two one-piece hubs. (The sixth part in the count is the clear leader tape.) The G-Zero videocassette shell is made from polypropylene and is therefore easily recycled. The G-Zero product itself can be made of recycled polypropylene as well. This videocassette shell seems ideally suited to the direct-mail promotion market and it has sold very well in that niche.
Here are some useful background facts about the industry and the production process for videocassette shells.
- The total market for 1/2" videocassettes is about 1 billion units annually. (This includes all standard formats like VHS, VHSC, Beta, etc.) The top 7 producers all have about 10% market share and swap the lead position almost monthly. The breakdown by region is approximately:
- North America 350M units
- Europe 250M units
- Japan 250M units
- Other 150M
- JVC licenses the VHS standard to several manufacturers. The VHS specification includes the shell dimensions and functions (e.g., hub lock, door latch), tape transport interfaces (e.g., splined hub, roller positions), and guidelines regarding the magnetic medium (e.g., minimum particle density).
- Design details governing how to achieve the functionality specified by JVC are left to the individual manufacturers. They can choose whatever mechanisms, springs, and fasteners they need. Differences in these details may represent the evolution of the design both within and across manufacturers. Innovations are patented or guarded closely but still transfer quickly.
- Manufacturers also choose the production and assembly methods, such as the winding technology, and manual versus automatic assembly. Many manufacturers choose automated assembly. Regular improvements are made to reduce cost and to improve quality. Some manufacturers produce complete, empty shells with leader tape only, and then wind in the desired length of the appropriate tape as a separate operation.
- Quality dimensions of interest include: tape "drop outs" (caused by dust in the manufacturing process), tape degradation (caused by roughness of the pins and rollers), and robustness to normal use and abuse.
- Tape quality generally exceeds the recording and playback capability of most consumer VCRs, making the higher quality (HQ grade) tapes unnecessary for most applications. Some manufacturers put the same tape in both (standard and high) grade products. Then based on test results from each batch (dropout density), they label the tapes with fewer defects as HQ tapes.
3. PaperMate Pen
This exercise takes at least 30 minutes of class time. We have used it successfully in undergraduate and graduate classes to introduce DFM and to emphasize two points:
- Many improvements are possible, even on simple, highly engineered, high-volume products.
- A detailed understanding of the relevant production processes is required.
Bring to class:
- a few boxes of PaperMate ball-point pens (one pen for each student)
PaperMate Pens
This exercise involves developing suggestions to improve a simple, high-volume product: the common PaperMate ball-point pen shown above, manufactured by Gillette. We start by supposing the following scenario:
Hand out the pens and ask the students to first take a few minutes to list the parts of the pen and their best guesses for the production processes employed for each component. Then they should consider how the assembly process takes place. Finally they should each develop several suggestions for improving the design and reducing the production cost.
When ten minutes are almost over, ask one student to list the pen’s components and their associated production processes on the board. After explaining these, ask each student to give one suggestion for improving the pen design. Several promising suggestions usually emerge.
Two related items of trivia that the students will appreciate as cocktail-party facts:
- You are taking a cross-country trip for a job interview and the airline mistakenly assigns you to a first-class seat. When you strike up a conversation with the person sitting next to you, you find that he is the vice president of engineering at Gillette. While you are thinking about their Sensor razor that you like, he mentions that Gillette sells on the order of one billion PaperMate pens a year. At this volume, even at 10 cents each, this is a huge business. You tell him that you have been exposed to DFM in school. His skeptical reply is: "Well if DFM could reduce the cost of my pen by just half a cent, I’d believe every word of it." Some time later, he gets up to stretch his legs. You grab for his pen and now you have ten minutes to develop some cost-saving suggestions. If you impress him, perhaps he will offer you a job!
Hand out the pens and ask the students to first take a few minutes to list the parts of the pen and their best guesses for the production processes employed for each component. Then they should consider how the assembly process takes place. Finally they should each develop several suggestions for improving the design and reducing the production cost.
When ten minutes are almost over, ask one student to list the pen’s components and their associated production processes on the board. After explaining these, ask each student to give one suggestion for improving the pen design. Several promising suggestions usually emerge.
Two related items of trivia that the students will appreciate as cocktail-party facts:
- Why do they put in so much ink? (Nobody ever uses it all.)
- One reason is so they can sell the pens to the US government, which has a durability specification.
- What is the meaning of the code stamped on the end of the barrel?
- This is the date code. The first digit (a number) corresponds to the year of manufacture. The second digit (a letter) indicates the month. The code 9C would translate to March 1999.