Tolerance Design: A Handbook for Developing Optimal Specifications

Description

• Copyright 1997
• Edition: 1st

• Book
• ISBN-10: 0-201-63473-2
• ISBN-13: 978-0-201-63473-0

Whether an engineer, designer, drafter, or technician, each member of the design engineering team has a valuable role to play in the development of product specifications. Tolerance Design recognizes this development process as the responsibility of the entire team and provides practical solutions that each team member can readily apply. The step-by-step details of analytical and experimental tolerance development methods are clearly explained, and as a result, you will be able to develop tolerances more economically.

The book is presented in four sections: Introductory topics to position the tolerance development process, Traditional Analytical and Computer-Aided Tolerance Development, Taguchi's Approach to Experimental Methods of Tolerance Development, as well as several actual industrial case studies illustrating the book's concepts. This book includes a major emphasis for Tolerance Design using Taguchi's Quality Loss Function in harmony with Motorola's famous methods for Six Sigma quality.

The blend of practical examples with substantive case studies provides a comprehensive process approach to tolerance development. Any company interested in properly developing tolerances for their manufacturing, assembly, or service communities will find this text to be a thorough and effective training resource and reference handbook. Students of design and engineering will find this book invaluable as they prepare to enter a competitive job market where practical design optimization skills are at a premium.

0201634732B04062001



Downloads

Downloads

Download replacement figures



Sample Content

Table of Contents

I. SETTING THE STAGE FOR UNDERSTANDING THE SCIENCE OF TOLERANCING.

1. Introduction to Tolerancing and Tolerance Design.

The Historical Roots of Tolerancing. The State of the Art in Tolerancing Techniques. Developing Tolerances: The Role of Engineers and Designers. Concepts, Definitions, and Relationships. Matching Design Tolerances with Appropriate Manufacturing Processes. Introduction to the Taguchi Approach to Tolerance Analysis. Summary. References.

2. The Relationship of Quality Engineering and Tolerancing to Reliability Growth.

The Three Initial Phases of Product Development. The Reliability Bathtub Curve and Tolerancing. Summary. References.

3. Introductory Statistics and Data Analysis for Tolerancing and Tolerance Design.

The Role of Data in Tolerance Analysis. Graphical Methods of Data Analysis. Introduction to the Fundamentals of Descriptive Statistics. The Use of Distributions. Introduction to the Fundamentals of Inferential Statistics. Manufacturing Process Capability Metrics. Six Sigma Process Metrics. The Relationship between the Quality-Loss Function, Cp and Cpk. Summary. References.

II. TRADITIONAL TOLERANCE ANALYSIS.

4. Using Standard Tolerance Publications and Manufacturer's Process Capability Recommendations.

Starting the Tolerance Design Process. The Three Sigma Paradigm. Processes for Establishing Initial Tolerances. Establishing Process Capability and Process Control for Identifying Initial Tolerances. Summary. References.

5. Linear and Nonlinear Worst-Case Tolerance Analysis.

Standard Worst-Case Methods. Summary. References.

6. Linear and Nonlinear Statistical Tolerance Analysis.

The Root Sum of Squares (RSS) Approach. Motorola's Dynamic Root Sum of Squares Approach. Motorola's Static Root Sum of Squares Approach. The Nonlinear RSS Case Method. Process Diagrams for the Statistical Methods of Tolerance Stackup Analysis. The Nonlinear RSS Example. Summary. References.

7. Sensitivity Analysis and Related Topics.

The Various Approaches to Performing Sensitivity Analysis. Using Sensitivity Analysis in Concept Design. An Example of the Use of Sensitivity Analysis in Concept Development. Summary. References.

8. Computer Aided Tolerancing Techniques.

Various Software and Platform Options to Support CAT Analysis. Monte Carlo Simulations in Tolerance Analysis. Characterizing Probability Distributions for Tolerance Analysis Applications. Sensitivity Analysis Using Crystal Ball. How to Use Crystal Ball. Running the Monte Carlo Simulation. Preparing Engineering Analysis Reports. Another Computer-Aided Tolerance Approach. Summary. References.

9. Introduction to Cost-Based Optimal Tolerancing Analysis.

Skills Required for Cost-Based Optimal Tolerancing Analysis. The Various Approaches to Cost-Based Optimal Tolerancing Analysis. Summary. References.

10. Strengths and Weaknesses of the Traditional Tolerance Approaches.

Using Standard Tolerance Publications and Manufacturer's Process Capability Recommendations. Worst-Case Tolerance Analysis. The Statistical Methods of Tolerance Analysis. Sensitivity Analysis. Computer-Aided Tolerancing. Cost-Based Optimal Tolerance Analysis. How the Six Processes Relate to the Overall Product Tolerancing Process. Summary. References.

III. TAGUCHI'S APPROACH TO TOLERANCING AND TOLERANCE DESIGN.

11. The Quality-Loss Function in Tolerancing and Tolerance Design.

Linking Cost and Functional Performance. An Example of the Cost of Quality. The Step Function: An Inadequate Description of Quality. The Customer Tolerance. The Quality-Loss Function: A Better Description of Quality. The Quality-Loss Coefficient. An Example of the Quality-Loss Function. Developing Quality-Loss Functions in a Customer's Environment. Constructing the Quality-Loss Economic Coefficient (A0/(D0)2). Summary. References.

12. The Application of the Quadratic Loss Function to Tolerancing.

The Difference between Customer, Design, and Manufacturing Tolerances. The Taguchi Tolerancing Equations. Relating Customer Tolerances to Engineering Tolerances. An Example of Tolerancing Using the Loss Function (Nominal-the-Best Case). Relating Customer Tolerances to Subsystem and Component Tolerances. The Linear Sensitivity Factor, b. Using the Loss Function for Multiple-Component Tolerance Analysis. An Example of Applying the Quality-Loss Function to a Multicomponent Problem. Setting Up the Problem. Identifying Critical Parameters. Converting the Traditional Tolerance Problem into a Quality-Loss Tolerance Problem. How to Evaluate Aggregated Low Level Tolerances. Using the Loss Function Nonlinear Relationships. Developing Tolerances for Deterioration Characteristics in the Design. Tolerancing the Deterioration Rate of a Higher Level Product Characteristic. Determining Initial and Deterioration Tolerances for a Product Characteristic. Summary. References.

13. General Review of Orthogonal Array Experimentation for Tolerance Design Applications.

Developing Tolerances Using a Designed Experiment. Use of Orthogonal Arrays in Tolerance Design. The Build-Test-Fix Approach. Introduction to Full Factorial Experiments. Methods to Account for Interactions within Tolerance Design Experiments. Summary. References.

14. Introducing Noise into a Tolerance Experiment.

Defining Noises and Creating Noise Diagrams and Maps. Summary. References.

15. Setting Up a Designed Experiment for Variance and Tolerance Analysis.

Preparing to Run a Statistical Variance Experiment. Using the ÷3s Transformation. A Comparison of Output Statistics. Conducting a Tolerance Experiment for Worst-Case Conditions. An L9 Experiment and Monte Carlo Simulation Using m ± 3s Levels Assuming Uniform Distributions. Metrology and Experimental Technique. Summary. References.

16. The ANOVA Method.

Accounting for Variation Using Experimental Data. A Note on Computer-Aided ANOVA. An Example of the ANOVA Process. Degrees of Freedom in ANOVA. Error Variance and Pooling. Error Variance and Replication. Error Variance and Utilizing Empty Columns. The F-Test. A WinRobust ANOVA Example. An ANOVA-TM Example. Summary. References.

17. The Tolerance Design Process: A Detailed Case Study.

The Steps for Performing the Tolerance Design Process. The ASI Circuit Case Study. Setting Up and Running the Experiment. Two-Level versus Three-Level Experiments. Techniques for Putting Noise into the Tolerance Experiment. Running the Experiment. Data Entry. Interactions in Tolerance Experiments. Analyzing the Data. Applying ANOVA. Relating the ANOVA Data to the Loss Function and Process Capability (Cp). Defining the Critica-To-Function (CTF) Factors. Defining the Cost Improvement Parameters (CIP). Identifying and Quantifying the Costs Associated with Improving Quality. Working with Suppliers to Lower Customer Losses through Reducing the Component Parameter Standard Deviations. Calculating New Variances and MSD Values Using the Variance Equation. Quantifying the Cost of Reducing the Parameter Standard Deviations. Identifying and Quantifying the Opportunities for Lowering Costs. Relaxing Tolerances and Material Specifications of CIPs to Balance Cost and Quality. New Loss and Cp after Upgrading the Critical-To-Function and Downgrading Cost Improvement Parameters. Using Tolerance Design to Help Attain Six Sigma Quality Goals. Using Tolerance Design to Improve System Reliability. Summary. References.

IV. INDUSTRIAL CASE STUDIES.

18. Drive System Case Studies.

The Drive System. The Drive Module. Case 1: Defining Tolerances for Standard Drive-Module Components. Case 2: Drive Module for Worst-Case Assembly Analysis. Case 3: Drive Module for RSS (Statistical) Assembly Analysis. Case 4: Drive Module for Computer-Aided Assembly Analysis. Case 5: Drive System Aided by the Use of a Designed Experiment. Summary.

Appendix A. The Z Transformation Tables and the t Transformation Table.
Appendix B. The Adjusted Z Transformation Tables.
Appendix C. The F Tables.
Appendix D. Additional References for Tolerance Design.
Appendix E: Suppliers for Tolerance Design.
Index.

Preface

This is the first book in the history of engineering science to comprehensively address the analytical and experimental methods available for the development of tolerances. This indicates, on one hand, an unfortunate legacy of neglecting a very important engineering topic. On the other hand, it provides an opportunity for engineering practitioners, students, and educators to fill a void in their skill sets.

This book is written for a fairly diverse audience. It is intended to be used as a handbook for industrial applications as well as a course text in technology and engineering science programs at the college and university level. It is structured so that undergraduate technology and engineering students at the associate and baccalaureate levels can equally share in its informational value. The author views the tolerance development process as the responsibility of the design engineering team. This requires that the book be suitable for engineers, designer/drafters, and technicians. Each of these engineering team members has a valuable role to play in the development of product specifications. Tolerance Design contains detailed information for each of these team members.

This book is broad in scope. It reviews the three initial phases of the product-development process where tolerance development resides as the final cost-versus-quality balancing process. It explains, in detail, how tolerance design relates to concept design and parameter design. It also relates the tolerance design process to many other engineering and product development tools and tasks, including reliability growth activities. It covers all the basic analytical methods that exist in modern and traditional tolerancing techniques. It also rigorously covers the latest experimental methods that can be employed in the tolerance design process.

This text details step-by-step the analytical and experimental tolerance-development methods so the reader can more easily and properly develop tolerances. Many examples and case studies are used to convey the process of defining the right tolerances so they can be properly communicated to the manufacturing, assembly, and service communities. The text illustrates how to balance product cost and quality through the tolerance development process. Great emphasis is placed on how the engineering and manufacturing teams must work harmoniously and concurrently to develop optimal tolerances. Tolerance Design instills a philosophy that tolerance development is as much about cost optimization as it is about functional integrity.

This book has been heavily critiqued and scrutinized, on your behalf, by many of the top experts in the U.S. engineering, tolerance analysis, and tolerance communication communities. Much of their work has been included in this text. This was done to make the depth and breadth of the book rock solid. Since the book is the first of its kind, the author and publisher believed it had to be thoroughly evaluated, perhaps even more rigorously than usual, so that it could stand as a major contribution to engineering literature for many years to come. The book was "engineered" to meet your needs so you can meet your customer's needs.

The world is so competitive now that there is simply no way to keep our industries economically viable by using the same ad hoc engineering processes that have evolved over the past 50 years. With this in mind, a major focus in this text is given to the disciplined methods of quality engineering as defined by Dr. Genichi Taguchi. Engineering processes are now being taught in universities and applied in industries all over the United States. This book is very process oriented. It strives to teach discipline in the development of tolerances. In fact, you will have difficulty succeeding in tolerance optimization if the engineering processes that follow the tolerancing process are void of rigor and discipline. Taguchi's approach to product development provides an excellent engineering process context for our discussions.

The literature of off-line quality engineering is primarily focused on the parameter design process. There are numerous books and papers available on this topic, which is frequently referred to as robust design. The other two phases of off-line quality engineering--concept design and tolerance design--have received much less attention in published material. This is particularly true of tolerance design. Until now there has been no single book dedicated to the engineering process of designing tolerances using the quality-loss function. Furthermore, there has not been a single text on how to establish six-sigma process capability in the context of the quality-loss function and upper and lower specification limits that have been produced from the output from orthogonal array matrix experiments.

The question of why so little has been written on this practical subject is relatively easy to answer. Taguchi and his colleagues (myself included) have focused the majority of their writing, teaching, and consulting on developing robustness (insensitivity to sources of variability) as early as possible in the product development process. Consequently, the time and energy spent on promoting tolerance design is relatively small in comparison to the work done to promote the proper application of parameter design. Taguchi has written several interesting chapters that show the power of continuing to improve the quality of a product or process through the methods of tolerance design.

His unique contributions to tolerance literature include the application of orthogonal array matrix experiments--complete with stressful noises, ANOVA data decomposition, and the quality-loss function. He has provided a way to bring the design into production with a deliberate, balancing process to be sure both the customer and the corporation have the right quality and cost designed into the product. Dr. Taguchi's work is the basis for the third section of this book.

In the United States, the traditional approach to tolerancing has been primarily in the domain of designers and drafters as opposed to that of engineers. The engineering focus has been on defining the nominal set points at which the part or design functions correctly. The designers are frequently left to finish the job by defining the tolerances and then properly communicate them to the appropriate manufacturing personnel through drawings or specification documents. Tolerancing occurs naturally as a two-step process: the determination of the tolerance and the communication of the tolerance. In his book Total Quality Development, Don Clausing states:

"In parameter design, the best nominal (target) values are selected for the critical design parameters. In tolerance design, we select the economical precision around the nominal values in two steps: (1) selecting the right production precision, and (2) putting tolerance values on drawings . . . The intrinsically more important part of tolerance design is the selection of the best production precision. Strictly speaking, this has nothing to do with tolerances at all. Tolerances are the numbers that go on the drawing (pp. 249-250)."

The point is that regardless of whether the engineer or the designer or both establish the tolerances, the tolerances are critical to the successful manufacture and performance of the product over its intended life cycle. This book is all about the process of determining production tolerances. The process of communicating the tolerances is left to other fine texts, such as Geometrics III, by Lowell Foster.

This book includes a focus on the economics of the tolerance development process. As Taguchi points out in his book Taguchi on Robust Technology Development, the average loss N products will incur over time, T, is derived from the following general relationship:

This integral equation captures the central concept of how, on average, money (L(y)) is lost over time (T, total life and t, any point in time) as a measurable parameter (y) in a design deviates from its intended target value. It is clear that tolerances can play a significant role in the loss (Li(t, y)) of dollars incurred over the life (T) of the product. Thus, tolerance design can also have an impact on the economics associated with product reliability. This book is designed to make this particular concept of accounting for loss and improving reliability understandable to any technical person with a basic understanding of algebra. The reader should not be alarmed that an integral equation is used to communicate the nature of loss. It is my intention to make the loss function clear to all readers-not just those with a background in advanced mathematics. One of Taguchi's finest professional attributes is that he simplifies complex optimization processes to the point where practicing engineers and technical personnel can rapidly learn and apply them. I took great pains to continue this approach in this text. This is not a book to dazzle the theoreticians; rather, it is a practical guide to solving real engineering problems using a disciplined process. It is also meant as a follow-up text to Engineering Methods for Robust Product Design, by William Y. Fowlkes and myself to help engineers define the optimum nominal (target) set points for a design. Tolerance Design will help you see your design through to the end of the design process by defining the tolerances that support the nominal set points.

As in all engineering tasks, the need for a disciplined process is essential to assure that the life cycle cost goals, including the customer's cost goals, are actually met. To this end, the overall tolerancing process is important. It is too important to be left as a detail that receives little or no serious thought until a project is nearly ready to go into production. It is a deliberate set of steps that are every bit as important as concept design and parameter design. The detailed procedures that are necessary to perform tolerance design on components that have first undergone parameter design is thoroughly documented and ,ow-charted, as is the process of performing tolerance design on components that have not been made robust and are in need of improvement.

This book is also about matching the designed tolerances with manufacturing processes that are capable of economically producing parts or components that maximize quality and minimize loss over the life of the product. Considerable material is presented to provide a strategy for attaining six-sigma-level part quality. The best tolerance design projects can fall far short of their intended goals if manufacturing process variability is not understood and optimized.

Optimizing both the numerator and the denominator of the process capability index (Cp) are considered in this text:

Three software packages are illustrated in this book. For personal-computer applications, I recommend the use of Crystal Ball--a ,exible Monte Carlo simulation package that is readily applicable to statistical tolerance analysis. For analysis of variance and computer-aided experimental design, ANOVA software and WinRobust software are recommended. Tolerance Design places a strong emphasis on the use of personal computers in both the analytical and experimental forms of tolerance development, and provides an overview of workstation-based computer applications.

The book concludes with an entire chapter dedicated to a related series of five case studies from Eastman Kodak Company. These case studies draw upon many of the topics and processes discussed in this text--including how and when to transition from analytical to experimental tolerance development methods.

It has been my pleasure to have been accorded the freedom and time to develop a complete course on tolerancing and tolerance design, and to apply the techniques at Eastman Kodak Company. It has been through these experiences that this text has taken form.

C. M. Creveling
June 1996

0201634732P04062001



Updates

Submit Errata

