4.2 bike lovers a solidify understanding task answers

Read The Second Chapter Of The Text BookPreview The DocumentView In A New Window " Understanding Mechanical Design?". Then Answer The Following Questions Using Your Own Wording: 2.1 Decompose A Simple System Such As A Home Appliance, Bicycle, Or Toy Into

Read the second chapter of the Text bookPreview the documentView in a new window " Understanding Mechanical
Design?". Then answer the following questions using your own wording:

2.1 Decompose a simple system such as a home appliance, bicycle, or toy into its assemblies,
components, electrical circuits, and the like. Figures 2.3 and 2.11 will help.
2.2 For the device decomposed, list all the important features of one component.
2.3 Select a fastener from a catalog that meets these requirements:
■ Can attach two pieces of 14-gauge sheet steel (0.075 in., 1.9 mm) together
■ Is easy to fasten with a standard tool
■ Can only be removed with special tools
■ Can be removed without destroying either base materials or fastener
2.4 Sketch at least five ways to configure two passengers in a new four-wheeled commuter
vehicle that you are designing.
2.5 You are a designer of diving boards. A simple model of your product is a cantilever beam.
You want to design a new board so that a 150-lb (67-kg) woman deflects the board 3 in.
(7.6 cm) when standing on the end. Parametrically vary the length, material, and thickness
of the board to find five configurations that will meet the deflection criterion
2.6 Find five examples of mature designs. Also, find one mature design that has been recently
redesigned. What pressures or new developments led to the change?
2.7 Describe your chair in each of the four languages at the three levels of abstraction, as was
done with the bolt in Table 2.2.

Knowledge about the design process is increasing rapidly. A goal in writing the fourth edition of the Mechanical Design Process was to incorporate this knowledge into a uni� ed structure – one of the strong points of the � rst three editions. Throughout the new edition, topics have been updated and integrated with other best practices in the book. This new edition builds on the earlier editions’ reputation for being concise, direct, and for logically developing the design method with detailed how-to instructions, while remaining easy and enjoyable to read.

What’s New in the Fourth Edition

Improved material to ensure team success

Over 20 templates that can be downloaded from the website to support activities throughout the design process

Improved material on project planning

Improved sections on Design for the Environment and Design for Sustainability

Improved material on making design decisions

A new section on using contradictions to generate ideas

New examples from industry: Irwin Tool, GE Medical, Marin Bicycles, and Jet Propulsion Laboratory are featured.

A book website provides additional resources to instructors and other readers. www.mhhe.com/ullman4e

The fourth edition of The Mechanical Design Process continues to evolve from David Ullman’s years of studying and teaching the design process and his extensive practical experience as a successful designer.

McGraw-Hill EngineeringCS.com

www.McGraw-HillEngineeringCS.com—Your one-stop online shop for all McGraw-Hill Engineering & Computer Science books, supplemental materials, content, & resources! For the student, the professor, and the professional, this site houses it all for your Engineering and CS needs.

Templates are available to both students and faculty for download from the website. These � ll-in-blank forms include:

Product decomposition Team contract Team health inventory Product proposal Pro/con analysis SWOT analysis Plastics part cost calculator Metals part cost calculator Reverse engineering DFAssy FMEA BOM Change order Patent prospects And more M

D D

A L

IM #996887 12/10/08 C

Y A

N M

A G

Y E

L O

B L

K

ullman-38162 ull75741_fm December 18, 2008 16:19

The Mechanical Design Process

ullman-38162 ull75741_fm December 18, 2008 16:19

McGraw-Hill Series in Mechanical Engineering

Alciatore/Histand Introduction to Mechatronics and Measurement System

Anderson Fundamentals of Aerodynamics

Anderson Introduction to Flight

Anderson Modern Compressible Flow

Barber Intermediate Mechanics of Materials

Beer/Johnston Vector Mechanics for Engineers

Beer/Johnston Mechanics of Materials

Budynas Advanced Strength and Applied Stress Analysis

Budynas/Nisbett Shigley’s Mechanical Engineering Design

Cengel Heat Transfer: A Practical Approach

Cengel Introduction to Thermodynamics & Heat Transfer

Cengel/Boles Thermodynamics: An Engineering Approach

Cengel/Clmbala Fluid Mechanics: Fundamentals and Applications

Cengel/Turner Fundamentals of Thermal-Fluid Sciences

Dieter Engineering Design: A Materials & Processing Approach

Doebelin Measurement Systems: Application & Design

Dorl/Byers Technology Ventures: From Idea to Enterprise

Dunn Measurement & Data Analysis for Engineering and Science

Fianemore/Franzial Fluid Mechanics with Engineering Applications

Hamrock/Schmid/Jacobson Fundamentals of Machine Elements

Heywood Internal Combustion Engine Fundamentals

Holman Experimental Methods for Engineers

Holman Heat Transfer

Hutton Fundamental of Finite Element Analysis

Kays/Crawford/Welgand Convective Heat and Mass Transfer

Meirovioeh Fundamentals of Vibrations

Norton Design of Machinery

Palm System Dynamics

Reddy An Introduction to Finite Element Method

Schey Introduction to Manufacturing Processes

Shames Mechanics of Fluids

Smith/Hashemi Foundations of Materials Science & Engineering

Turns An Introduction to Combustion: Concepts and Applications

Ugural Mechanical Design: An Integrated Approach

Ullman The Mechanical Design Process

White Fluid Mechanics

White Viscous Fluid Flow

Zeid CAD/CAM Theory and Practice

Zeid Mastering CAD/CAM

ullman-38162 ull75741_fm December 18, 2008 16:19

The Mechanical Design Process

Fourth Edition

David G. Ullman Professor Emeritus, Oregon State University

ullman-38162 ull75741_FM December 30, 2008 9:25

THE MECHANICAL DESIGN PROCESS, FOURTH EDITION

Published by McGraw-Hill, a business unit of The McGraw-Hill Companies, Inc., 1221 Avenue of the Americas, New York, NY 10020. Copyright © 2010 by The McGraw-Hill Companies, Inc. All rights reserved. Previous editions © 2003, 1997, and 1992. No part of this publication may be reproduced or distributed in any form or by any means, or stored in a database or retrieval system, without the prior written consent of The McGraw-Hill Companies, Inc., including, but not limited to, in any network or other electronic storage or transmission, or broadcast for distance learning.

Some ancillaries, including electronic and print components, may not be available to customers outside the United States.

This book is printed on acid-free paper.

1 2 3 4 5 6 7 8 9 0 DOC/DOC 0 9

ISBN 978–0–07–297574–1 MHID 0–07–297574–1

Global Publisher: Raghothaman Srinivasan Senior Sponsoring Editor: Bill Stenquist Director of Development: Kristine Tibbetts Senior Marketing Manager: Curt Reynolds Senior Project Manager: Kay J. Brimeyer Senior Production Supervisor: Sherry L. Kane Lead Media Project Manager: Stacy A. Patch Associate Design Coordinator: Brenda A. Rolwes Cover Designer: Studio Montage, St. Louis, Missouri Cover Image: Irwin clamp: © Irwin Industrial Tools; Marin bike: © Marin Bicycles; MER: © NASA/JPL. Senior Photo Research Coordinator: John C. Leland Compositor: S4Carlisle Publishing Services Typeface: 10.5/12 Times Roman Printer: R. R. Donnelley Crawfordsville, IN

Library of Congress Cataloging-in-Publication Data

Ullman, David G., 1944- The mechanical design process / David G. Ullman.—4th ed.

p. cm.—(McGraw-Hill series in mechanical engineering) Includes index. ISBN 978–0–07–297574–1—ISBN 0–07–297574–1 (alk. paper) 1. Machine design. I. Title.

TJ230.U54 2010 621.8′15—dc22 2008049434

www.mhhe.com

http://www.mhhe.com
ullman-38162 ull75741_fm December 18, 2008 16:19

ABOUT THE AUTHOR

David G. Ullman is an active product designer who has taught, researched, and written about design for over thirty years. He is president of Robust Decisions, Inc., a supplier of software products and training for product development and decision support. He is Emeritus Professor of Mechanical Design at Oregon State University. He has professionally designed fluid/thermal, control, and transporta- tion systems. He has published over twenty papers focused on understanding the mechanical product design process and the development of tools to support it. He is founder of the American Society Mechanical Engineers (ASME)—Design Theory and Methodology Committee and is a Fellow in the ASME. He holds a Ph.D. in Mechanical Engineering from the Ohio State University.

ullman-38162 ull75741_fm December 18, 2008 16:19

ullman-38162 ull75741_fm December 18, 2008 16:19

CONTENTS

Preface xi

CHAPTER 1 Why Study the Design Process? 1

1.1 Introduction 1 1.2 Measuring the Design Process with Product

Cost, Quality, and Time to Market 3

1.3 The History of the Design Process 8 1.4 The Life of a Product 10 1.5 The Many Solutions for Design

Problems 15

1.6 The Basic Actions of Problem Solving 17 1.7 Knowledge and Learning During Design 19 1.8 Design for Sustainability 20 1.9 Summary 21 1.10 Sources 22 1.11 Exercises 22

CHAPTER 2 Understanding Mechanical Design 25

2.1 Introduction 25 2.2 Importance of Product Function, Behavior,

and Performance 28

2.3 Mechanical Design Languages and Abstraction 30

2.4 Different Types of Mechanical Design Problems 33

2.5 Constraints, Goals, and Design Decisions 40

2.6 Product Decomposition 41 2.7 Summary 44

2.8 Sources 44 2.9 Exercises 45 2.10 On the Web 45

CHAPTER 3 Designers and Design Teams 47

3.1 Introduction 47 3.2 The Individual Designer: A Model of Human

Information Processing 48

3.3 Mental Processes That Occur During Design 56

3.4 Characteristics of Creators 64 3.5 The Structure of Design Teams 66 3.6 Building Design Team Performance 72 3.7 Summary 78 3.8 Sources 78 3.9 Exercises 79 3.10 On the Web 80

CHAPTER 4 The Design Process and Product Discovery 81

4.1 Introduction 81 4.2 Overview of the Design Process 81 4.3 Designing Quality into Products 92 4.4 Product Discovery 95 4.5 Choosing a Project 101 4.6 Summary 109 4.7 Sources 110 4.8 Exercises 110 4.9 On the Web 110

vii

ullman-38162 ull75741_fm December 18, 2008 16:19

viii Contents

CHAPTER 5 Planning for Design 111

5.1 Introduction 111 5.2 Types of Project Plans 113 5.3 Planning for Deliverables—

The Development of Information 117

5.4 Building a Plan 126 5.5 Design Plan Examples 134 5.6 Communication During the

Design Process 137

5.7 Summary 141 5.8 Sources 141 5.9 Exercises 142 5.10 On the Web 142

CHAPTER 6 Understanding the Problem and the Development of Engineering Specifications 143

6.1 Introduction 143 6.2 Step 1: Identify the Customers:

Who Are They? 151

6.3 Step 2: Determine the Customers’ Requirements: What Do the Customers Want? 151

6.4 Step 3: Determine Relative Importance of the Requirements: Who Versus What 155

6.5 Step 4: Identify and Evaluate the Competition: How Satisfied Are the Customers Now? 157

6.6 Step 5: Generate Engineering Specifications: How Will the Customers’ Requirement Be Met? 158

6.7 Step 6: Relate Customers’ Requirements to Engineering Specifications: How to Measure What? 163

6.8 Step 7: Set Engineering Specification Targets and Importance: How Much Is Good Enough? 164

6.9 Step 8: Identify Relationships Between Engineering Specifications: How Are the Hows Dependent on Each Other? 166

6.10 Further Comments on QFD 168 6.11 Summary 169 6.12 Sources 169 6.13 Exercises 169 6.14 On the Web 170

CHAPTER 7 Concept Generation 171

7.1 Introduction 171 7.2 Understanding the Function of Existing

Devices 176

7.3 ATechnique for Designing with Function 181 7.4 Basic Methods of Generating Concepts 189 7.5 Patents as a Source of Ideas 194 7.6 Using Contradictions to Generate Ideas 197 7.7 The Theory of Inventive Machines, TRIZ 201 7.8 Building a Morphology 204 7.9 Other Important Concerns During Concept

Generation 208

7.10 Summary 209 7.11 Sources 209 7.12 Exercises 211 7.13 On the Web 211

CHAPTER 8 Concept Evaluation and Selection 213

8.1 Introduction 213 8.2 Concept Evaluation Information 215 8.3 Feasibility Evaluations 218 8.4 Technology Readiness 219 8.5 The Decision Matrix—Pugh’s Method 221 8.6 Product, Project, and Decision Risk 226

ullman-38162 ull75741_fm December 18, 2008 16:19

Contents ix

8.7 Robust Decision Making 233 8.8 Summary 239 8.9 Sources 239 8.10 Exercises 240 8.11 On the Web 240

CHAPTER 9 Product Generation 241

9.1 Introduction 241 9.2 BOMs 245 9.3 Form Generation 246 9.4 Materials and Process Selection 264 9.5 Vendor Development 266 9.6 Generating a Suspension Design for the

Marin 2008 Mount Vision Pro Bicycle 269

9.7 Summary 276 9.8 Sources 276 9.9 Exercises 277 9.10 On the Web 278

CHAPTER 10 Product Evaluation for Performance and the Effects of Variation 279

10.1 Introduction 279 10.2 Monitoring Functional Change 280 10.3 The Goals of Performance Evaluation 281 10.4 Trade-Off Management 284 10.5 Accuracy, Variation, and Noise 286 10.6 Modeling for Performance Evaluation 292 10.7 Tolerance Analysis 296 10.8 Sensitivity Analysis 302 10.9 Robust Design by Analysis 305 10.10 Robust Design Through Testing 308 10.11 Summary 313

10.12 Sources 313 10.13 Exercises 314

CHAPTER 11 Product Evaluation: Design For Cost, Manufacture, Assembly, and Other Measures 315

11.1 Introduction 315 11.2 DFC—Design For Cost 315 11.3 DFV—Design For Value 325 11.4 DFM—Design For Manufacture 328 11.5 DFA—Design-For-Assembly

Evaluation 329

11.6 DFR—Design For Reliability 350 11.7 DFT and DFM—Design For Test and

Maintenance 357

11.8 DFE—Design For the Environment 358 11.9 Summary 360 11.10 Sources 361 11.11 Exercises 361 11.12 On the Web 362

CHAPTER 12 Wrapping Up the Design Process and Supporting the Product 363

12.1 Introduction 363 12.2 Design Documentation and

Communication 366

12.3 Support 368 12.4 Engineering Changes 370 12.5 Patent Applications 371 12.6 Design for End of Life 375 12.7 Sources 378 12.8 On the Web 378

ullman-38162 ull75741_fm December 18, 2008 16:19

x Contents

APPENDIX A Properties of 25 Materials Most Commonly Used in Mechanical Design 379

A.1 Introduction 379 A.2 Properties of the Most Commonly Used

Materials 380

A.3 Materials Used in Common Items 393 A.4 Sources 394

APPENDIX B Normal Probability 397

B.1 Introduction 397 B.2 Other Measures 401

APPENDIX C The Factor of Safety as a Design Variable 403

C.1 Introduction 403

C.2 The Classical Rule-of-Thumb Factor of Safety 405

C.3 The Statistical, Reliability-Based, Factor of Safety 406

C.4 Sources 414

APPENDIX D Human Factors in Design 415

D.1 Introduction 415 D.2 The Human in the Workspace 416 D.3 The Human as Source of Power 419 D.4 The Human as Sensor and

Controller 419

D.5 Sources 426

Index 427

ullman-38162 ull75741_fm December 18, 2008 16:19

PREFACE

I have been a designer all my life. I have designed bicycles, medical equipment,furniture, and sculpture, both static and dynamic. Designing objects has comeeasy for me. I have been fortunate in having whatever talents are necessary to be a successful designer. However, after a number of years of teaching mechanical design courses, I came to the realization that I didn’t know how to teach what I knew so well. I could show students examples of good-quality design and poor- quality design. I could give them case histories of designers in action. I could suggest design ideas. But I could not tell them what to do to solve a design problem. Additionally, I realized from talking with other mechanical design teachers that I was not alone.

This situation reminded me of an experience I had once had on ice skates. As a novice skater I could stand up and go forward, lamely. A friend (a teacher by trade) could easily skate forward and backward as well. He had been skating since he was a young boy, and it was second nature to him. One day while we were skating together, I asked him to teach me how to skate backward. He said it was easy, told me to watch, and skated off backward. But when I tried to do what he did, I immediately fell down. As he helped me up, I asked him to tell me exactly what to do, not just show me. After a moment’s thought, he concluded that he couldn’t actually describe the feat to me. I still can’t skate backward, and I suppose he still can’t explain the skills involved in skating backward. The frustration that I felt falling down as my friend skated with ease must have been the same emotion felt by my design students when I failed to tell them exactly what to do to solve a design problem.

This realization led me to study the process of mechanical design, and it eventually led to this book. Part has been original research, part studying U.S. in- dustry, part studying foreign design techniques, and part trying different teaching approaches on design classes. I came to four basic conclusions about mechanical design as a result of these studies:

1. The only way to learn about design is to do design. 2. In engineering design, the designer uses three types of knowledge: knowl-

edge to generate ideas, knowledge to evaluate ideas and make decisions, and knowledge to structure the design process. Idea generation comes from ex- perience and natural ability. Idea evaluation comes partially from experience and partially from formal training, and is the focus of most engineering ed- ucation. Generative and evaluative knowledge are forms of domain-specific knowledge. Knowledge about the design process and decision making is largely independent of domain-specific knowledge.

3. A design process that results in a quality product can be learned, provided there is enough ability and experience to generate ideas and enough experi- ence and training to evaluate them.

xi

ullman-38162 ull75741_fm December 18, 2008 16:19

xii Preface

4. A design process should be learned in a dual setting: in an academic envi- ronment and, at the same time, in an environment that simulates industrial realities.

I have incorporated these concepts into this book, which is organized so that readers can learn about the design process at the same time they are developing a product. Chaps. 1–3 present background on mechanical design, define the terms that are basic to the study of the design process, and discuss the human element of product design. Chaps. 4–12, the body of the book, present a step-by-step development of a design method that leads the reader from the realization that there is a design problem to a solution ready for manufacture and assembly. This material is presented in a manner independent of the exact problem being solved. The techniques discussed are used in industry, and their names have become buzzwords in mechanical design: quality function deployment, decision-making methods, concurrent engineering, design for assembly, and Taguchi’s method for robust design. These techniques have all been brought together in this book. Although they are presented sequentially as step-by-step methods, the overall process is highly iterative, and the steps are merely a guide to be used when needed.

As mentioned earlier, domain knowledge is somewhat distinct from process knowledge. Because of this independence, a successful product can result from the design process regardless of the knowledge of the designer or the type of design problem. Even students at the freshman level could take a course using this text and learn most of the process. However, to produce any reasonably realistic design, substantial domain knowledge is required, and it is assumed throughout the book that the reader has a background in basic engineering science, material science, manufacturing processes, and engineering economics. Thus, this book is intended for upper-level undergraduate students, graduate students, and professional engineers who have never had a formal course in the mechanical design process.

ADDITIONS TO THE FOURTH EDITION

Knowledge about the design process is increasing rapidly. A goal in writing the fourth edition was to incorporate this knowledge into the unified structure—one of the strong points of the first three editions. Throughout the new edition, topics have been updated and integrated with other best practices in the book. Some specific additions to the new edition include:

1. Improved material to ensure team success. 2. Over twenty blank templates are available for download from the book’s web-

site (www.mhhe.com/ullman4e) to support activities throughout the design process. The text includes many of them filled out for student reference.

3. Improved material on project planning.

http://www.mhhe.com/ullman4e
ullman-38162 ull75741_fm December 18, 2008 16:19

Preface xiii

4. Improved sections on Design for the Environment and Design for Sustainability.

5. Improved material on making design decisions. 6. A new section on using contradictions to generate ideas. 7. New examples from the industry, with new photos and diagrams to illustrate

the examples throughout.

Beyond these, many small changes have been made to keep the book current and useful.

ELECTRONIC TEXTBOOK

CourseSmart is a new way for faculty to find and review eTextbooks. It’s also a great option for students who are interested in accessing their course materials digitally and saving money. CourseSmart offers thousands of the most commonly adopted textbooks across hundreds of courses from a wide variety of higher education publishers. It is the only place for faculty to review and compare the full text of a textbook online, providing immediate access without the environmental impact of requesting a print exam copy. At CourseSmart, students can save up to 50% off the cost of a print book, reduce their impact on the environment, and gain access to powerful Web tools for learning including full text search, notes and highlighting, and email tools for sharing notes between classmates. www.CourseSmart.com

ACKNOWLEDGMENTS

I would like to thank these reviewers for their helpful comments:

Patricia Brackin, Rose-Hulman Institute of Technology William Callen, Georgia Institute of Technology Xiaoping Du, University of Missouri-Rolla Ian Grosse, University of Massachusetts–Amherst Karl-Heinrich Grote, Otto-von-Guericke University, Magdeburg, Germany Mica Grujicic, Clemson University John Halloran, University of Michigan Peter Jones, Auburn University Mary Kasarda, Virginia Technical College Jesa Kreiner, California State University–Fullerton Yuyi Lin, University of Missouri–Columbia Ron Lumia, University of New Mexico Spencer Magleby, Brigham Young University Lorin Maletsky, University of Kansas

http://www.CourseSmart.com
ullman-38162 ull75741_fm December 18, 2008 16:19

xiv Preface

Make McDermott, Texas A&M University Joel Ness, University of North Dakota Charles Pezeshki, Washington State University John Renaud, University of Notre Dame Keith Rouch, University of Kentucky Ali Sadegh, The City College of The City University of New York Shin-Min Song, Northern Illinois University Mark Steiner, Rensselaer Polytechnic Institute Joshua Summers, Clemson University Meenakshi Sundaram, Tennessee Technical University Shih-Hsi Tong, University of California–Los Angeles Kristin Wood, University of Texas

Additionally, I would like to thank Bill Stenquist, senior sponsoring editor for mechanical engineering of McGraw-Hill, Robin Reed, developmental editor, Kay Brimeyer, project manager, and Lynn Steines, project editor, for their interest and encouragement in this project. Also, thanks to the following who helped with examples in the book:

Wayne Collier, UGS Jason Faircloth, Marin Bicycles Marci Lackovic, Autodesk Samir Mesihovic, Volvo Trucks Professor Bob Paasch, Oregon State University Matt Popik, Irwin Tools Cary Rogers, GE Medical Professor Tim Simpson, Penn State University Ralf Strauss, Irwin Tools Christopher Voorhees, Jet Propulsion Laboratory Professor Joe Zaworski, Oregon State University

Last and most important my thanks to my wife, Adele, for her never ques- tioning confidence that I could finish this project.

ullman-38162 ull75741_01 December 17, 2008 11:15

1C H A P T E R Why Study the Design Process?

KEY QUESTIONS ■ What can be done to design quality mechanical products on time and within

budget? ■ What are the ten key features of design best practice that will lead to better

products? ■ What are the phases of a product’s life cycle? ■ How are design problems different from analysis problems? ■ Why is it during design, the more you know, the less design freedom you

have? ■ What are the Hanover Principles?

1.1 INTRODUCTION Beginning with the simple potter’s wheel and evolving to complex consumer products and transportation systems, humans have been designing mechanical objects for nearly five thousand years. Each of these objects is the end result of a long and often difficult design process. This book is about that process. Regardless of whether we are designing gearboxes, heat exchangers, satellites, or doorknobs, there are certain techniques that can be used during the design process to help ensure successful results. Since this book is about the process of mechanical design, it focuses not on the design of any one type of object but on techniques that apply to the design of all types of mechanical objects.

If people have been designing for five thousand years and there are literally millions of mechanical objects that work and work well, why study the design process? The answer, simply put, is that there is a continuous need for new, cost-effective, high-quality products. Today’s products have become so complex that most require a team of people from diverse areas of expertise to develop an idea into hardware. The more people involved in a project, the greater is the need for assistance in communication and structure to ensure nothing important

1

ullman-38162 ull75741_01 December 17, 2008 11:15

2 CHAPTER 1 Why Study the Design Process?

is overlooked and customers will be satisfied. In addition, the global marketplace has fostered the need to develop new products at a very rapid and accelerating pace. To compete in this market, a company must be very efficient in the design of its products. It is the process that will be studied here that determines the efficiency of new product development. Finally, it has been estimated that 85% of the problems with new products not working as they should, taking too long to bring to market, or costing too much are the result of a poor design process.

The goal of this book is to give you the tools to develop an efficient design process regardless of the product being developed. In this chapter the important features of design problems and the processes for solving them will be introduced. These features apply to any type of design problem, whether for mechanical, elec- trical, software, or construction projects. Subsequent chapters will focus more on mechanical design, but even these can be applied to a broader range of problems.

Consider the important factors that determine the success or failure of a product (Fig. 1.1). These factors are organized into three ovals representing those factors important to product design, business, and production.

Product design factors focus on the product’s function, which is a description of what the object does. The importance of function to the designer is a major topic of this book. Related to the function are the product’s form, materials, and manufacturing processes. Form includes the product’s architecture, its shape, its color, its texture, and other factors relating to its structure. Of equal importance to form are the materials and manufacturing processes used to produce the product. These four variables—function, form, materials, and manufacturing processes—

Business

Production

Product design

Product form Price

Promotion

Distribution coverage

Sales forecast

Target market

Manufacturing processes

Production planning/ sourcing

Production system

Cost/risk

Facilities

Materials

Product function

Figure 1.1 Controllable variables in product development.

ullman-38162 ull75741_01 December 17, 2008 11:15

1.2 Measuring the Design Process with Product Cost, Quality, and Time to Market 3

are of major concern to the designer. This product design oval is further refined in Fig. 9.3.

The product form and function is also important to the business because the customers in the target market judge a product primarily on what it does (its function) and how it looks (its form). The target market is one factor important to the business, as shown in Fig. 1.1. The goal of a business is to make money— to meet its sales forecasts. Sales are also affected by the company’s ability to promote the product, distribute the product, and price the product, as shown in Fig. 1.1.

The business is dependent not only on the product form and function, but also on the company’s ability to produce the product. As shown in the production oval in Fig. 1.1, the production system is the central factor. Notice how product design and production are both concerned with manufacturing processes. The choice of form and materials that give the product function affects the manufacturing processes that can be used. These processes, in turn, affect the cost and hence the price of the product. This is just one example of how intertwined product design, production, and businesses truly are. In this book we focus on the product design oval. But, we will also pay much attention to the business and production variables that are related to design.As shown in the upcoming sections, the design process has a great effect on product cost, quality, and time to market.

1.2 MEASURING THE DESIGN PROCESS WITH PRODUCT COST, QUALITY, AND TIME TO MARKET

The three measures of the effectiveness of the design process are product cost, quality, and time to market. Regardless of the product being designed—whether it is an entire system, some small subpart of a larger product, or just a small change in an existing product—the customer and management always want it cheaper (lower cost), better (higher quality), and faster (less time).

The actual cost of designing a product is usually a small part of the manufac- turing cost of a product, as can be seen in Fig. 1.2, which is based on data from Ford Motor Company. The data show that only 5% of the manufacturing cost of a car (the cost to produce the car but not to distribute or sell it) is for design activities that were needed to develop it. This number varies with industry and product, but for most products the cost of design is a small part of the manufacturing cost.

However, the effect of the quality of the design on the manufacturing cost is much greater than 5%. This is most accurately shown from the results of a detailed study of 18 different automatic coffeemakers. Each coffeemaker had the same function—to make coffee. The results of this study are shown in Fig. 1.3. Here the effects of changes in manufacturing efficiency, such as material cost, labor wages, and cost of equipment, have been separated from the effects of the design process. Note that manufacturing efficiency and design have about the same influence on the cost of manufacturing a product. The figure shows that

ullman-38162 ull75741_01 December 17, 2008 11:15

4 CHAPTER 1 Why Study the Design Process?

15% Design Labor

Material

5%

50%

30% Overhead

Figure 1.2 Design cost as fraction of manufacturing cost.

$4.98 Good design

Efficient manufacturing

$9.72 Good design

Inefficient manufacturing

$8.17 Average design

Average manufacturing

$8.06 Poor design

Efficient manufacturing

$14.34 Poor design

Inefficient manufacturing

Figure 1.3 The effect of design on manufacturing cost. (Source: Data reduced from “Assessing the Importance of Design through Product Archaeology,” Management Science, Vol. 44, No. 3, pp. 352–369, March 1998, by K. Ulrich and S. A. Pearson.)

Designers cost little, their impact on product cost, great.

good design, regardless of manufacturing efficiency, cuts the cost by about 35%. In some industries this effect is as high as 75%.

Thus, comparing Fig. 1.2 to Fig. 1.3, we can conclude that the decisions made during the design process have a great effect on the cost of a product but cost very little. Design decisions directly determine the materials used, the goods

ullman-38162 ull75741_01 December 17, 2008 11:15

1.2 Measuring the Design Process with Product Cost, Quality, and Time to Market 5

Product cost is committed early in the design process and spent late in the process.

purchased, the parts, the shape of those parts, the product sold, the price of the product, and the sales.

Another example of the relationship of the design process to cost comes from Xerox. In the 1960s and early 1970s, Xerox controlled the copier market. However, by 1980 there were over 40 different manufacturers of copiers in the marketplace and Xerox’s share of the market had fallen significantly. Part of the problem was the cost of Xerox’s products. In fact, in 1980 Xerox realized that some producers were able to sell a copier for less than Xerox was able to manu- facture one of similar functionality. In one study of the problem, Xerox focused on the cost of individual parts. Comparing plastic parts from their machines and ones that performed a similar function in Japanese and European machines, they found that Japanese firms could produce a part for 50% less than American or European firms. Xerox attributed the cost difference to three factors: materials costs were 10% less in Japan, tooling and processing costs were 15% less, and the remaining 25% (half of the difference) was attributable to how the parts were designed.

Not only is much of the product cost committed during the design process, it is committed early in the design process. As shown in Fig. 1.4, about 75% of the manufacturing cost of a typical product is committed by the end of the conceptual phase process. This means that decisions made after this time can influence only 25% of the product’s manufacturing cost. Also shown in the figure is the amount of cost incurred, which is the amount of money spent on the design of the product.

100

80

60

40

20

Sp ec

if ic

at io

n de

ve lo

pm en

t

C on

ce pt

ua l

de si

gn

Pr od

uc t

de si

gn

Cost com mitted

Cost incurred

Time

Pe rc

en ta

ge o

f pr

od uc

t c os

t c om

m itt

ed

0

Figure 1.4 Manufacturing cost commitment during design.

ullman-38162 ull75741_01 December 17, 2008 11:15

6 CHAPTER 1 Why Study the Design Process?

Table 1.1 What determines quality

1989 2002

Works as it should 4.99 (1) 4.58 (1) Lasts a long time 4.75 (2) 3.93 (5) Is easy to maintain 4.65 (3) 3.29 (5) Looks attractive 2.95 (4–5) 3.58 (3–4) Incorporates latest technology/features 2.95 (4–5) 3.58 (3–4)

Scale: 5 = very important, 1 = not important at all, brackets denote rank. Sources: Based on a survey of consumers published in Time, Nov. 13, 1989, and a survey based on quality professional, R. Sebastianelli and N. Tamimi, “How Product Quality Dimensions Relate to Defining Quality,” International Journal of Quality and Reliability Management, Vol. 19, No. 4, pp. 442–453, 2002.

It is not until money is committed for production that large amounts of capital are spent.

The results of the design process also have a great effect on product quality. In a survey taken in 1989, American consumers were asked, “What determines quality?” Their responses, shown in Table 1.1, indicate that “quality” is a compos- ite of factors that are the responsibility of the design engineer. In a 2002 survey of engineers responsible for quality, what is important to “quality” is little changed. Although the surveys were of different groups, it is interesting to note that in the thirteen years between surveys, the importance of being easy to maintain has dropped, but the main measures of quality have remained unchanged.

Note that the most important quality measure is “works as it should.” This, and “incorporates latest technology/features,” are both measures of product function. “Lasts a long time” and most of the other quality measures are dependent on the form designed and on the materials and the manufacturing process selected. What is evident is that the decisions made during the design process determine the product’s quality.

Besides affecting cost and quality, the design process also affects the time it takes to produce a new product. Consider Fig. 1.5, which shows the num- ber of design changes made by two automobile companies with different design philosophies. The data points for Company B are actual for a U.S. automobile manufacturer, and the dashed line for Company A is what is typical for Toyota. Iteration, or change, is an essential part of the design process. However, changes occurring late in the design process are more expensive than those occurring ear- lier, as prior work is scrapped. The curve for Company B shows that the company was still making changes after the design had been released for production. In fact, over 35% of the cost of the product occurred after it was in production. In essence, Company B was still designing the automobile as it was being sold as a product. This causes tooling and assembly-line changes during production and the possibility of recalling cars for retrofit, both of which would necessitate significant expense, to say nothing about the loss of customer confidence. Com- pany A, on the other hand, made many changes early in the design process and finished the design of the car before it went into production. Early design changes require more engineering time and effort but do not require changes in hardware or documentation. A change that would cost $1000 in engineering time if made

ullman-38162 ull75741_01 December 17, 2008 11:15

1.2 Measuring the Design Process with Product Cost, Quality, and Time to Market 7

C ha

ng es

Time

St ar

t p ro

du ct

io n Company B

Actual project hours

Company A Ideal effort

Figure 1.5 Engineering changes during automobile development. (Source: Data from Tom Judd, Cognition Corp., “Taking DFSS to the Next Level,” WCBF, Design for Six Sigma Conference, Las Vegas, June 2005.)

Fail early; fail often.

early in the design process may cost $10,000 later during product refinement and $1,000,000 or more in tooling, sales, and goodwill expenses if made after production has begun.

Figure 1.5 also indicates that Company A took less time to design the auto- mobile than Company B. This is due to differences in the design philosophies of the companies. Company A assigns a large engineering staff to the project early in product development and encourages these engineers to utilize the latest in design techniques and to explore all the options early to preclude the need for changes later on. Company B, on the other hand, assigns a small staff and pres- sures them for quick results, in the form of hardware, discouraging the engineers from exploring all options (the region in the oval in the figure). The design ax- iom, fail early, fail often, applies to this example. Changes are required in order to find a good design, and early changes are easier and less expensive than changes made later. The engineers in Company B spend much time “firefighting” after the product is in production. In fact, many engineers spend as much as 50% of their time firefighting for companies similar to Company B.

An additional way that the design process affects product development time is in how long it takes to bring a product to market. Prior to the 1980s there was little emphasis on the length of time to develop new products, Since then competition has forced new products to be introduced at a faster and faster rate. During the 1990s development time in most industries was cut by half. This trend

ullman-38162 ull75741_01 December 17, 2008 11:15

8 CHAPTER 1 Why Study the Design Process?

has continued into the twenty-first century. More on how the design process has played a major role in this reduction is in Chap. 4.

Finally, for many years it was believed that there was a trade-off between high-quality products and low costs or time—namely, that it costs more and takes more time to develop and produce high-quality products. However, recent experience has shown that increasing quality and lowering costs and time can go hand in hand. Some of the examples we have discussed and ones throughout the rest of the book reinforce this point.

1.3 THE HISTORY OF THE DESIGN PROCESS During design activities, ideas are developed into hardware that is usable as a product. Whether this piece of hardware is a bookshelf or a space station, it is the result of a process that combines people and their knowledge, tools, and skills to develop a new creation. This task requires their time and costs money, and if the people are good at what they do and the environment they work in is well structured, they can do it efficiently. Further, if they are skilled, the final product will be well liked by those who use it and work with it—the customers will see it as a quality product. The design process, then, is the organization and management of people and the information they develop in the evolution of a product.

In simpler times, one person could design and manufacture an entire product. Even for a large project such as the design of a ship or a bridge, one person had sufficient knowledge of the physics, materials, and manufacturing processes to manage all aspects of the design and construction of the project.

By the middle of the twentieth century, products and manufacturing processes had become so complex that one person no longer had sufficient knowledge or time to focus on all the aspects of the evolving product. Different groups of people became responsible for marketing, design, manufacturing, and overall management. This evolution led to what is commonly known as the “over-the- wall” design process (Fig. 1.6).

In the structure shown in Fig. 1.6, the engineering design process is walled off from the other product development functions. Basically, people in market- ing communicate a perceived market need to engineering either as a simple, written request or, in many instances, orally. This is effectively a one-way com- munication and is thus represented as information that is “thrown over the wall.”

Customers Marketing Engineering design

Production

Figure 1.6 The over-the-wall design method.

ullman-38162 ull75741_01 December 17, 2008 11:15

1.3 The History of the Design Process 9

Engineering interprets the request, develops concepts, and refines the best concept into manufacturing specifications (i.e., drawings, bills of materials, and assembly instructions). These manufacturing specifications are thrown over the wall to be produced. Manufacturing then interprets the information passed to it and builds what it thinks engineering wanted.

Unfortunately, often what is manufactured by a company using the over-the- wall process is not what the customer had in mind. This is because of the many weaknesses in this product development process. First, marketing may not be able to communicate to engineering a clear picture of what the customers want. Since the design engineers have no contact with the customers and limited communi- cation with marketing, there is much room for poor understanding of the design problem. Second, design engineers do not know as much about the manufacturing processes as manufacturing specialists, and therefore some parts may not be able to be manufactured as drawn or manufactured on existing equipment. Further, manufacturing experts may know less-expensive methods to produce the prod- uct. Thus, this single-direction over-the-wall approach is inefficient and costly and may result in poor-quality products. Although many companies still use this method, most are realizing its weaknesses and are moving away from its use.

In the late 1970s and early 1980s, the concept of simultaneous engineering began to break down the walls. This philosophy emphasized the simultaneous development of the manufacturing process with the evolution of the product. Simultaneous engineering was accomplished by assigning manufacturing repre- sentatives to be members of design teams so that they could interact with the design engineers throughout the design process. The goal was the simultaneous development of the product and the manufacturing process.

In the 1980s the simultaneous design philosophy was broadened and called concurrent engineering, which, in the 1990s, became Integrated Product and Process Design (IPPD). Although the terms simultaneous, concurrent, and inte- grated are basically synonymous, the change in terms implies a greater refinement in thought about what it takes to efficiently develop a product. Throughout the rest of this text, the term concurrent engineering will be used to express this refinement.

In the 1990s the concepts of Lean and Six Sigma became popular in manu- facturing and began to have an influence on design. Lean manufacturing concepts were based on studies of the Toyota manufacturing system and introduced in the United States in the early 1990s. Lean manufacturing seeks to eliminate waste in all parts of the system, principally through teamwork. This means eliminating products nobody wants, unneeded steps, many different materials, and people waiting downstream because upstream activities haven’t been delivered on time. In design and manufacturing, the term “lean” has become synonymous with min- imizing the time to do a task and the material to make a product. The Lean philosophy will be refined in later chapters.

Where Lean focuses on time, Six Sigma focuses on quality. Six Sigma, some- times written as (6σ) was developed at Motorola in the 1980s and popularized in the 1990s as a way to help ensure that products were manufactured to the highest

ullman-38162 ull75741_01 December 17, 2008 11:15

10 CHAPTER 1 Why Study the Design Process?

Table 1.2 The ten key features of design best practice

1. Focus on the entire product life (Chap. 1) 2. Use and support of design teams (Chap. 3) 3. Realization that the processes are as important as the product (Chaps. 1 and 4) 4. Attention to planning for information-centered tasks (Chap. 4) 5. Careful product requirements development (Chap. 5) 6. Encouragement of multiple concept generation and evaluation (Chaps. 6 and 7) 7. Awareness of the decision-making process (Chap. 8) 8. Attention to designing in quality during every phase of the design process (throughout) 9. Concurrent development of product and manufacturing process (Chaps. 9–12)

10. Emphasis on communication of the right information to the right people at the right time (throughout and in Section 1.4.)

standards of quality. Six Sigma uses statistical methods to account for and manage product manufacturing uncertainty and variation. Key to Six Sigma methodology is the five-step DMAIC process (Define, Measure, Analyze, Improve, and Con- trol). Six Sigma brought improved quality to manufactured products. However, quality begins in the design of products, and processes, not in their manufacture. Recognizing this, the Six Sigma community began to emphasize quality earlier in the product development cycle, evolving DFSS (Design for Six Sigma) in the late 1990s.

Essentially DFSS is a collection of design best practices similar to those introduced in this book. DFSS is still an emerging discipline.

Beyond these formal methodologies, during the 1980s and 1990s many de- sign process techniques were introduced and became popular. They are essential building blocks of the design philosophy introduced throughout the book.

All of these methodologies and best practices are built around a concern for the ten key features listed in Table 1.2. These ten features are covered in the chapters shown and are integrated into the philosophy covered in this book. The primary focus is on the integration of teams of people, design tools and techniques, and in- formation about the product and the processes used to develop and manufacture it.

The use of teams, including all the “stakeholders” (people who have a concern for the product), eliminates many of the problems with the over-the-wall method. During each phase in the development of a product, different people will be important and will be included in the product development team. This mix of people with different views will also help the team address the entire life cycle of the product.

Tools and techniques connect the teams with the information.Although many of the tools are computer-based, much design work is still done with pencil and paper. Thus, the emphasis in this book is not on computer-aided design but on the techniques that affect the culture of design and the tools used to support them.

1.4 THE LIFE OF A PRODUCT Regardless of the design process followed, every product has a life history, as described in Fig. 1.7. Here, each box represents a phase in the product’s life.

ullman-38162 ull75741_01 December 17, 2008 11:15

1.4 The Life of a Product 11

. . .

Use Operate in sequence 1

Operate in sequence N

Clean

Maintain Diagnose

Test Repair

Use

Manufacture

Assemble

Distribute

Install

Production and delivery

Identify need

Plan for the design process

Develop engineering specifications

Develop concepts

Develop product

Product development

Retire

Disassemble

Reuse or recycle

End of life

Figure 1.7 The life of a product.

These phases are grouped into four broad areas. The first area concerns the development of the product, the focus of this book. The second group of phases includes the production and delivery of the product. The third group contains all the considerations important to the product’s use. And the final group focuses on what happens to the product after it is no longer useful. Each phase will be introduced in this section, and all are detailed later in the book. Note that design- ers, responsible for the first five phases, must fully understand all the subsequent phases if they are to develop a quality product.

The design phases are:

Identify need. Design projects are initiated either by a market requirement, the development of a new technology, or the desire to improve an existing product.

ullman-38162 ull75741_01 December 17, 2008 11:15

12 CHAPTER 1 Why Study the Design Process?

The design process not only gives birth to a product but is also responsible for its life and death.

Plan for the design process. Efficient product development requires plan- ning for the process to be followed. Planning for the design process is the topic of Chap. 4. Develop engineering requirements. The importance of developing a good set of specifications has become one of the key points in concurrent engi- neering. It has recently been realized that the time spent evolving complete specifications prior to developing concepts saves time and money and im- proves quality. A technique to help in developing specifications is covered in Chap. 6. Develop concepts. Chapters 7 and 8 focus on techniques for generating and evaluating new concepts. This is an important phase in the development of a product, as decisions made here affect all the downstream phases. Develop product. Turning a concept into a manufacturable product is a ma- jor engineering challenge. Chapters 9–12 present techniques to make this a more reliable process. This phase ends with manufacturing specifications and release to production.

These first five phases all must take into account what will happen to the product in the remainder of its lifetime. When the design work is completed, the product is released for production, and except for engineering changes, the design engineers will have no further involvement with it.

The production and delivery phases include:

Manufacture. Some products are just assemblies of existing components. For most products, unique components need to be formed from raw materials and thus require some manufacturing. In the over-the-wall design philoso- phy, design engineers sometimes consider manufacturing issues, but since they are not experts, they sometimes do not make good decisions. Concur- rent engineering encourages having manufacturing experts on the design team to ensure that the product can be produced and can meet cost require- ments. The specific consideration of design for manufacturing and product cost estimation is covered in Chap. 11. Assemble. How a product is to be assembled is a major consideration dur- ing the product design phase. Part of Chap. 11. is devoted to a technique called design for assembly, which focuses on making a product easy to assemble. Distribute. Although distribution may not seem like a concern for the design engineer, each product must be delivered to the customer in a safe and cost- effective manner. Design requirements may include the need for the product to be shipped in a prespecified container or on a standard pallet. Thus, the

ullman-38162 ull75741_01 December 17, 2008 11:15

1.4 The Life of a Product 13

design engineers may need to alter their product just to satisfy distribution needs. Install. Some products require installation before the customer can use them. This is especially true for manufacturing equipment and building industry products. Additionally, concern for installation can also mean concern for how customers will react to the statement, “Some assembly required.”

The goal of product development, production, and delivery is the use of the product. The “Use” phases are:

Operate. Most design requirements are aimed at specifying the use of the product. Products may have many different operating sequences that describe their use. Consider as an example a common hammer that can be used to put in nails or take them out. Each use involves a different sequence of operations, and both must be considered during the design of a hammer. Clean. Another aspect of a product’s use is keeping it clean. This can range from frequent need (e.g., public bathroom fixtures) to never. Every consumer has experienced the frustration of not being able to clean a product. This inability is seldom designed into the product on purpose; rather, it is usually simply the result of poor design. Maintain. As shown in Fig. 1.7, to maintain a product requires that problems must be diagnosed, the diagnosis may require tests, and the product must be repaired.

Finally, every product has a finite life. End-of-life concerns have become increasingly important.

Retire. The final phase in a product’s life is its retirement. In past years de- signers did not worry about a product beyond its use. However, during the 1980s increased concern for the environment forced designers to begin con- sidering the entire life of their products. In the 1990s the European Union enacted legislation that makes the original manufacturer responsible for col- lecting and reusing or recycling its products when their usefulness is finished. This topic will be further discussed in Section 12.8. Disassemble. Before the 1970s, consumer products could be easily disas- sembled for repair, but now we live in a “throwaway” society, where disas- sembly of consumer goods is difficult and often impossible. However, due to legislation requiring us to recycle or reuse products, the need to design for disassembling a product is returning. Reuse or recycle. After a product has been disassembled, its parts can either be reused in other products or recycled—reduced to a more basic form and used again (e.g., metals can be melted, paper reduced to pulp again).

This emphasis on the life of a product has resulted in the concept of Prod- uct Life-cycle Management (PLM). The term PLM was coined in the fall of 2001 as a blanket term for computer systems that support both the definition or authoring of product information from cradle to grave. PLM enables management

ullman-38162 ull75741_01 December 17, 2008 11:15

14 CHAPTER 1 Why Study the Design Process?

of this information in forms and languages understandable by each constituency in the product life cycle—namely, the words and representations that the engineers understand are not the same as what manufacturing or service people understand.

Apredecessor to PLM was Product Data Management (PDM), which evolved in the 1980s to help control and share the product data. The change from “data” in PDM to life cycle in PLM reflects the realization that there is more to a prod- uct than the description of its geometry and function—the processes are also important.

As shown in Fig. 1.8, PLM integrates six different major types of information. In the past these were separate, and communications between the communities

Layout

MCAD

ECAD

Design Automation

Systems Engineering

Bills of Materials

Solid models

Features

Functions

Architecture

Signals and connections

Simulations

Needs Customer

Environment

Regulations

Drawings

Software

DFA

DFM

Manufacturing Engineering

Service, Diagnosis, Warrantee

Portfolio Planning

Product Life-cycle Management (PLM)

Assembly

Detail

Figure 1.8 Product Life-cycle Management.

ullman-38162 ull75741_01 December 17, 2008 11:15

1.5 The Many Solutions for Design Problems 15

was poor (think of the over-the-wall method, Fig. 1.6). Whereas Fig. 1.7 focuses on the activities that happen during a product’s life, PLM, Fig. 1.8 focuses on the information that must be managed to support that life. What PLM calls “Sys- tems Engineering” is support for the technical development of the function of the product. The topics listed under Systems Engineering are all covered in this book.

What historically was called CAD (Computer-Aided Design) is now often referred to as MCAD for Mechanical CAD to differentiate it from Electronic CAD (ECAD). These two, along with software are all part of design automation. Like most of PLM, this structure grew from the twigs to the root of the tree. Traditional drawings included layout and detailed and assembly drawings. The advent of solid models made them a part of an MCAD system.

Bills Of Materials (BOMs) are effectively parts lists. BOMs are fundamen- tal documents for manufacturing. However, as product is evolving in systems engineering so does the BOM; early on there may be no parts to list. In manufac- turing, PLM manages information about Design For Manufacturing (DFM) and Assembly (DFA).

Once the product is launched and in use, there is a need to maintain it, or as shown in Fig. 1.7, diagnose, test, and repair it. These activities are supported by service, diagnosis, and warrantee information in a PLM system. Finally, there is need to manage the product portfolio—namely, of the products that could be offered, which ones are chosen to be offered (the organization’s portfolio). Portfolio decisions are the part of doing business that determines which products will be developed and sold.

This description of the life of a product and systems to manage it, gives a good basic understanding of the issues that will be addressed in this book. The rest of this chapter details the unique features of design problems and their solution processes.

1.5 THE MANY SOLUTIONS FOR DESIGN PROBLEMS

Consider this problem from a textbook on the design of machine components (see Fig. 1.9):

What size SAE grade 5 bolt should be used to fasten together two pieces of 1045 sheet steel, each 4 mm thick and 6 cm wide, which are lapped over each other and loaded with 100 N?

Figure 1.9 A simple lap joint.

ullman-38162 ull75741_01 December 17, 2008 11:15

16 CHAPTER 1 Why Study the Design Process?

Design problems have many satisfactory solutions but no clear best solution.

In this problem the need is very clear, and if we know the methods for analyzing shear stress in bolts, the problem is easily understood. There is no necessity to design the joint because a design solution is already given, namely, a grade 5 bolt, with one parameter to be determined—its diameter. The product evaluation is straight from textbook formulas, and the only decision made is in determining whether we did the problem correctly.

In comparison, consider this, only slightly different, problem:

Design a joint to fasten together two pieces of 1045 sheet steel, each 4 mm thick and 6 cm wide, which are lapped over each other and loaded with 100 N.

The only difference between these problems is in their opening clauses (shown in italics) and a period replacing the question mark (you might want to think about this change in punctuation). The second problem is even easier to understand than the first; we do not need to know how to design for shear failure in bolted joints. However, there is much more latitude in generating ideas for potential concepts here. It may be possible to use a bolted joint, a glued joint, a joint in which the two pieces are folded over each other, a welded joint, a joint held by magnets, a Velcro joint, or a bubble-gum joint. Which one is best depends on other, unstated factors. This problem is not as well defined as the first one. To evaluate pro- posed concepts, more information about the joint will be needed. In other words, the problem is not really understood at all. Some questions still need to be an- swered: Will the joint require disassembly? Will it be used at high temperatures? What tools are available to make the joint? What skill levels do the joint manu- facturers have?

The first problem statement describes an analysis problem. To solve it we need to find the correct formula and plug in the right values. The second statement describes a design problem, which is ill-defined in that the problem statement does not give all the information needed to find the solution. The potential solutions are not given and the constraints on the solution are incomplete. This problem requires us to fill in missing information in order to understand it fully.

Another difference between the two problems is in the number of potential solutions. For the first problem there is only one correct answer. For the second there is no correct answer. In fact, there may be many good solutions to this problem, and it may be difficult if not impossible to define what is meant by the “best solution.” Just consider all the different cars, televisions, and other products that compete in the same market. In each case, all the different models solve essentially the same problem, yet there are many different solutions. The goal in design is to find a good solution that leads to a quality product with the least commitment of time and other resources. All design problems have a multitude of satisfactory solutions and no clear best solution. This is shown graphically

ullman-38162 ull75741_01 December 17, 2008 11:15

1.6 The Basic Actions of Problem Solving 17

Design process knowledge

Design need

Design process paths

Resulting products that meet the need

Physics

Materials science

Engineering science

Engineering economics

Domain knowledge

Manufacturing processes

Welding design

Thermodynamics

Kinematics Pumps

Electric motors

Figure 1.10 The many results of the design process.

in Fig. 1.10 where the factors that affect exactly what solution is developed are noted. Domain knowledge is developed through the study of engineering physics and other technical areas and through the observation of existing products. It is the study of science and engineering science that provides the basis on which the design process is based. Design process knowledge is the subject of this book.

For mechanical design problems in particular, there is an additional char- acteristic: the solution must be a piece of working hardware—a product. Thus, mechanical design problems begin with an ill-defined need and result in an object that behaves in a certain way, a way that the designers feel meets this need. This creates a paradox. A designer must develop a device that, by definition, has the capabilities to meet some need that is not fully defined.

1.6 THE BASIC ACTIONS OF PROBLEM SOLVING

Regardless of what design problem we are solving, we always, consciously or unconsciously, take six basic actions:

1. Establish the need or realize that there is a problem to be solved. 2. Plan how to solve the problem.

ullman-38162 ull75741_01 December 17, 2008 11:15

18 CHAPTER 1 Why Study the Design Process?

3. Understand the problem by developing requirements and uncovering existing solutions for similar problems.

4. Generate alternative solutions. 5. Evaluate the alternatives by comparing them to the design requirements and

to each other. 6. Decide on acceptable solutions.

This model fits design whether we are looking at the entire product (see the product life-cycle diagram, Fig. 1.7) or the smallest detail of it.

These actions are not necessarily taken in 1-2-3 order. In fact they are often in- termingled with solution generation and evaluation improving the understanding of the problem, enabling new, improved solutions to be generated. This iterative nature of design is another feature that separates it from analysis.

The list of actions is not complete. If we want anyone else on the design team to make use of our results, a seventh action is also needed:

7. Communicate the results.

The need that initiates the process may be very clearly defined or ill-defined. Consider the problem statements for the design of the simple lap joint of two pieces of metal given earlier (Fig. 1.9). The need was given by the problem statement in both cases. In the first statement, understanding is the knowledge of what parameters are needed to characterize a problem of this type and the equations that relate the parameters to each other (a model of the joint). There is no need to generate potential solutions, evaluate them, or make any decision, because this is an analysis problem. The second problem statement needs work to understand. The requirements for an acceptable solution must be developed, and then alternative solutions can be generated and evaluated. Some of the evaluation may be the same as the analysis problem, if one of the concepts is a bolt.

Some important observations:

■ New needs are established throughout the design effort because new design problems arise as the product evolves. Details not addressed early in the process must be dealt with as they arise; thus, the design of these details poses new subproblems.

■ Planning occurs mainly at the beginning of a project. Plans are always updated because understanding is improved as the process progresses.

■ Formal efforts to understand new design problems continue throughout the process. Each new subproblem requires new understanding.

■ There are two distinct modes of generation: concept generation and product generation. The techniques used in these two actions differ.

■ Evaluation techniques also depend on the design phase; there are differ- ences between the evaluation techniques used for concepts and those used for products.

■ It is difficult to make decisions, as each decision requires a commitment based on incomplete evaluation. Additionally, since most design problems

ullman-38162 ull75741_01 December 17, 2008 11:15

1.7 Knowledge and Learning During Design 19

are solved by teams, a decision requires consensus, which is often difficult to obtain.

■ Communication of the information developed to others on the design team and to management is an essential part of concurrent engineering.

We will return to these observations as the design process is developed through this text.

1.7 KNOWLEDGE AND LEARNING DURING DESIGN

When a new design problem is begun, very little may be known about the solution, especially if the problem is a new one for the designer. As work on the project progresses, the designer’s knowledge about the technologies involved and the alternative solutions increases, as shown in Fig. 1.11. Therefore, after completing a project, most designers want a chance to start all over in order to do the project properly now that they fully understand it. Unfortunately, few designers get the opportunity to redo their projects.

Throughout the solution process knowledge about the problem and its po- tential solutions is gained and, conversely, design freedom is lost. This can also be seen in Fig. 1.11, where the time into the design process is equivalent to ex- posure to the problem. The curve representing knowledge about the problem is a learning curve; the steeper the slope, the more knowledge is gained per unit time. Throughout most of the design process the learning rate is high. The second curve in Fig. 1.11 illustrates the degree of design freedom. As design decisions are made, the ability to change the product becomes increasingly limited. At the beginning the designer has great freedom because few decisions have been made and little capital has been committed. But by the time the product is in production,

Time into design process

Pe rc

en ta

ge

0

20

40

60

80

100

Design freedom

Knowledge about the design problem

Figure 1.11 The design process paradox.

ullman-38162 ull75741_01 December 17, 2008 11:15

20 CHAPTER 1 Why Study the Design Process?

A design paradox: The more you learn the less freedom you have to use what you know.

any change requires great expense, which limits freedom to make changes. Thus, the goal during the design process is to learn as much about the evolving prod- uct as early as possible in the design process because during the early phases changes are least expensive.

1.8 DESIGN FOR SUSTAINABILITY It is important to realize that design engineers have much control over what products are designed and how they interact with the earth over their lifetime. The responsibility that goes with designing is well summarized in the Hannover Principles. These were developed for EXPO 2000, The World’s Fair in Hannover, Germany. These principles define the basics of Designing For Sus- tainability (DFS) or Design For the Environment (DFE). DFS requires awareness of the short- and long-term consequences of your design decisions.

The Hannover Principles aim to provide a platform on which designers can consider how to adapt their work toward sustainable ends.According to the World Commission on Environment and Development, the high-level goal is “Meeting the needs of the present without compromising the ability of future generations to meet their own needs.”

The Hannover Principles are:

1. Insist on rights of humanity and nature to coexist in a healthy, supportive, diverse, and sustainable condition.

2. Recognize interdependence. The elements of human design interact with and depend on the natural world, with broad and diverse implications at every scale. Expand design considerations to recognizing even distant effects.

3. Accept responsibility for the consequences of design decisions on human well-being, the viability of natural systems and their right to coexist.

4. Create safe objects of long-term value. Do not burden future generations with requirements for maintenance or vigilant administration of potential danger due to the careless creation of products, processes, or standards.

5. Eliminate the concept of waste. Evaluate and optimize the full life cycle of products and processes to approach the state of natural systems in which there is no waste.

6. Rely on natural energy flows. Human designs should, like the living world, derive their creative forces from perpetual solar income. Incorporate this energy efficiently and safely for responsible use.

7. Understand the limitations of design. No human creation lasts forever and design does not solve all problems. Those who create and plan should practice

ullman-38162 ull75741_01 December 17, 2008 11:15

1.9 Summary 21

You are responsible for the impact of your products on others.

humility in the face of nature. Treat nature as a model and mentor, not as an inconvenience to be evaded or controlled.

8. Seek constant improvement by the sharing of knowledge. Encourage di- rect and open communication between colleagues, patrons, manufacturers, and users to link long-term sustainable considerations with ethical responsi- bility, and reestablish the integral relationship between natural processes and human activity.

9. Respect relationships between spirit and matter. Consider all aspects of human settlement including community, dwelling, industry, and trade in terms of existing and evolving connections between spiritual and material consciousness.

We will work to respect these principles in the chapters that follow. We intro- duced the concept of “lean” earlier in this chapter as the effort to reduce waste (Principle 5). We will revisit this and the other principles throughout the book. In Chap. 11, we will specifically revisit DFS as part of Design for the Environ- ment. In Chap. 12, we focus on product retirement. Many products are retired to landfills, but in keeping with the first three principles, and focusing on the fifth principle, it is best to design products that can be reused and recycled.

1.9 SUMMARY The design process is the organization and management of people and the infor- mation they develop in the evolution of a product.

■ The success of the design process can be measured in the cost of the design effort, the cost of the final product, the quality of the final product, and the time needed to develop the product.

■ Cost is committed early in the design process, so it is important to pay par- ticular attention to early phases.

■ The process described in this book integrates all the stakeholders from the beginning of the design process and emphasizes both the design of the product and concern for all processes—the design process, the manufacturing process, the assembly process, and the distribution process.

■ All products have a life cycle beginning with establishing a need and ending with retirement. Although this book is primarily concerned with plan- ning for the design process, engineering requirements development, concep- tual design, and product design phases, attention to all the other phases is important. PLM systems are designed to support life-cycle information and communication.

ullman-38162 ull75741_01 December 17, 2008 11:15

22 CHAPTER 1 Why Study the Design Process?

■ The mechanical design process is a problem-solving process that transforms an ill-defined problem into a final product.

■ Design problems have more than one satisfactory solution. ■ Design for Sustainability embodied in the Hannover Principles is becoming

an increasingly important part of the design process.

1.10 SOURCES Creveling, C. M., Dave Antis, and Jeffrey Lee Slutsky: Design for Six Sigma in Technology

and Product Development, Prentice Hall PTR, 2002. A good book on DFSS.

Ginn, D., and E. Varner: The Design for Six Sigma Memory Jogger, Goal/QPC, 2004. A quick introduction to DFSS

The Hannover Principles, Design for Sustainability. Prepared for EXPO 2000, Hannover, Germany, http://www.mcdonough.com/principles.pdf

Product life-cycle management (PLM) description based on work at Siemens PLM supplied by Wayne Embry their PLM Functional Architect. http://www.plm.automation.siemens.com/en_us/products/teamcenter/index.shtml http://www.johnstark.com/epwl4.html PLM listing of over 100 vendors.

Ulrich, K. T., and S. A. Pearson: “Assessing the Importance of Design through Product Archaeology,” Management Science, Vol. 44, No. 3, pp. 352–369, March 1998, or “Does Product Design Really Determine 80% of Manufacturing Cost?” working paper 3601–93, Sloan School of Management, MIT, Cambridge, Mass., 1993. In the first edition of The Mechanical Design Process it was stated that design determined 80% of the cost of a product. To confirm or deny that statement, researchers at MIT performed a study of automatic coffeemakers and wrote this paper. The results show that the number is closer to 50% on the average (see Fig. 1.3) but can range as high as 75%.

Womack, James P., and Daniel T. Jones: Lean Thinking: Banish Waste and Create Wealth in Your Corporation, Simon and Schuster, New York, 1996.

1.11 EXERCISES 1.1 Change a problem from one of your engineering science classes into a design problem.

Try changing as few words as possible.

1.2 Identify the basic problem-solving actions for

a. Selecting a new car

b. Finding an item in a grocery store

c. Installing a wall-mounted bookshelf

d. Placing a piece in a puzzle

1.3 Find examples of products that are very different yet solve exactly the same design problem. Different brands of automobiles, bikes, CD players, cheese slicers, wine bot- tle openers, and personal computers are examples. For each, list its features, cost, and perceived quality.

1.4 How well do the products in Exercise 1.3 meet the Hannover Principles?

1.5 To experience the limitations of the over-the-wall design method try this. With a group of four to six people, have one person write down the description of some object that is

http://www.mcdonough.com/principles.pdf
http://www.plm.automation.siemens.com/en_us/products/teamcenter/index.shtml
http://www.johnstark.com/epwl4.html
ullman-38162 ull75741_01 December 17, 2008 11:15

1.11 Exercises 23

not familiar to the others. This description should contain at least six different nouns that describe different features of the object. Without showing the description to the others, describe the object to one other person in such a manner that the others can’t hear. This can be done by whispering or leaving the room. Limit the description to what was written down. The second person now conveys the information to the third person, and so on until the last person redescribes the object to the whole group and compares it to the original written description. The modification that occurs is magnified with more complex objects and poorer communication. (Professor Mark Costello of Georgia Institute of Technology originated this problem.)

ullman-38162 ull75741_01 December 17, 2008 11:15

ullman-38162 ull75741_02 December 17, 2008 16:16

2C H A P T E R Understanding Mechanical

Design

KEY QUESTIONS ■ What is the difference between function, behavior, and performance? ■ Why does mechanical design flow from function to form? ■ What are the languages of mechanical design? ■ Are all design problems the same? ■ What can you learn from dissecting products?

2.1 INTRODUCTION For most of history, the discipline of mechanical design required knowledge of only mechanical parts and assemblies. But early in the twentieth century, elec- trical components were introduced in mechanical devices. Then, during World War II, in the 1940s, electronic control systems became part of the mix. Since this change, designers have often had to choose between purely mechanical sys- tems and systems that were a mix of mechanical and electronic components and systems. These electronic systems have matured from very simple functions and logic to the incorporation of computers and complex logic. Many electrome- chanical products now include microprocessors. Consider, for example, cameras, office copiers, cars, and just about everything else. Systems that have mechanical, electronic, and software components are often called mechatronic devices. What makes the design of these devices difficult is the necessity for domain and design process knowledge in three overlapping but clearly different disciplines. But, no matter how electronic or computer-centric devices become, nearly all products require mechanical functions and a mechanical interface with humans. Addition- ally, all products require mechanical machinery for manufacture and assembly

25

ullman-38162 ull75741_02 December 17, 2008 16:16

26 CHAPTER 2 Understanding Mechanical Design

and mechanical components for housing. Thus, no matter how “smart” products become, there will always be the need for mechanical design.

To explore systems that have significant mechanical components consider two examples that will be used throughout the book, the Irwin Quick-Grip clamp (Fig. 2.1) and the drive wheel assembly for the NASA Mars Exploration Rover (MER) developed by Cal Tech’s Jet Propulsion Laboratory (JPL) (Fig. 2.2).

Figure 2.1 Irwin Quick-Grip clamp. (Reprinted with permission of Irwin Industrial Tools.)

Figure 2.2 The Mars Exploration Rover being tested by JPL engineers. (Reprinted with permission of NASA/JPL.)

ullman-38162 ull75741_02 December 17, 2008 16:16

2.1 Introduction 27

Irwin is one of the largest manufacturers of one-handed bar clamps. What makes the model shown in Fig 2.1 unique is that it can generate over 550 lb (250 kg) of force with the strength of only one hand. Irwin introduced this product in 2006 and sells many tens of thousands of them a month. In contrast to the purely mechanical, high-production-volume Quick-Grip, only two MERs were made and they are highly mechatronic.

The two MERs were launched toward Mars on June 10 and July 7, 2003, in search of answers about the history of water on Mars. They landed on Mars January 3 and January 24, 2004. They were designed for 90 Sol (Martian days, about 40 min longer than an Earth day) and were still operating in 2008, over 1300 Sols (over 3.5 years) past their design life. One of the Rovers, Opportunity, had traveled over 11 km (7.1 mi) during its five years of life.

Each Rover is a six-wheeled, solar-powered robot that stands 1.5 m (4.9 ft) high and is 2.3 m (7.5 ft) wide and 1.6 m (5.2 ft) long. They weigh 180 kg (400 lb) on Earth, 35 kg (80 lb) of which is the wheel and suspension system. Mars has only 38% the gravitational pull of Earth. So they weigh 68.4 kg (152 lb) on Mars. As shown in Fig. 2.3, a very simplified diagram of the MER’s systems, propulsion and steering are two of the subsystems. Later in this chapter, we delve further into the MER, and in later chapters we will detail the wheels.

In general, during the design process the function of the system and its de- composition are considered first.After the function has been decomposed into the finest subsystems possible, assemblies and components are developed to provide these functions. For mechanical devices, the general decomposition is system– subsystem–assembly–component. Figure 2.3 shows the MER propulsion system, within which the motor and transmission are two subsystems. The wheel is a component. Systems, subsystems, and components all have features, specific at- tributes that are important, such as dimensions, material properties, shapes, or functional details. For the MER propulsion system, an important feature is that it can propel the MER at 5 cm/sec. For the transmission, a feature is that it has a 1500:1 reduction ratio. For the MER wheel, some of the important features are its diameter, tread pattern, and flexibility.

We must also note that many systems have both electrical and mechanical subsystems and components. Electrical systems generally provide energy, sens- ing, and control functions. The function of these electrical systems is fulfilled by circuits (electrical assemblies) that can be decomposed into electrical com- ponents (e.g., switches, transistors, and ICs), much as with mechanical objects. Finally, some of the control functions are filled by microprocessors. Physically, these are electric circuits, but the actual control function is provided by software programs in the processor. These programs are assemblies of coding modules composed of individual coding statements. Note that the function of the micro- processor could be filled by an electrical or possibly even a purely mechanical system. During the early phases of the design process, when developing systems is the focus of the effort, it is often unclear whether the actual function will be met by mechanical assemblies, electrical circuits, software programs, or a mix of these elements.

ullman-38162 ull75741_02 December 17, 2008 16:16

28 CHAPTER 2 Understanding Mechanical Design

Solar collectors

Battery

IMU (Inertial

Measurement Unit)

REM (Rover Electronics

Module)

Drive motor (1 of 6)

Steering motor (1 of 4)

Transmission (1 of 4)

Encoder (1 of 4)

Transmission (1 of 6)

SteeringPropulsion

Wheel (1 of 6)

Figure 2.3 The MER Propulsion System showing some of the sub-systems and components.

2.2 IMPORTANCE OF PRODUCT FUNCTION, BEHAVIOR, AND PERFORMANCE

What is the function of the Irwin clamp? How does it behave? Does it have good performance? These three questions revolve around the terms “function,” “behavior,” and “performance”—similar, but different attributes of the clamp.

There are many synonyms for the word function. In mechanical engineering, we commonly use the terms function, operation, and purpose to describe what a device does. A common way of classifying mechanical devices is by their function. In fact, some devices having only one main function are named for

ullman-38162 ull75741_02 December 17, 2008 16:16

2.2 Importance of Product Function, Behavior, and Performance 29

Function determines form and form, in turn, enables function.

that function. For example, a screwdriver has the function of enabling a person to insert or remove a screw. The terms drive, insert, and remove are all verbs that tell what the screwdriver does. In telling what the screwdriver does, we have given no indication of how the screwdriver accomplishes its function. To discover how, we must have some information on the form of the device. The term form relates to any aspect of physical shape, geometry, construction, material, or size. As we shall see in Chap. 3, one of the main ways engineers mentally index their knowledge about the mechanical world is by function. Now reread this paragraph and replace the screwdriver example with the Quick-Grip clamp.

In Fig. 2.3, we physically decomposed the Mars Rover propulsion and steer- ing systems into subsystems and components at its physical boundaries. Func- tional decomposition is often much more difficult than physical decomposition, as each function may use part of many components and each component may serve many functions. Consider the handlebar of a bicycle. The handlebar is a bent piece of tubing, a single component that serves many functions. It enables the rider to “steer the bicycle” (“steer” is a verb that tells what the device does), and the handlebar “supports the rider” (again, a function telling what the handle- bar does). Further, it not only “supports the brake levers” but also “transforms (another function) the gripping force” to a pull on the brake cable. The shape of the handlebar and its relationship with other components determine how it provides all these different functions. The handlebar, however, is not the only component needed to steer the bike. Additional components necessary to per- form this function are the front fork, the bearings between the fork and the frame, the front wheel, and miscellaneous fasteners. Actually, it can be argued that all the components on a bike contribute to steering, since a bike without a seat or rear wheel would be hard to steer. In any case, the handlebar performs many different functions, but in fulfilling these functions, the handlebar is only a part of various assemblies. Similarly, the steering on the MER cannot actually steer it without the wheels in the propulsion system. The coupling between form and function makes mechanical design challenging.

Many common devices are cataloged by their function. If we want to specify a bearing, for example, we can search a bearing catalog and find many different styles of bearings (plain, ball, or tapered roller, for example). Each “style” has a different geometry—a different form—though all have the same primary func- tion, namely, to reduce friction between a shaft and another object. Cataloging is possible in mechanical design as long as the primary function is clearly defined by a single piece of hardware, either a single component or an assembly. In other words, the form and function are decomposed along the same boundaries. This is true of many mechanical devices, such as pumps, valves, heat exchangers, gearboxes, and fan blades, and is especially true of many electrical circuits and components, such as resistors, capacitors, and amplifier circuits.

ullman-38162 ull75741_02 December 17, 2008 16:16

30 CHAPTER 2 Understanding Mechanical Design

?????? To be designed

(a) Function

Known input

(b) Behavior

Physical properties of

system

Desired performance

Known input Actual performance

Figure 2.4 Function and behavior.

Two other terms often related to function are behavior and performance. Function and behavior are often used synonymously. However, there is a subtle difference, as shown in Fig. 2.4. In this figure there are two standard system blocks with an input represented by an arrow into the box, the system acted on by the input represented by the box, and the reaction of the system to the input represented by the arrow out of the box. The box in the upper part of the figure shows that function is the desired output from a system that is yet to be designed. When we begin to design a device, the device itself is unknown, but what we want it to do is known. If the system is known, as in the second part of the figure, then the behavior of the system can be found. Behavior is the actual output, the response of the system’s physical properties to the input energy or control. Thus, the behavior can be simulated or measured, whereas function is only a desire.

Performance is the measure of function and behavior—how well the device does what it is designed to do. When we say that one function of the handlebar is to steer the bicycle, we say nothing about how well it serves this purpose. Before designing a handlebar, we must develop a clear picture of its desired performance. For example, one design functional goal is that the handlebar must “support 50 kg,” a measurable desired performance for the handlebar. The development of clear performance measures is the focus of Chap. 6. Further, after designing the handlebar we can simulate its strength analytically or measure the strength of a prototype to find the actual performance for comparison to that desired. This comparison is a major focus of Chap. 10.

2.3 MECHANICAL DESIGN LANGUAGES AND ABSTRACTION

Many “languages” or representations can be used to describe a mechanical object. Consider for a moment the difference between a detailed drawing of a component and the actual hardware that is the component. Both the drawing and the hardware represent the same object; however, they each represent it in a different language.

ullman-38162 ull75741_02 December 17, 2008 16:16

2.3 Mechanical Design Languages and Abstraction 31

A skilled designer speaks many languages.

Extending this example further, if the component we are discussing is a bolt, then the word bolt is a textual (semantic or word) description of the component, a third language. Additionally, the bolt can be represented through equations (the final language) that describe its functionality and possibly its form. For example, the ability of the bolt to “carry shear stress” (a function) is described by the equation τ = F/A; the shear stress τ is equal to the shear force F on the bolt divided by the stress area A of the bolt.

Based on this, we can use four different representations or languages to describe the bolt. These four can be used to describe any mechanical object:

Semantic. The verbal or textual representation of the object—for example, the word bolt, or the sentence, “The shear stress on the bolt is the shear force divided by the stress area.” Graphical. The drawings of the object—for example, scale representations such as solid models, orthogonal drawings, sketches, or artistic renderings. Analytical. The equations, rules, or procedures representing the form or func- tion of the object—for example, τ = F/A. Physical. The hardware or a physical model of the object.

In most mechanical design problems, the initial need is expressed in a se- mantic language as a written specification or a verbal request by a customer or supervisor. The result of the design process is a physical object. Although the de- signer produces a graphical representation of the product, not the hardware itself, all the languages will be used as the product is refined from its initial, abstract semantic representation to its final physical form.

Further complicating how we refer to objects being designed, consider two drawings for a MER wheel, as shown in Fig. 2.5. Figure 2.5a is a rough sketch, which gives only abstract information about the component. It centers on the

(a)

Figure 2.5 Abstract sketch and solid model of a MER wheel.

ullman-38162 ull75741_02 December 17, 2008 16:16

32 CHAPTER 2 Understanding Mechanical Design

function of the wheel’s spokes to act like springs. Figure 2.5b is a solid model of the same component, focused on the final form of the wheel. In progressing from the sketch to the solid model, the level of abstraction of the device is refined.

Some design process techniques are better suited for abstract levels and others for levels that are more concrete. There are no true levels of abstractions, but rather a continuum on which the form or function can be represented. Descriptions of three levels of abstraction in each of the four languages are given in Table 2.1. The object we call a bolt is used as an example in Table 2.2.

Another term that is often used in describing the analytical row in Table 2.1 is simulation fidelity. As analytical models or simulations increase in fidelity, their representation of the actual object or system becomes a more accurate represen- tation of reality. Simulation fidelity will be further refined in Chap. 10.

The process of making an object less abstract (or more concrete) is called refinement. Mechanical design is a continuous process of refining the given needs

Table 2.1 Levels of abstraction in different languages

Level of abstraction

Language Abstract −−−−−−→ Concrete Semantic Qualitative words Reference to Reference to the values

(e.g., long, fast, specific parameters of the specific parameters lightest) or components or components

Graphical Rough sketches Scale drawings Solid models with tolerances

Analytical Qualitative relations Back-of-the-envelope Detailed analysis (e.g., left of) calculations

Physical None Models of the product Final hardware

Table 2.2 Levels of abstraction in describing a bolt

Level of abstraction

Language Abstract −−−−−−→ Concrete Semantic A bolt A short bolt A 1′′ 1/4−20

UNC Grade 5 bolt

Graphical

Length of bolt

Length of thread

Body diameter

-UNC-2A58

Analytical Right-hand rule τ = F/A τ = F/A

Physical — —

ullman-38162 ull75741_02 December 17, 2008 16:16

2.4 Different Types of Mechanical Design Problems 33

to the final hardware. The refinement of the bolt in Table 2.4 is illustrated on a left-to-right continuum. In most design situations, the beginning of the problem appears in the upper left corner and the final product in the lower right. The path connecting these is a mix of the other representations and levels of abstractions.

2.4 DIFFERENT TYPES OF MECHANICAL DESIGN PROBLEMS

Traditionally, we decompose mechanical engineering by discipline: fluids, ther- modynamics, mechanics, and so on. In categorizing the types of mechanical design problems, this discipline-oriented approach is not appropriate. Consider, for example, the simplest kind of design problem, a selection design problem. Selection design means picking one (maybe more) item from a list such that the chosen item meets certain requirements. Common examples are selecting the cor- rect bearing from a bearings catalog, selecting the correct lenses for an optical device, selecting the proper fan for cooling equipment, or selecting the proper heat exchanger for a heating or cooling process. The design process for each of these problems is essentially the same, even though the disciplines are very dif- ferent. The goal of this section is to describe different types of design problems independently of the discipline.

Before beginning, we must realize that most design situations are a mix of various types of problems. For example, we might be designing a new type of consumer product that will accept a whole raw egg, break it, fry it, and deliver it on a plate. Since this is a new product, there will be a lot of original design work to be done. As the design process proceeds, we will configure the various parts. To determine the thickness of the frying surface we will analyze the heat conduction of the frying component, which is parametric design. And we will select a heating element and various fasteners to hold the components together. Further, if we are clever, we may be able to redesign an existing product to meet some or all of the requirements. Each of the italicized terms is a different type of design problem. It is rare to find a problem that is purely one type.

2.4.1 Selection Design

Selection design involves choosing one item (or maybe more) from a list of similar items. We do this type of design every time we choose an item from a catalog. It may sound simple, but if the catalog contains more than a few items and there are many different features to the items, the decision can be quite complex.

To solve a selection problem we must start with a clear need. The catalog or the list of choices then effectively generates potential solutions for the problem. We must evaluate the potential solutions with respect to our specific requirements to make the right choice. Consider the following example. During the process of designing a product, an engineer must select a bearing to support a shaft. The known information is given in Fig. 2.6. The shaft has a diameter of 20 mm (0.787 in.). There is a radial force of 6675 N (1500 lb) on the shaft at the bearing,

ullman-38162 ull75741_02 December 17, 2008 16:16

34 CHAPTER 2 Understanding Mechanical Design

Bearing

Shaft

2000 rpm

Housing

6675 N

20 m

m

Figure 2.6 Load on a shaft.

and the shaft rotates at a maximum of 2000 rpm. The housing to support the bearing is still to be designed. All we need to do is select a bearing to meet the needs. The information on shaft size, maximum radial force, and maximum rpm given in bearing catalogs enables us to quickly develop a list of potential bearings (Table 2.3). This is the simplest type of design problem we could have, but it is still incompletely defined. We do not have enough information to select among the five possible choices. Even if a short list is developed—the most likely candidates being the 42-mm-deep groove ball bearing and the 24-mm needle bearing—there is no way to make a good decision without more knowledge of the function of the bearing and of the engineering requirements on it.

2.4.2 Configuration Design

A slightly more complex type of design is called configuration or packaging design. In this type of problem, all the components have been designed and the problem is how to assemble them into the completed product. Essentially, this type of design is similar to playing with an Erector set or other construction toy, or arranging living-room furniture.

Consider packaging of the assemblies in the MER. The body of the MER is made up of a Rover Equipment Deck (RED) where all the experiments are mounted, a Rover Electronics Module (REM), an Inertial Measurement Unit (IMU), a Warm Electronics Box (WEB), a battery, a UHF radio, an X-band telecom HW, and a Solid-State Power Amplifier (SSPA), as shown in Fig. 2.7. Each of these assemblies is of known size and has certain constraints on its position. For example, the RED must be on top and the WEB on the bottom, but

ullman-38162 ull75741_02 December 17, 2008 16:16

2.4 Different Types of Mechanical Design Problems 35

Table 2.3 Potential bearings for a shaft

Outside Load Speed Catalog Type diameter (mm) Width (mm) rating (lb) limit (rpm) number

Deep-groove 42 8 1560 18,000 6000 ball bearing 47 14 2900 15,000 6204

52 15 3900 9000 6304

Angular-contact 47 14 3000 13,000 7204 ball bearing 37 9 1960 34,000 71,904

Roller 47 14 6200 13,000 204 bearing 52 15 7350 13,000 220

Needle 24 20 1930 13,000 206 bearing 26 12 2800 13,000 208

Nylon 23 Variable 290 10 4930 bushing

... ...

8 500

“Front”

+Y +Z

+X

RED

X-band telecom HW

REM

SSPA

WEB

Battery

UHF radio

IMU

Figure 2.7 The major assemblies in the MER.

ullman-38162 ull75741_02 December 17, 2008 16:16

36 CHAPTER 2 Understanding Mechanical Design

many of the other major assemblies can be anywhere inside the envelop defined by these two.

Configuration design answers the question, How do we fit all the assemblies in an envelop? or Where do we put what? One methodology for solving this type of problem is to randomly select one component from the list and position it so that all the constraints on that assembly are met. We could start with the REM in the middle, then we select and place a second component. This procedure is continued until either we run into a conflict or all the components are in the MER. If a conflict arises, we back up and try again. For many configuration problems, some of the components to be fit into the assembly can be altered in size, shape, or function, giving the designer more latitude to determine potential configurations and making the problem solution more difficult. There are other methods to configure assemblies. They will be covered in Chap. 11.

2.4.3 Parametric Design

Parametric design involves finding values for the features that characterize the object being studied.This may seem easy enough—just find some values that meet the requirements. However, consider a very simple example. We want to design a cylindrical storage tank that must hold 4 m3 of liquid. This tank is described by the parameters r, its radius, and l, its length and its volume is determined by

V = π r2l Given a volume equal to 4 m3, then

r2l = 1.273 We can see that an infinite number of values for the radius and length will

satisfy this equation. To what values should the parameters be set? The answer is not obvious, nor even completely defined with the information given. (This problem will be readdressed in Chap. 10, where the accuracy to which the radius and the length can be manufactured will be used to help find the best values for the parameters.)

Let us extend the concept further. It may be that instead of a simple equation, a whole set of equations and rules govern the design. Consider the instance in which a major manufacturer of copying machines had to design paper-feed mechanisms for each new copier. (A paper feed is a set of rollers, drive wheels, and baffles that move a piece of paper from one location to another in the machine.) Many parameters—the number of rollers, their positions, the shape of the baffles, and the like—characterize this particular design problem, but obviously there are certain similarities in paper feeders, regardless of the relative positions of the beginning and end points of the paper, the obstructions (other components in the machine) that must be cleared, and the size and weight of the paper. The company developed a set of equations and rules to aid designers in developing workable paper paths, and using this information, the designers could generate values for parameters in new products.

ullman-38162 ull75741_02 December 17, 2008 16:16

2.4 Different Types of Mechanical Design Problems 37

2.4.4 Original Design

Any time the design problem requires the development of a process, assembly, or component not previously in existence it calls for an original design. (It can be said that if we have never seen a wheel and we design one, then we have an original design.) Though most selection, configuration, and parametric problems are represented by equations, rules, or some other logical scheme, original de- sign problems usually cannot be reduced to any algorithm. Each one represents something new and unique.

In many ways the other types of design problems—selection, configuration, and parametric—are simply constrained subsets of an original design. The po- tential solutions are limited to a list, an arrangement of components, or a set of related characterizing values. Thus, if we have a clear methodology for perform- ing original design, we should be able to solve any design problem with a more limited set of potential solutions.

2.4.5 Redesign

Most design problems solved in industry are for the redesign of an existing prod- uct. Suppose a manufacturer of hydraulic cylinders makes a product that is 0.25 m long. If the customer needs a cylinder 0.3 m long, the manufacturer might lengthen the outer cylinder and the piston rod to meet this special need. These changes may require only parameter changes, or they may require something more extensive. What if the materials are not available in the needed length, or cylinder fill time becomes too slow with the added length? Then the redesign effort may require much more than parameter changes. Regardless of the change, this is an example of redesign, the modification of an existing product to meet new requirements.

Many redesign problems are routine; the design domain is so well understood that the method used can be put in a handbook as a series of formulas or rules. The parameter changes in the example of the hydraulic cylinder are probably routine for the manufacturer.

The hydraulic cylinder can also be used as an example of a mature design, in that it has remained virtually unchanged over many years. There are many examples of mature designs in our everyday lives: pencil sharpeners, hole punches, and staplers are a few found on the average desk. For these products, knowledge about the design problem is high. There is little more to learn.

However, consider the bicycle. The basic configuration of the bicycle—the two tensioned, spoked wheels of equal diameter, the diamond-shaped frame, and the chain drive—was fairly refined late in the nineteenth century. While the 1890 Humber shown in Fig. 2.8 looks much like a modern bicycle, not all bicycles of this era were of this configuration. The Otto dicycle, shown in Fig. 2.9, had two spoked wheels and a chain; stopping and steering this machine must have been a challenge. In fact, the technology of bicycle design was so well developed by the end of the nineteenth century that a major book on the subject, Bicycles and Tricycles: An Elementary Treatise on Their Design and Construction, was

ullman-38162 ull75741_02 December 17, 2008 16:16

38 CHAPTER 2 Understanding Mechanical Design

Figure 2.8 1890 Humber bicycle.

Figure 2.9 The Otto dicycle.

ullman-38162 ull75741_02 December 17, 2008 16:16

2.4 Different Types of Mechanical Design Problems 39

Figure 2.10 The Marin Mount Vision. (Reprinted with permission of Marin Bicycles.)

Most design problems are redesign problems since they are based on prior, similar solutions. Conversely, most design

problems are original as they contain something new that makes prior solutions inadequate.

published in 1896.1 The only major change in bicycle design since the publication of that book was the introduction of the derailleur in the 1930s.

However, in the 1980s the traditional bicycle design began to change again. For example, the mountain bike shown in Fig. 2.10 no longer has a diamond- shaped frame. Why did a mature design like a bicycle begin evolving again? First, customers are always looking for improved performance. Bicycles of the style shown in Fig. 2.10 are better able to handle rough terrain than traditional bikes. Second, there is improved understanding of human comfort, ergonomics, and suspensions. Third, customers are always looking for something new and exciting even if performance is not greatly improved. Fourth, materials and components have improved.

The point is that even mature designs change to meet new needs, to attract new customers, or to take advantage of new materials. Part of the design of a new bicycle like the Marin Mount Vision is routine, and part is original. Additionally,

1The book, written by Archibald Sharp, has recently (1977) been reprinted by the MIT Press, Cambridge, Mass.

ullman-38162 ull75741_02 December 17, 2008 16:16

40 CHAPTER 2 Understanding Mechanical Design

many subproblems were parametric problems, selection problems, and configu- ration problems. Thus, the redesign of a product, even a mature one, may require a wide range of design activity.

2.4.6 Variant Design

Sometimes companies will produce a large number of variants as their products. A variant is a customized product designed to meet the needs of the customer. For example, when you order a new computer from companies such as Dell, you can specify one of three graphics cards, two battery configurations, three communication options, and two levels of memory. Any combination of these is a variant that is specifically tuned to your needs. Also, Volvo trucks estimates that of the 50,000 parts it has in its inventory it annually supplies over 5000 variants, different truck models specifically assembled to meet the needs of the customer.

2.4.7 Conceptual Design and Product Design

Two other terms that will be used throughout the book are conceptual design and product design. These are catchall terms for two parts of the product development process. First, you must develop a concept and then refine the concept into a product. The activities during the conceptual and product development phases may make use of original, parametric, and selection design and redesign as needed.

2.5 CONSTRAINTS, GOALS, AND DESIGN DECISIONS

The progression from the initial need (the design problem) to the final product is made in increments punctuated by design decisions. Each design decision changes the design state. The state of a product is a snapshot of all the information known about it at any given time during the process. In the beginning, the design state is just the problem statement. During the process, the design state is a collection of all the knowledge, drawings, models, analyses, and notes thus far generated.

Two different views can be taken of how the design process progresses from one design state to the next. One view is that products evolve by a continuous comparison between the design state and the goal, that is, the requirements for the product given in the problem statement. This philosophy implies that all the requirements are known at the beginning of the design problem and that the difference between them and the current design state can be easily found. This difference controls the process. This philosophy is the basis for the methods in Chap. 6.

Another view of the design process is that when a new problem is begun, the design requirements effectively constrain the possible solutions to a subset of all possible product designs. As the design process continues, other constraints are added to further reduce the potential solutions to the problem, and potential solutions are continually eliminated until there is only one final design. In other

ullman-38162 ull75741_02 December 17, 2008 16:16

2.6 Product Decomposition 41

Constraints are often opportunities in disguise.

words, design is the successive development and application of constraints until only one unique product remains.

Beyond the constraints in the original problem specifications, constraints added during the design process come from two sources. The first is from the designer’s knowledge of mechanical devices and the specific problem being solved. If a designer says, “I know bolted joints are good for fastening together sheet metal,” this piece of knowledge constrains the solution to bolted joints only. Since every designer has different knowledge, the constraints introduced into the design process make each designer’s solution to a given problem unique. The second type of constraint added during the design process is the result of de- sign decisions. If a designer says, “I will use 1-cm-diameter bolts to fasten these two pieces of sheet metal together,” the solution is constrained to 1-cm-diameter bolts, a constraint that may affect many other decisions—clearance for tools to tighten the bolt, thickness of materials used, and the like. During the design pro- cess, a majority of the constraints are based on the results of design decisions. Thus, the individual designer’s ability to make well-informed decisions through- out the design process is essential. Decision-making techniques are emphasized in Chap. 8.

2.6 PRODUCT DECOMPOSITION We will conclude this chapter with a method that can is the basis for understand- ing existing products. As such, it can serve as a starting point whether doing redesign, original design, or some other type of design, whether at the system or subsystem level. This product decomposition or “benchmarking” method helps us understand how a product is built, its parts, its assembly, and its function. It cannot be overemphasized how important it is to do decomposition and how it is the starting place for all design. In this chapter, we will decompose to under- stand the parts and assembly. In Chap. 7, the decomposition begun here will be extended to understand function.

Figure 2.11 shows a template that can be used to organize the decomposition. It is partially filled in for a pre-2003 version of the Irwin Quick-Grip. This version is the starting point for the redesign effort that resulted in the product shown in Fig. 2.1

The template begins with a brief description of the product and how it works— its function. This follows with a section showing each part. Only a selection of the parts is shown for the clamp in Fig. 2.11. Each part is given a name, the number required, its material, and the manufacturing process.

Often it can be hard to determine the material and manufacturing process. For plastics, there is a set of simple experiments for rough identification. Over the last few years, handheld devices have been developed that can identify materials

ullman-38162 ull75741_02 December 17, 2008 16:16

42 CHAPTER 2 Understanding Mechanical Design

Product Decomposition

Design Organization: Example for the Mechanical Design Process Date: Aug. 14, 2007

Product Decomposed: Irwin Quick Grip—pre 2007

Description: This is the Quick-Grip Product that has been on the market for many years

How it works: Squeeze the pistol grip repeatedly to move the jaws closer together and increase the clamping force. Squeeze the release trigger to release the clamping force. The foot (the part on the left in the picture that holds the face that is clamped against) is reversible so the clamping force can be made to push apart rather than squeeze together.

Part # Part Name # Req’d. Material Mfg. Process Image

1 Main body 1 PPO or PVC Injection molded

2 Trigger 1 PVC Injection molded

4 Face plate, 1 Polyethylene Injection molded left

Parts:

ullman-38162 ull75741_02 December 17, 2008 16:16

Part # Part Name # Req’d. Material Mfg. Process Image

8 Pad 2 ?? Injection molded

13 Power spring 1 Steel Wound wire

14 Jam plates 2 Steel Stamped sheet

Step # Procedure Part #s removed Image

1 Take off left face plate 4

12 Remove jam plates 13, 14, 1 and power spring from main body assembly

13 Remove trigger from 2 main body assembly

14 Pry off pad from main 8 body assembly

The Mechanical Design Process Designed by Professor David G. Ullman

Copyright 2008, McGraw-Hill Form # 1.0

Disassembly:

Figure 2.11 Product decomposition samples for an older version of the Irwin Quick-Grip. (Photos reprinted with permission of Irwin Industrial Tools.)

43

ullman-38162 ull75741_02 December 17, 2008 16:16

44 CHAPTER 2 Understanding Mechanical Design

just by pointing the device at a sample of the material. While the main market for these devices is recycling, they are very useful when decomposing a product. Details on these are given in the Sources section at the end of this chapter.

The final section of the template is for the disassembly of the product. To build this section of the Product Decomposition report, remove one part at a time. Document the procedure needed to remove the part and the part numbers for those parts removed. Document what was done with a photograph. Figure 2.11 shows only a couple of the steps. Usually disassembly and part naming occur at the same time. Disassembly step 1 shows the left face plate, Part #4, was removed from the product. The internal parts of the clamp can now be seen in the photo. As this is a digital image in the actual template, it can easily be rescaled and studied as needed. Steps 12–14 are shown using a single image. The first one shows the removal of two parts, #13 and #14, at the same time as they come out together. Note how each procedure begins with a verb or verb phrase to tell what has to be done to remove the parts. Make these as descriptive as possible.

2.7 SUMMARY ■ A product can be divided into functionally oriented operating systems. These

are made-up of mechanical assemblies, electronic circuits, and computer programs. Mechanical assemblies are built of various components.

■ The important form and function aspects of mechanical devices are called features.

■ Function and behavior tell what a device does; form describes how it is accomplished.

■ Mechanical design moves from function to form. ■ One component may play a role in many functions, and a single function may

require many different components. ■ There are many different types of mechanical design problems: selection,

configuration, parametric, original, redesign, routine, and mature. ■ Mechanical objects can be described semantically, graphically, analytically,

or physically. ■ The design process is a continuous constraining of the potential product de-

signs until one final product evolves. This constraining of the design space is made through repeated decisions based on comparison of design alternatives with design requirements.

■ Mechanical design is the refinement from abstract representations to a final physical artifact.

■ Product dissection is a useful way to understand the structure of a product.

2.8 SOURCES Good books on designing new products

Clausing, Don, and Victor Fey: Effective Innovation: Development of Winning Technologies, ASME Press, 2004.

ullman-38162 ull75741_02 December 17, 2008 16:16

2.10 On the Web 45

Cooper, Robert G.: Winning at New Products, 3rd ed., Perseus Publishing, 2001.

Vogel, C.M., J. Cagan, and P. Boatwright: The Design of Things to Come, Wharton School Publishing, 2005.

Plastics identification

The PHAZIR is a handheld, battery-powered, point-and-shoot plastic identifier. It weighs only 4 lb (1.8 kg) and takes 1–2 sec to determine the makeup of the sample. www.polychromix.com

Metals identification

The iSort is a handheld, battery-powered, point-and-shoot spectrometer for on-site identifica- tion and analysis of all common metal alloys. Metal identification just requires pointing the gun-shaped iSort at a clean metal sample. The iSort is fairly expensive. http://www.spectro.com/pages/e/p010101.htm

An inexpensive method uses the color of a chemical deposition to identify the metal. The process requires putting a drop of solution on the sample, then using a battery-powered electric charge through the solution to cause a chemical deposition on a piece of blotter paper. The color of the resulting deposit identifies the metal. http://www.alloyid.com

2.9 EXERCISES 2.1 Decompose a simple system such as a home appliance, bicycle, or toy into its assemblies,

components, electrical circuits, and the like. Figures 2.3 and 2.11 will help.

2.2 For the device decomposed, list all the important features of one component.

2.3 Select a fastener from a catalog that meets these requirements:

■ Can attach two pieces of 14-gauge sheet steel (0.075 in., 1.9 mm) together

■ Is easy to fasten with a standard tool

■ Can only be removed with special tools

■ Can be removed without destroying either base materials or fastener

2.4 Sketch at least five ways to configure two passengers in a new four-wheeled commuter vehicle that you are designing.

2.5 You are a designer of diving boards. A simple model of your product is a cantilever beam. You want to design a new board so that a 150-lb (67-kg) woman deflects the board 3 in. (7.6 cm) when standing on the end. Parametrically vary the length, material, and thickness of the board to find five configurations that will meet the deflection criterion.

2.6 Find five examples of mature designs.Also, find one mature design that has been recently redesigned. What pressures or new developments led to the change?

2.7 Describe your chair in each of the four languages at the three levels of abstraction, as was done with the bolt in Table 2.2.

2.10 ON THE WEB A template for the following document is available on the book’s website: www.mhhe.com/Ullman4e

■ Product Decomposition

http://www.polychromix.com
http://www.spectro.com/pages/e/p010101.htm
http://www.alloyid.com
http://www.mhhe.com/Ullman4e
ullman-38162 ull75741_02 December 17, 2008 16:16

ullman-38162 ull75741_03 December 17, 2008 14:55

3C H A P T E R Designers and Design Teams

KEY QUESTIONS ■ Why is it important to know how people do design? ■ How is your ability to design dependent on your cognitive preferences? ■ What are the characteristics of creators? ■ How do individual cognitive abilities interact with the abilities of others during

team activities? ■ Why is a team more than a group of people? ■ What can you do to help teams be successful? ■ How can you measure team health?

3.1 INTRODUCTION Since the time of the early potter’s wheel, mechanical devices have become in- creasingly complex and sophisticated. This sophistication has evolved without much concern for how humans solve design problems. Throughout history peo- ple who were just naturally good at design were trained, through an apprentice program, to be masters in their art. The design methods they used and the knowl- edge of the domain in which they worked was refined through their personal experiences and passed, in turn, to their apprentices. Much of this experience was gained through experiments, through building prototypes and then going “back to the drawing board” to iterate toward the next product. The results of these exper- iments taught the designers what worked and what did not and pointed the way to the next refinement. With this methodology, products took many generations to be refined to the point of mature design.

However, as systems grew more complex and the world community grew more competitive, this mode of design became too time-consuming and too ex- pensive. Designers recognized the need to find ways to deal with larger, more complex systems; to speed the design process; and to ensure that the final design

47

ullman-38162 ull75741_03 December 17, 2008 14:55

48 CHAPTER 3 Designers and Design Teams

be reached with a minimum use of resources and time. In this book we discuss design techniques that meet these goals. To understand how these techniques help streamline the design process, it is important to understand how designers and design teams progress from abstract needs to final, detailed products.

To put this chapter in context, it is important to realize that design is the confluence of technical processes, cognitive processes, and social processes. We begin our discussion of how humans design mechanical objects by describing a cognitive model of how memory is structured in the individual designer. The types of information that are processed in this structure are explored, and the term knowledge is defined. Once we understand the information flow in human memory, we develop the different types of operations that a designer must perform in memory during the design process, and we explore creativity.

Based on this model of the individual’s cognitive process, the chapter moves to the social aspect of design—working in teams. First, the structure of design teams is developed. This includes descriptions of the members of teams and how they are managed. Further, beyond the formal titles that people have, there is a more subtle, cognitive role that people play on teams. Second, an entire section is devoted to building and maintaining a design team. This includes how to start a team, inventory its health, and resolve problems as they develop. Supporting this chapter is a series of templates available at the book’s website.

3.2 THE INDIVIDUAL DESIGNER: A MODEL OF HUMAN INFORMATION PROCESSING

The study of human problem-solving abilities is called cognitive psychology. Although this science has not yet fully explained the problem-solving process, psychologists have developed models that give us a pretty good idea of what hap- pens inside our heads during design activities. A simplification of a generally ac- cepted model is shown in Fig. 3.1. This model, called the information-processing system and developed in the late 1950s, describes the mental system used in the solution of any type of problem. In discussing that system here, we give special emphasis to the solution of mechanical design problems.

Information processing takes place through the interaction of two environ- ments: the internal environment (information storage and processing inside the human brain) and the external environment. The external environment comprises paper and pencil, catalogs, computer output, and whatever else is used outside the human body to extend the internal environment.

In the internal environment, that is, within the human mind, there are two different types of memory: short-term memory, which is similar to a computer’s operating memory (its random access memory or RAM), and long-term memory, which is like a computer’s disk storage. Bringing information into this system from the external environment are sensors, such as the eyes, ears, and hands. Taste and smell are less often used in design. Information is output from the body with the use of the hands and the voice. There are other means of output, such

ullman-38162 ull75741_03 December 17, 2008 14:55

3.2 The Individual Designer: A Model of Human Information Processing 49

External environment

Internal environment

Sensors

Sight

Touch

Recorders

Notes and drawings

Short- term

memory

Long- term

memory

Controller

Figure 3.1 The human problem solver.

as body position, that are less often used in design. Additionally, as part of the internal processing capability, there is a controller that manages the information flow from the sensors to the short-term memory, between the short-term and the long-term memory, and between the short-term memory and the means of output.

Before describing short-term and long-term memory and the control of infor- mation flow, we need to describe the information that is processed in this system. In a computer the information is in terms of bits, or binary digits (0s and 1s), but in the human brain, information is much more complex.

In recent experiments, an orthographic drawing of a power transmission sys- tem consisting of shafts, gears, and bearings was shown to mechanical engineering students and professional engineers. The students were lower-level undergradu- ates who had not studied power transmission systems. The drawing was shown briefly and then removed, and the subjects were asked to sketch what they had seen. The students tended to reconstruct the drawings from the line segments and simple shapes they had seen in the original drawing. Not understanding the complexities of geared transmissions, they could not remember anything more complicated. They remembered and drew only the basic form of the components. On the other hand, the professional engineers were able to remember compo- nents grouped together by their function. In recalling a gear set, for example, the experts knew that two meshed gears and their associated shafts and bearings pro- vide the function of changing the rpm and torque in the system. They also knew

ullman-38162 ull75741_03 December 17, 2008 14:55

50 CHAPTER 3 Designers and Design Teams

what geometry or line segments were needed to represent the form of a gear set. Thus, the experienced engineers using functional groupings were able to include substantially more information than the students in their sketches.

The line segments remembered by the students and the functional groupings remembered by the experienced designers are called chunks of information by cognitive psychologists.The greater the expertise of the designer, the more content there is in the chunks of information processed. Exactly what types of information are in these chunks, however, is not always clear. Types of knowledge that might be in a chunk include

■ General knowledge, information that most people know and apply without regard to a specific domain. For example, red is a color, the number 4 is bigger than the number 3, an applied force causes a mass to accelerate—all exemplify general knowledge. This knowledge is gained through everyday experiences and basic schooling.

■ Domain-specific knowledge, information on the form or function of an in- dividual object or a class of objects. For example, all bolts have a head, a threaded body, and a tip; bolts are used to carry shear or axial stresses; the proof stress of a grade 5 bolt is 85 kpsi. This knowledge comes from study and experience in the specific domain. It is estimated that it takes about ten years to gain enough specific knowledge to be considered an expert in a domain. Formal education sets the foundation for gaining this knowledge.

■ Procedural knowledge, the knowledge of what to do next. For example, if there is no answer to problem X, then decomposing X into two indepen- dent easier-to-solve subproblems, X1 and X2, would illustrate procedural knowledge. This knowledge comes from experience, but some procedu- ral knowledge is also based on general knowledge and some on domain- specific knowledge. We must often make use of procedural knowledge to solve mechanical design problems.

In mechanical engineering the term feature is synonymous with chunks of information. Since a design feature is some important aspect of a component, assembly, or function, the gear set discussed in the preceding example is both a chunk and a feature.

The exact language in which chunks of information are encoded in the brain is unknown. They might be dealt with as semantic information (text), graphi- cal information (visual images), or analytical information (equations or relation- ships). Psychologists believe that most mechanical designers process information in terms of visual images and that these images are three-dimensional and are readily manipulated in the short-term memory.

All design and decision making is limited by human cognitive capabilities.

ullman-38162 ull75741_03 December 17, 2008 14:55

3.2 The Individual Designer: A Model of Human Information Processing 51

3.2.1 Short-Term Memory

The short-term memory is the main information processor in the human brain. It has no known specific anatomic location, yet it is known to have very specific attributes.

One important attribute of the short-term memory is its quickness. Informa- tion chunks can be processed in the short-term memory in about 0.1 second. The term processed implies such actions as comparing one chunk of information to another, modifying a chunk by decomposing it into smaller parts, combining two or more chunks into one new one, changing a chunk’s size or distorting its shape, and making a decision about the chunk. It is unknown how much of the short-term memory is actually used to process the information. We do know that the harder it is to solve the problem, the more short-term memory is used for processing.

The capacity of the short-term memory was first described in a paper ti- tled “The Magical Number Seven, Plus or Minus Two” (see Section 3.8), which reported that the short-term memory is effectively limited to seven chunks of information (plus or minus two). This is like having a computer RAM with only seven memory locations. These approximately seven chunks—these seven unique things—are all that a person can deal with at one time. For example, let us say we are working on a design problem and have an idea (a chunk of information, maybe just a word or maybe a visual image) that we want to compare to some constraints on the design (other chunks of information). How many constraints can we compare to the idea in our head? Only two or three at a time, since the idea itself takes one slot in the short-term memory and the constraints take two or three more. That does not leave much memory to do the processing necessary for comparison. Add any more constraints and the processing stops; the short-term memory is simply too full to make any progress on solving the problem.

A couple of quick experiments are convincing about the limits of the short term memory. Open a phone book and randomly choose a phone number in which the seven digits are unrelated to each other. (A number such as 555-2000 is not acceptable because the last four digits can be lumped together as a single chunk—two thousand.)After looking at the number briefly, close the phone book, walk across the room, and dial the number. Most people can manage to do this task if they are not interrupted or do not think about anything else. The same experiment can be tried with two unrelated phone numbers. Few people are able to remember them long enough to dial them both since they require dialing 14 pieces of information, which is beyond the capacity of the short-term memory. Granted, these 14 digits can be memorized, or stored in long-term memory, but that would take some study time.

Another example of the size limitations of short-term memory is more mechanical in nature. Consider the four-bar linkage of Fig. 3.2. It is made up of four elements: the driver A–B, the link B–C, the follower C–D, and the base D–A.

It is not difficult for most engineers to visualize the follower C–D rocking back and forth as the driver A–B is rotated. Point B makes a circle, and point C moves in an arc about point D. An expert on linkages would only use a single

ullman-38162 ull75741_03 December 17, 2008 14:55

52 CHAPTER 3 Designers and Design Teams

B

A D

E

C

F

?

?

Figure 3.2 A four-bar linkage.

chunk to encode this mechanism. But a novice in the domain of four-bar linkages would need to visualize four line segments, using four chunks plus others for processing the motion. To make the task more difficult, trace the path of point E on the link. This requires more short-term memory. Harder still is tracing the path of point F. In fact, this requires so many different parameters to track that only a few linkage experts can visualize the path of point F.

Another feature of the short-term memory is the fading of information stored there. The phone number remembered earlier is probably forgotten within a few minutes. To keep from forgetting short-term information, like the phone number, many people keep repeating the information over and over. With such continuous refreshing, it is possible to retain certain objects or parts of objects within the short-term memory and to let only the unimportant information fade to make room for the processing of new chunks of information.

Last, it is impossible for us to be aware of what is happening in our short- term memory while we are solving problems. To follow our own thoughts, we need to use some of that memory to monitor and understand the problem-solving process, making that space no longer available for problem solving. Thus, you can not really observe what you are doing during problem solving without affecting what you are trying to observe.

3.2.2 Long-Term Memory

The long-term memory was earlier compared with the disk storage in a com- puter; like disk storage, it is for permanent retention of information. Let us look at the four major characteristics of long-term memory. First, long-term memory has seemingly unlimited capacity. Despite the cartoon in Fig. 3.3, there is no

ullman-38162 ull75741_03 December 17, 2008 14:55

3.2 The Individual Designer: A Model of Human Information Processing 53

Figure 3.3 Long-term memory problems.

documented case of anybody’s brain becoming “full,” regardless of head size. It is hypothesized that as we learn more we unconsciously find more efficient ways to organize the information by reorganizing the chunks in storage. Reconsider the difference between the student’s and the expert’s ways of remembering informa- tion about the power transmission system. The expert’s information storage was more efficient than the student’s.

The second characteristic of the long-term memory is that it is fairly slow in recording information. It takes 2 to 5 min to memorize a single chunk of information. This explains why studying new material takes so long.

The third characteristic is the speedy recovery of information from long term memory. Retrieval is much quicker than storage, the time depending on the complexity of the information and the recentness of its use. It can be as fast as 0.1 sec per chunk of information.

ullman-38162 ull75741_03 December 17, 2008 14:55

54 CHAPTER 3 Designers and Design Teams

The fourth characteristic is that the information stored in the long-term mem- ory can be retrieved at different levels of abstraction, in different languages, and with different features. For example, consider the knowledge an average engineer can retrieve about a car (Fig. 3.4). The sample data ranges from images of entire vehicles to semantic rules and equations for diagnosing problems. Hu- man memory is very powerful in matching the form of the data retrieved to that which is needed for processing in the short-term memory.

The body of a Corvette

is fiberglass.

Car=drive train+body +interior

65 mph; 0 to 60 in 6 seconds

If engine is running rough,

then spark plugs might be fouled.

If car won’t start, then

It’s fast. It handles well.

Car

Parts list: #80312—Floatbody #87426—32 jet

Tire size= 195/60

hp= Tn

5252

1. Check fuel level. 2. Check battery.

Figure 3.4 Knowledge stored in memory about cars.

ullman-38162 ull75741_03 December 17, 2008 14:55

3.2 The Individual Designer: A Model of Human Information Processing 55

3.2.3 Control of the Information-Processing System

During problem solving, the controller (Fig. 3.1) enables us to encode outside information obtained through our senses or retrieve information from long-term memory for processing in the short-term memory. Some of the information in the short-term memory is allowed to fade, and new information is input as it is needed and becomes available. Additionally, the controller can help extend the short-term memory by making notes and sketches; these need to be done quickly so that they do not bog down the problem-solving process. When we have completed manipulating the information, the controller can store the results in long-term memory, or in the external environment by describing it in text, verbally, or in graphic images.

3.2.4 External Environment

The external environment—paper and pencil, computers, books—plays a number of roles in the design process: it is a source of information; it is an analytical capability; it is a documentation/communication facility; and, most importantly for designers, it is an extension for the short-term memory. The first three of these roles seem evident; however, the last role, as an extension for the short-term memory, needs some discussion.

Because the short-term memory is a space-limited central processor, human problem solvers utilize the external environment as a short-term memory exten- sion, much as a computer extends RAM by using cache memory. This is accom- plished by making notes and sketches of ideas and other information needed in problem solving. In order to be useful to the short-term memory, any extension must share the characteristics of being very fast and having high information content. Watch any design engineer trying to solve a problem. He or she will make sketches even when not trying to communicate. These sketches serve as aids in generating and evaluating the ideas by serving as additional chunks of information to be processed. Sketches are fast to make and are information-rich.

3.2.5 Implications of the Model

One of the implications of the information-processing model of human problem solving is that the size of the short-term memory is a major limiting factor in the ability to solve problems. To accommodate this limitation we break down problems into finer and finer subproblems until we can “get our mind around it”—in other words, manage the information in our short-term memory. Typically, these fine-grained subproblems are worked on for about 1 minute before going to the next one. Thus, design of even a simple problem is the solution of many thousands of subproblems. Further, our thinking process has evolved so that, as we solve problems, our expertise about the constraints and potential solutions increases and our configuration of chunks becomes more efficient. This helps offset the “magic number” seven, but human designers are still quite limited. It would almost seem that these limitations would preclude our ability to solve

ullman-38162 ull75741_03 December 17, 2008 14:55

56 CHAPTER 3 Designers and Design Teams

If you try to think about what you are doing while you are doing it, you stop doing it. If you don’t reflect on what you

just did, you are doomed to repeat it.

complex problems. As discussed in the upcoming sections, processing speed and flexibility of information storage and recovery enable designers to develop very complex products.

3.3 MENTAL PROCESSES THAT OCCUR DURING DESIGN

We can now describe what happens when a designer faces a new design problem. The problem may be the design of a large, complex system or of some small feature on a component. We will focus on how a designer understands new in- formation such as the problem statement, how ideas are generated, and how they are evaluated.

In Section 1.6 we introduced seven basic actions of problem solving. The core actions—understand, generate, evaluate, and decide—are refined here.

3.3.1 Understanding the Problem

Consider what happens when a new problem is broached. If we think of its design state as a blackboard on which is written or drawn everything known about the device being designed, then the blackboard is initially blank, i.e., the design state is empty. Let us return to the fastening problem presented in Chap. 1 (see Fig. 1.9):

Design a joint to fasten together two pieces of 1045 sheet steel, each 4 mm thick and 6 cm wide, that are lapped over each other and loaded with 100 N.

Before any information about the problem is put on the design-state black- board, the problem statement must be understood. If the problem is outside the realm of experience (the designer does not know what the term lapped means, for example), then the problem cannot be understood.

But how do we “understand” a problem? Most likely in this way: As the problem is read, it is “chunked” into significant packets of information. This happens in the short-term memory, where we naturally parse the sentence into phrases like “design a joint,” “to fasten together,” and so on. These chunks are compared with long-term memory information to see if they make sense, and then most are allowed to fade. The goal of this first pass through the problem is to try and retain only the major functions of the needed device. Usually a problem will be read or sensed a number of times until the major function(s) is identified. Unfortunately there is no guarantee that, from the usually incomplete data that

ullman-38162 ull75741_03 December 17, 2008 14:55

3.3 Mental Processes That Occur During Design 57

exist at the beginning of a design problem, the most important functions will be identified. In our example there is no ambiguity. The prime function is to transfer a load from one sheet of steel to another through a lapped joint.